Making Things Better Competing in Manufacturing

Making Things Better: Competing inManufacturing

February 1990

OTA-ITE-443NTIS order #PB90-205469

Recommended Citation:

U.S. Congress, Office of Technology Assessment, Making Things Better: Competing inManufacturing, OTA-ITE-443 (Washington, DC: U.S. Government Printing Office, February1990).

For sale by the Superintendent of DocumentsU.S. Government Printing Office, Washington, DC 20402-9325

(order form can be found in the back of this report)

Foreword

U.S. manufacturing is in trouble. That spells trouble for the Nation, because manufacturingprovides well-paid jobs, pays for most privately funded research and development, anddominates international trade. In industry after industry, U.S. manufacturers have lost out tocompetitors who are able to make things better—products with better features and morereliable quality, at lower cost. The key to this better performance is technology, which includesnot only new products and advanced manufacturing equipment but also efficient organizationof work and effective use of people. Once, U.S. manufacturers led the world in technology.Now, in one field after another-first radios and TV, then automobiles, now semiconductors—Japanese manufacturers are passing us by. Other Asian countries like Korea and Taiwan arecoming up fast, and Europe is mounting new challenges.

In a sense, these changes are welcome. Since World War II, U.S. policy has aimed tostrengthen the economies of advanced nations and help poorer ones develop, so it should beno surprise that the world is now full of able manufacturers. It was not part of the plan for theUnited States to fall to second rank, but that is what is happening.

This report considers ways to promote the restoration of American leadership inmanufacturing technology. Some of the things that most need doing are up to industry—especially in handling people, from managers to engineers to shopfloor workers, and informing stable, productive relationships between different segments of an industry complex.Government also has a critical role to play. The first essential is to create an economicenvironment that supports manufacturing and encourages long-term investment in technol-ogy. This means higher national savings rates and a declining Federal deficit. Other lesstraditional activities (at least for the U.S. Government) also deserve consideration-forexample, collaboration with industry on supporting R&D for strategic technologies.

For many years, national security was almost the only acceptable reason for governmentsupport of commercial technologies and industrial excellence, but as the Cold War windsdown, this reason becomes less compelling. In an era of more secure peace but toughereconomic competition, national security is taking on new meanings. To preserve our longtradition of industrial success and rising living standards requires continuing innovation andsuccessful adaptation of existing technology, and that is a task for industry, government, andAmerican citizens.

This report is the second in a series of three in OTA’s assessment of Technology,Innovation, and U.S. Trade. The assessment was requested by the Senate Committee onFinance; the Senate Committee on Banking, Housing and Urban Affairs; and the HouseCommittee on Banking, Finance and Urban Affairs. The first report in the series, Paying theBill: Manufacturing and America’s Trade Deficit, concluded that the stubbornly high U.S.trade deficits of the 1980s and many other signs pointed to genuine weakness in Americanmanufacturing and lags in technology, compared to our best competitors. This report looks forsome of the reasons for the weakness, and suggests policies aimed specifically at repairing it.The last report will examine the trade and industrial policies of Japan, other East Asiancountries, and Europe; and their possible relevance to the competitive position of the UnitedStates.

~f~AA’ ‘ 2

JOHN H. GIBBONSU D i r e c t o r

Assessment on Technology, Innovation, and U.S. TradeAdvisory Panel

Lewis Branscomb, ChairmanHarvard University

Michael AhOCouncil on Foreign Relations

Ralph GomoryThe Sloan Foundation

Howard GreisKinefac Corp.

Joseph GrunwaldUniversity of California, San Diego

Thomas HoutBoston Consulting Group

Ramchandran JaikumarHarvard Business School

Franklin P. Johnson, Jr.Asset Management Co.

Lester C. KroghMinnesota Mining and Manufacturing

Paul R. KrugmanNational Bureau of Economic Research

Alvin P. LehnerdSteelcase Inc.

Ann MarkusenRutgers University

Ray MarshallUniversity of Texas

Regis McKennaRegis McKenna, Inc.

Richard S. Morse*Consultant

David MoweryUniversity of California, Berkeley

Paula StemThe Stern Group

Brian TurnerAFL-CIO

Gus TylerInternational Ladies Garment Worker Union

Lewis C. VeraldiFord Motor Co.

Ezra F. VogelHarvard University

NOTE: OTA appreciates and is grateful for the valuable assistance and thoughtful critiques provided by the advisory panel members.The panel does not however, necessarily approve, disapprove, or endorse this report. OTA assumes full responsibility for thereport and the accuracy of its contents.

iv

OTA Project Staff-Making Things Better: Competing in Manufacturing

Lionel S. Johns, Assistant Director, OTAEnergy, Materials, and international Security Division

Audrey B. Buyrn, Managerindustry, Technology, and Employment Program

Julie FOX Gorte, Project Director

Katherine Gillman, Deputy Project Director

Robert Weissler, Senior Analyst Robin Gaster, OTA Fellow

Samuel F. Baldwin George Eberstadt Joan Gutta

Sebastian Remoy Philip Shapira

Contributors

Brenda Brockman Robert WadeDorothy Robyn

Contractors

Elinor Horwitz Robert R. MillerDavid Sheridan D. Hugh WhittakerYoshitaka Kurosawa Quick, Finan, and Associates, Inc.Industry and Trade Strategies Massachusetts Institute of TechnologyTokaji and Hirsch, Inc. BSI Consulting, Inc.

Administrative Staff

Christopher N. Clary, Office Administrator

Diane D. White, Administrative Secretary

Publishing Staff

Katie Boss, Publishing Officer

Cheryl Davis, Desk-top Publishing

Dorinda Edmondson, Desk-top Publishing

Christine Onrubia Graphic Designer

Susan Zimmerman, Graphics

Contents

Page

Chapter l. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Chapter 2. Strategies To Improve U.S. Manufacturing Technology:Policy Issues and Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

Chapter 3. Financing Long-Term Investments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

Chapter 4. Human Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115

Chapter 5. Links Between Firms and Industries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

Chapter 6. Technology Transfer and Diffusion: Some International Comparisons . . . . . . . . . . 151

Chapter 7. Where We Stand: Public Policy and Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173

.

vi

Chapter 1

Summary

CONTENTSPage

INVESTMENT . . . . . . . . . . . . . . . . ... ...+.... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9PROMOTING COOPERATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13TRANSFERRING KNOWLEDGE .. .. .. .. .. .. .. .. .. ... .+. ....,... .. ... ,.. ...+...+ 18POLICY ISSUES AND OPTIONS . . . . . . . . . . . . . . . . . . . . . .. ............+.=..+....,+. 20

Improving the Financial Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21Human Resources . . . . . . . . . . . . . . . . . . . . . .. . .. .. ... . ... .....9. .. .. A T, ..+. ..+...... 23Diffusing Manufacturing Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25Strategic Technology Policy.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

FiguresFigure Page

l-1. Merchandise Trade Balance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3l-2. Average Annual Productivity Growth in Manufacturing. . . . . . . . . . . . . . . . . . . . . . . . . 4l-3. Merchandise and Manufacturing Trade Balances, 1960-88 . . . . . . . . . . . . . . . . . . . . . . 4l-4. GDP Per Capita in 1988 U.S. Dollars . . . . . . . ......** . . . . . . . . . . . . . . . . . . . . . . . . . . . 6l-5. Fixed Investment in Machinery and Equipment as a Percentage of GNP/GDP . . . . 7l-6. Productivity Performance, World Auto Manufacturers ● .*.... +...,,.. . . . . . . . . . . . 71-7. Quality Performance, World Auto Manufacturers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81-8. Cost of Capital for Equipment and Machinery With 20-Year Physical Life . . . . . . . 101-9. Cost of Capital for R&D Project With 10-Year Payoff Lag . . . . . . . . . . . . . . . . . +... 10

1-10. Cost of Capital for Factory With 40-Year Physical Life . . . . . . . . . . . . . . . . . . . . . . . . 101-11. Real Long-Term Interest Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111-12. Twelfth Grade Achievement Scores in Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121-13. Twelfth Grade Achievement Scores in Advanced Algebr a . . . . . . . . . . . . . . . . . . . . 131-14. Spending for Education Grades K-12, Percent of GDP, 1985 . . . . . . . . . . . . . . . . . . . . 141-15. Spending for Education Grades K-12, Per Pupil, 1985 . . . . . . . . . . . . . . . . . . . . . . . . . . 141-16. Shift in Market Shares for Wafer Steppers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161-17. U.S. Market Shares of Selected Semiconductor Equipment . . . . . . . . . . . . . . . . . . . . . 171-18. World Semiconductor Equipment Sales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

TablesTable Page1-1. Work Force Involved in Manufacturing and Average Full-Time Equivalent

Compensation, 1984 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51-2. Top Ten Semiconductor Equipment Suppliers, World Sales . . . . . . . . . . . . . . . . . . . . . . 16

Chapter 1

Summary

American manufacturing has never been inmore trouble than it is now. Its biggest challengeis from Japan, where, more than in any othernation, well-designed products are manufac-tured with great reliability, while costs arerigorously controlled. Other nations, developedand developing, are rising to the Japanesechallenge in creative ways. The importantdifference is that many of those nations areresponding as nations, with the support andparticipation of government. While some Amer-ican companies and institutions have redoubledefforts to improve manufacturing, the govern-ment is dozing at the switch. Certainly, there aremany problems that manufacturers must solvethemselves. But some of the problems aregenerated by the American people and govern-ment. As a nation, we owe it to ourselves to helpwith their solution.

Symptoms of America’s problems are clearlyvisible: the merchandise trade deficit remainsstubbornly high, despite significant downwardadjustment of the dollar against major curren-

Figure 1-1--Merchandise Trade Balance

Billions of dollars

‘“~o 1-

-50 -

-100 -

-150

I

m

-200 ‘ i 1 I 1 I 1 1 1 1 I 1 1 1 1 1 I 1 1 1 i 1 1 I 1 1 1 1 1 1 I1960 1964 1968 1972 1976 1980 1984 1988

SOURCE: U.S. Department of Commer~, Bureau of Economic Analysis,Business Conditions Digest, S@ember 1939 (VVashington,DC: U.S. Government Printing Office, September 1989), table822.

cies (figure l-l). Productivity growth is slug-gish compared with that of many other advancedand developing nations, including our ablestcompetitors (figure 1-2). U.S. manufacturers areincreasingly dependent on foreign producers fora wide range of machinery and tools of produc-tion. Even the microelectronics industry, oncethe standard bearer for American competenceand inventiveness, is losing sales and marketshare to Japanese, Korean, and Taiwaneseproducers.

The weaknesses in U.S. manufacturing tech-nology must be cured if the Nation is to enjoyrising living standards together with a strong,stable position in international trade. Most of theU.S. trade deficit is in manufactured goods(figure 1-3). The most constructive way to rightthe deficit is to manufacture products that theworld will buy because the products are well-made and reasonably priced (not just because alow dollar makes them cheap). More fundamen-tally, manufacturing is valuable to the Nation asa direct source of productive, well-paid jobs andthe indirect source of many better-than-averagejobs in the service sector (table l-l). Manufac-turing also supports most of this country’scommercial research and development.l

There is no single solution, but all the signspoint in one direction: U.S. manufacturingtechnology must improve—in everything fromproduct design to manufacturing process devel-opment and refinement. For industrial nations,technology is the key to competitive success.Nations that rely on low wages to sell theirgoods in the world market are, by definition,poor, whereas superior technology raises pro-ductivity and thus supports rising standards ofliving. Moreover, technology is a steady, pre-dictable source of advantage, while others mayshift with political currents. For example, anation’s fiscal and monetary policies affect the

IFor mom de~l~ discussion of the place of manufacturing in international trade and the national economy, see Office of ‘kChIIOIOgY As=sammPaying the Bill: Man@acturing and Americans Trade Deficit, OTA-ITE-390 (Sprin@leld, VA: National Technical Information Service, 1988).

- 3 -

4 ● Making Things Better: Competing in Manufacturing

Figure 1-2—Average Annual Productivity Growthin Manufacturing

Percent growth in output/hour

9

7

5

3

1

5.5

3.7

4.4

7.8

U.S. Canada U.K. FRG France Italy JapanSOURCE: U.S. Department of Labor, Bureau of Labor Statistics, “interna-

tional Comparisons of Manufacturing Productivity and LaborTrends, 1986,” June 1989, table 1.

value of its currency, which in turn affects thesalability of its manufactured goods in the worldmarket. But macroeconomic policies are change-able, and are far beyond the control of privatef i n s .

Americans are used to thinking of their nationas leading the world in technology—with theselect company perhaps of a few other devel-oped countries or a few foreign industries. Butthe realization has dawned that we are no longerat the forefront.2 Several major U.S. industrieshave not only fallen behind in technology, butwill be hard put to catch up even if they adopt awhole catalog of changes needed to reverse theslide. Not all American industries are lagging,but trends in many sectors, from computers toaircraft, indicate that our ablest competitors cannow or soon will match our technology, and areaccelerating faster.

Figure l-3-Merchandise and ManufacturingTrade Balances, 1960-88

Billions of dollars50

0

-so -

-100 -

-150 -

-200 f , 1 1 ! 1 1 1 , , i I 1 1 1 1 1 I 1 t 1 , I , , 1 i I1960 1964 1968 1972 1976 1980 1984 1988

— Merchand ise — Manufacturing

SOURCES: U.S. Department of Commerce, Bureau of Economic Analysis,Business Conditions Digest, September 1989 (Washington,DC: U.S. Government Printing Office, September 1969), table622; U.S. Department of Commerce, International TradeAdministration, Office of Trade and Information Analysis,unpublished data, 1989; and President of the United Statesand the Council of Economic Advisers, Economic Report of thePresident (Washington, DC: U.S. Government Printing Office,January 1987), table B-102.

Is this a problem? We have long accepted (inprinciple, if not in fact) that our technologicallead across a wide range of industries was fatedto narrow or disappear as developed countriesrecovered from war damage and poorer coun-tries advanced. But we did not expect the gap toclose so rapidly, and we certainly never ex-pected to fall behind.

The toughest challenge is coming from theFar East. At the close of the 1980s, Japan hasemerged as the world’s premier industrial com-petitor. The United States is still the richest ofnations, with gross domestic product per capitaconsiderably higher than most others (onlyCanada is close; see figure 1-4). Several Euro-pean countries are strong performers in one oranother manufacturing sector or product—especially Germany, which excels in metal-working and machinery, and consistently runslarge trade surpluses. But Japan’s record isunique. It has led all major industrial countriesin productivity growth for decades—not just inthe early postwar years when it was rising fromthe ashes, but also right through the 1970s and

2For ~wep ~~caton of ~erica’s relative technological performance, ibid., PP. 26-35.

Chapter I-Summary ● 5

Table l-l—Work Force Involved in Manufacturing and Average Full-Time Equivalent Compensation, 1984

Average annualPercent of full-time

Wage and sector equivalentsalary workers employment compensation

involved in involved in (thousands ofmanufacturing manufacturing dollars)

Agriculture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Mining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

All public and private services . . . . . . . . . . . . . . . . . . . . . .Ail private services . . . . . . . . . . . . . . . . . . . . . . . . . .Wholesale trade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Transportation and warehousing . . . . . . . . . . . . . . . . . . . .Business services . . . . . . . . . . . . . . . . . . . . . . . . . . . .Radio and TV broadcasting . . . . . . . . . . . . . . . . . . . . . .Electric, gas, water and sanitary services . . . . . . . . . . . . .Communications, except radio and television . . . . . . . . .Automobile repair and services . . . . . . . . . . . . . . . . . . . .Retail, except eating and drinking . . . . . . . . . . . . . . . . . .Finance and insurance . . . . . . . . . . . . . . . . . . . . . . .Hotels, personal and repair services (exe. auto) ,, .Eating and drinking places . . . . . . . . . . . . . . . . . . . . . . .Real estate and rental* . . . . . . . . . . . . . . . . . . . . . . .Amusements . . . . . . . . . . . . . .Health, educ. & social sew. and nonprofit org.....,

Government . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Total . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

792

443

575

19,396

6,4926,3431,501

7041,276

5017112979

1,176413207428

724689

149

27,697

50.4%

45.5

13.3

100.0

9.411.926.324.222.821,821.411.611.610.39.08.57.96.74,50.90.9

29.0%

$11.3

37.026.8

28.7

24624,427.630.324.729.637.539.717.817.127.415.711,021.119.920.2

31.1

$27.4

SOURCE: Workersinvolved in manufacturing dataisderived from OTA input-OutputModel (1980 technicalcoeflicients, 1984 estimated demand, 1984BLSemployment, djustedforcapital flows, imports andduties). Compensation dataderived from Bureau of EconomicAnalysis, National lncomeandProduct Accounts, electronicdata, mappedto input-output industry classifications.

1980s, despite the oil shocks and two periods ofa steeply rising yen. Alone among advancedindustrial countries, Japan managed in the 1980sto combine great productivity growth in manu-facturing with rising manufacturing employ-ment, rising wages and benefits, and greatlyrising output.

These singular achievements suggest somesystematic advantages in Japan that are wellworth examining. There are of course elementsof superiority in other countries too (includingthe United States) and things to be learned fromthem. But Japan’s sustained improvement inproductivity and its pre-eminence in severalindustries that were once nearly an Americanpreserve (e.g., computers, semiconductors) makeJapanese manufacturing a subject of specialinterest. Thus this assessment on the contribu-tion technology makes to competitiveness in

manufacturing concentrates quite heavily—though not exclusively--on Japan.

The Japanese accomplishment rests to a greatextent on technology. Broadly defined, manu-facturing technology covers not only the genera-tion of new products but also know-how in usingequipment, organizing work, and managingpeople to make the products. Where U.S. firmshave fallen down in recent years is in themanufacturing process. The American system,including our great universities as well asindustrial labs, still excels at making technicaldiscoveries and inventing new products. Butforeign companies (especially Japanese compa-nies) have repeatedly beaten U.S. firms ingetting new, improved versions of a great manyproducts to market while keeping costs compet-itive and quality high.


Figure 14-GDP Per Capita in 1988 U.S. Dollars(Purchasing Pourer Parity Exchange Rates)

Thousands25

20

15

10

5

0U.S. Canada Norway Japan FRG France UK Italy

-1988 1987

SOURCE: U.S. Department of Labor, Bureau of Labor Statistics, Offioe ofProductivity and Technology, unpublished data, August 1989.

Over the past decade or so, we have learnedmuch about the sources of the Japanese manu-facturing superiority. We have become familiarwith features such as kaizen, or continualimprovement in every detail of manufacturing;the training of workers to participate in /wizen,learn multiple skills, and work in teams; andkanban, the just-in-time delivery system forparts that depends on reliable high quality andreveals failures to achieve it. These features areall part of the “lean” production system that ispracticed by the leading Japanese manufacturersand is widely credited with keeping costs lowand quality high. The “buffered” system,common in U.S. plants, depends on having largestocks of parts and work in progress, so thatfaulty items can be replaced, and sizable repairareas for fining up defects in the finishedproduct.3 The lean system, by contrast, isdesigned to expose problems while the work isin process, solve the problems, and from thereon do it right the first time. If this report givesonly passing attention to some of these aspectsof Japanese manufacturing, that is not because

they lack importance, but because they are verywell-known.

Greater investment in advanced equipment isanother advantage of leading Japanese indus-tries. From 1976 through 1987, Japanese invest-ment in machinery and equipment consistentlyran from 14.9 to 20.6 percent of GNP; inAmerica, it ranged from 7.5 to 9.0 percent ofGNP4 (figure 1-5). Japanese capital investmentin the late 1980s was especially high, postingdouble-digit increases in both 1988 and 1989. Inmanufacturing, the rate of increase was evengreater--over 25 percent for both years. Animportant reason for these whopping investmentincreases was a shift in production to highervalue added goods.5 Capital investment inAmerican manufacturing rose only 9 percentfrom 1988 to 1989 (less in real terms).

It is not simply advanced hardware that givesJapanese manufacturers the edge, however.Their genius lies at least as much in theemployment of people in relation to the hard-ware. This effective use of people is also a factorin the Japanese ability to shorten the productdevelopment cycle—to repeatedly incorporatestate-of-the-art improvements in their productsand bring them to market quickly. For example,it takes Japanese auto companies about 3 1/2years to get a new model from design tofull-scale production, compared to over 5 yearsfor American and European auto makers.6 A keydifference is the Japanese emphasis on simulta-neous rather than sequential engineering. Thepeople doing research, development and designof the new model are in constant communicationwith the people responsible for manufacture.Other factors are involved too, such as thereliance of the major manufacturers on a trustedgroup of suppliers to do part of the productdevelopment work. The result is a headstart over

3- -s ~ ~M ~ JOhn F. IQticik, ‘‘A New Diet for U.S. Manufacturing, ’ Technology Review, J~. 28, 1989.

4~~m~m~ ~~w ~, wor~&-om~c o#/ook, April 1989. me Jap- fips exclude public investtnmt, while those for the unitedStatea do not,

SW J- Development Bank, “The Japan Development Bank Reports on Capital Spending: Suwey for Fiscal Year 1988 to 1990,” Economicand Industrial Reseamh Department, September 1989.

6W B. ~SIIC ad Takahiro Fujimoto, “OverlappingRoblem Solving in Product Development” working paper 87-048, Harvard Business School,revised April 1988.

Chapter 1--Summary . 7

Figure 1-5-Fixed investment in Machinery andEquipment as a Percentage of GNP/GDP

Percent

“~20

15

10

5

01976-80 1981-85 1986 1987 1988*

= U n i t e d S t a t e s J a p a n * * U West Germany

= France = United Kingdom

● January to Junew Figures for Japan exclude public investment.

SOURCE: International Monetary Fund (IMF), Wodd Economk OuUook(Washington, DC: April 1989), table 17.

slower competitors in responding to consumerpreferences and, perhaps even more important,in incorporating the latest technologies.

Some American managers are now adoptingJapanese-style approaches, or versions of them,to turn out better goods at lower cost. Forexample, in the early 1980s, it took twice asmany hours to assemble a standard car in anaverage American auto plant as in the averageJapanese plant. By 1988, U.S. assembly plantshad improved enough that the Japanese advan-tage was down from 100 percent to about 50percent (25.1 hours for assembly in the averageU.S.-owned and managed plant v. 16.8 hours inthe Japanese). The best Japanese plant had anadvantage of 5.4 hours over the best Americanplant7 (figure 1-6).

In quite a few other industries (e.g., textilesand steel), well-managed U.S. firms have shownthat they are able to turn some of the Japanese-style approaches to good account. But that is noreason for complacency. For one thing, thetarget is moving. Faced with the high yen, whichraises the prices of goods they export, the

Figure 1-6-Productivity Performance, World AutoManufacturers

Productivity: hours/auto60

50

40

30

20

10

0

Japanese plants,North America

n

U S plants,North America

n

= Best Average m Worst

● Includes foreign owned.● * Includes East Asia, Mexico, and Brazil.

SOURCE: John F. Krafcik and John Paul MacDuffie, “Explaining HighPerformance Manufacturing: The International Automotive As-sembly Plant Study,” paper prepared for the International MotorVehicles Program International Policy Forum, May 1989.

Japanese Government and Japanese manufac-turers redoubled their own efforts to improvetechnology and competitiveness. For example,the best Japanese plant shaved assembly time fora standard model car from 16 to 13.2 hours injust one year, 1987 to 1988, and the averageplant improved from 19.1 hours to 16.8.8 TheJapanese were also holding onto a lead in betterquality. In 1988, the average Japanese assemblyplant was turning out cars with less thanthree-quarters of the defects of cars produced inAmerican plants. U.S. plants stacked up verywell against the Europeans, however, as shownin figures 1-6 and 1-7.

The reasons for Japanese success are broadand complex. Public as well as private actions,and the interrelation between them, are verymuch involved. The issues selected for analysisin this assessment include both, and may begrouped into a few broad areas: 1) the cost andavailability of capital, and its influence onbusiness decisions to invest for the long pull inproduct and process improvements; 2) the use of

TJohnF. wcfi ad JoM pad Mac ~ffie, ‘Explaining High performance Manufacturing: The International Automotive Assembly PISnt Study,”working paper of the International Motor Vehicles program of the Massachusetts Institute of Technology, May 1989.

‘Ibid., p. 5.


Figure 1-7--Quallty Performance, World AutoManufacturers

Quality: assembly defects/100 vehicles250

200

150

100

50

0

1European plants,

Europe In I

U.S. plants,North Amsrlca I I I

Japanese plants,North America n New entrants”

_ Best = Average m Worst

● Includes East Asia, Mexico, and BrazilNOTE: Data are derived from 1988 J.D. Powers International Quality

Survey and corporate data.

SOURCE: John F. Krafcik and John Paul MacDuffie, “Explaining HighPerformance Manufacturing: The International Automotive As-sembly Plant Study,” paper prepared for the International MotorVehicles Program International Policy Forum, May 1989.

human resources to contribute to manufacturingexcellence, with special emphasis on engineers;3) relations between supplier and customerfirms within an industry complex, in particularthe benefits of close, cooperative links; 4) waysto diffuse new technologies from outside sourcesto private companies, and especially to smallermanufacturers; and 5) existing governmentprograms-Federal, State, regional and local—that help (or in some cases hinder) U.S. manu-facturing firms in using technology to improvetheir competitive performance.

Lessons from the successes of other countriesare not always easy to apply. Some elements inthe Japanese system may be quite adaptable toU.S. companies that are enterprising enough totry them-for example, close relations betweendifferent segments of an industry complex (e.g.,chemical companies that make textile fibers,textile producers, apparel makers, designers,and retailers) in which suppliers are attuned andresponsive to the needs of their customer firms,and purchasers are willing to form stable,cooperative relations with their suppliers. Otherpractices and policies of foreign nations wouldbe much harder to translate into Americanterms-for example, the century-old system of

vocational education that trains half the youngpeople of West Germany in good work habitsand a variety of skills. And some policies ofother nations are quite foreign to our traditionsand outlook-for example, centralized directionof trade and industrial policy as practiced inKorea (until recently, when controls have loos-ened somewhat).

One way or another, however, the UnitedStates must regain excellence in the manufactur-ing process. That is key to raising income for theNation. No longer can U.S. industries count onprofiting from new inventions for years beforecompetitors begin to produce them. Manytechnical inventions cannot be protected fromskillful imitators-and the world is now full ofmanufacturers who can quickly and ably pro-duce things that were invented elsewhere (just asU.S. manufacturers themselves have often donewith foreign inventions). Over the long run, acountry and its citizens cannot control or profitfrom what they cannot produce competently.

Restoring or creating excellence is no easytask. U.S. manufacturers who once were themasters of mass-production grew complacent inthe years of American domination. Many stillcling to wasteful production systems that take anarrow view of cost reduction, and do it at theexpense of reliability, flexibility, and customerservice. Many smaller manufacturers are farbehind the times technologically. Federal tech-nology policy is still aimed much more atresearch and the generation of new inventionsthan at quickly diffusing new technologies(whatever their source) and putting them intopractice. Some government policies run counterto manufacturers’ efforts to improve their perform-ance, although that is not their intention. Mostimportant is the government’s inability to elimi-nate the budget deficit, which increases pressureto raise interest rates and the value of the dollar,and directly diminishes manufacturers’ abilityto sustain long-term, risky investments. TheFederal Government, along with many State andlocal governments, has initiated some newprograms to help manufacturers improve com-petitiveness and technology, but these are mod-

Chapter 1--Summary . 9

est at best. The dampening effects of macro-economic and foreign policies can easily over-whelm them.

With will and effort, a nation’s industries canchange. Forty years ago, Japan was a poornation, backward in manufacturing technology,lacking engineers and scientists, relying mostlyon low labor costs to make products attractiveenough for export. Between that Japan and theJapan we know today are years of heavyinvestment in people, technology, and machin-ery, and a great deal of sacrifice on the part ofconsumers. The United States today is in a farstronger position than Japan was then but,ironically, this may make it harder to undertakethe sacrifices and changes needed to rebuild ourcompetitiveness. We are still a wealthy nation,and there is no widespread feeling that we are inor approaching a crisis. Under such circum-stances, it would take extraordinary leadershipto summon the energies to make significantchanges. One hopeful sign is that the nations ofthe European Community—also wealthy andwith no apparent crisis—have pulled together tocreate a new economic order, with the SingleMarket Act.

The European Community’s efforts to har-monize internal markets beginning in 1992 haveseveral things in common with the measuresJapan took to industrialize two decades ago.They also have much in common with measuresthe United States needs to consider to improveour competitive performance. Broadly speak-ing, they are measures to promote investment inpeople, technology, and equipment; to dissemi-nate information and know-how; and to encour-age cooperative efforts to solve common prob-lems.

INVESTMENTInvestments in technology require patience.

Researchers, inventors, and designers oftenmust wait years—sometimes decades—for theirefforts to pay off. Although investments in

equipment are more predictable and less risky,even these may not break even for years.

America’s financial climate is not conduciveto long-term investments in technology andequipment, compared with Japan, Germany, andthe most rapidly developing Asian nations.Several things contribute to this relativelyunfriendly environment. High U.S. capital costsshorten the time horizons of investors, so do thepressures exerted on companies by the stockmarket, particularly by institutional investorsand takeover specialists. In sum, both govern-ment policies and business practices reinforcean excessive concern with short-term profit inAmerica. If these conditions persist, it will beincreasingly difficult to keep up with technolog-ical advances made elsewhere.

U.S. capital costs have been and remain highcompared with those in Japan, the nation thatprovides the greatest contrast with U.S. short-term thinking. There is some disagreement overjust how large (or small) the differences are, butmost recent studies estimate significantly highercapital costs in the United States9 (figures 1-8,1-9, and 1-10). On the high side, the estimatesrange up to 13 percentage points difference,while the difference at the low end is on theorder of 1 or 2 percentage points. Even relativelymodest differences of a few percentage points incapital costs can be a significant disadvantage inmaking investments that take many years to payoff.

U.S. capital costs are high for many reasons.Interest rates rose in the 1980s and remain highprincipally because of the enormous pressure ofthe budget deficit, which is a large drain onsavings, and the fall in other savings rates. Butthere is a great deal more to capital costs thaninterest rates. The price a firm pays for capital isalso a function of its relationships with creditorsand equity holders, and the taxes it pays. DuringJapan’s high growth period, which lasted until1973-74, heavy reliance on debt financing frommain banks kept capital costs down for Japanese

%fich~] L. lkto~s,~ckmd K. kster, and Robert M. SoloW, Made in America: Rega”ning the Productive Edge (Cambridge, MA me ~ ~ss,1989).


Figure 1-8--Cost of Capital for Equipment andMachinery With %)-Year Physical Life

Cost of capital141 !

12 -

10 -

8 -

6-

;~1977 197819791980 1981 19821983198419851986 19871988

‘— United States ‘-+ Japan

SOURCE: Robert N. McCauley and Steven A. Zimmer, “ExplainingInternational Differences in the Cost of Capital,” Federa/Reeerve Bank of New York C?uwteriy R&view, Summer 1969,table 2.

Figure l-9-Cost of Capital for R&D Project With10-Year Payoff Lag

20 -

15 -

10 -

5 -

o~1977197819791980 1981 19821983198419851986 19871988

‘– United States ‘ Japan

SOURCE: Robert N. MoCauley and Steven A. Zimmer, “ExplainingInternational Differences in the Cost of Capital,” Federa/Reserve Bank of New York Quarterly Review, Summer 1989,table 2.

manufacturing firms (particularly in favoredindustries). A variety of Japanese Governmentpolicies encouraged the banks to lend heavily atlow rates to large corporations. These policiesincluded direct government lending through theJapan Development Bank (which is a signal toprivate banks), administrative guidance fromthe Ministry of Finance, and close regulation of

Figure I-l O-Cost of Capital for Factory With40-Year Physical Life

Cost of capital14

12 -

10

8 -

6 -

4 ‘

2 -

0 1 1 1 1 I 1 1 1 I 1

1977 197819791980 1981 19821983198419851986 19871988

‘ -- United States + Japan

SOURCE: Robert N. McCauley and Steven A. Zimmer, “ExplainingInternational Differences in the Cost of Capital,” FededReserve Bank of New York (2uarterly Review, Summer 1989,table 2.

every aspect of banking and finance, includingthe disposition of household savings.

Today, Japan has enormous capital reserves,and most major corporations finance all theirinvestment with retained earnings and deprecia-tion. Moreover, Japan is deregulating its finan-cial markets, and large Japanese companies aregetting more of their external capital in foreignmarkets. Most estimates of U.S. and Japanesecapital costs still show American firms at asubstantial disadvantage-one study, for in-stance, reports U.S. cost of capital at 20.3percent, compared with 8.7 percent in Japan.10

But even if nominal costs were the same,differences in the financial environments in thetwo countries would still favor Japanese firms.Most of the stock of large Japanese corporationsis held by other corporations, often in the samekeiretsu (industry group), who agree to hold thestock for long periods with few demands inreturn. This system, known variously as crossshareholding, mutual shareholding, or stableshareholding, is in marked contrast with U.S.practice. Here, shareholders must be given farmore attention; corporations pay larger divi-dends, and corporate managers are under heavy

lmo~~ N. McCauley and Steven A. Zirnmer, “Explaining International Differences in the Cost of Capital,’ Federul Reserve Bank o~fVew YorkQwtertyReview, summer 1989, pp. 7-28. These figures apply to investments in research and development. Other investments, such as equipment andmachinery and factories, axe also shown to be more expensive in America than in Japan and West Germany.

Chapter 1--Summary ● 11

pressure to show a profit each quarter. In the1980s, new financial instruments have made itmuch easier for outsiders to mount takeoverbids, and managers in U.S. companies feel thatthey must show profits or become vulnerable totakeover attempts. American managers’ increas-ing preoccupation with the short-term bottomline in the 1980s is in part due to that vulnerabil-ity.

Several other factors tend to reinforce short-term bias in America. None by itself is conclu-sively important, but together they have aconsiderable effect. Company size and structuremay account for some of the short term focus ofthe semiconductor industry, in particular. Theleading Japanese semiconductor producers arelarge, integrated, stable companies making avariety of products, from semiconductors tocomputers and consumer goods. The U.S. indus-try has a few large, integrated producers, makingchips mostly for their own use, but the merchantfirms that sell semiconductors to systems mak-ers are mostly smaller, entrepreneurial compa-nies. Such companies have been highly innova-tive, but also highly unstable. Personnel turn-over (especially defections to start new firms) ishigh, as are rates of entry and exit. Theirrelatively small size, instability, and irregularcash flow make it especially hard for them toraise the large amounts of capital required forsemiconductor production. These factors exag-gerate the short-term focus that is endemic inU.S. financial markets.11

Government policies that increase uncer-tainty also aggravate the problem. For example,in the 1980s, American business managers werefaced with a very high dollar, which made itharder to sell goods abroad and to competeagainst foreign goods at home. The dollar finallybegan falling in 1985. But throughout the highdollar period of the early 1980s, the U.S.Government made no provision for firms work-ing under that disadvantage. In contrast, theJapanese Government put in place special loanand loan guarantee programs to help Japanese

firms cope with endaka (high yen) after theinternational accords that brought the dollardown in 1985.

The single most important step the govern-ment could take to improve the financial envi-ronment is to greatly reduce the Federal budgetdeficit, and eventually eliminate it. That wouldhelp to lower interest rates and allow the dollarto find a level that more accurately reflects thecompetitiveness of American industry (Figure1-11 shows real long-term interest rates in the1970s and 1980s.) It would also be a powerfulsignal to the business community that govern-ment could be relied on to provide somestability.

None of this means that American manufac-turing is entirely a victim of circumstancesbeyond its control. U.S. companies are hobbled,but not crippled, by a financial environment thatundervalues long-term investment. Some of themyopia of U.S. firms could be overcomethrough the will of top management. Against thegeneral background of short-term decisionmak-ing, a few firms standout as long-term investors.Many of these firms have done well. But thepower of finance and accounting in Americancorporations has lifted financial specialists to

Figure 1-1 l—Real Long-Term Interest Rates

Percent per annum

‘“~

2 I-2

F---J-4~1974 1976 1978 1980 1982 1984 1986

SOURCE: Organization for Economic Cooperation and Development(OECD), /iistorica/Statistics 1960- 1967 (Paris, France: OECD,1989), table 10.10.

IIM~umS, et al., 1989, op. cit.


many top decisionmaking spots, and their biasescould be difficult to overcome, especially if therewards for managing for the short-term bottomline do not start to dwindle.

In the discussion so far, investment in tech-nology has been defined as investment in capitalequipment, research, and development. TheUnited States also needs well-educated andtrained people to make the best use of sophisti-cated technology. Currently, the investments wemake in human resources have disappointingresults.

Success in manufacturing depends on thecompetence and inventiveness of people at alllevels. Increasingly, workers from the produc-tion line to the executive suite must be comfort-able with advanced technology. Productionworkers are responsible for implementing statis-tical process control procedures; designers, linemanagers, and workers must interact frequentlyand productively; and everyone must assumebroader responsibility for making high-qualityproducts effectively. The skills demanded forthese tasks are those of analysis and problem-solving. The days when most factory workersused their hands more than their heads aredisappearing.

American workers are poorly equipped tocope with these changes, in part because ourpublic schools do not educate many of ourchildren adequately, and in part because firmshave been slow to adopt production systems thatdemand higher order skills, and to train workersto use them. Firms, in turn, are often reluctant toinvest heavily in training for fear that they willnot be able to recoup their investments.

U.S. educational deficiencies are great inscience and mathematics. In the mid-1980s,American junior high school students ranked10th in arithmetic, 12th in algebra, and 16th ingeometry in tests of mathematics competence in20 countries. In a another comparison of stu-dents in 14 nations, American 12th gradersranked 12th in geometry and 13th in advancedalgebra (figures 1-12 and 1-13). In the 1960s,American students performed as well as stu-

Figure l-12—Twelfth Grade Achievement Scores—

Hong KongJapan

England\ WalesSwedenFinland

New ZealandBelgium(Flemish)

ScotlandCanada(Ont)

Belgium(French)Israel

Us.Hungary

Canada(B.C.)Thailand

in Geometry

1 1 1 ( 1 1 1

0 10 20 30 40 50 60 70 80Mean score

SOURCE: International Association for the Evaluation of EducationalAchievement, The Underachieving Curriculum: Assessing U.S.School Mathematics From an International Perspective (Cham-paign, IL: Stipes Publishing Co., 1987).

dents anywhere in the world. Further evidenceof deterioration is the decline in ScholasticAptitude Test scores over the last quarter of acentury.

Workers who cannot cope with mathematicsor problem-solving are a liability in advancedmanufacturing. For example, Motorola deter-mined that workers in its Factories of the Futureneeded math skills equivalent to seventh gradeproficiency to get by. Even this modest require-ment has obliged Motorola to invest tens ofmillions of dollars in training.

Not only is our general public educationinadequate, our vocational education systemfalls far short of the standards set by othercountries. It certainly does not match theapprenticeship training taken by more than halfthe young people of West Germany. This systemgets much of the credit for the broad diffusion oftechnical competence throughout German man-ufacturing.

While there are small indications of im-provement—a recent turnup in SAT scores, forexample-there is need for a great deal more.The fact that American students are behind those


Figure l-13-Twelfth Grade Achievement Scores in

Hong Kong

Japan

Finland

England\ Wales

Belgium(Flemish)

Israel

Sweden

Canada(Ont)New Zealand

Belgium(French)

ScotlandCanada(BC)

Hungary

Us.Thailand

Advanced Algebra

1 1 1 1

0 20 40 60 60 100Mean score

SOURCE: lnternationaJ Association for the Evaluation of EducationalAchievement, The Underachieving Curriculum: Assessing U.S.School Mathematics From an International Perspeti”ve (Charn-paigrt, IL: Stipes Publishing Co., 1987).

of other advanced nations—and of severaldeveloping nations as well—makes it harder forthe United States to keep up in manufacturing.Another worrisome trend is demographic. In thepast, most engineers and scientists were whitemales; they now comprise a shrinking portion ofthe population of school-age children. Minori-ties and women have historically performedmuch less well than white males in math andscience, for reasons that are only partly under-stood. To avoid a future scarcity of technolo-gists, the Nation must devote particular effortsto improving math and science proficiencies—of students of both sexes and all races—all theway from grammar school to employer-provided training.

Do we need to invest more money? It is awidely held belief that the United States investsmore in educating its children than other na-tions, both per capita and as a share of gross

domestic product.12 This is clearly true only ifpost-secondary education is included. A recentstudy that separated out education past highschool found that U.S. public and privatespending on schooling from kindergartenthrough 12th grade, as a share of GDP, is lowerthan in most industrialized countries-tied for12th among 16 (figure 1-14). In spending perstudent in grades K-12, the United States rankshigher—5th of the 16 (figure 1-15).13 T h eUnited States has some special educationalproblems: our population is much more diversein culture and language than that of most of ourcompetitors. It could well take heavier invest-ments in human resources to solve our uniqueproblems.

PROMOTING COOPERATIONPartly because of American traditions—the

emphasis on individual initiative, for example—and partly because of public policies that limitcooperation, U.S. firms tend to be isolated fromcustomers, suppliers, and competitors comparedwith Japanese and many European firms. Japa-nese firms, in particular, are knitted into anetwork of mutual obligation and cooperation.This is not to say they don’t compete; competi-tion is fierce, but is often greater in productquality and features than in price.14 The bondsof cooperation and obligation, together withrelatively limited price competition in the Japa-nese market, provide Japanese firms with twoadvantages: access to a wider array of informa-tion and support than they would have alone,and enough stability to encourage investment inequipment, knowledge, and people.

U.S. companies, on the whole, do not formstrong collaborative links. The typical relation-ship between supplier and customer is distant,even adversarial. Price has been the major basisfor dealings with both suppliers and competi-

IZF~eX~ple, ~Sident BA told the “Education Summit’ in September that the United States “lavishes unsurpassed resources IOUr children’s]schooling.”

ISM. Mh -11 and Lawrenee Mi.shel, “Shortchanging E4hXtiOn,” Economic Policy Institute briefing paper, Washington, DC, January I%Xl.14~ fwt ~ Pnws of may Conwer g- made in Japm we lower ~ tie u~t~ Stare and other f~ign countries than in Japan iIXhCateS that

Japaneaemanufacturersdo not always compete vigorously on price, Japan’s complex distribution system amounts for some but not all of thehigherretailprice for many goods.


Figure 1-14--Spending for Education Grades K-12.

SwedenAustria

Switzerland

NorwayBelgium

DenmarkJapan

CanadaFRG

France

Netherlands

U.K.Italy

Us.Australia

Ireland

percent of GDP, 1085

0 1 2 3 4 5 6 7 8Percent of GDP

SOURCE: Lawrence Mishel and M. Edith Raaell, “Shortchanging Educa-tion,” Eoonomic Policy Institute briefing paper, Washington, DC,January 1980. -

-.

tors. This is not invariable, nor is it withoutadvantages. In some industries-notably, theairline and aircraft industries-relationships be-tween suppliers and customers are traditionallystrong; and in some where relations used to bedistant or hostile--such as textiles and apparel—they are becoming stronger. Moreover, pricecompetition among suppliers or between com-petitors, is desirable and healthy. But taken toofar, narrow reliance on price competition cansever close links between customer and sup-plier, and reduce incentives to improve qualityand timeliness, Close and stable relationshipswith customer firms are incentives for supplierfirms to invest in human resources and inequipment that may take several years to payoff. To illustrate the point, in a recent study ofmetalworking companies, about half the firmsthat had not bought numerically controlled (NC)or computer numerically controlled (CNC) ma-chine tools cited lack of stable demand for theirproduct as the reason.15

Both parent and supplier companies in Japanbenefit from close, cooperative relationships.Without having to manage every detail, theparent company is able to demand favorable

Figure 1-15--Spending for Education Grades K-12,

Sweden

Switzerland

Norway

Canada

Us.Austria

Denmark

Japanm

FRG

Belgium

France

Netherlands

Australia

U.K.

Italy

Ireland

.Per Pupil, 1985

0 1,000 2,000 3,000 4,000 5,000

SOURCE: Lawrence Mishel and M. Edith Raaell, “Shortchanging Educa-tion,” Economio Policy Institute briefing paper, Washington, DC,January 1990; arrd U.S. Department of Labor, Bureau of LaborStatistic& unpublished data, August 1989.

terms for costs, quality, and delivery times. Thesupplier has the advantage of a reliable customerwho can provide assistance with technicalproblems and occasionally with finance if needed.While these relationships are often quite stress-ful for the supplier companies, they havepromoted the diffusion of technology and know-how to Japan’s myriad of small companies withremarkable effectiveness, aided by an abun-dance of Japanese Government technology dif-fusion programs. (See the following section inthis chapter on Transferring Knowledge and ch. 6).

In contrast, American companies have tradition-ally opted for one of two strategies: verticalintegration, or arms’ -length dealing with com-peting suppliers. While vertical integration couldbe thought of as the ultimate in close relation-ships, the control over cost that a company canexercise with an outside supplier may be sacri-ficed. And pitting suppliers against each other—making them compete for every contract onprice with no assurance of ever getting anotherone—makes it more difficult to transfer technol-ogy and design responsibilities. The Japanesesystem has been a remarkably effective com-promise. A measure of its effectiveness is that

15Mqell~ R. Kelley @ H~ey Brooks, ‘‘The State of Computerized Automation in U.S. Manufacturing,” Joh F. Kenn~y SC~l of&VtXIIIDUlt, Harvard hiveraity, 1988.


many American industries-the motor vehicleindustry, as well as the textile and apparelindustries-are making similar arrangementswith their own suppliers.

Close relations between capital equipmentsuppliers and their customer firms are especiallyimportant to technological prowess, particularlyin fast-moving industries like microelectronics.In the past two decades, American industry hasbecome steadily more dependent on foreignmanufacturers for its production machinery.Japanese suppliers have come to dominate themarket for workhorse CNC machine tools;Swiss, German, Japanese, and other Europeanmakers lead the market for textile and paperindustry machinery; and U.S. producers ofsemiconductor production equipment are fastlosing the lead to Japanese rivals.

In textile industry machinery, where thedomestic market share fell from 93 percent in1960 to less than half in 1986, the reasons for thedemise of most U.S. producers are instructive.The industry’s decline, which began in the1960s, was due largely to its unresponsivenessto customer needs and to a short-term perspec-tive, reflected in scanty spending on researchand development compared with foreign com-petitors. The neglect of R&D spending wasmade worse by the merger mania of the 1960s.Most of the U.S. textile machinery companieswere bought by conglomerates.

Although the decline of the American textilemachinery industry has not, it seems, crippledAmerican textile makers. Nearly all reportsatisfactory service from their foreign suppliers.However, the situation is different in the semi-conductor industry. As recently as 10 years ago,American firms held more than three-fourths ofsemiconductor production equipment world mar-ket. By 1988, the U.S. share was down to 47percent and still dropping (table 1-2, figures1-16, 1-17, and 1-18). This year, Perkin-Elmer,one of the major remaining U.S. manufacturers

of lithography equipment, dropped out of thatmarket, which had become a money loser for thecompany.

Already, losses in the American semiconduc-tor equipment industry are a handicap for U.S.semiconductor producers. U.S. producers saythat, for some critical production equipment,they are unable to buy the latest model fromJapanese makers only after it has been in wideuse by Japanese chipmakers for months. ManyU.S. chipmakers are concerned that their abilityto get state-of-the-art equipment will declinefurther in the future. The next generation oflithography equipment is expected to use X-rays, and the Japanese are well ahead of U.S.companies in developing X-ray lithographyequipment. If commercial use of X-ray lithogra-phy equipment begins, as expected, in the1990s, it is likely that the first use will be inJapan. That development would add to thealready substantial number of microelectronicstechnologies dominated by Japanese producers.

Sematech, the U.S. industry-led consortiumto develop a process to manufacture a 16-megabit DRAM, has given top priority toimproving relations between chipmakers andequipment producers. Sematech’s directors seebetter relations as essential to develop a range ofhigh-quality, affordable equipment and materi-als for American producers.

U.S. producers of supercomputers also riskdependence on Japanese suppliers of compo-nents. Significantly, many of those suppliers arealso competitors, making supercomputers them-selves, or else are closely aligned with competi-tors. For example, the highest performancememory and bipolar logic components forsupercomputers come only from Japan. Themanagement of Cray, a U.S. manufacturer ofsupercomputers, has at times been told that thelatest and best of these components are ‘‘not yetavailable for export” from Japan.l6 They are,however, available to Japanese supercomputer

161EF4CJSAB tit~ on -~ications and Information Policy, “U.S. SuperComputer Vulnerability,’ report to the IMhXe Of ~wtricd mdElectronics Engineers, Inc., prepared by the Scientific Supercomputer Subcommittee, Committee on Communications and Information Policy, UnitedStates Activities Board (Washington, DC, August 1988).


Table 1-2—Top Ten Semiconductor Equipment Suppliers, World Sales(millions of dollars)

1982 1988

Perkin Elmer . . . . . . . . . . . . . . . . . . . . ... .$162 Nikon . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..$521Varian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 Tokyo Electron (TEL) . . . . . . . . . . . . . . . . . 508Schlumberger . . . . . . . . . . . . . . . . . . . . . . . 96 Advantest . . . . . . . . . . . . . . . . . . . . . . . . . . . 385Takeda Riken(Advantest). . . . . . . . . . . . . 84 Applied Materials.. . . . . . . . . . . . . . . . . . . . 382Applied Materials.,.. . . . . . . . . . . . . . . . . . 84 General Signal... . . . . . . . . . . . . . . . . . . . . 375Eaton . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 Canon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290Teradyne . . . . . . . . . . . . . . . . . . . . . . . . . . 79 Varian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211Canon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 Perkin Elmer . . . . . . . . . . . . . . . . . . . . . . . . 205General Signal . . . . . . . . . . . . . . . . . . . . . . . 77 Teradyne . . . . . . . . . . . . . . . . . . . . . . . . . . 190Nikon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 LTX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180(Japanese Firms Italicized)

SOURCE: VLSIResearch,lnc.

Figure l-16-Shift in Market Shares forWafer Steppers

I

60 -

40 -Japan

\

Europeo I I I I I I I

‘83’84 ‘85 ’86’87’88 ’89est

NOTE: The wafer stepper is a device central to manufacturing semlcon-ductors.

SOURCE: VLSIResearch,lnc.

makers, and the Japanese supercomputers them-selves are ready for export. Closer relations withU.S. suppliers is not just an advantage but anecessity for maintaining market share, in aworld where a firm’s major suppliers are itsfiercest competitors.

Another wellspring of Japanese technicalprowess is cooperative research and develop-ment, which has the advantages of sharedexpenses and synergism. Participants in consor-tia to develop new products or techniques cangain access to research results they could notafford on their own, and have the chance to workwith scientists or engineers from other firms andinstitutions.

Complex manufacturing processes and sophis-ticated products demand increasing inputs ofresearch and development. The higher the costof R&D, the riskier the investment-too risky,perhaps, for all but the largest and most stablefins. For example, it is costing billions ofdollars to develop X-ray lithography, an amountthat strains the resources of even giant firms. InJapan, a government-sponsored consortium ishelping to share the risk and effort involved indeveloping commercial X-ray lithography.

R&D consortia have other attractions. Forexample, they are often effective at diffusingtechnology to participants; they help to avoidproblems of redundancy, or wasting of re-sources on reinventing wheels; and they can bevaluable training grounds for researchers. Espe-cially when government is a participant, censor-tia can help to provide adequate investment intechnologies that have a great many externalities—where the rewards cannot be captured by asingle firm. In this way, consortia can helplengthen the short time horizons of Americanmanagement.

Consortia are not,of course, a panacea.Theyseem to work best when there are clear goals andleast potential for conflict among members--forexample, in catch-up projects, where no firm canhope to get a monopoly on a new technology. IfAmerica were in the competitive position itoccupied two decades ago, we might wellconclude that the case for stimulation of consor-tia (especially ones with government participa-tion) is dubious. But that is not the situation. The

Chapter I-Summary ● 1 7

Figure 1-17—U.S. Market Shares of SelectedSemiconductor Equipment

\ 1

-.. -1 US. resist- - - - - - -

\ processing80

K - .\ . - - -

t ~

\ \/\ \

60- - U.S. integrated ‘ - ‘*H ‘%-?-.circuit testers -\. —.

\ - - -N.

4’‘ \\ “ \40

/\ \ \.

U.S. stepping ‘~ . . \aligners \ \

20-— — — — — — — — ~

o1979 1980 1981 1982 1983 1984 1985 1986 1987 1988

Year

SOURCE: VLSI Research, Inc.

United States has serious competitive problemsto solve. The European Community, moreover,has opted to support a profusion of new scienceand technology consortia. These consortia arelargely aimed at overcoming what are perceivedas substantial foreign leads in a wide variety oftechnologies. While the EC’s technology con-sortia probably will never amount to more than10 percent of all the Community’s expenditureon R&D, that small percent is viewed as critical,both because it adds to the amount spent, andbecause it gives the EC an important strategiclever for guiding European manufacturing tech-nology.

American industry and government havemoved cautiously toward collaborative R&D inthe past few years. The Federal Government putup half the funding for Sematech, and it hascontributed $5 million per year for 3 years to theNational Center for Manufacturing Sciences, aconsortium designed to do generic research onmetalworking and other manufacturing technol-ogies. The National Science Foundation’s Engi-neering Research Centers offer another ap-proach. ERCs are university-based centers that

Figure 1-18--World Semiconductor Equipment Sales

Billions of dollars5

4

3

2

1

01984 1985 1986 1987 1988 1989 1990 ‘

= United States J a p a n

m Rest Of world m Joint ventures

● Forecast

SOURCE: VLSI Research, inc.

get half their funds from NSF, one-third fromindustry—which must cooperate with the uni-versity in generating and running the researchprogram-and the rest from university, State,and local funds. This small program encouragesinterdisciplinary engineering research and edu-cation and promotes cooperation among univer-sity and industry researchers.

One obstacle that sometimes hinders greatercollaboration-more in downstream activitieslike manufacturing than in R&D--is our anti-trust law. The discouragement comes not be-cause all collaborative projects would actuallyviolate antitrust law, but because the law israther unclear, its penalties can be harsh, andtrials are expensive. Antitrust law can alsointerfere with U.S. firms’ merging to facecompetition from larger foreign fins.

These problems suggest the need for modestchanges in our antitrust laws. They need to bemade carefully, so as to preserve the laws’protection against price-fixing and other anti-competitive practices. Possible approaches in-clude clarifying that conduct should be judgedwith full consideration of the long-term benefitsof cooperation, reducing harsh penalties, andproviding for advance approval of cooperativeprojects.


TRANSFERRING KNOWLEDGE

Public and private institutions for diffusingnew technologies across the manufacturingsector are thin in the United States. In particular,there is little technical assistance available tosmall manufacturing enterprises. While somesmall manufacturers are on the cutting edge oftechnology-the Silicon Valley startup springsto mind-most are not. Many cannot afford todevote the time and attention to keeping up withtechnological developments made in the UnitedStates, to say nothing of technical advancesmade abroad.

It is uncommon for large manufacturers inAmerica to lend technical assistance to theirsuppliers—still less for the first line suppliers topass along technical help to smaller subcontrac-tors. Both are everyday practice in Japan. Thereis little in this country to compare with Japan’sdense nationwide network of free, public tech-nology extension services for small and medium-size firms. Nor do we have anything like thehuge programs of financial assistance thataccompany technical assistance to small andmedium-size firms in Japan. In 1988, low-costdirect loans to smaller firms from Japanesenational government financial institutionsamounted to more than $27 billion, not to speakof $56 billion in loan guarantees, plus additionaltechnical and financial assistance from prefectu-ral and local governments.

In the United States, aside from some smallprograms for disadvantaged individuals, theFederal government makes no direct loans tosmall business. In fiscal year 1989, the SmallBusiness Administration underwrote guaran-teed loans worth $3.6 billion. A few States haveindustrial extension services to help smallmanufacturers make informed decisions aboutimproving their production methods and imple-menting new technology. No accurate count isavailable, but these State programs are probablyfunded at about $25 to $40 millions per year.

Federal involvement in industrial extension issketchy, although Congress has recently takensome steps to strengthen it. The Federal programof technology extension consists mainly of threeManufacturing Technology Centers created inthe Omnibus Trade and Competitiveness Act of1988 and funded at $7.5 million in fiscal year1990; three more centers are planned. Alto-gether, current industrial extension programs,State and Federal, reach only a small fraction—probably less than 2 percent per year-of theNation’s small manufacturing firms.

Government technical assistance to smallmanufacturers in Japan far outpaces similarprograms in America. Because financial andtechnical assistance programs are interrelated,an estimate of the size of these programs is notavailable, but they are large. For example,Japan’s national government provides half thefunding for the nationwide system of 185technology extension centers with the other halfprovided by prefectural governments. Totalfunding for the centers is over $470 million peryear. Local governments support additionaltechnology extension centers as well. But gov-ernment assistance is not the only or even themajor form of technical assistance. In a recentsurvey done by MITI’s Small and Medium SizeEnterprise Agency, 45 percent of respondents(small and medium-size businesses) reportedthat they received technical assistance from aparent company, 37 percent got information, 28percent were loaned or leased equipment, and 24percent got training for their employees.17 Insome cases, vertical transfer of technologywithin Japanese supplier groups is effectiveenough to allow major manufacturers to dele-gate other functions to their suppliers. BothToyota and Nissan, for instance, have delegatedassembly of some of their cars to formerfirst-tier suppliers.

American companies—including all the BigThree motor vehicle companies—have insti-tuted similar programs recently, becoming both

17D.H. ~tt~er, ‘New TahIIoIogy ~q~isi~ion in srn~l JapaneW Enterprises: Government Assistance and Private hIitiative, ” COnWactm repoflprepared for the Office of Technology Assessment, May 1989, p. 23.


more demanding and more supportive of theirsuppliers. They have pared down the numbers ofsuppliers, given more technical assistance, andare moving towards performance-based stand-ards. But U.S. firms are still far behind Japanesemanufacturers in diffusing technology and know-how along supplier chains, or among firmswithin an industry, or from public institutions toprivate fins.

Because of Japanese direct investment, someAmerican firms are experiencing the Japanesesystem firsthand. According to a recent GAOstudy, U.S. auto parts producers who work withJapanese transplant assembly firms report thattheir Japanese customers keep in closer contactthan their U.S. customer firms, and send manymore staff on site visits to the supplier’s plant.They characterized the Japanese companies theywork with as ‘‘preventative’ in solving prob-lems, rather than “reactive,” like Americanfirms.18

Large U.S. firms as well as small ones sufferfrom isolation. Their customary arm’s-length,adversarial relation with suppliers deprivesthem of the back-and-forth collaborative workon new technologies that takes place betweenlarge firms and first-line suppliers in much ofJapanese manufacturing. This collaboration isimportant to innovation in Japan. Japanesemanufacturers of all kinds of products, fromautomobiles to office copier machines, are quickto make incremental changes in products andbring new models embodying the latest technol-ogy to market ahead of their competitors.

Another contributing factor to firms’ compet-itive position is their readiness to scan the world,find out what new technologies are available andplug them into new products. American firmsseem much less inclined to exploit technologiesthat originate outside the fro-a stance oftencalled the not invented here (NIH) syndrome.One study of 50 large Japanese firms and 75large American firms found that Japanese firms

spent considerably less time and money than theU.S. firms in developing new products andprocesses, mostly because the Japanese wereadept at exploiting innovations made elsewhere,while American firms were trying to generatemore of their innovation internally .19 The abilityto make effective use of external technology isalso related to short product cycles. Japanesefirms in automobiles and electronics have man-aged to pare product cycles so that they areshorter than those of American competitors.Since new ideas, from inside or outside, aremost likely to be adopted at the beginning of aproduct cycle, shorter cycles mean moreinnovation—and they do, for many Japaneseindustries.

The reluctance of U.S. firms to adopt andwork with outside ideas has not underminedtheir ability to apply big-bang, fundamentallynew technologies that can be exploited commer-cially. American companies in general havebeen good at this, and many small startup firmshave found venture capitalists to stake them.NIH applies more to technologies that are goodfor incremental improvements. It may also helpto explain why U.S. firms take curiously littleadvantage of new technologies developed inFederal laboratories. However, another reason isthat the labs, short of money for technologytransfer and hampered by red tape, do not reachout to industry.

Most of the $21 billion per year spent onR&D in Federal labs is for defense or basicresearch-missions not directly relevant to com-mercial manufacturing. Some of this R&Dcould be made useful to civilian manufacturing,both by transferring lab technology to industryfor further development and by lab-industrycooperative R&D on subjects of mutual interest.Although Congress passed several bills in the1980s to encourage commercialization of tech-nology from the Federal labs, such commerciali-zation has been modest.

18u.s. G~~ ~w~g ~lce, Foreign Investment: Growing Japanese Presence in the U.S. AJUO l?tdW?Y, GAO~SIAD-88-l 1! M~h 1988”

lg~w~ -field, “Industrid bovation in Japan and the United States, ” Science, Sept. 30, 1988.

20 . Making Things Better: Competing in Manufacturing

One main reason is that the labs’ efforts atencouraging commercialization have not beenadequately funded. Without line-item funding,such efforts are often considered by personnel atthe labs and their parent agencies to be meredistractions from their primary missions. Interac-tions between the Federal labs and privateindustry require a new philosophy and newprocedures, and the resolution of some difficultissues (e.g., potential conflicts of interest).Resolving them is more difficult when agencyofficials, such as the general counsel, putforward objections and there is no strongcountervailing voice to push the process along.In addition, some provisions of the law hinderthe labs from granting a firm exclusive rights totechnology. Without those rights, firms may notfind it worth their while to commercializetechnology coming out of the labs.

Concerns about exclusive rights extend toR&D in general. Many American firms com-plain that in the United States and, especiallyabroad, their new products and manufacturingprocesses are copied by imitators who did notpay to develop them. They desire strongerintellectual property protection for new tech-nology-chiefly patent rights, and copyrightsfor software. Without it, they assert, they faceunfair competition and cannot pay for theirR&D.

This argument has some merit, and somemeasures to increase protection would help. Themost promising ones include strengtheningpatent enforcement in the United States andJapan, and negotiating to harmonize and eventu-ally unify the patent systems of different coun-tries. However, there are limits to the benefits tobe expected from beefing up intellectual prop-erty protection. Developing countries may beinduced to add some protection but, on thewhole, they do not see stronger measures as intheir interest. More generally, strong protection,while encouraging creation of technology, caninhibit its diffusion and, in the long run, cannotmake up for disadvantages in manufacturingquality and cost. Therefore, while strongerintellectual property protection can help U.S.

manufacturing competitiveness somewhat, meas-ures to improve manufacturing quality and costwill help more.

POLICY ISSUES AND OPTIONSIn building a stronger technological base for

American manufacturing, both industry andgovernment have important parts to play. Manyof the things that must be done are squarely upto manufacturers themselves. Company manag-ers have to learn to use their people moreeffectively by promoting a back-and-forth flowof people and ideas between research (or design)and production, insisting on design for easymanufacture, pushing simultaneous engineeringof improved products and the processes to makethem, and giving shopfloor workers the trainingand responsibility for improving efficiency andproduct quality. Likewise, it is managers’ job toget the fat out of the American productionsystem—for example, by trimming inventoriesthat cost money and hide problems, and byorganizing work for reduction of waste. And itis largely up to managers to make the most offorming cooperative relationships between largefirms and their smaller suppliers, or betweendifferent segments of an industry complex.

There is also much that government can do.Traditional U.S. R&D support, mainly fordefense and science, has been beneficial to theNation as a whole and often to industry inparticular, but it is not enough to maintaintechnological leads, or even parity, in mostindustries-especially since most of the otherOECD nations are making greater efforts toadvance civilian technology.

First, government policies that create anenvironment more conducive to manufacturingmake it easier for companies to concentrate onthe things that only they can do to improvetechnology. For example, if government poli-cies succeed in lowering the cost of capital tobusiness, or lifting some of the pressure forshort-term profits, they are “preparing theground’ (as the Japanese say) for business to doits job well.


Government can also take more direct ac-tions, some within traditional U.S. policy, andothers less so. Starting with broad policiesaffecting the financial environment and humanresources, they could go on to stepped-upprograms for active diffusion of technology toprivate firms and, still further, to a strategicapproach that would target government R&Dsupport to critical technologies.

The possibilities for government action donot stop there. Many governments throughoutthe world use means beyond R&D support topromote industries they consider strategicallyimportant. For instance, they may favor certainindustries with low-cost capital or government-guaranteed purchases, and they may add furthersupport with trade policies designed to managecompetition from dominant foreign producersduring developmental phases. In building up itsindustrial might, Japan relied heavily on coordi-nated technology, industry, and trade policies topromote key industries. Korea and Taiwanfollowed Japan’s lead, and the European Com-munity is using many of the same industrial andtrade policy tools as it prepares for the Europeansingle market in 1992.

Whether the United States should or evencould try to use such comprehensive govern-ment policies to bolster competitiveness will beconsidered in another report, the final one inOTA’s assessment of Technology, Innovation,and U.S. Trade. That report will discuss industryand trade policies of Europe and the Asian rimcountries, and in what way they might berelevant to the United States.

In this assessment, the spotlight is on technol-ogy. The policy options analyzed in chapter 2and summarized below are directed toward fourprincipal strategic aims:

. Improving the financial environment forU.S. manufacturing firms. This meanslowering capital costs and relieving otherpressures in the financial markets to showhigh short-term profits every quarter. Thegoal is a more hospitable environment for

●

●

●

long-term investment in new technologiesand productive equipment.Upgrading education and training of theworkers, managers, and engineers neededin manufacturing. U.S. manufacturing suf-fers from the failings of our public schools,but also from failures of managers inorganizing work and training people to useadvancing technologies effectively. Be-sides continuing efforts to improve educa-tion generally, government can help withthe retraining of active workers and thebetterment of manufacturing engineering.Diffusing technologies throughout the manu-facturing sector. Government can be muchmore active than it has been up to now inhelping manufacturers acquire up-to-dateproduction equipment and learn to use iteffectively. Options might include stepped-up Federal support for technology exten-sion services and a subsidized equipmentleasing system. Such things as easier ac-cess to technologies coming from Federallabs or foreign countries could benefit allU.S. manufacturing.Supporting R&D for commercially impor-tant technologies. Some technologies ofgreat potential benefit to society do not getadequate private backing because the pay-off for individual firms is too small,uncertain, and far in the future. The U.S.Government has sometimes given specialsupport to R&D for commercially impor-tant technologies, but in an ad hoc ratherthan proactive way. A coherent, strategictechnology policy require having an agencyin charge that can set goals and choosetechnologies to support that fit the goals.

Improving the Financial Environment

To keep up with the competition, U.S.manufacturing firms need two basic things thatare mainly the province of government tosupply: well-educated workers and capital coststhat are not so high as to be disabling. As mattersstand, government in this country is not doingwell at supplying either of these necessities.


The combination of massive governmentdissaving (the Federal budget deficit, at historichighs in the 1980s) and anemic personal andbusiness saving (at historic lows in the 1980s) isa powerful force driving up interest rates and thecost of capital to business. Congress has madesome progress in reducing the Federal budgetdeficit but it remains high. Some combination ofhigher revenue and lower spending over severalyears will be needed to reduce the budget deficit,and this poses a problem. Many of the policyoptions suggested in this report would, all otherthings being equal, have a contrary effect,because they would entail either increased taxexpenditure or reduced revenue. If these or otherpolicies are not to have the perverse effect ofincreasing the deficit, even stronger measureswould be needed to reduce it. If the UnitedStates succeeds in restoring its strong competi-tive position, then economic growth will help toshrink budget problems in the future. There willbe a price to pay in the short run for improvingmanufacturing, but if it restores our ability toraise standards of living for the great majority ofAmericans in the long run, it will be worth it.

The budget deficit is a significant source ofupward pressure on interest rates, but not theonly one. To make capital less costly, the supplyavailable for capital formation must also beexpanded. That means raising domestic savingsrates. Although the personal savings rate hasrisen from its extreme low in 1987—less than 2percent-it is still below the U.S. norm of 6 to8 percent, and far below the rates in Japan andmost European countries. Some analysts arguethat the United States can continue to rely onforeign capital to make up the difference be-tween domestic investment and domestic sav-ings, but that is inconsistent with loweringcapital costs. It takes high interest rates to attractforeign capital.

To encourage household savings, Congresscould consider a national savings initiative,which would reward increases in regular savings(e.g., payroll savings) for households in all taxbrackets. To be effective, such a campaignwould need to include several substantial sav-

ings inducements, such as guaranteed interestrates, high enough to be attractive, on widelyavailable savings instruments. One suggestion isfor anew type of government bond with a fixedcoupon rate. Reducing taxes on the interestincome to regular savings could also be consid-ered.

Inducements to save may not be sufficient toraise savings rates or promote capital formationin industry without some accompanying meas-ure to discourage consumption. Congress maywish to consider a consumption tax, scaled to taxluxury items most heavily, or with substantialexemptions to avoid the severe regressivity of aflat consumption tax. Another possibility is tolimit the deductibility of interest paid on homemortgages more severely. There are some limitsnow, but they are set very high; this encouragesconsumption of housing and builds equity forhouseholds, but the capital tied up in housing isnot available for industrial capital formation.

The measures suggested above could help tobring down interest rates generally, and thatwould tend to lower capital costs. Interest ratesand capital costs are not synonymous, however;capital costs are also a function of taxes and ofrelationships between capital suppliers and com-panies. Several measures could help to lower thecost of capital to U.S. companies even if generalinterest rates remain high. One set of optionsCongress might wish to consider is special taxinducements for technology development andcapital investments. The United States has trieda few such measures in the past. For example,the Accelerated Cost Recovery System (ACRS)and the Investment Tax Credit (ITC) weredesigned to promote capital investment, and theresearch and development tax credit to increaseR&D spending. While the effectiveness of thesemeasures is debated, there is enough substanceto the arguments in their favor that they (ormeasures like them) are worth considering. Andthey should be considered separately, for theyare very different. ITC and ACRS were veryexpensive (costing tens of billions of dollars,when they were in full force); such measurescould, if designed carefully, promote mainly


improvements in manufacturing technique. TheR&D tax credit is far less expensive, and hasmore effect on new technology developmentthan on current practice.

Another set of forces affecting capital costs,especially for long-term investment in technol-ogy development and capital equipment, is thecurrent wave of hostile takeover activity andspeculative turnover of stock. This activity, andthe threat of it, reinforces the effect of highcapital costs in impelling managers to focus onshort-term profits. The relative influence of thetakeover boom and high capital costs is acontroversial matter which OTA does not re-solve. Nonetheless, it is reasonable to concludethat takeover activity is a significant damper onmanagers’ willingness to commit resources tolong-term projects, or to retain earnings forreinvestment. The pressure from this sourcemight be manageable if overall capital costswere lower, or if there were enough effectivecountervailing measures to promote higher lev-els of investment in R&D and capital equipmentpurchase.

As it is, Congress might wish to considermitigating the pressures of hostile takeoveractivity by means of incentives for investors tohold investments longer. This might be done byadjusting the capital gains tax rate to favorlong-term gains and penalize short-term assetturnover. This measure would have most effectif the tax were extended to pension and otherfunds that are currently tax-free, but account formore than half the transactions in the financialmarkets. Another option is to tax securitiestransactions, which would penalize those whoseturnover is greatest. However, without real,steady progress toward eliminating the budgetdeficit, all of these other measures taken to-gether will probably have only a marginal effect.

Finally, the financial environment of theUnited States is unstable and unpredictable,compared with our premier international com-petitors, Japan and West Germany. In Germany,in particular, macroeconomic policymaking con-centrates on maintaining stability in prices and

exchange rates and controlling inflation. Suchstability is an enormous asset to business,especially in a country that is heavily dependenton foreign trade (like West Germany), andespecially when supplier-manufacturer-customer links are increasingly likely to spannational borders (as the 1992 European SingleMarket approaches).

Japan’s financial environment is also verystable. Policymakers there are highly sensitiveto how macroeconomic developments affectbusiness, and they take steps to help the privatesector adjust. For example, after the interna-tional financial accords were reached to raise thevalue of the yen (and other currencies) againstthe dollar in the mid-1980s, the JapaneseGovernment put in place loan programs to helpfirms (small ones, in particular) adapt to therising yen (endaka). Japan’s economy did slowdown in 1985 and 1986, at the beginning ofendaka, but the adjustment was swift. Muchmore painful were the circumstances faced byAmerican manufacturers in the early 1980swhen the dollar began its long climb, and nogovernment policy was in place to ease theadjustment. In sum, a major difference betweenJapanese and U.S. policies is that little concernis evident in the United States about the effectsof macroeconomic, trade, and other policies onthe competitiveness of U.S. firms in general ormanufacturers in particular. In Japan and WestGermany, competitiveness is customarily takeninto account. It plays a prominent role in makingand implementing those governments’ policies.

Human Resources

Human resources, like capital costs, have apervasive effect on manufacturing. In the past,most manufacturing workers learned their jobsby the sides of more experienced workers, andan ordinary grammar school or high schooleducation was plenty of preparation for aproduction worker in manufacturing. Today,with automation affecting more workplaces andless automated work being exported overseas,production jobs in manufacturing require moreconceptual knowledge-and often competence


in statistical process control and managingcomputerized equipment. Jobs typically encom-pass more diverse tasks than in the past, andworkers must grasp the relationships of differentparts of production to each other in ways neverrequired before. In other words, more is de-manded of manufacturing workers. At the sametime, the typical American education is leavingyoung people less well prepared for theirworklives. Managers have remarked for yearsthat young people could be better prepared, butthe situation now is commonly described as acrisis. And it is likely to get worse before it getsbetter. About half the new entrants to the workforce between now and the turn of the centurywill be members of minority groups, and abouttwo-fifths of minority children live in poverty.Poor children typically drop out of school indisproportionate numbers, and many grow uplacking the skills they need to be productiveworkers.

There is a broad consensus that the Nation’spublic school system needs help. But even ifhelp arrived tomorrow, the results would bemany years coming. A more immediate ap-proach is to help people already in the workforce to acquire needed skills. While some largecompanies are providing education and trainingthemselves, the financial burden of such programs-good ones can run into hundreds of thousands tomillions of dollars-is another drain on limitedfinancial resources, Most small companies can-not afford extensive training programs.

Congress could consider several options tohelp workers raise their educational levels andimprove their work skills. One is to offerfederally guaranteed student loans to employedpeople taking classes part-time; another is to letemployers and employees deduct the costs oftraining and education (the present tax lawalready allows this, subject to some limitations).Another possibility is to tailor military trainingprograms, which are already extensive, to fitmore closely the skills required of workers incivilian jobs. The Federal Government providesless than $10 million to a program that partiallyfunds demonstration projects for literacy teach-

ing in workplaces. There is ample evidence ofadditional demand for such projects; increasedfunding could be used effectively and immedi-ately. Training could also be made a part of anytechnology extension services offered by theFederal Government or funded in part byFederal money, (See the section below onTechnology Extension.)

These suggestions do not constitute a com-plete list of options for training active workers.A fuller examination of the possibilities forcongressional action will appear in a forthcom-ing OTA report, Worker Training: Implicationsfor U.S. Competitiveness.

Although well-educated and trained produc-tion workers are essential to improving manu-facturing efficiency and quality, there are other,equally critical needs for highly trained people.Production workers are a steadily falling per-centage of manufacturing employment; profes-sional and technical employees are a growingshare. Engineers, in particular, are essential forexcellence in manufacturing. It could be moredifficult in the future to maintain an adequatesupply of engineers to sustain manufacturing.

There is not now an obvious shortage ofengineers in manufacturing; about as manyengineers are employed per thousand workers inthe United States as in Japan and Germany,whose manufacturing is justly famous for itsexcellence. But Japan is graduating more engi-neers per capita than the United States, andGermany has what is probably the world’s finestset of training institutions to provide technicalpeople for manufacturing, from the shopfloor tothe engineering workstation. Meanwhile, in theUnited States, the demographic group mostinclined to enter engineering—white males-isshrinking as a proportion of young people.

This trend is not new. Several Federal pro-grams, are already in place to encourage womenand minorities to enter engineering. Largerprograms support the recruitment and trainingof students generally in scientific and technicalcareers, and special training for teachers.) Manyof these programs are producing good results


and could be expanded. But without improve-ment in math and science education in theelementary and high schools, their effects arebound to be limited. Children who performpoorly in elementary school arithmetic and mathare unlikely to choose engineering careers.General education improvement, especially inmath, is the first necessity for keeping theengineering pipeline filled.

Some possible programs could help shore upthe supply of engineering talent for the next fewyears, before improvements in education (if theyare made) begin to yield results. If defenseprograms wind down as expected over the nextfew years, Federal programs might help retrainand equip engineers who have been working inthe military sector to enter civilian manufactur-ing. More generally, programs to encourage orfund midcareer training of engineers whoseknowledge needs updating might be considered.

The effective use of engineers is at least asimportant as an adequate supply. There areindications that U.S. manufacturers could makebetter use of their engineers. Elitism amongengineering staffs, and their aloofness towardshopfloor problems in producing their designs,are often cited as a peculiarity of Americanmanufacturing. This kind of problem is bestsolved by manufacturers themselves, but theFederal government could encourage manufac-turers to recognize and correct the problem,through support of education and research inmanufacturing engineering. One option is toincrease Federal support of manufacturing engi-neering, possibly through the creation of aManufacturing Sciences Directorate in the Na-tional Science Foundation.

Diffusing Manufacturing Technology

Making the financial and human resourceenvironment more conducive to improved manu-facturing quality, efficiency, and technologymay not be enough. American manufacturershave lost too much ground to foreign manufac-turers, in nearly every sector. Even with lowercapital costs and more competent people, somemanufacturers may still lack the resources or the

knowledge to find or develop and implement thebest technologies.

Congress might consider an array of optionsto promote technology diffusion and transfermore widely, or remove obstacles to diffusion.None, by itself, will make a great deal ofdifference; patience and an experimental ap-proach will be required to make any of themwork. Some may fail. Yet it is likely that somecombination of policies to promote technologytransfer could pay off handsomely, given time,the commitment to adapt to changing circum-stances, and the willingness to learn fromexperience.

Technology Extension

Large firms generally have the resources todevelop or acquire technologies they need,although they may neglect to take what theycould from outside the firm. But many smallfirms have a hard time staying abreast ofadvancing technology. Americans like to cher-ish the notion that all small firms are like SiliconValley startups—technically and scientificallyadvanced, staffed and run by entrepreneurialinnovators—but the image is hardly typical ofsmall manufacturing fins. For many of Amer-ica’s 355,000 small and medium-sized manufactur-ing fins, exposure to new technologies ishaphazard, and the effort to keep informed isbeyond their means.

To contribute to the competitiveness of U.S.manufacturing, small firms need to keep up withtechnology as much as large ones. While smallenterprises are usually not heavily involved inforeign markets themselves, their performanceis important to the ability of larger manufactur-ers, who are their customers, to compete. Largeauto companies, for instance, depend on theability of their myriad suppliers, some of whichare quite small, to deliver the right components,well made, on time. As specifications becomemore exacting, and the tolerance for defectsdecreases, the demands for small firms to usenew technologies effectively grow. America’smost adept competitors, Japan and West Ger-many, have broad, deep institutions that support


technology diffusion and transfer to smallenterprises.

Large firms can transfer technology to smallercompanies quite effectively themselves. Even inJapan, however, an extensive network of govern-ment programs and institutions to support technol-ogy diffusion and training supplements theseprivate efforts. The United States, in contrast,has a few State programs and, until recently,very little at the Federal level. In 1988, thecombined technology transfer and technology/management assistance programs of the 30States that had such programs came to $58million, and that included assistance to all kindsof business, not just manufacturing. State indus-trial extension programs, giving one-on-onetechnical advice to individual firms, probablyadd up to about $25 to $40 million per year.

The Federal programs include: 1) three exist-ing and three more planned ManufacturingTechnology Centers to demonstrate advancedtechnologies and provide extension; 2) someassistance to State programs; and 3) the Ad-vanced Technology Program, a mechanism forFederal guidance and participation in joint R&Dventures with private firms. Together, the threeprograms have funding of less than $19 millionfor fiscal year 1990. A smattering of otherFederal programs also offer some technologyextension services; the largest of these is TradeAdjustment Assistance for firms and industries,funded at less than $10 million in fiscal year1990. These small, scattered programs contrastwith billions of dollars’ worth of financial andtechnical assistance to small and medium-sizedenterprises in Japan, plus Japanese Governmentparticipation in dozens of R&D efforts. Whileprecise comparisons of funding for technicalassistance to small manufacturing enterprisesare impossible, it is certain that Japan’s commit-ment to upgrading the level of technical abilityin small firms is more than an order ofmagnitude greater than that of the United States.(See chs. 6 and 7 for details of the Japanese andU.S. programs.)

If Congress wishes to deepen its commitmentto upgrading technology in small and medium-sized manufacturing enterprises, it could in-crease funding for the Manufacturing Technol-ogy Centers, provide more money for Stateindustrial extension services, or some of both.Manufacturing Technology Centers are man-aged by the National Institute for Standards andTechnology, as authorized by the OmnibusTrade and Competitiveness Act of 1988. Theyare responsible for transferring technologiesdeveloped at NIST to manufacturers, makingnew technologies usable by small firms, provid-ing technical and management information tosmall firms, demonstrating advanced produc-tion technologies, and making short-term loansof advanced manufacturing equipment to manu-facturing firms with fewer than 100 employees.Although funding was authorized at $20 millionper year, appropriations have been much smaller:$5 million in fiscal year 1988, $6.85 million in1989, and $7.5 million in 1990. These amountscover administration as well as technologyextension activities. The three existing Centershave each received $1.5 million per year fortheir first 2 years, and must match the Federalfunding. Federal funding starts to decline after3 years, and drops to zero after 6 years.

In addition to the Manufacturing TechnologyCenters, the 1988 trade act authorized a programof Federal assistance to State technology agen-cies, administered by NIST. This programreceived no funding until fiscal year 1990, whenCongress gave it $1.3 million to help States withindustrial extension programs expand thoseprograms. States receiving Federal money fromthis program must match it with their ownfunding.

Only a few States have real industrial exten-sion services. (NIST, in a nationwide study,found only 13 that met their definition of“technology extension services,” but morehave since been established.) Several of thoseare quite new. Nonetheless, State programs aregenerally better developed than Federal ones,and a very few have years of experience.

Chapter I-Summary ● 2 7

There is room for expansion of both State andFederal efforts in technology extension. Statesmay do abetter job of service delivery, being inbetter touch with the needs of local manufactur-ers. But there may be some things the FederalGovernment can provide that States cannot. Bytheir nature, industrial extension offices special-ize in the industries most prominent in theirservice delivery area. And industries tend to beregionally concentrated, spanning State lines;Federal services are often better suited to serveregional concentrations of industries. Also,while some State programs are excellent, othersare less so; a Federal service could help ensureconsistent quality of service, or at least mini-mum standards. If Congress wishes to considerexpanding efforts in support of industrial exten-sion, financial support for good State programs,as well as technical and financial support forStates which are new to the effort, would be aneffective combination with support of Federalextension services.

If Congress were to set a minimum goal ofextending industrial extension services to 24,000small firms per year nationwide (7 percent of thenation’s 355,000 small manufacturers), the totalcost would be $120 to $480 million, dependingon the level of service. If the Federal share offunding were 30 percent, as it is in the Agricul-tural Extension Service, the cost to the U.S.Government would be $36 to $144 million. Thatwould provide a modest level of service, onethat might easily be overwhelmed by requestsfor assistance. The State of Georgia’s experi-enced, effective industrial extension programserves a roughly similar proportion of itsmanufacturers, and Georgia Tech, which oper-ates the service, reports that it does not advertisebecause it would be swamped with requests itcould not meet. However, considering the inex-perience of State and Federal Governments inproviding industrial extension, moderate annualincreases may be all that could be handled now.

Financial Aid for Modernizing Manufacturing

Technical assistance to small business isoften most useful if it is accompanied by

21-700 0- 90 - 2

financial aid. Improving the general financialclimate for investment or offering special incen-tives to invest in research, development, andcapital equipment, will help all businesses. Butsmall businesses still have special problemsraising capital. They usually must pay more forboth debt and equity capital, and they often donot have enough retained earnings to financemodernization programs or training on theirown. Without help in financing, small firms maynot be able to implement the advice of industrialextension services.

In fiscal year 1989, the Federal Governmentmade 47 million dollars’ worth of direct loans tosmall businesses run by special groups (disabledveterans, the handicapped, and others), andguaranteed $3.6 billion in commercial loans tosmall businesses. It contributed $154 million toinvestment corporations, which make equityinvestments and long-term loans to small busi-nesses. These programs are not confined tomanufacturing. None is aimed at improving thepractice of manufacturing in general.

These programs are in striking contrast, bothin funding and in purpose, with Japan’s financialassistance to small and medium enterprises(SMEs). Japan’s SME programs spend $27billion annually indirect loans and an additional$56 billion in loan guarantees. Again, thisfunding is not confined to manufacturing (whichmakes it comparable to the figures given abovefor American programs). Much of the Japanesefunding is tied to technical assistance, and someis directly targeted to technology improvement.Part of the reason for such heavy support toSMEs in Japan is that for many years, smallbusiness was a technological backwater. Thesame is true in many sectors of American smallbusiness.

There are, of course, important differences inmanufacturing in Japan and the United States.One is that small firms play a bigger role inJapan’s manufacturing sector-74 percent ofmanufacturing employment is in small andmedium-size firms in Japan, compared to 35percent in the United States. However, because


of the larger size of the U.S. economy, thedifference is less in absolute terms. SmallJapanese manufacturing firms employ 10.7million people, compared to 6.8 million in theUnited States. In both countries, small manufactur-ing plants play key roles as suppliers to the largecorporations that are major actors in the worldeconomy. And in both countries, small manufactur-ing fins’ needs for technical and financialassistance have much in common.

Congress might consider several options toencourage firms to invest more in advancedtechnology and in training support required inthe service and to use the technology well. Oneoption is an equipment leasing system thatwould make new production equipment availa-ble to manufacturers on below-market terms. Ifthe system bought U.S.-made equipment couldserve two related purposes: besides enablingfirms to get advanced equipment on easierterms, it could also help assure U.S. makers ofproduction machinery a market for at least partof their output. In both ways, the program wouldhelp American manufacturers to focus more onlong-term investment and improvement. Theprogram could be open only to small manufac-turing business, or to all manufacturers, possiblywith more favorable terms for smaller firms.

Another option to encourage technologicalimprovement in small business is more directfinancial support. As noted above, the govern-ment’s financial support (loans, loan guaran-tees, and investments in development corpora-tions) was about $3.8 billion in 1989. Thiscompares with $487 billion in fixed investment(structures and producers’ durable equipment)by all private business in the same year. Whileexact comparisons with Japan are not possible,we do know that Japanese loan and loanguarantee programs to small firms area at least 20times greater than those of the United States, andthe level of subsidy in Japan is more substantial.For example, even a federally guaranteed loan to

a small business in the United States may be acouple of percentage points above the primerate, while in Japan, government-guaranteedloans to small business are typically substan-tially below market rates, and in some casesinterest-free. While Japanese policies clearly arenot a template for American action, they domake a difference in the competitiveness ofJapanese industry at all levels.

Greater financial aid for small manufacturerscould offer an opportunity to upgrade technol-ogy. One qualifying condition for financial aid(either direct loans or loan guarantees) could bethat the firm receive a technical assessment,possibly from an extension service, and that iteither follow the guidance of the assessor orwork out an alternative plan. This is notnecessarily intrusive. From the late 1970sthrough 1989, hundreds of small U.S. firmsinjured by import competition received techni-cal help from a small U.S. Government pro-gram, Trade Adjustment Assistance for firms.20

(The program, formerly funded at about $15 or$16 million per year, including assistance toindustries as well as firms, has been substan-tially reduced. Its fiscal year 1990 funding was$9.9 million in new and carryover funds.) Anassessment was a precondition for assistanceunder the program, and many participating firmsfound it a valuable service. Many small firms inJapan voluntarily undergo assessments eachyear in order to learn of new techniques andmarkets, and to get an independent (though notdetailed) assessment of the directions competi-tors are taking. This option presupposes anindustrial extension service that could delivercompetent, timely service nationwide.

Another possibility is to target financial aid toinvestments in advanced equipment, as Japanhas done several times. Recently, for example,producers were allowed to depreciate automatedelectronic ("mechatronic" equipment very rap-idly, encouraging many small and medium-

~or a description and analysis of the program, and the larger and better known program of Trade Adjustment Assistance for workers, see U.S.Congress, Office of Technology Assessment, Trude Adjustment Assistance: New Meusjbr an Ofd Progmm, OTA-ITE-346 (Sprin#leld, VA: NationalTechnical Information Service, 1987).


sized firms to acquire numerically controlledmachine tools. A drawback is that some firmsmight invest in equipment they aren’t preparedto use properly; but a technical assessment orindustrial extension service could help here.

Financial and technical assistance to smallfirms could be explicitly extended to coopera-tives of small firms as well. Managers of smallfins, with too few staff to dedicate even oneperson to keeping up with technology-or forthat matter, with competitors, customers, orsuppliers-often have to depend on a few ad hocsources for information about changes thataffect their business. Cooperative networks canhelp these managers in many ways, by poolingthe time and resources needed to keep up withtechnology, changing markets, customers’ needs,and competitors’ doing by obtaining quantitydiscounts on equipment that individual firmsbuy in ones or twos; and by providing anindependent source of information on newtechnologies that does not have its own commer-cial interests at stake, as vendors do.

If Congress wishes to support the formationof cooperative associations, it could considermaking the services of federally funded indus-trial extension services available to coopera-tives, or extending financial assistance to coop-eratives as well as firms. Congress might alsowant to make provision for small firms tocooperate in marketing and manufacturing with-out risking violation of the antitrust law.

Commercialization of Technology FromFederal Laboratories

The Federal Government spends $21 billioneach year on R&D in Federal laboratories, ofwhich three-fifths goes to defense applications.Some of the defense R&D could be useful tocivilian industry, along with some of the basicresearch done for nondefense applications inDepartment of Energy (DOE) laboratories. Forexample, industry has benefited from usingspecialized facilities at some DOE labs, such asthe Synchrotrons Light Source at BrookhavenNational Laboratory and the Combustion Re-search Facility of Sandia National Laboratories.

On the whole it has not been easy for industryto take advantage of the labs’ technology—despite legislation enacted throughout the 1980sto facilitate the process. There are still obstacleson the Government’s side, and further measuresby Congress could help, although ultimatelysuccess will depend on industry’s willingness totap into the labs.

Congress could consider earmarking some ofthe labs’ R&D appropriation for promotingcommercialization. This would include identifi-cation and marketing of promising technologies,patenting when appropriate, and cooperativeR&D projects to bridge the gap between thelaboratory and commercial exploitation. Ear-marking some funds for cooperative R&D couldbe particularly beneficial. (DOE’s high-temperature superconductivity pilot centers inthree national laboratories are examples ofcooperative R&D projects, planned from thestart with industry and funded 50-50 by industryand the labs.) Congress might begin by mandat-ing that a few percent of the labs’ budgets be setaside for cooperative projects as appropriate.This would encourage labs to seize opportuni-ties for cooperative work promptly.

Another possibility is increased funding forthe Federal Laboratory Consortium. The FLC,with a small central staff and volunteer represen-tatives from over 300 labs, tries to matchinquiries from firms with the appropriate labresearcher. Additional funding could help theFLC to perform this function, and also toincrease its projects demonstrating new meansof technology transfer.

In addition, Congress could consider meas-ures to remove several obstacles to technologytransfer and cooperative R&D. For example,DOE’s national labs have sometimes beenstalled by Agency red tape when they wish tolicense technology to firms. Congress has al-ready taken some steps to give the labs moreindependence in this regard and could go further(e.g., by extending to all labs the power to takeautomatic title to patents from lab research,removing the necessity to wait for extended


agency review). To make cooperative projectsmore attractive to industry. Congress could alsoclarify DOE’s right to keep information devel-oped in the projects proprietary, and allowcopyright of computer software created bygovernment employees involved in such proj-ects.

Lab-industry cooperation raises several otherissues, such as possible conflicts between labemployees’ duties at work and their desire to getconsulting work or royalties from the commerciali-zation of their work. Congress could considerforming an interagency legal task force whichcould give a broader perspective than a singleagency has on these and other legal issues raisedby lab-industry cooperation. The task force’sapproval would not be required; but an agency’sgeneral counsel, if concerned about an issue,could seek the task force’s advice.

Tapping Into Japanese Technology

American firms are often faulted for notmaking greater efforts to investigate and importtechnologies developed elsewhere—sometimes,even in another division of the same firm. WhenU.S. firms were technologically dominant inmost industries, this parochial attitude was nogreat handicap. Now, with technological advan-tage more evenly distributed around the globe,it is a significant hindrance. Many firms haveresponded to the challenge to keep up withtechnology developed abroad, but they facespecial difficulties getting access to Japanesetechnologies. One is simply the language. Euro-pean languages are enough like English, andenough Americans know some European lan-guage, that it is not too hard to get the gist oftechnical articles or to have them translated. Butthe Japanese language poses much more serioustranslation problems. Another difficulty is thatmuch of Japanese technology is developed in theindustrial sector and thus is inherently lessaccessible than technical expertise and knowl-edge freely available at universities and otherpublic or quasi-public institutions.

A sprinkling of U.S. programs promotetechnology transfer from Japan. A few universi-

ties have fellowship programs that send gradu-ates in science and engineering to Japanesecompanies and research institutions; and theNational Science Foundation and the Govern-ment of Japan sponsor several new programs tosupport long-term research by U.S. engineersand scientists in Japan. The NSF-Japan pro-grams were not fully subscribed as this reportwas written, although there is reason to believethat they will attract more applicants as theybecome better known. Congress may wish tomonitor the progress of government-supportedprograms, and provide additional funds whenand if they become overcrowded.

Other options are to establish a CongressionalU.S.-Japanese Fellowship Program, and to encour-age government researchers working in Federallabs or elsewhere in the Federal Government toundertake long-term projects in Japan. Post-doctoral or midcareer fellowships for profes-sionals other than scientists and engineers couldalso be useful, not in directly transferringtechnology, but in helping more people tounderstand the workings of Japanese manage-ment and government-industry relations.

Congress might wish to consider increasingthe funding for the Office of Japanese TechnicalLiterature. While demand for the office’s prod-ucts has been disappointing, expanding theservices available could create more interest.Finally, the government could promote Japa-nese language instruction in public schools,possibly by examining the critical foreign lan-guages program in the 1988 education act to seeif it gives sufficient weight to Japanese. Anotheroption is to fund expansion of Japanese lan-guage programs in post-secondary and post-doctorate education, especially for scientists andengineers.

Antitrust

Antitrust law and enforcement have beenrelaxed in the past decade, but fear of runningafoul of antitrust statutes is still a potent force inindustry, because the law is complex and oftenvague, penalty for violation can be stiff, andprivate parties as well as the government can


bring suit under the laws. There is good reasonfor firm enforcement of antitrust law; for manyyears, it has served this country well in main-taining competition. However, some kinds ofcooperation among firms could help Americancompetitiveness, and some modest changes inantitrust law and enforcement could help pro-mote them.

Congress has already amended and clarifiedthe law to make some joint activities easier.Among other provisions, the National Coopera-tive Research Act of 1984 clarified that jointR&D (as defined in the Act) will be judgedunder the rule of reason if suit is brought. Thisrule ensures full consideration of the activity’spro-competitive effects. Congress might wish toconsider extending this provision to joint manu-facturing and standards-setting. The 1984 Actalso reduced private treble damages to singledamages for registered R&D projects. Congressmight wish to consider reducing treble damagesin other circumstances as well.

Advance certification for some kinds of jointactivities is another option. Firms could apply tothe Justice Department for a determination thata proposed project complies with antitrust law.Private parties could challenge that determina-tion in court but could not collect damages foractivity covered by it. Another possibility is toestablish safe harbor market shares, belowwhich firms would not be in violation. Finally,Congress could make findings that joint ven-tures or mergers between U.S. firms are some-times necessary to fend off foreign competition,and could instruct courts to evaluate suchactivity based on long-term effects.

Whether modifying the antitrust laws or theirenforcement would unleash a great deal ofcooperative work, and whether such changeswould substantially improve manufacturing competi-tiveness, is unknown. It is also unknown whetherchanges such as those suggested would havesubstantial negative effects from lessening thefear of antitrust suits-effects such as increasedhostile takeover activity or more price-fixing.Changes in antitrust law and enforcement

should be made cautiously, but they deserveserious consideration.

Innovation and Intellectual Property

Improvement of intellectual property protec-tion could well start at home. Within the UnitedStates, the greatest complaint is that patentenforcement is slow. Patent cases that go to trialtake, on average, more than 2 1/2 years before adecision. Congress could consider several waysto speed up enforcement of patent infringementstatutes. It might designate special judges forpatent cases, or increase judicial manpowerdevoted to hearing patent cases. In a way, thereis already extrajudicial manpower available; theInternational Trade Commission employs fouradministrative law judges to hear cases underSection 337 of the Tariff Act of 1930. Undersection 337, a U.S. firm whose patent isinfringed by imported goods can apply for anorder to stop the goods from entering thecountry. The procedures for hearing and settlingcases brought under Section 337 have beenfound to violate GATT, however, and theAdministration is considering how to amendSection 337 to satisfy GATT while keeping itsadvantages of a quick trial and enforcement atthe border.

Effective domestic intellectual property protec-tion is not sufficient, however. U.S. firms needadequate protection in foreign markets as well.To many innovative companies, the Japanesepatent system is a particular problem. It isslower than ours in issuing and enforcingpatents, and it strongly favors licensing ofpatents-something U.S. firms do not alwayswish to do. The Administration is pursuingnegotiations to fix these problems. Anotherproblem for American firms is that they don’tunderstand the Japanese system very well, andcan’t easily find out more. The language barrieradds to the difficulties. Congress might wish toestablish a program in the U.S. Patent andTrademark Office to collect and disseminateinformation about the Japanese system.

Differing patent systems throughout the worldpresent a general problem. Usually, a firm must


apply for a patent in each country in which itwants protection; this is expensive and time-consuming. One option is to harmonize interna-tional patent law and application procedures, atleast among nations that trade heavily in high-technology products. The United States hasbeen negotiating with Japan and the countries ofthe European Community to this end. Anyagreement will probably require substantialchanges in the U.S. patent system. While suchchanges--e.g., changing to a first-to-file systemrather than first-to-invent-will be controver-sial, Congress might give any such proposalserious consideration, since a harmonized (andeventually unified) system could take much ofthe time and expense out of obtaining interna-tional patent protection.

Strategic Technology Policy

With few exceptions, the U.S. Governmenthas been reluctant to adopt proactive policies tobuild competitiveness. For generations, mostAmerican academics and policymakers havebeen convinced that market mechanisms werebetter than government planners at identifyingpromising technologies. There are examples offailures of central planning that reinforce thesebeliefs, and for several decades, the economicperformance of the American economy alsojustified that faith.

There are reasons to challenge this ideologynow. First is the simple fact that many Americanindustries are having great trouble in worldcompetition, and some of the ablest interna-tional competitors assuredly do not have freermarkets or lighter government involvement insupporting industrial technologies than the UnitedStates does. The governments of many Euro-pean nations, Japan, Korea, and Taiwan, have allactively promoted manufacturing technologyacquisition, development, and diffusion; andwhile they have had their failures, they have hadmany outstanding successes. This is not proof,of course. Many other nations with less thanadmirable economic performance have alsosupported technology development and diffu-sion.

America’s own history provides examples ofsuccessful commercial industries building onabundant government support of technology.Some of this has been an indirect effect; theDepartment of Defense’s support of the earlydevelopment of semiconductors and computerspaved the way for substantial investments incommercial technologies by the private sector.But the United States has sometimes beenwilling to make exceptions to the tenet thatdirect government support should be limitmostly to basic research and national security.The development of a U.S. civilian aircraftindustry can be linked directly to government-supported research on airframe and propulsiontechnologies in the early part of the century.This support was justified on patriotic grounds,and was not drawn so narrowly as to includeonly military security needs. Government sup-port of agricultural technology through the landgrant universities and the Cooperative Exten-sion Service has been a key to the rapidproductivity growth of American agriculture inthe 20th century. Government support of thespace program from the 1950s onward rested asmuch on national pride as on defense needs, andhas had some important commercial payoffs.

Still, the argument most often put forward forFederal support of technology developmentremains rooted in national security. The Depart-ment of Defense depends on the civilian microe-lectronics and other high technology industriesof its procurement needs. This was a key factorin the consideration of whether and how muchto support Sematech, high-temperature super-conductivity, and lately, high-definition televi-sion. But the idea that only the direct, immediateneeds of the military justify government supportof technology development is wearing thin. Thetime is ripe for reopening the question of howthe Federal government could support develop-ment of civilian industrial technology proactively—i.e., before the industry is so weakened thatnational security is threatened.

Many people still reject this strategy. Theyargue that selective government support of keytechnologies or industrial sectors amounts to

Chapter 1--Summary ● 3 3

“picking winners” and that government bu-reaucrats are ill-equipped to make these choices.This argument rests mostly on politics. TheAmerican political system is too subject tomanipulation by special interests, it is argued, tomake rational choices among all the potentialindustries and technologies that might meritgovernment support. This is a forceful argu-ment, one to be taken seriously. Another pillarof the argument is the simple claim that themarket, with its imperfections, is better thangovernment interference.

The other side has a powerful justification aswell. That is, that some technologies are so riskyor involve such large investments over the longterm that little or no development will beundertaken unless society, which stands tobenefit, shares the risk of development. In theU.S. financial environment, with its burden-some penalties on long-term investment, theargument takes on special force.

The debate over “picking winners” hasresolved little. Those who argue that govern-ment cannot make consistently rational choicescan point to failures, such as the money pouredinto the Synfuels Corporation in the early 1980sto make wood-based, coal-based, and shale-based substitutes for petroleum. Japanese poli-cies have not been invariably successful either.Examples of projects that did not achieve theirinitial objectives include efforts to jump-startbiotechnology development, the fifth-genera-tion computer project, and entry in the civilianair transport industry.

There have been some notable successes aswell. U.S. Government support for aircrafttechnology development, through both civilianand military agencies, and agriculture are exam-ples. These industries, which have had muchgreater government support than most, areadvanced technologically and successful inter-nationally. Both can boast large trade surpluses.Successes in Japan encompass the major indus-tries on which that nation’s astounding postwareconomic achievements rest—first, steel, chem-icals, and shipbuilding; then automobiles; and

now microelectronics, computers, and telecom-munications.

More to the point, the argument cannot (andshould not) be resolved by counting up suc-cesses and failures. Any sustained effort tosupport new technology development will in-clude some failures, and some industries mightsucceed more in spite of government supportthan because of it. The fact is, the U.S.Government is increasingly being asked tosupport technology development, and it isbecoming ever more obvious that the reason isto build civilian industrial competitiveness. It ispossible to take the best from the “pickingwinners’ debate by focusing on how to designinstitutions that are open to counsel from andcollaboration with industry and other interests,but avoid becoming their captives. Anotherlesson is that a crisis is a poor crucible formaking such decisions. The failure of theSynfuels project can be traced largely to theatmosphere of crisis in which it was born.

A Civilian Technology Agency

Efforts to support industrial technology willrequire commitment and money. Both have theirlimits. Public initiatives to help private manufactur-ing improve its performance cannot afford toplunge into repairing and developing everyindustry and technology. Yet the Federal Gov-ernment has no institutional ability to dis-criminate between technologies and industriesthat are most promising for the Nation’s eco-nomic future, and those that have some appealbut are less important. While the U.S. Govern-ment has acted to support certain key tech-nologies, the responses to declining competi-tiveness have been ad hoc, and are usuallyjustified by the seriousness of potential losses inmilitary security. If Congress wishes to considerways of responding to pleas for support oftechnology toward the goal of economic secu-rity, one option is to create a civilian technologyagency.

One approach is to build on existing institu-tions. NIST’s Advanced Technology Program,created in the 1988 trade act and funded for the


first time in fiscal year 1990, at $10 million hasthe potential to develop into a CTA. A bill thatpassed the Senate in 1989 authorized $100million for the program to support industry-ledjoint R&D in economically critical technolo-gies. Five such technologies were spelled out inthe Act.

Other bills in both the 100th and the 101stCongress proposed the creation of a CivilianTechnology Agency (CTA) within a new De-partment of Industry and Technology taking theplace of the Department of Commerce. Theagency would make grants or cooperative agree-ments with private performers of R&D onhigh-risk projects that could have exceptionalvalue to the civilian economy. The closestanalogy among existing agencies to a CTA is theDefense Advanced Research Projects Agency,or DARPA, which supports development intechnologies and industries considered criticalto the nation’s defense. This small agency (staffof 150, funding of nearly $2 billion per year) hasgained a reputation for placing intelligent bets inserving U.S. military technology needs. It makeslong-term commitments that have added up todecades for some of its projects. DARPA has attimes interpreted its mission broadly, support-ing technology development that will benefit thecommercial sector because the military dependson that sector. A CTA could learn a good dealfrom DARPA’s experience on how to evaluatethe potential benefits and risks of investments innew technology, and how to balance the pres-sures of industrial and parochial interests inmaking such decisions. The CTA might besubject to greater special-interest pressures, butthe difference is likely to be one of degree ratherthan kind.

In some ways, a CTA would be quite differentfrom DARPA. Most important, a CTA wouldinteract closely with industry in choosing whattechnologies to support and designing the R&Dprojects. Until recently, DARPA did not fundprojects jointly with industry; a CTA wouldprobably finance most of its projects withcontributions from industry that are at leastequal to if not greater than the government

share. This joint funding is essential as assur-ance that industry is genuinely committed thatand the projects are really promising commer-cially, in the opinion of industry. Thus, theproblem of government’s “picking winners”would be greatly diminished.

Where in the Federal bureaucracy the CTA isplaced may not matter very much. There aresome advantages to its being an independentagency like the National Science Foundation.With the right mandate, independent agencies,even small ones, can wield influence beyondwhat their size would indicate. (NSF is fundedat less than $2 billion per year.) However,DARPA demonstrates that a small agencywithin an enormous bureaucracy can be effec-tive and powerful. With the right design, suffi-cient funds, top-notch staff, and a strong man-date from Congress, a CTA could probablyfunction well either within the Department ofCommerce (or a successor department) or inde-pendently.

Other issues are more important to a CTA’sperformance. Judging by the difference betweenDARPA’s performance and the record of otherDoD technology development and acquisition,it is clear that the agency should not beconstrained by detailed rules and procedures.Giving the agency staff a large degree offreedom and responsibility could help to attractand keep technically first-rate people, which isincreasingly difficult as salaries for scientistsand engineers rise faster than government sala-ries.

One of CTA’s first tasks would be to developguidelines for the selection of industries ortechnologies to consider for support. Here,much can be learned from the debates overwhether to support specific technologies orprojects like HDTV, superconductivity, andSematech. There is an obvious preference forindustries that are high-tech, provide well-paidjobs, and have high growth potential. In addi-tion, CTA would need to consider entire techno-logical systems, not just particular technologies.For example, if it chose some semiconductor


technologies, it would have to be sensitive toR&D needs throughout the system, starting withimproved materials, and continuing throughthings like lithography for etching chips, auto-mated techniques for packaging, and soon. CTAcould also look for technologies important tomore than one application or industry down-stream.

One of the surest ways to doom the effortwould be to subject a CTA to unrealisticexpectations. If CTA is expected to makestrategic choices of high-risk technologies, itwould have to be given time for its investmentsto play out, and some leeway to make less thanperfect choices. The ability to make multi-year

funding could also be critical. As it is, Americanbusiness regards government support as volatileand undependable. The fact that Silicon Valleycompanies took very seriously recent rumorsthat the Administration proposed to abandonfunding for Sematech illustrates the point. If theagency is to succeed at pushing technology, itwould need to provide steady support for severalyears to many different technologies. Even then,it should not be expected to turn Americanindustrial competitiveness around singlehandedly.Coordinated support in other policy areas liketrade and macroeconomic policy will be neededto do that.

Chapter 2

Strategies To ImproveUS. Manufacturing Technology:

Policy Issues and Options

CONTENTSPage

FINANCING LONG-TERM INVESTMENT.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Capital Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Relationships With Providers of Capital . . . . . . . . . . . . . . . . . . . . . . . . . . . +... . . . . . . . . . . .Environmental Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

HUMAN RESOURCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Training the Active Work Force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . .Supply of Engineers: Keeping the Pipeline Filled . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Manufacturing Education and Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

DIFFUSING MANUFACTURING TECHNOLOGY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Technology Extension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Financial Aid for Modernizing Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Cooperative Networks of Small Manufacturing Firms . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Commercialization of Technology From Federal Laboratories . . . . . . . . . . . . . . . . . . . . . .University-Industry Collaborations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Tapping Into Japanese Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Antitrust Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Innovation and Intellectual Property . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

STRATEGIC TECHNOLOGY POLICY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Picking Winners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Creating a Civilian Technology Agency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Designing a Civilian Technology Agency . . . . . . . . . +... . . . . . . . . . . . . . . . . . . . . . . . . . . . .Defining Goals, Choosing Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Government-Industry Collaboration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Beyond Technology Policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .One Example of Technology Policy: The Case of Advanced Television . . . . . . . . . . . . .

424649505152525353576161646566697171737677787980

Box2-A.2-B.2-c.

Page

Government Backing for the Civilian Aircraft Industry . . . . . . . . . . . . . . . . . . . . . . . . . . 74Digital and Analog Data: Television Transmission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81Technology Spillovers From Consumer Electronics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

Chapter 2

Strategies To Improve U.S. Manufacturing Technology:Policy Issues and Options

Only 5 years ago, the idea that American manu-facturing was in trouble was not widely accepted.Many people-including manufacturers themselves—blamed the spiraling trade deficits on nothing morethan the overvalued dollar and unfair trading prac-tices by other nations. As the 1990s begin, a sobererview has taken over. Yes, the high dollar didinterfere with U.S. exports. Five years after thedollar started down, the merchandise trade deficithad dropped one-third from its peak and exportswere at anew high. But the deficit was still runningat over $100 billion a year and 2.1 percent of GNPin 1989, and that is still very high by historicalstandards; moreover, the dollar had started climbingagain. And yes, many of our trading partners,including some of the richest, discriminate subtly oropenly against imports. Nevertheless, the Japaneseare beating us in our own home market in things likeautos and semiconductors, where not so long ago wewere the world’s best; the Koreans and Taiwanesehave become adept competitors in some kinds ofsemiconductors and computers; and the Europeansuccess with the Airbus threatens our top remainingexport industry.

Today, there is much greater agreement that U.S.manufacturing has to improve to keep up with thecompetition, and that technology is key to theimprovement. It is not so clear that, as a nation, weare ready to make the commitment-or the sacrifices—that are required to reinvigorate U.S. manufacturing.Much of the effort has to come from inside industry,with better management and better relations be-tween managers and workers. But some involves allof us, as savers and consumers, teachers and studentsand families of students, taxpayers and citizens.

For example, a major reason for the notoriousshortsightedness of American industry is high inter-est rates. The high cost of capital discouragesinvestment in new plant and equipment, and has aneven more dampening effect on research and devel-opment, with its more distant and uncertain payoff.

The massive U.S. Government budget deficits of the1980s, combined with low personal savings rates,are prime reasons for high interest rates. So far, thereis little sign that either political leaders or voters areready to make the disagreeable choices—highertaxes or cuts in popular government programs orboth-that would make a real dent in the budgetdeficit.

Despite the decline in real wages and stagnationin family income over the past decade, Americansare still the richest people in the world; only Canadarivals the United States in income per capita.l Wegot a free ride in rising consumption throughout the1980s because foreign investors remained willing tofinance our budget and merchandise trade deficits.And rising consumption led a record peacetimeexpansion of the economy. It is a real questionwhether such a nation-still comfortable, not reallyhurting--can summon the energies needed to regaintechnological leadership in an increasingly competi-tive world.

Traditionally, U.S. Government policy on tech-nology for manufacturing has been to support basicresearch, allowing private companies to help them-selves to whatever items of commercial interestcome out of that research. Federal R&D aimed atapplications has mostly been limited to defense andspace (areas in which the government itself iscustomer), health, energy (mainly nuclear), andagriculture. On occasion in the past, Department ofDefense spending for both R&D and procurementhas given commercial industries a vital boost, insuch things as semiconductors, computers, andaircraft. But these spinoffs are less common thanthey used to be. Military systems have become moreesoteric and more are secret; differing businesspractices in the military and civilian sectors erectbarriers to the transfer of technology; the processesfor manufacturing a few copies of a custom item (forthe military) has little in common with high-volumelow-cost manufacture (for commercial markets);

IFi~ 1.3 ShOWS grOSS domestic product per capita for the United States and other countries. in the United States, from 1977 tO 1988, tie avera~family income was virtually unchanged in constant dollars. However, there were marked changes in distribution; in every income decile up through theeighth, family income declined over the 12 years, and the lower the income the greater the decline. Only in the top decile was there a significant increasein family income (16 percent), and the top 1 percent racked up an increase of 49 percent. (U.S. Congress, Congressional Budget Office, The ChangingDistributwn of Federal Taxes: 1975-1990 (Washington, DC: Congressional Budget Office, 1987), p. 39.

-39-


and in many high-technology areas, the defensesector is lagging behind the commercial.2 Today, ifthe government wants to support industry in com-mercializing new technologies, it must usually do somore directly.

Some changes are occurring. Recently, the U.S.Government has shown increased interest in positiveactions to help American industry restore its techno-logical edge. Part of the reason is defense-related.Loss of competitiveness in the commercial sector(especially in semiconductors) is a worry to thedefense establishment, because it means that weap-ons systems may either have to rely on foreignsuppliers or else take second best.3 More broadly,the idea that economic performance is at least asimportant to the United States as military securityhas gained some ground.

Congress has taken several initiatives to offermore government support for improving U.S. manu-facturing performance. Legislation passed in the1980s promotes technology transfer from the Fed-eral laboratories. In the 1988 trade act, Congresscreated regional centers to transfer advanced manu-facturing technology to industry. It has appropriatedspecial funds to advance R&D in high-temperaturesuperconductivity. In a distinct departure fromtraditional U.S. policy, it is providing $100 milliona year for 5 years to Sematech, the government-industry consortium for R&D in semiconductormanufacturing technology. New ideas for a moreaggressive, commercially oriented technology pol-icy are getting an attentive hearing in Congress.

Real change in this direction is by no meanscertain, however. According to press reports in late1989, the Administration was ready to rein in anyDoD support for technologies that are not strictlymilitary, and continued funding for Sematech was inquestion. 4 Responding to protests from Congress,the Administration denied the reports, but also tookpains to announce opposition to any increasedfunding for Sematech or similar ventures.

The government programs actually undertaken sofar to improve technology in manufacturing havebeen modest, and spending for them is low (Sema-tech’s $100 million a year is by far the mostexpensive of the new initiatives). The costs couldrise considerably if the government sticks with theprograms already started, and possibly enlargesthem as it gains experience. Still more costly wouldbe real efforts to change some of the basic factorsaffecting U.S. competitiveness-getting the cost ofcapital down, doing afar better job of educating andtraining the work force.

As the arms race with the Soviet Union dwindles,the prospects are good for reducing military spend-ing. Some of those savings could go for deficitreduction, or for measures to improve education andtraining, or for technology advance and diffusion.Some might go for other social purposes. On theother hand, the savings might be spent on furtherlowering of taxes and the resulting increase inprivate consumption. These are public policy choices.National leaders, guided by the voters, will ulti-mately make them.

Quite a few government actions are worth consider-ing as ways to promote a stronger technological basefor American manufacturing, some of them tradi-tional and others with little precedent in this country.These actions can be directed toward four somewhatoverlapping strategic targets.

. Financial Policies. These shape the financialenvironment for industry, including the cost ofcapital. The broadest of these policies--ontaxes, spending, and the Federal budget—affectthe whole sweep of the economy and aresubjects of intense national debate; they arediscussed in general terms in this report. Closerattention is given to specific policies that mighthelp firms take a longer term view thanquarterly profit performance and invest moreheavily in technology development and up-to-date production equipment.

W.S. ~n%ss, office of Technology Assessment, Commercializing High-Temperature Superconductivity, OTA-~-388 (Sprinsleld, VA:National Technical Information Service, 1988), pp. %-98; and Holding the Edge: Maintaining the Defense Technology Base, OTA-ISCA20(Washington, DC: U,S, Government Printing Office, 1989), passim and esp. pp. 174-178.

3u.s. &pwmentof Defense, ‘Bolstering Defense Industrial Competitiveness,” report to the Secretary of Defense by the Under Secretary of Defensefor Acquisition, July 1988; Defense Science Board, ‘‘Report of the Defense Science Board Task Force on Defense Semiconductor Dependency, ’ reportto the office of the Under Secretary of Defense for Acquisition, February 1987; U.S. Congress, Office of Technology Assessment, Hofding the Edge,op. cit.

4New TeC&~gy Week, NOV. 6, 1989.

Chapter 2-Strategies To Improve U.S. Manufacturing Technology: Policy Issues and Options ● 41

●

s

●

Human Resource Policies. These affect theavailability of well-qualified people to fillmanufacturing jobs, and thus have a powerfulinfluence on competitiveness. Education of theNation’s children and the reeducation andtraining of adult workers are subjects too broadto fit completely within a report on manufactur-ing technology, but some policies with specialrelevance to manufacturing performance areselected for consideration.Technology Diffusion Policies. These arepositive, deliberate government actions to helpfirms improve their manufacturing processesand commercialize new or improved products.They support technological advance across theboard for all manufacturers, with no distinctionin kind (i.e., no special support for particulartechnologies or industries). Congress has al-ready taken a few steps in this direction. Furtheroptions include such measures as easier accessto new technologies coming out of Federal labsand stepped-up Federal support for technologyextension services to manufacturers.Strategic Technology Policy. This includes acoherent set of actions that would promotegeneral technology advance and also targetsupport to technologies that are seen as vital toeconomic growth. The U.S. Government hasused some of the tools of strategic technologypolicy in the past, sometimes quite success-fully, but usually in an ad hoc way. Currentexamples include the Federal funding for Se-matech and the special collaborations on high-temperature superconductivity R&D in threenational labs. If a consensus develops in favorof forming a coherent strategic technologypolicy, an agency or institution would need tobe in charge, to define goals and choosetechnologies for government backing that fitthe goals.

This list does not by any means exhaust thepossibilities for government actions to bolster thecompetitiveness of manufacturing. Many nationshave used broader instruments of policy than theseto promote industries they consider essential to theircountries’ well-being. Japan, other East Asian na-tions and, increasingly, the European Communityhave used a full range of technology, industry andtrade policies in support of the strategic industriesthey wish to develop. Policy tools include suchthings as preferential low-cost loans, government-

guaranteed purchases, and trade protection againstpowerful foreign competitors during the infancy anddevelopment of native industries.

Improving the financial environment and upgrad-ing education and training are the fundamentals forany set of policies to improve technology. It may be,however, that the addition of policies to step uptechnology diffusion and target government R&Dsupport to critical commercial technologies will notgo far enough to boost American manufacturing toworld class competitive level, when other nationsare doing much more. Whether the United States canor should employ more comprehensive policies tobolster competitiveness is an open question, onlytouched on in this report. Industry and trade policiesof the Asian rim nations and the European commu-nity, and their possible relevance to U.S. policy, willbe considered in the final report of OTA’s assess-ment of Technology, Innovation, and U.S. Trade.

FINANCING LONG-TERMINVESTMENT

American business managers have been lesswilling than their Japanese and German competitorsto make investments in technology development orequipment that requires many years to begin yield-ing a return. Paying attention to the bottom line inthe short term is obviously important, but too muchof it can be costly in a world where manufacturers inother developed nations pay less attention to short-term profit and more to long-term growth and marketshare.

American shortsightedness will be hard to over-come. If it were mostly due to culture—the waymanagers and decisionmakers are socialized andtaught to think about problems—some good mightbe accomplished by progressive business schoolsrevamping their curricula. Also, if the problem weremerely cultural, experience would prove that man-agers who concentrate on long-term gain outperformthose who do not, and the problem would beself-correcting. But the myopia is long-standing;experience has not remedied it. And our bestbusiness schools have led-not resisted—the effortto analyze and propose solutions to the shortsighted-ness of American management. Undoubtedly, somecultural changes are needed, but without changes inthe underlying financial environment, simply en-lightening managers on the potential gains of longerterm vision probably will have little effect.

42 Making Things netter: Competing in Manufacturing

The underlying financial environment that makesour undue emphasis on short-term profit a consis-tently rational choice for American managers con-sists of many parts. The most straightforward is thecost of capital: even in the absence of other factors,the fact that American manufacturers have facedconsistently higher capital costs than their Japaneseand West German competitors will shorten therequired payback period for American investments.5

Another factor is the relationship between providersof capital and companies. Providers of both debt andequity capital have pushed American corporations topay more attention to short term gains than to longterm market share. Japanese and West Germanbanks and other creditors and equityholders havemore incentives to focus on long-term growth ratherthan short-term payout. American managers, partic-ularly those most responsible for strategic decisions,may also be encouraged personally to focus onshort-term profit. According to the MIT Commis-sion on Industrial Productivity, there is ‘no shortageof executive bonuses geared to yearly or evensemiannual performance."6 Finally, the uncertaintyof the business environment could also lead manag-ers to be cautious about long-term investments.Analysts point to uncertainties such as moneyexchange rates, regulatory and fiscal policies, andtrade. 7 These factors all play some role in howmanagers view long-term investment. Making sig-nificant changes in any of them will be difficult,even where they are sensitive to Federal policyintervention.

Capital Costs

Our ablest international competitors have madearrangements to provide capital to industry on morefavorable terms than the market provides. So,however, have a number of Third World countriesthat are regarded as prime examples of the pitfalls ofbungled, state-led planning. Even in cases wherechanneling of capital has rather clearly promotedindustrial development—Japan is most often cited—the policy involved a heavy price to consumers. If

Congress wishes to overcome the disadvantage ofour capital costs, it should be known at the outsetthat this cannot be done without sacrifice.

There are two basic approaches to the problem.One is to make capital more available to everyone;the other is to use selective policies to reduce itscosts for certain sectors or activities. The firstapproach is to increase the pool of savings fromwhich capital is formed. This includes increasinggovernment saving, which means reducing theFederal budget deficit in one way or another (raisingrevenues or cutting spending). The second approachinvolves the use of tax instruments to reduce the costof capital investment, R&D, and other productivity-enhancing activities.

The obvious step is to reduce or eliminate theFederal budget deficit. The tax cuts and increasedgovernment spending of the 1980s were an enormousfiscal stimulus to the American economy. To avoidexcessive inflation, the Federal Reserve has pursueda very tight monetary policy. This, in turn, keepsupward pressure on interest rates, which does helpcontrol inflation but also gradually robs industries ofthe capital they need to improve real wages andproductivity. 8 The alternative—a less restrictivemonetary policy—would not drive interest rates upso high in the short run, but the resulting inflationcould result in disaster too. With high inflation andlower interest rates, foreign investors who are nowfinancing a large share of American investmentcould find investments here less attractive and mighteven lose confidence in the soundness of invest-ments in the United States, which could result in asevere recession. Charles Schultze characterizes thisscenario-which he thinks unlikely to happen—as‘‘the wolf at the door. Most experts agree that it isimpossible to eliminate the budget deficit rapidly,that is, in a couple of years, but that some combina-tion of higher revenue and lower spending over adecade or so will be needed. Lest we forget, it wasa combination of lower taxes (revenue) and greater

5Robefl N. McCauley and Steven A Zimmer, “Explaining International Differences in the Cost of Capital,” Fe&ral Reserve Bank of New YorkQuarterly Review, summer 1989, pp. 7-28.

b~chael L. WrtOUZOS, Richard K, Lester, and Robert M. Solow, Made in America: Regainhg the Productive Edge (Cambridge, MA: The ~ ~ess,1989), p. 62.

TIbid., p, 61.s~is ~~ent is set out in Charles L. Schultze, “Of Wolves, Termites and Pussycats: Or, Why We Should Worry About the Budget Deficit,” The

Brookings Review, summer 1989, pp. 26-33.gIbid., p. 26.


Federal spending in the 1980s that caused theFederal budget deficit to balloon after 1981.

Encouraging Savings

Saving is the source of capital. At any given levelof demand for capital, if domestic savings rates fall,capital formation must fall unless foreign sourcesmake up the difference. Some dependence onforeign capital is probably acceptable for any nation,but excessive reliance on it is worrisome. Americansavings rates have fallen in the 1980s, partly becausethe budget deficit comprises a large chunk ofdissaving, but also partly because household andbusiness savings rates have dropped.

Of the two, the drop in household savings is muchgreater,. Household savings averaged nearly 8 per-cent of GNP over the 1970s, and dropped to 2.1percent by the mid-1980s, partially recoveringthereafter, to about 5 percent by the end of thedecade. To raise the household savings rate, Con-gress could consider incentives to save, such aspreferential tax treatment of interest income ordeferred taxation on income that is saved. The latterhas been tried in the form of deferred tax on moneyplaced in Individual Retirement Accounts (IRAs),with disappointing results. The fact that the house-hold savings rate fell while IRA tax incentives werein place led many economists to conclude thatsavings incentives by themselves are ineffective,and that discouraging consumption must be a part ofany package to increase private savings rates. Thisis not a universally accepted conclusion, however.For example, Hatsopoulos, Krugman and Poterbaargue that a national savings initiative that wouldencourage savings in all tax brackets and rewardregular savings rather than portfolio reshufflingcould be effective.10 Congress could consider anational savings initiative, based on these principlesand accompanied by a public campaign to encouragesavings, as was done in Japan after World War II.

Before the war, Japanese household savings rateswere lower than American rates.

Clearly, Japan’s savings rates were a response tomuch more than just a national savings initiative,and it may well be that even heavy incentives and apublic campaign are not enough to raise savings tothe levels needed to sustain competitiveness (i.e.,above the rates of the 1970s and 1960s). Americanshave been encouraged in various ways to consume,and consumption reached all-time highs as a percentof GNP in the 1980s. Congress may also wish toconsider some measures to discourage consump-tion. ll The classic device is the consumption tax,which has been rejected before because of itsregressivity.

12 However, by scaling consumptiontaxes to tax most lightly (or not at all) those itemsregarded as necessities and most heavily thoseconsidered luxuries, several European countrieshave shown that consumption taxes are not necessar-ily overly regressive.

Another option might be additional limitations onconsumers’ ability to deduct mortgage interestpayments. Mortgage interest payments are 100percent deductible up to the generous limit of twohomes (primary and secondary). Because of thisdeductibility, and because Americans are allowed tomake relatively low downpayments, Americansconsume more housing and save less than people inJapan, Germany, and many other advanced nations.While home equity is a form of savings for ahousehold, the money tied up in housing is not thesame as savings accounts and other forms of savingsfrom society’s point of view, because of its illiquid-ity. It is again to the household, but not available forother investments. Moreover, the buildup of homeequity may substitute for other kinds of savings formany households. Limiting mortgage interest de-ductibility to one home could also help to raisehousehold savings rates. In fact, the current limitsmay be doing so. While the deductions allowed onmortgage interest payments are still substantial, they

l~mrge N. Ha~~~os, Paul R. mWm, and J~es M. Poterba, Overconsurqption: The challenge to U.S. hmmaic polio (w*@JKxIs Dc:American Business Conference and Thermo Electron Corp., 1989), p. 14.

llsme steps have ~n ~en. ~ he 1986 t= act, con~= ~gan top- out the d~~tibili~ of interest on m~y tyWs of cxmsunwr cl@t~1990, 10 pereent of eonsurner interest paid can be dedueted, and none after that-and placed additional (though not very restrictive) limits on mortgageinterest deductibility. These had little effect on the propensity to consume, however, because consumers can still deduct substantial amounts of intereston home equity loans, which have substituted to some extent for other typ of cxmumercredit, Indeed, as consumer interest deductibility has diminished,the value of home quity lines of credit has mushroomed. In 1986, the year of the Tax Reform Act, home quity loans totaled $35 billion; by 1989, thetotal was $100 billion, Source: David Olson, SMR Researeh, personal communication, January 1989.

12pm exmple, tie late Jo~ph pa~m, a Prominent tax ex~rt at The Brookings Institution, maintained that c0n5UmPti0n trees would favor thewealthy, and argued for a more progressive income tax. See Hobart Rowe~ ‘‘Joseph Pechman’s Simple Solution for Fairer Taxes, ’ The WusM”ngtonpost, Dec. 31,1989.


are more limited than they were. This may be onecause of the partial recovery of personal saving fromits nadir of 1987, and the slowdown in the rate ofgrowth of housing prices.

Selective Lowering of Capital Costs

Progress in budget-balancing and stimulatingsaving would result in moderation of interest ratesand encourage more longer term investment. Thatmay not be enough to support the kind of changeneeded to reverse the relative slide of Americanmanufacturing technology and productivity. Afterall, before the run-up of interest rates brought on bythe burgeoning deficit, American manufacturingproductivity was still advancing at a slow pace,compared with its earlier performance and comparedwith Japan. Congress might wish to consider othermeasures to lower the cost of investment for specificsectors or purposes.

The United States has tried using tax instrumentsto stimulate additional investment in technologydevelopment and application in the private sector.These measures include accelerated depreciationallowances and tax credits or deductions for thepurchase of equipment and facilities, and researchand development tax credits. Different policiesaffect different activities in the spectrum of technol-ogy development, implementation, and diffusion.The rationale behind all of these measures is that themarket does not provide strong enough incentives toinvest in the supported activities, considering thetotal of private and social benefits that stem frominvestments in plant and equipment or research anddevelopment.

Right now, the case for underinvestment inequipment-particularly advanced equipment toproduce state-of-the-art products--probably is strongerthan arguments that we have underinvested in R&D.However, both Japan and West Germany spend ahigher percentage of their GNP on civilian R&D,13

and the European Community is topping that offwith about $1.5 billion a year on R&D through theFramework program.14 Most of the R&D performed

by these key competitors is dedicated to improvingcivilian science and technology. The United Statesspends more money on R&D, but much of it isgeared towards military technologies. About half thetotal R&D spending in the United States is funded bythe Federal Government, and 70 percent of that byDoD. In contrast, less than 5 percent of Japan’sgovernment R&D is spent on defense, and about 12percent of West Germany’s.15 While lagging R&Dspending has not been a major competitive problemfor American industry in the past, it is becoming one.

Capital investment is probably a greater problemat the moment. Particularly in high-technologyindustries, capital equipment investment is a keypart of technical competitiveness, and America’shigh capital costs have damped investment. IfCongress wishes to provide incentives to stimulatethe development, commercialization, and imple-mentation of new technology, it might considerreauthorizing some form of rapid depreciation orinvestment tax credit, both of which were eliminatedin the 1986 Tax Reform Act.

Both accelerated depreciation and investment taxcredits (ITC) can be aimed at encouraging busi-nesses to acquire new capital equipment. Investmenttax credits have been applied, on and off, since theearly 1960s, most recently in the Economic Recov-ery Tax Act of 1981. This tax credit was eliminatedin 1986, in favor of an overall reduction in thecorporate tax rate. The Accelerated Cost RecoverySystem (ACRS) was also eliminated (althoughcertain classes of assets still enjoy fast depreciation).

Some argue that the ITC and ACRS were inappro-priate in the frost place—that because a firm can reapall (or nearly all) the benefit of investing in newcapital equipment, it is inappropriate for society tosubsidize such purchases. It is also unclear howeffective the subsidy was, at least in the 1980s, atstimulating capital investment. according to variousestimates, for every dollar of revenue the Treasuryforegoes as a result of the investment tax credit,

13BY 1$)8$ Jwan’5 to~ R&D spending was slightly above U.S. total R&D spending by 0.1 pefeent of G~, accor~g to a PmliminW fi- ‘mthe National ScieneeFoundation. See National Science Fomtitim,fnter~tiow/ ScUwe and Technology Duru Updute 1987, NSF 87-319 (Washington,DC: 1987).

14~e EC mn~bu~ about 5 percent of all government-funded R&D in the countries of the Eurwan Community.

15NSF, op. cit, p. 9.

Chapter 2--Strategies To Improve U.S. Manufacturing Technology: Policy Issues and Options ● 45

industry invests $0.12 to $0.80 in equipment, abovewhat would have been invested without the credit. l6

Despite the apparently modest results of the ITC,there are arguments in favor of tax stimuli forinvestment. Investment in durable equipment wasrobust in the recovery from the 1982 recession, eventhough real interest rates were high, so without theITC investment might have been smaller than it was.If the intent of the ITC was to stimulate equipmentpurchases to raise productivity, that, too, could beclaimed as a modest success. Productivity growth inAmerican manufacturing averaged 3.5 percent annu-ally from 1979 to 1986, a substantial pickup from its1.4 percent average annual increase in 1973-79, andeven higher than the 3.2 percent annual average of1960-73, the heyday of American manufacturing. Towhat extent this is causally related to investmentincentives in the 1980s is not known. For example,some of the productivity growth of the period camefrom the closure of inefficient plants, rather thanfrom new investments in plant and equipment.However, the coincidence of high productivitygrowth and investment stimulation is worth exami-nation.

The effect of investment tax incentives on produc-tivity improvement and the diffusion of best practicein American manufacturing will require additionalanalysis. Congress may wish to initiate such a studyin one of its analytical agencies, or by a panel ofexperts. This is a topic of great importance, butconsiderable uncertainty. Some analysis suggeststhat investment tax incentives are inefficient, andthey are certainly expensive. Between 1979 and1987, the ITC cost between $13 billion and $37billion each year in tax expenditure; ACRS’ costvaried from $8 billion in 1982 to $64 billion at itspeak in 1987.17 Unless Congress can find anotherway to raise revenue, or effect other substantialspending cuts, reinstating investment tax incentiveswill only worsen the deficit and increase the pressure

to keep interest rates up. Yet in view of the pressingimportance of raising productivity and diffusingstate-of-the-art technology in manufacturing, thesetax changes deserve consideration.

R&D tax credits are less controversial, at least inprinciple, and are a great deal less expensive. It iswidely agreed that there are many societal benefitsfrom the generation of new knowledge that individ-ual firms cannot capture. As for the cost, in 1985,before the provision for R&D tax credits in Eco-nomic Recovery Tax Act (ERTA) and Tax Equityand Fiscal Reform Act expired and before the newtax law, the tax revenues foregone because of theR&D tax credit were estimated at $700 million bythe Joint Committee on Taxation.18 Estimates of theamount of additional R&D generated by each dollarof foregone revenue range from $0.35 and $0.99.19

The high estimate, if correct, indicates that the R&Dtax credit is quite efficient, compared with manyother tax instruments; but if the low estimate iscorrect, the impact is modest.

One possible explanation for a moderate impact atthe low range of estimates is that the tax credit forR&D is only one stimulus. R&D costs can beexpensed--educted from revenues to yield taxableincome—in the year they are incurred, which is theultimate in fast depreciation. While the R&D taxcredit has repeatedly been subject to sunset provi-sions, expensing has been an option for decades.With a powerful stimulus already in place, we wouldexpect the additional impact of a tax credit to bemodest. Also, it is possible that the impact of theR&D tax credit in the early 1980s was affected bythe ITC and ACRS, which made other competinginvestments more attractive.

The R&D tax credit survived the Tax Reform Actof 1986, but in a form that many agree is not aseffective as it could be. One often-mentionedcriticism is that the R&D tax credit has never been

16Joseph J. Cordes, ‘‘The Effect of Tax Policy on the Creation of New Technical Knowledge: An Assessment of the Evidence, ’ in Richard M Cyert&David C. Mowe~ (eds.) The impact of Technological Change on Employment and Economic Growth (Cambridge, MA: Ballinger, 1988), and RobertChirinko and Robert Eisner, “Tax Policy and Investment in Major U.S. Macroeconomic Models, ” Journal of Public Economics, March 1983. Theseestimates were developed using econometric simulations, and varying assumptions in the simulations account for the large range of the estimates.

ITJoint committee on Taxation, Estimates of Federal Tu Expend@res, Fiscal Years, A.IUNMl.l~,sm ~Wss, CmWs610n~ Budget office, Feder~ Support for R&D and Innovation (Washington, DC: Congressional Budget offices APril

1984).lgIbid., ~. 780 Bo~ the R&D tu c~t and the ~vestment ~ cr~it were design~ to elicit ~dition~ ~nding on R&D md investment. While the

R&D tax credit was designed to apply only to incremental spending above a base level, there is little doubt that some of the credit was claimed for R&Dthat would have been done anyway by companies increasing R&D; and many assert that corporations redefine certain activities as R&D in order to claimthe credit. llese are some of the considerations that are taken into account when estimating how much additional R&D was done as a result solely ofthe tax credit.


made permanent, and it is therefore not somethingbusiness planners can count on. While there has beenno lapse in its availability, the form of the tax credithas been changed twice since it was enacted in theERTA in 1981. It was reauthorized in the OmnibusBudget Reconciliation Act of 1989, with a fewchanges.20 So, while some analysts have pointed outthat R&D tax credits have not clearly stimulatedsignificant increases in R&D spending, uncertaintyover the form and duration of the credit itself maybepartially responsible. Congress might wish to con-sider making the tax credit permanent.

When investment and R&D tax credits are sub-jected to tests of efficiency or effectiveness, bothseem to have only a modest impact.21 While analysisof the effects of such measures can give someinsight, it is impossible to predict accurately theresponses of business to these stimuli to developnew technology or diffuse best practice. Although afew different combinations of stimuli have beentried a few times, the possibilities of these measureshave not been exhausted. Meanwhile, there is strongevidence that something is needed to stimulatetechnology development and diffusion. Under theseuncertainties, it may be tempting to try somethingsmall or temporary, as the R&D tax credit hasalways been in the past, and as the ITC proved to be.If the tools are used tentatively, however, modestimpacts should be expected. We may have to relymore on informed judgment than economic analysis,and make a stronger commitment to tax or otherstimuli to investment and R&D.

Relationships With Providers of Capital

In addition to high capital costs, there are otherpressures in the American financial environment tofocus on short-term gain and avoid long-term orrisky investments. Heavy turnover of stock inmarket trading and the pressures on institutionalmoney mangers to show short-term gains in excessof market averages are important factors affecting

the outlook of publicly owned American companies.In Japan, these problems have been avoided throughstable (or mutual) shareholding, an arrangementwhich permits a company to cache most of itsstock--estimates of 70 percent are common-in thehands of other companies, where it is not often soldor traded, and there is little pressure to pay largedividends. 22 While Japan had a long tradition ofmutual shareholding within its prewar zaibatsu andpostwar keiretsu company groups, incentives to findstable shareholders were increased in the early 1970swhen government agencies began to worry thatJapan’s heavy dependence on outside expertise wasbringing with it too much foreign investment. Therenewed zeal with which company managers soughtstable shareholders, then, was a response to thethreat of foreign takeovers.23

If Japanese companies can afford to treat theirshareholders as peripheral to the decisionmakingprocess, American companies have come underincreasing pressure to do just the opposite. Americanfirms have always had to pay more attention to thedemands of their shareholders than Japanese firms.However, recently, the demands of shareholdershave focused more than ever on short-term gains,and as a consequence American fins’ concern withshort-term performance has become a preoccupa-tion.

The change has come about in part because of thewave of merger and acquisition (M&A) activity inthe 1980s. Mergers and acquisitions go on con-stantly, occasionally rising to peaks; however, theactivity seen at peak periods differs in kind as wellas magnitude from ordinary M&A. In the 1980s, thedifference was that far more institutions and individ-uals could become acquirers, even with relativelysmall resources. In the past, M&A was characterizedby large firms acquiring-in friendly or hostilefashion-smaller ones. The change resulted fromrelatively loose antitrust law enforcement, and theavailability of short-term, high-interest capital from

specifically, the credit now applies to R&D spending over a fixed base, which is calculated as the ratio of a firm’s R&D expenses to gross receiptsfrom 1984 to 1988. In addition, the new law aIlows fmns to claim R&D on prospective lines of business, rather than limiting qualified credits to R&Din current lines as the old law did. Source: David L. Brumbaugh, “The Research and Experimentation Tax Credit,’ CRS Issue Brief, updated Dec. 21,1989.

21~e h@ e~ata of ~itim~ ex~ndi~e cauwd by tie ITC and tie R&D t~ c~it *OW hat bo~ could ~SO k Rgmdti ~ qUite eff~tivt?,but most analysts seem to think that the true impact is well below the high end.

22s=, fm Cxmple, HidW Ishiluua, ‘Japan’s Compliant Sllmholkrs, “ The Asian WaflStreetJournal Weekfy, June 13, 1988; “Backof the Queue,Please,” The Economist, Apr. 29, 1989; Robert J. Ballon and Iwao Tomita, The Financial Behavwr of Japanese Colorations (lbkyo: KodanshaInternational, 1988), pp. 50-53.

23Bwm ~d ~mi~, op. Cit., pp. 5@51.


high-risk bonds (“junk” bonds) for financing.Finally, another characteristic of the 1980s peak wasthe rise of the bustup takeover, where the acquirerquickly split up the acquired company and soldmany of its pieces in order to reduce the debtincurred in the takeover.

A great deal of effort has been spent trying tounderstand the consequences of M&A activity, butthere are few areas of consensus. Some maintain thatthe M&A peak in the 1980s was mostly positive,correcting excesses of the 1960s wave of M&A,when large firms tried to diversify their business andstabilize their overall cash flows by buying smallercompanies. Longitudinal studies of many transac-tions show evidence of increased productivity andprofitability in acquired companies. Detractors pointout that the 1980s M&A wave resulted in muchincreased corporate debt levels, which in turn forcedcompanies to curtail current or planned spending onR&D, capital equipment, marketing, or other itemsconsidered discretionary in the short run. Whilesome reductions in capital equipment purchase andR&D may be taken without severe damage for atime, prolonged reductions will cost a firm its abilityto compete technically.

According to the evidence, M&A overall has hadlittle or no direct effect on things like R&Dspending. However, National Science Foundation(NSF) data show that high debt—which is character-istic of the 1980s-style takeovers, but not of friendlymergers and acquisitions—was strongly associatedwith a drop in R&D funding, while companies thatdid not undergo high-debt restructuring increasedR&D funding. If the focus is narrowed from allM&A to hostile takeovers and defenses againstthem, the argument that takeovers are having delete-rious effects on technology development, capitalequipment spending, and general willingness tomake long-term investments becomes stronger.

Institutional investors-mostly pension funds—account for the lion’s share of the new short-termpressure. Pension funds and other institutionalinvestors hold about a third of all outstanding stock,but are believed to account for more than half of alltrading. 24 Pension and institutional fund managers,in turn, keep or lose their jobs depending on whether

their stock portfolios have done as well as themarket. Firms, responding to these powerful inves-tors, feel pushed to maximize their own short-termprofits, believing that the market will penalize themfor long-term investments that dilute those profits.25

The penalty is the threat of a hostile takeover. Whileonly a few companies have actually experienced atakeover attempt, the possibility of facing one isviewed with great consternation by many business-men, and many CEOs devote valuable time andresources to the problem. The irony is that somecompanies have acted to avoid hostile takeovers byplunging into debt to buy out shareholders, whichcan have an effect on the company’s long-termperformance similar to that of a hostile takeoveritself, or the attempt to fight one off.

In some cases, hostile takeovers have had benefi-cial effects, replacing ineffective management andrestoring control to managers whose companieswere swallowed by large conglomerates unfamiliarwith the business of their subsidiaries and uninter-ested in measures of performance other than profit.Few people, even the harshest critics of the wave ofhostile, bustup takeovers of the 1980s, wouldadvocate a cessation of all merger and acquisitionactivity; most agree that some threat of a hostiletakeover is an important disciplinary force. Yet therelative ease of hostile takeovers in the 1980s—brought about principally by the availability ofhigh-risk bonds for financing, and also by lessstringent antitrust enforcement—has made the fi-nancial environment even less conducive to long-term investment than in the past.

Mitigating the pressure for short-term profits isnot simple. Any policy change would have to becarefully crafted to have a substantial effect onmarket behavior yet avoid working too well andblunting the ability of shareholders to oust badmanagement.

Most of the proposals for changing investors’short-term time horizons are tax proposals. One thatis often advanced is that Congress provide incen-tives for holding stocks for a longer period byreducing the rate of capital gains tax on gains fromthose stocks. Currently, capital gains are taxed likeincome, with a top rate that is, in effect, 33 percent.

24Alan Murray, “Capital-Gains Tax Bill Would Spur Asset Sales More than Investment,” The Wall Street Journal,” Sept. 28, 1989.ZsMich=l L. ~~ouzos, Wchwd K. ~Ster, and Ro~fi M. Solow, Made /n America: Regaining the prod~tive Edge (Cambridge, MA: The MIT ~eSS,

1989), p. 62.


In the 1989 debates over lowering the capital gainstax rate, one of the options before Congress was togive preferential tax treatment-an effective rate of20 percent over the next 2 years-for assets held fora year or more, rising to a top rate of 28 percent ratethereafter, but with the gain indexed to net out theeffect of inflation. Another included a two-stepschedule of capital gains tax: assets held for a yearor more would qualify for indexing, and those heldfor 5 or more years would qualify for alternativepreferential treatment, with the option of calculatingtaxable gains on 25 percent of the sale price ratherthan the full indexed value.26 Another proposal, thePackwood-Roth bill, would allow individuals toexclude from capital gains tax a percentage of thegain, depending on how long the asset is held.Investors who have capital gains on assets held forless than 1 year could exclude 5 percent of the gainfrom tax; the amount excludable increases by 5percent for assets held each additional year up to amaximum excludable gain of 35 percent, after 7years. Earlier, the President proposed atop rate of 15percent on all capital gains.

Some proposals would make more fundamentalchanges in the tax treatment of capital gains than anyso far considered in the 101st Congress. Forexample, one scheme is to have a seven-stepschedule of capital gains taxes, taxing very heavily(at 50 percent) those held for a year or less, andlightly (at 10 percent) those held for 6 years ormore.27 Another proposal would tax capital gains onsecurities held less than a year at 50 percent, andreduce the rate to 15 percent on gains on securitiesheld for more than 5 years.28 Both proposals alsobroaden the base for taxing all capital gains, toinclude institutional investors as well.

The above proposals, or a similar steeply variableschedule of capital gains tax to reward long-terminvestment, could help to lessen the pressures onmanagers to show short-term profits. It would alsobean incentive for investors of all types to evaluateand monitor more carefully the performance andprospects of companies they invest in. That is all tothe good; inattentive investors with short time

horizons contribute little if anything to the manage-ment of business, and much less to technologydevelopment in the private sector. However, thereare some potential problems as well. For example,investors might be unduly influenced by tax consid-erations to leave their money in companies withmediocre performance, blunting the signals themarket is expected to give to managers. But thedamage done by the short-term outlook to Americanmanufacturing is severe enough to warrant seriousconsideration of significant changes in capital gainsrates.

Although a variable capital gains tax schedulewould encourage investors to hold assets longer, itwould not by itself affect the group of investors mostoften cited as engaging in speculative turnover.Institutional investors-pension funds and invest-ment funds for nonprofit institutions like universities—pay no capital gains tax. In order to quell thespeculative turnover on the stock market, therefore,Congress might consider additional measures tochange the incentives of either institutional fundmanagers or investment bankers who handle trans-actions. One proposal is to charge an excise fee onthe pension funds’ gains on stock turnovers if thestock is held for 180 days or less. Another is tosubject these institutional investors to capital gainstaxes.

Another possibility is to charge a transactions taxon all stock trading, or a securities transfer excisetax. 29 This would raise the costs of stock transac-tions, but would disproportionately discourage rapid,speculative turnover; the greater the turnover ofstock, the greater the disincentive caused by thetransactions tax. The securities transactions taxwould also raise the cost of capital, but according toone analysis, not enough to match the beneficialeffects of increasing corporate time horizons andreducing “the diversion of resources into the econ-omy’s financial sector. ”30 An added benefit of asecurities tax is the revenue it raises; Summers andSummers estimate that a 0.5 percent tax would raiseabout $10 billion annually. Japan’s securities trans-action tax raised $12 billion last year. All of these

MEliz&~ Wek, “fiamis~ Fault WVaI plans for Capital Gains cuu” Congresswnuf Qwterly, Aug. 19, 1989.27D~@j p, Ba~n, U@d~~B~~on /n~~~tm~~~epO~, Vol, Lxx~, Noc 1,, J~, 3, 1989‘/%Felix G< ROMP, “~~ituion~ ‘~ves~r’ or ‘Speculator’’?” The Wull Street Journal, June 24, 1988.z~awence He s~em and Victoria P. Smers, “When Fimncial Markets Work Too Well: A Cautious Case for a Securities Transactions ‘Ihx,”

paper presented at the Annenberg Conference on Technology and Financial Markets, Washington, DC, Feb. 28, 1989.3qbid., p. 1.


proposals favor long-term investments, and coulddiscourage those leveraged buyouts (LBOs), hostiletakeovers, and junk bond transactions aimed atshort-term speculation.

So far, none of these proposals has been subjectedto thorough examination and public debate. Most ofthe legislative proposals made so far would confer abenefit to those who hold stock for more than acertain time (6 months or 1 year, in differentproposals made before Congress), hardly long termby the standards our strongest international competi-tors have set. None of the legislative proposalsconsidered so far would penalize those holding astock for less than 6 months to 1 year, beyond taxingthe gain at the marginal rates for ordinary incomeand retaining limitations on the deductibility of aloss. The potential risk-possibly reducing theliquidity of investments in securities, and therebyreducing the ability of the market to give appropriatesignals to company managers—is real, but we do notyet know how great a risk this is. The issue centerson just how important taxes are, relative to otherconsiderations, in the investment decisions of allkinds of investors. That is one of the most importantquestions to address in order to craft policies thatcontinue to encourage investors to make theirsavings available to companies, but favor companiesthat are managed for long-term gain as well asshort-term profit.

In addition to tax measures, a menu of othermeasures could be considered to return hostiletakeovers to the role they played in the past—namely, a disciplinary force on poor management.They include extending the minimum duration oftender offers, outlawing greenmail and goldenparachutes, shortening disclosure time when aninvestor has acquired more than 5 percent of acompany’s stock, and requiring tender offers inexcess of 110 percent of share value to be made toall stockholders.31 These are aimed specifically athostile takeovers. But by most accounts such raidingis on the wane. If the flurry of junk-bond financingand hostile tender offers is subsiding, Congress hasan opportunity to assess the effects of the bubble ofrestructuring activity, without the sense of urgencythat caused many of the anti-takeover proposals tobe raised. Some limits on the ability to make andfinance hostile tender offers may therefore be worthconsidering, even though such limits will have to be

balanced against the healthy and indeed necessaryeffects of takeovers on managerial performance. Inan important sense, takeovers are the fundamentalenforcer of market forces on individual fins; thetrick is to keep the pressure on while ensuring that itdoesn’t get out of hand.

Environmental Uncertainty

American managers have long had to contendwith a macroeconomic and political environmentthat was managed less for their welfare than for otherpurposes. Foreign policy, macroeconomic policy,international finance, and trade policy have at manytimes been conducted with scant consideration forthe effects of different choices on the competitiveposition of American producers. When America wasthe world’s dominant maker of most goods and hadthe best technology and manufacturing practice in awide variety of industries, this was not a debilitatinghandicap. Now we must take it more seriously.

Although the process of making macroeconomic,foreign, and trade policy is not manifestly moreindifferent to business (or manufacturing) competi-tiveness than it once was, the consequences of thosepolicies are now more important. In many areas ofobvious importance to the economy (e.g., parts ofthe semiconductor industry), American manufactur-ing is struggling to survive. Changes in the generaleconomic and political environment that would havebeen inconvenient in the past could be cripplingnow.

This is not meant to suggest that the conduct of allour most important domestic and internationalpolicies be guided solely by the wish lists ofAmerican manufacturers. But we might considerbuilding institutions that could advise policymakersin key areas on the effects of their choices onAmerican competitiveness. Foreign policymaking,for example, is often at odds with the commercialinterests of U.S. manufacturers. The Department ofState has just one office that concerns itself with acommodity, rather than a country or region. That isthe Textile Division of the Bureau of Economic andBusiness Affairs, the purpose of which is to keeptrade frictions in textiles and apparel from interfer-ing with the foreign policy aims of the Department.The U.S. Trade Representative’s office and theDepartment of Commerce sometimes champion the

31Rmd V. WOg, “HOW I Fought Off the Raiders,” FORTUNE, Feb 27, 1989, p. 118.


competitiveness of American manufacturing, butthis is more a matter of the political persuasion of theappointees and administration currently in officethan a standard practice.

There are many approaches to solving this prob-lem, and various forms have appeared in legislativeproposals over the past several years. One approach,often proposed, is to create anew, powerful voice inthe cabinet for competitiveness interests—a Depart-ment of Trade and Industry, loosely patterned afterJapan’s Ministry of International Trade and Indus-try. Another and more difficult approach would be tocreate institutions within existing departments torepresent competitiveness and manufacturing inter-ests, and to build sensitivity to those concerns intoall departmental decisionmaking. This, in fact, maybe more like the Japanese approach than creating ourown version of MITI. Nearly every Japanese minis-try has strong incentives to consider the competitive-ness of Japanese companies under its jurisdiction increating and implementing policies. If Congresswishes to consider this approach, a thorough studyof what those incentives are in Japan and otherdeveloped nations would be a good starting point.

HUMAN RESOURCES

Manufacturing managers, having grumbled foryears about the shortcomings of American publicschools and a poorly educated work force, havebegun to speak of a crisis. Semi-literate machinefixers who used to repair machinery by looking athow it worked are baffled by computerized equip-ment stuffed with invisible electronic components;these machines need repairers who can read manualsand diagrams. Young people leaving school withmeager math skills are not prepared to deal withcomputer printouts and digital analyzers to monitorquality on the assembly line.

Some large companies are trying to deal with theproblem by educating employees themselves. Mo-torola, for example, estimates that from 1989 to 1993it will have spent $35 million teaching its workersreading and arithmetic. Motorola is committed toeducating workers already on its payroll, but hasbecome more selective in hiring; it no longer takespeople who cannot do fifth-grade math and seventh-grade reading. At that, said a company vice-president, ‘‘ We’ve had situations where we couldn’topen the factory because we didn’t have the workforce.”32

The situation threatens to get worse before it getsbetter. More than half the net growth of the workforce from 1986 to 2000 will be from minoritygroups, 33 and a great many minority children (38 to

45 percent) are growing up poor. Poor children dropout of school in disproportionate numbers, and manyemerge sadly lacking in the skills they need foreconomic survival. David Kearns, chairman of theXerox Corp., sees in this the ‘makings of a nationaldisaster." 34

Few issues on the domestic front have received asmuch attention in the past few years as the sorryresults of American public schooling. Indeed, it ishard to overstate the importance of better educationin the basics, not only for national competitivenessbut also for a peaceful and prosperous society--onewhich gives most people a chance at decent jobs anda middle class livelihood. However, this reportconcentrates on the factory rather than the schoolroom, and thus does not attempt to add much to themany recent analyses and proposals for improve-ment in our children’s basic education. Other OTAassessments, examining various aspects of educa-tion and training, have analyzed some public policyissues that are particularly relevant to manufacturingperformance. 35 The discussion below flags some ofthese issues and describes them briefly, withoutanalyzing specific policy options.

s~tidy Shycki, “me Company as Educator: Firms Teach Workers to Read, Write,” The Washington Post, Sept. 22, 1989, p. G1.33~cor~gto tie B-au of Labor s~tistic~, 57 ~rcent of fie 20.9 mi]llon net ~o~ in tie labor force from 1986 to 2000 will come from minority

groups (6 million Hispanic, 3.6 million Black, and 2.4 million Asian and other). Ronald E. Kutscher, ‘‘Overview and Implications of the Projectionsto 2CM)0, ” Monthly Lubor Review, September 1987, pp. 34.

sQEdw~d B. Fiske, “~pnding U.S. Jobs ‘Disaster’: Work Force Unqualified to Work, ” The New York Times, Sept. 25, 1989, p. 1.35u,s. ~n~~, of ficeof Tec~oloW As=ssment, E~cating scie~~ts and E@neers: Gr& SChOOl to Gr~schoo[, OTA-SET-377 (WUhhgtOIt,

DC: U.S. Government Printing Office, June 1988); Technology and Structural Unemployment: Reenjoying Displaced Aaldfi, OTA-ITE-250(Sprin@leld, VA: National Technical Information Service, 1986); [international Competitwn in Services, OTA-ITE-328 (Spnngfleld, VA: NationalTechnical Information Semice, 1987), chs. 7, 8, and 10; and the forthcoming assessment “Worker Training: Implications for U.S. Competitiveness”(publication expected fall 1990). OTA has also conducted a several assessments of technology and public school education; two recent ones are PowerOn/New Twhfor Teaching and Learning, OTA-SET-379 (Washington, DC: U.S. Government Printing Office, 1988) and Linkingforharning: A NCWCourse for Education, OTA-SET430 (Washington, DC: U.S. Government Printing Office, November 1989).


Training the Active Work Force

Essential though it is, improvement of publicschooling is a longrun proposition. Children enter-ing the first grade in 1990 will leave high school in2002, and effective education often begins sooner(as in the Headstart program, which starts at age 3)and ends later. Thus, even if we improved publiceducation radically, starting tomorrow, the fullresults would not show up in the work force untilwell into the 21st century.

A more immediate approach to improving humanresources for manufacturing is to help peoplealready in the work force gain the skills they need formodern jobs. ‘‘Skills training” covers abroad range,from upgrading basic math and reading abilities tomastery of a complex craft. Often the most urgentlyneeded skills are the basics, so that workers canunderstand operating manuals and take part instatistical process control for quality. In addition,worker training is only one aspect of improvinghuman resources for manufacturing. Managers alsoneed training in organizing work and using peopleeffectively in relation to new technologies. Givingshopfloor workers a genuine stake in the companyand real responsibilities for better quality and greaterefficiency; promoting team work (among engineersas well as operatives); organizing work to make themost of people’s abilities-all these things add up toskillful management of human resources.

The Federal Government has had a long butgenerally not very close or direct involvement intraining of adult workers who want to upgrade theirskills. The most pervasive Federal influences areindirect: in government-guaranteed student loans,which workers can use for taking part-time courseswhile they hold down jobs; and in the tax laws thatlet employers deduct the costs of employee trainingfrom taxable income and, in some cases, allowworkers to deduct what they pay. 36 The biggest

direct Federal involvement is in the armed forces,where training and R&D in how to provide it havebeen major concerns since World War II. Some

computer-based training technologies developed forthe armed forces have found their way intoworkplace training on the civilian side.37 Aside fromthe military sphere, Federal activity is minor. Asmall program that partially funds demonstrationprojects for teaching literacy at workplace sites isgreatly oversubscribed. Congress provided $9.5million for it in 1988, and a flood of proposals camein, requesting a total of nearly $100 million; theprogram was funded at $11.9 million in fiscal years1989 and 1990. Another small effort on the FederalGovernment’s part is encouragement and technicalassistance for employee involvement projects, pro-vided by the Federal Mediation and ConciliationService and the Labor Department’s Bureau ofLabor-Management Relations and Cooperative Serv-ices.

Some of the States are far more active than theFederal Government in supporting workplace train-ing. Illinois, for example, in its Prairie State programpays half the direct cost of worker training coursesfor companies that are in trouble (as shown by theirtax returns). Typically, the companies are small onesand the training is very often in statistical processcontrol-something that larger companies are in-creasingly demanding of their suppliers. SeveralStates that run industrial extension programs, offer-ing technical assistance to small manufacturers,have found that training is an absolutely essentialingredient in the adoption of new technologies.38 Atleast one program, the Michigan ModernizationService, systematically pairs training with technol-ogy extension. In supporting State technology exten-sion programs or developing Federal centers thatprovide such services (see the discussion below), theFederal Government might insist that training beprovided along with advice and assistance in acquir-ing advanced equipment.

A full examination of policy issues surroundingthe retraining of active workers will appear in aforthcoming OTA report, Worker Training: Impli-cations for U.S. Competitiveness.

sb~~ti~s f~ individ~s ~ limi@ to WO&-rCl~ mining, and can be taken only if the amount spent for training plus all other mkdaWWSdeductions is more than 2 percent of the taxpayer’s adjusted gross income, Material hereon the Federal role in workplace training is abstracted fromwork in progress on OTA’s forthcoming assessment of “Worker Training: Implications for U.S. Competitiveness. ”

s7SP~g ~ ~ ~p~mt of ~fenw on R&D for educat.iond technoiogks is ei@ times the combined spending of W National ScienceFoundation and the Department of Education ($56 million a year, on average, v. $7 million). Charles Blaschke et al., “Support for EducationalTedmology R&D: The Fe&ml Role,” contractor report prepared for OTA Sept. 30, 1987, p. vi., for the as sessment Power On! (op. cit.)

38s= tie &scWion of this pint in the section entitled “Industrial Extension” in ch. 7.


Supply of Engineers: Keeping thePipeline Filled

In the next decade or so, it could become muchharder than it is today to maintain an adequatesupply of technically competent people for manufac-turing, especially engineers. In the mid to late 1980s,most analysts found that there was little evidence ofa real shortage of engineers in the United States—yet.39 Also, the United States was about on a par withJapan, Germany, and other advanced countries in theproportion of engineers in the work force (see ch. 4).But it looks as though this parity will not last long;Japan is now graduating far more engineers percapita than the United States.

Demographic facts suggests that maintainingeven the present level of supply could become moredifficult over the next 10 or 15 years. A growingproportion of the young people coming through theeducational pipeline are from minority groups, andup to now minorities have been very much under-represented among engineers. Blacks are 12 percentof the population and Hispanics 9 percent; each werebelow 2 percent of all employed engineers in 1986.Women, too, are underrepresented in engineering;they are 45 percent of the Nation’s work force, butonly 4.1 percent of employed engineers. That raterose from 1.6 percent in 1976, however, and willcontinue to rise, since nearly 15 percent of engineersgraduating with a bachelor’s degree in 1986 werewomen. The proportion of blacks among employedengineers rose more slowly over the 10 years, from1.2 to 1.7 percent.40

Public policy has not been heedless of the fact thatwhite males-predominant in science and engineer-ing in the past-are a dwindling proportion of newentrants to the labor force. Several Federal agenciesoffer special scholarships and grants to encourageminority students, or women, or both, to studyscience and engineering in college or graduateschool; 41 some also offer programs such as summer

internships to stimulate interest in science and mathamong minority high school students.42 Many ofthese programs have scored good results, anddeserve support. But they are inevitably limited. Theinclination toward a choice of science or engineeringusually comes early. Children who decide in ele-mentary school that they don’t like or can’t learnmath are not likely to see themselves as engineerswhen they grow up. This means that, to really openwider opportunities to all children to choose engi-neering careers, we must do a better job of teachingmath and science from the beginning.

Meanwhile, retraining of midcareer engineers,like the retraining of adult workers in general, couldhelp to shore up the supply of engineers available tomanufacturing in the next few years. If funding forthe Department of Defense declines as expected withthe melting of the Cold War, some of the engineersdoing military work will likely lose their jobs. Partof a U.S. Government program for easing thetransition from military to civilian production andemployment could be providing retraining opportu-nities specifically designed for engineers Withgovernment support, retraining courses might bedeveloped to fit the needs of manufacturing—something that is generally neglected in universityengineering departments.

Manufacturing Education and Research

The quality of engineering is as important tomanufacturing performance as the quantity. Theelitism of design engineers and their remotenessfrom problems of manufacturing (“throwing thedesign over the wall”) are well-known failings inAmerican manufacturing. Insofar as these are prob-lems of management, there is little that governmentcan do about them directly. However, efforts toencourage more interaction between the designcenter and the shop-floor (such things as designingfor manufacturability and simultaneous product andprocess engineering) also involve education and

39u.s. Cm=ess, Offiu of TWhnoIoU Asses~ent, Demographic Trends and the Scient#ic and Engineering Work Force--A TechnicalMemorandum, OTA-TM-SET-35 (Springfield, VA: National Technical Information Service, 1985), pp. 92-109; Higher Education for Science andEngineerin~ Background Paper, OTA-BP-SET-52 (Washington, DC: U.S. Government Printing Office, 1989), p. 14 ff.

%ational ScieneeFoundation, ProjiletilectricallElectronicsEngineering: HwnanResourcesandFunding, NSF 88-326 (Washington, DC: U.S.Government Printing Offke, 1988).

ql~e MARC progr~ (M.inotity Access to Research Careers) of the National Institutes of Health is a good example of such programs. It h donewell at bringing minority students into science careers, and currently provides 410 undergraduate scholarships and 69 graduate and faculty fellowships.For a brief description, see OTA, Educating Scientists and Engineers, op. cit., p. 54.

QF~~ ~ncia ~m provi~ rn~ and science internships for high school students, college scholarships, and teaeher training sessions d modelcourses that are open to everyone.

Chapter 2--Strategies To Improve U.S. Manufacturing Technology: Policy Issues and Options .53

research. For example, simultaneous engineering isa difficult technical as well as management chal-lenge. The technical problems might eventually besolved with more R&D attention and more powerfulcomputers. In education and research, the govern-ment does have some leverage.

Few American universities have departments ofmanufacturing engineering, nor do they offer mucheducation and research relevant to manufacturing intheir other engineering departments. This is partly amatter of money. Manufacturing R&D gets littleFederal funding; it probably received well under 1percent of the total $65 billion the U.S. Governmentspent for R&D in 1989, and nearly all of that camefrom the Department of Defense.43 Other Federalsupport for manufacturing R&D is truly meager. TheCenter for Manufacturing Engineering of the Na-tional Institute of Standards & Technology wasfunded at $6.2 million in fiscal year 1989. Technol-ogy awards by the National Science Foundation’sManufacturing Systems Division were about $6.5million, out of NSF’s total of $1.5 billion grants andawards. The NSF-sponsored Engineering ResearchCenters at 18 universities received about $33 mil-lion; some (not all) of these centers emphasizemanufacturing R&D, and are giving engineeringstudents cross-disciplinary training that is valuableto manufacturing companies (see the discussion ofERCs below). One option for raising attention tomanufacturing in universities beyond the presentlevel would be to elevate the NSF’s ManufacturingSystems Division to a Manufacturing SciencesDirectorate. This would provide a solid, prestigiousbase for government support of research and educa-tion specifically focused on manufacturing.

DIFFUSING MANUFACTURINGTECHNOLOGY

Throughout the 1980s, Congress has taken anumber of actions to transfer advanced technologiesfrom labs to factories, bring smaller firms up to datein manufacturing technology, and modify laws thatmay interfere with technology advancement inmanufacturing. Some of these actions are wellalong; others have barely begun. Not one of them, byitself, is likely to have any very dramatic effect,certainly not overnight, Some, after a fair try, will

pan out and others will not. Given patience and anopen-minded experimental approach, it is likely thatsome combination of these measures could make anappreciable difference in improving manufacturingperformance.

Some of the most promising options are similar toJapanese government programs (national and local)that have served that country’s manufacturing firmsfor years. There are of course many economic,social, and political differences between the UnitedStates and Japan; not everything that works therewould work here. However, as discussed below,several of these Japanese programs do seem to bequite adaptable to American conditions.

Technology Extension

One way for government to help manufacturersadopt improved technologies is through variouskinds of technology extension services. A few Statesare providing services of this kind with a good dealof success. This is one of the programs that workswell in Japan. The nationwide network of technol-ogy extension services in Japan is much used bysmall and medium-size manufacturers. (See chs. 6and 7 for discussions of the importance of smallermanufacturers to U.S. competitiveness and descrip-tions of government programs in Japan and theUnited States that offer small firms technologyassistance.)

Until very recently, Federal involvement in tech-nology extension was minimal. The States havedone more, but even so, in 1988 the combinedtechnology transfer and technology/managementassistance programs of 30 States added up to only$58 million-and this figure overstates technologyextension to manufacturers, since it includes man-agement assistance of various kinds to all sorts ofbusinesses (see ch. 7 for details). The total for Statetechnology extension services was probably be-tween $25 million and $40 million.

In 1988 Congress created a framework for abroader Federal program of technology extension.The Omnibus Trade and Competitiveness Act of1988 authorized several kinds of technology assis-tance to manufacturers, including ManufacturingTechnology Centers to demonstrate advanced tech-

qsF~er~ s~ndingon R&D relate.dto manufacturing was no more than about $400 million in fiscal year 1989, and may have been less; preci= fi~sare not available. U.S. Congress, Office of Technology Assessment, “U.S. Manufacturing: Problems and Opportunities in Defense and CommercialIndustries,” staff paper, December 1989.


nology and provide extension services, especially tosmaller firms; Federal assistance to State technologyextension programs; and the Advanced TechnologyProgram, a mechanism for Federal guidance andparticipation in joint R&D ventures with privatebusiness. The actual performance of these programshas been modest so far. In fiscal year 1990, Congressappropriated $7.5 million for the ManufacturingTechnology Centers and, for the first time, fundedaid to State programs, at $1.3 million.44 A smatteringof older Federal programs also provide some tech-nology extension services.

At current levels, the combined Federal and Statetechnology extension programs cannot reach morethan a small fraction of the country’s 355,000 smalland medium-size manufacturing firms-those thatare most likely to need technical assistance. Asnoted in chapter 7, one of the most valuable kinds oftechnology extension is customized advice to indi-vidual manufacturers. Giving that service to just 7percent of smaller manufacturers would cost a totalof $120 million to $480 million a year, depending onthe level and quality of service.

If Congress wishes to deepen its commitment totechnology extension, several choices are open. Itcould provide more funds for Manufacturing Tech-nology Centers under the Federal aegis. It could setup a more generous program of Federal matchingfunds to State industrial extension services than thepresent law authorizes. Or it could do some of both.These choices are discussed below.

The Federal Program: ManufacturingTechnology Centers

The Omnibus Trade and Competitiveness Act of1988 gave the National Bureau of Standards newresponsibilities for technology transfer to manufac-turers and renamed it the National Institute forStandards and Technology (NIST). One part of thelaw directed NIST to help create and supportnon-profit regional centers for the transfer of manu-facturing technology, especially to small and medium-size firms. The tasks of the Manufacturing Technol-ogy Centers (MTCs) are to transfer technologiesdeveloped at NIST to manufacturing companies;make new manufacturing technologies usable tosmaller firms; actively provide technical and man-agement information to these fins; demonstrateadvanced production technologies; and make short-

term loans of advanced manufacturing equipment tofirms with fewer than 100 employees.

The trade act authorized $20 million a year forNIST technology extension, but appropriations havebeen much less—$5 million in fiscal year 1988,$6.85 million in 1989, and $7.5 million in 1990.NIST has signed 6-year agreements with threeregional MTCs, giving each $1.5 million per year for2 years in succession, through calendar year 1990.(The remainder is for administrative expenses andother technology extension activities.) The Centersmust match at least half the Federal dollars for thefirst 3 years and an increasing share thereafter; underthe law, the Federal share declines to zero at the endof 6 years.

Japan’s nationwide network of public testing andresearch centers, which provide technology exten-sion services to smaller manufacturers, has manyfeatures in common with the NIST centers but is farmore extensive. In 1985, there were 185 of thesetesting and research centers; they had 7,000 employ-ees and annual funding of 66 billion yen ($470million at 140 yen to the dollar), half from thenational government and half from the prefectures.In addition, many Japanese cities, wards, and otherlocalities have industrial halls that offer much thesame kind of services. (See ch. 6 for details.)

In running the new manufacturing technologyprogram, NIST officials say they are not just passingalong Federal money but are taking an active handin advising the Centers and learning along withthem. Centers are encouraged to work with Stateprograms and take advantage of State resources andexperience. One of the criteria for selecting opera-tors of the Centers is that they have previous linkswith State and local extension programs. NIST hasalso set up monthly meetings of all the Centers sothey can learn from each other.

A key question about the future of the NISTtechnology extension program is how it can best bemeshed with State extension programs that aim to domuch the same thing, with as much coverage and aslittle overlap and re-invention of the wheel aspossible. The 1988 trade act made some provisionfor Federal support of State technology extensionprograms, but in quite limited ways, as the nextsection describes.

~The Mv- Technology Rogram, discussed in a later section of this chapter, also got its first funding, $10 million in fiscid Y= 1990.

Chapter 2--Strategies To Improve U.S. Manufacturing Technology: Policy Issues and Options .55

Federal Assistance to State Programs

The 1988 Omnibus Trade and CompetitivenessAct also set up a limited program of Federalassistance to State technology extension programs.Included was a nationwide study of State technologyextension services; technical advice on how totransfer Federal manufacturing technology to firms;and a clearinghouse for information about Statetechnology programs. The act also authorized asmall program of Federal financial aid to Statetechnology extension programs that already existand want to expand. States would have to increasetheir own funding by the same amount as the Federalcontribution. Their proposals would be judged byhow many new firms they proposed to help under thecooperative Federal-State agreement, whether theycould maintain service after the agreement expired,and to what extent they intended to demonstrate newand expanded uses of Federal technology.

As this report was written, NIST’s State technol-ogy extension program had just begun, havingreceived its first finding of $1.3 million in fiscalyear 1990. On reprogrammed funds, NIST hadalready done the study of State technology extensionservices 45 and started a small, one-man effort toacquaint State agencies with NIST services andresources. The clearinghouse was just getting organ-ized, and Federal financial aid to State programs wasin the planning stages.

In its study of State programs, NIST defined“technology extension services’ as programs whoseprimary purpose is to provide direct consultation tomanufacturers for technology deployment. It foundonly 13 State-supported organizations in 9 Statesthat fit the definition. More and more States,however, are taking an interest in technology exten-sion, and at least one new program (Nebraska’s) wascreated shortly after the survey was done.

Although the State programs are few, scattered,and mostly quite new, they are, on the whole, betterdeveloped than technology extension services at the

Federal level. One or two have years of experiencebehind them and have built up outstanding reputa-tions. For example, Georgia Tech’s statewide indus-trial extension service dates back over a quarter of acentury and is so much in demand that it refrainsfrom any advertisement (see ch. 7). The MichiganModernization Service is less than 5 years old, butit has gained a solid reputation and demands for itsservices are growing; its budget rose 40 percent in1989.

Getting the Job Done: Federal or State Programs,or Both?

Despite the present flurry of State and Federalinterest in technology extension to manufacturers,the actual coverage of such services is still verysmall. It doesn’t begin to compare with the Agricul-tural Extension Service, with its funding of morethan $1.2 billion (31 percent Federal), its offices innearly every county in the 50 States, its 9,650 countyagents, and its 4,650 specialist scientific and techni-cal staff.46 To put this in perspective, consider thatagriculture contributes 2 percent to the gross na-tional product, and manufacturing 19 percent.

Before taking up the question of who can bestprovide technology extension services, it is worthstepping back and considering what a comprehen-sive nationwide system might look like. Sincemanufacturing industries are regionally concen-trated, technology extension centers would not beevenly distributed across the country. In areas ofsufficient concentration, some centers could focuson technologies for just one industry or group ofindustries (e.g., electronics suppliers, auto parts andcomponents makers), while others would be moreeclectic.

If the average center served about 200 clients peryear, and if 24,000, or just 7 percent, of the Nation’s355,000 small and medium-size manufacturing firmstook advantage of the services, then about 120centers might be needed. This is a modest number,based on the experience of the Georgia Tech

Qs’rhe Nation~ Governors’ Associ~ion conduct~ the study under contract for NIST. Results were published in Marianne K. Clarke and Eric N.Dobson, Promoting Technological Excellence: The Role of State and Federal Extension Activities (Washington, DC: National Governors’ Association,1989.)

~wo studies have found high rates of return on investments in agricultural research, extension, and farmers’ schooling. One study estimated internalrates of return (value of agricultural product/research and extension expenditures) of 27 percent on such public investments in the State of Virginia(George W. Norton, Joseph D. Coffey, and E. Berner Frye, ‘‘Estimating Returns to Agricultural Research, Extension, and Teaching at the State bvel,”Southern Journal ofAgricultura/ Economics, July 1984). The other study found asocial internal rate of return to public crop research of 62 percent, and15 percent to farmers’ schooling (Wallace E. Huffman and Robert E. Evenson, ‘‘Supply and Demand Functions for Multiproduct U.S. Cash Grain Farms:Biases Caused by Research and Other Policies,’ American Journal of Agricultural Economics, August 1989.)


industrial extension service. The Georgia Techservice, with 13 small offices statewide and a staffof 26 professionals, makes site visits to about 480clients per year, usually limits service to 5 days, and,as noted, does not advertise, for fear of attractingmore clients than it can serve.47 Georgia has about 2percent of the manufacturing establishments in theUnited States. If other areas provided industrialextension at only the same limited level, and eachcenter served about 200 clients per year, the centerswould number 120, the staff 3,120, and the clientsabout 24,000.48

These figures are based on the assumption that thetechnology extension services do a good job andprove to be worth what they cost. Assuming that theydo, a nationwide technology extension serviceobviously cannot arise overnight. There is room forexpansion of both State and Federal centers, and itwill take time. The question is whether one or theother is better suited to provide the services. It isoften thought that States, being in closer touch withtheir own citizens, do a better job of providingbusiness and technical services. On the other hand,regional concentrations of industries cross Statelines, and it is usually difficult for States to combineforces and provide services on a regional basis. Stillmore important, some States simply do a better jobthan others, and the interest in improving manufac-turing competitiveness is more than parochial; it isnational.49

A combination of State and Federal programsmight best serve the national interest. (It is worthnoting that Japan’s technology extension networkcombines national, regional and local support, withthe national government and prefectures sharingequally the funding 185 centers nationwide, andlocal governments funding more centers on theirown.) Federal grants to support expansion of experi-enced, high quality State programs and technicalassistance to bring newer ones along could be anefficient use of resources. At the same time, there arebenefits in having Federal programs as well. Federalofficials who supervise technology extension have

the advantage of frost-hand knowledge, which isvaluable in evaluating State programs. Federaltechnology extension centers may be especiallyuseful in places where concentrations of one indus-try or allied industries cross State lines, or in areasthat are otherwise underserved.

If Congress decides to support the expansion ofState programs, it might consider raising the presentauthorization of $2 million in Federal matchinggrants. That sum would not go far toward buildinga comprehensive nationwide network of technologyextension services. Suppose that within 5 years theU.S. Government is contributing to the support of 60State programs, each with total funding of $1 millionto $4 million a year, depending on the level ofservice. If the Federal share were 30 percent (as it isin the Agricultural Extension Service), that wouldamount to $18 million to $72 million a year. Theseare extremely modest assumptions. If a nationwideprogram were even as large, in proportion, as theGeorgia Tech extension service, it would include120 centers and cost the Federal Government $36million to $144 million a year.

Congress might also consider removing the con-dition that State programs, to receive funding, mustdemonstrate methods to increase uses of Federaltechnology. Helping U.S. manufacturers make bet-ter use of technology, whatever the origin of thetechnology, is in the national interest.

As for Federal Manufacturing Technology Cen-ters, Congress may wish to reconsider the law’ssunset provision, under which Federal funding stopsafter 6 years. NIST officials expect that the Centerswill generate some income themselves by chargingsome fees for service, but that they will rely mainlyon State funds as Federal funds are phased out. IfCongress considers technology extension a matter ofcontinuing interest, it may want to extend Federalfunding at some level beyond the 6 years. Stabilityand predictability is an important ingredient in thesuccess of institutions like these, and continuedFederal funding is a factor in stability.

d’l~e G~gia TWh pro- seines the same number of clients without site viSitS-a totzd Of tibout 960 Per y=.AsThe e~mate of tie sim of a m~~~ n~ionwi~ extension service is based on the lower number, i.e., the 480 clients receiving site visits.4~ome F~er~ ~rogms hat offer ~mts t. s~tes, ~th ve~ 1i~e in the way of oversight or guitince, ~ve run into the problem of uneven level

and quantity of senice in different States. An example is the displaced worker reemployment and retraining program of the Job Training PartnershipAct, See U.S. Congress, Office of Technology Assessment, Technology and Structural Unemphyment: Reemploying Displaced Adulti, OTA-ITE-250(Springfield, VA: National Technical Information Service, 1986).

Chapter 2-Strategies To Improve U.S. Manufacturing Technology: Policy Issues and Options .57

Financial Aid for ModernizingManufacturing

Technical assistance is one part of the prescriptionfor improving the technology base in Americanmanufacturing, especially for small and medium-size enterprises that do not have a large or diversetechnical staff. Another part is money. Unless asmall firm has an outstanding track record, it willgenerally have a harder time raising money forpurchase of new production equipment than will alarge one. It is hard enough for large U.S. firms tomatch the capital investment rates and R&D spend-ing of their best foreign competitors, in view of thehigh interest rates in the United States and a financialclimate that rewards short-term profits more thanIong-term improvement in market share (see ch. 3).For smaller fins, the difficulties are often com-pounded.

There are many U.S. laws on the books that givespecial breaks to small business.50 For example, theBuy American laws governing purchases by U.S.Government agencies give American firms a 6percent price advantage (the agency must buyAmerican unless the price of the foreign-made goodis at least 6 percent lower); but for small businesses,the price advantage is 12 percent. Another exampleis the Small Business Innovation Research program,which sets aside about $350 million of Federal R&Dmoney per year for small businesses (see ch. 7).

Also, there are special guaranteed loan andsubsidized capital programs for small businesses.Direct Federal loans to small business are limited tospecial groups (disabled veterans, the handicapped,low-income people), and totaled only $47 million infiscal year 1989. (Direct Federal loans to smallbusiness were virtually abolished in the Reaganyears, on the philosophical grounds that governmentloans were an interference with efficient allocationof resources through the free market.) Federallyguaranteed commercial loans to small businessamounted to $3.6 billion.51 In addition, the FederalGovernment subsidizes the Small Business Invest-ment Corporation and the Minority Small BusinessInvestment Corporation, which make equity invest-ments as well as long-term loans to small fins.

Congress appropriated $154 million for these twoprograms in fiscal year 1989, and the corporationsmade investments amounting to $715 million. All ofthese financial programs, it should be noted, are forall kinds of small and mid-size businesses, not justmanufacturers.

The point of most of these programs is to givegeneral support to smaller businesses on the groundsthat they are dynamic and entrepreneurial, andcontribute to economic growth and flexibility. Theprograms have rarely been designed for the specificpurpose of promoting effective use of manufacturingtechnologies. This contrasts with the Japaneseapproach. In Japan, financial aid to small firms is notonly very much larger-some $27 billion in directloans from national government programs and anadditional $56 billion in loan guarantees (again, toall kinds of small and mid-size businesses, includinga great many in the service sector)--but also, muchof the financial aid is tied to technical assistance andsome is directly targeted to technology improve-ments (see ch. 6).

Some options for linking government financialaid to manufacture with technological improve-ments, and possibly raising the amount, are dis-cussed below.

Equipment Leasing

To encourage the adoption of modern manufac-turing equipment, Congress might consider creatinga government-supported equipment leasing systemthat would: 1) make available to manufacturers(especially small companies) new production equip-ment on easy terms; and 2) provide an assuredmarket for at least part of the output of companiesmaking production machinery.

The Japanese government’s equipment leasingsystem, under which small and mid-size companiescan lease new equipment or buy it on the installmentplan at less than market rates, is a key technology-promoting measure, and one that seems reasonablyadaptable to the United States. The Japanese systemwas frost created in 1966, but a new part was addedin 1986 that applies specifically to computers and‘‘mechatronics" —such things as numerically con-trolled (NC) machine tools and robots. Both the

sqn tie Unitd stat=, tie km *Csm~l business” USually means firms with fewer than 500 employees, and thUS includes medium-sti business ~well. In Japan, the tam small and medium-size enterprise (SME) usually means fms with fewer than 300 employees.

SIF~r~ly ~uant~loas Werekept ~ nmin~ly he me level from fisc~ yews 1980 ~CI@ 1989 (about $3.5 billion per year), although pricesrose by 47 pereent over the period, redueing the amount of real dollars.


national government and the prefectures contributefunds to the system; in 1987, leases and installmentsales worth 49 billion yen ($350 million, at 140 yento the dollar) were made under the system. Besidessupporting this frankly subsidized system, the Japa-nese Government has also provided capital forquasi-public leasing corporations that serve larger aswell as smaller companies. One of these is for leaseof computers, another for robots (see ch. 6).

Small companies benefit from the leasing systemin several ways. If they are strapped for cash, theydon’t need a downpayment; if they are not sure of theeconomic benefits of a new piece of equipment theycan try it out without committing to it; and thesystem provides technical consultations and guid-ance on what equipment they need. Besides thesebenefits for users, the system also provides asubstantial, stable market for manufacturers ofproduction equipment (e.g., machine tools).

If Congress wishes to create and support such asystem for the benefit of users only, the country oforigin of the equipment does not matter. But if thesystem is designed to build up the capacity of U.S.makers of production equipment as well, then itwould be necessary to define what a U.S. companyis. The limited American experience with providinggovernment help to private industry in improvingmanufacturing technology does not offer muchguidance on this question. The answer might varydepending on practical circumstances. If one mainpurpose of a government-subsidized leasing pro-gram were to rebuild the U.S. toolmaking industry,it might make sense to restrict the purchases tomachine tools made in this country, perhaps byU.S.-owned companies. (Such a requirement mightbe phased in, since it might be against the interestsof machine tool users if U.S.-made machines werenot as good as foreign-made machines.)

A government-supported leasing system could beset up in various ways. It might be open only to smallfirms or to all firms without regard to size. If open toall, it might give more favorable terms to small firmsif it were open only to small fins, the governmentcould also support in a less direct manner (i.e.,provision of capital on favorable terms) a quasi-public leasing company that would be open to all.

Should Congress be interested in creating anequipment leasing system, an opportune place tostart might be in the effort just getting underway to

develop a next-generation controller for machinetools to be made in the United States. The NationalCenter for Manufacturing Sciences (made up ofabout 90 manufacturing firms, large and small) andthe U.S. Air Force are sponsoring a 3- to 5-year jointproject to promote the development of a new,U.S.-made, single-standard computer controller forNC machine tools. A government-supported leasingsystem could provide some assurance of a market forU.S.-made machine tools using the new controller,and could add impetus to the R&D effort. IfCongress wants to start small, on an experimentalbasis, with a government-supported leasing system,this could be a place to begin.

An equipment leasing system for NC machinetools could start with quite modest funds. Total salesof NC machine tools in the United States amountedto $1.7 billion in 1988; one-quarter of that ($425million) was spent for U.S.-made machines. U.S.producers of machine tools (all kinds, not just NC)lost an average of 11 percent per year in sales from1981 through 1988. Suppose they regained sales ofNC machines at an average of 10 percent per year;in the first year, their sales would rise by $43 million.Suppose the government leasing system boughtroughly 30 percent of the incremental output, or 13million dollars worth, and leased it at a subsidizedrate of about 80 percent of the sales price (i.e., a 20percent subsidy). Then the cost of the programwould be $2.6 million for that year, plus a modestsum for administrative expenses, less the taxes firmswould pay on their increased profits.

A question that is always asked about schemessuch as this is whether they really encourage widerdiffusion of manufacturing technologies, or whetherthe government is simply subsidizing purchases thatcompanies would make anyway. No certain answercan be given, but it seems likely that there would besome real encouragement. First, experience suggeststhat government purchases are a genuine factor inpromoting the development and manufacture ofnew, advanced products; this incentive applies to themakers of the machinery. As for users of themachinery, a 1987 survey of representative metal-working companies found that uncertainty aboutdemand for the companies’ products and lack offinancial resources were the biggest obstacles toinvestment in new plant and equipment. In plantswithout any NC machines (or other programmableautomated equipment), managers gave as a leading


reason that the payback period was too long.52

Leasing the equipment could help managers copewith the uncertainty about demand, and subsidiesembedded in the leasing program would lessenconcern about financial resources and paybackperiods.

Tying Technical Assistance to Financial Aid

In the United States, government financial aid tosmall businesses is not necessarily aimed at techno-logical improvement. But it could be shaped to servethat purpose. For example, Congress might wish torequire a technical assessment as a condition for afirm’s getting a federally guaranteed loan or capitalfrom one of the federally subsidized small businessinvestment corporations.53 But this requirementmakes sense only if a government-supported exten-sion service exists and is able to supply competentpeople to make the assessment. Any such require-ment would probably have to wait for the develop-ment of a much more extensive network of technol-ogy extension services than the United States hastoday.

Another caveat is that government-supportedloans and capital investments in small business arecurrently a minor source of business financing—about $3.8 billion in 1989. To put this in someperspective, all freed investment (in structures, plantand equipment) by all private business was $487billion in 1988. Moreover, since only about 9percent of small American enterprises are in manu-facturing, it is unlikely that more than a smallportion of the U.S. financial aid to small businessesgoes to manufacturers. Furthermore, the aid proba-bly reaches very few fins. In fiscal year 1988,16,469 federally guaranteed loans were made tosmall businesses, and the quasi-public small busi-ness corporations made a total of 4,137 financing.If small manufacturers got a proportionate share ofthese guaranteed loans and subsidized financing,then 1,915 small manufacturing firms benefited—

about one-half of one percent of the 355,000 smallmanufacturing firms in the country. Even if techni-cal assessment were a condition forgetting financialhelp, not many small manufacturers would get eitherone.

This raises the question of whether U.S. Govern-ment financial aid to encourage the adoption of newtechnologies, especially by small fins, is tooskimpy. Recognizing that there is no exact parallelbetween the two countries, it is still notable thatJapanese loans and loan guarantees to small firmsare at least 20 times as high as U.S. Federal financialaid to small business.54 Moreover, the amount ofsubsidy in the Japanese loan programs is oftengreater. Some examples: In the United States, theterms for federally guaranteed loans are negotiatedbetween the borrower and private lender, but interestrates can be as high as 2 3/4 percent above prime. InJapan, interest charges on such loans are generallywell below the market rate. For instance, theEquipment Modernization Loan Program (whichmade direct loans of about $300 million in 1988)lends up to half the amount of the equipmentpurchase, and charges no interest.

In many ways, Japanese and American smallmanufacturing are not really comparable. Manufac-turing in Japan is much more weighted to smallfins, which account for 74 percent of Japanesemanufacturing employment but only 35 percent inthe United States.55 Although total manufacturingemployment is higher in the larger U.S. economy(19.4 million vs. 14.5 million in Japan) the numberof employees in small and mid-size manufacturingfirms is nonetheless greater in Japan (10.7 million v.6.8 million in the United States).

Considering the political and economic differ-ences between the two countries, Japanese policiesobviously cannot be a template for U.S. policies. Yetthe great disparity in assistance to small businessesdoes suggest that some higher level of aid to small

szM~ellen R. Kelley and H~ey Brooks, The S@te of Computerized Autornarh in U.S. Manufacturing, Hwtid university. John F. Kerm~ySchool of Government (Cambridge, MA: October 1988). Managers of plants with no programmable automation also gave technological reasons fornon-adoption, the major one being that there were too few repeat runs to make the initial programming worthwhile.

sqM~Wement ~istmce is av~lable from tie Small Business Investment Corporation ~d the ~noriv En@Pfi* sm~l Bus~ess ~v~tm~tCorporation, but is not a eortdition of getting capital funds from the corporations.

54Mmy SW= have s~i~ lea, gr~t, or capit~ investment programs for small businesses. OTA is nOt WSR of tiny estimate for~e tot~ of fi~ci~aid to small business in all States, nor of any similar estimate of financial aid from prefectures, cities, or other local governments to Japanese small firms.

551n JWn, SMES ~ defm~ ~ fires M* fewer ~~ 3~ employ~s; in he unit~ st~s, fewer tian 5W employ~s. AIso, SMES contribute 56pereent of value added to manufacturing in Japan, 21 percent in the United States. Employment, rather than value added, is used here as an indicatorof the importance of SMEs in manufacturing because the biggest component of value added is wages, and in both countries wages are substantially lowerin small manufacturing firms than in large ones.

2 1 - 7 0 0 0 - 9 0 - 3


U.S. manufacturing firms is worth considering as away to raise their technological level and make themmore competitive. Government help to small U.S.manufacturers could be especially significant, sinceit is uncommon in this country for large customerfirms to give financial or technical aid to theirsuppliers. By contrast, many Japanese subcontrac-tors get some financial support from their customerfirms and a great deal of technical assistance.

If Congress wishes to consider an option ofgreater financial aid to small manufacturers, expan-sion of guaranteed loans, which takes advantage ofthe existing private banking system, probably hasmore appeal than resurrection of direct loans. Thedisaster of the 1980s with Federal savings and loaninsurance might argue against any new or expandedprogram of Federal financial guarantees. However,other loan guarantee programs, such as the FederalHousing Administration’s guarantees for home mort-gage loans have a better record. With the backing ofthe government guarantee, banks can offer lowerthan market rates for FHA mortgages and lowerrequirements for downpayments and borrowers’incomes. At the same time, an FHA inspectionprovides some assurance that the property subject tothe loan is sound. This program can be reckoned asuccess. At least until the great inflation in real estateof the 1970s, FHA-backed loans made it possible forpeople of quite modest means to own a home.Although the default rates on FHA loans have risensomewhat in recent years, they have generally beenmoderate. Default rates on the quite limited programof federally guaranteed loans to small business arealso moderate.56

If Congress should decide to raise the amount ofFederal loan guarantees for small manufacturers,options for tying financial aid to technologicalimprovement assume greater importance. One op-tion would be to target new financial aid toinvestments in advanced equipment. The JapaneseGovernment has done this through its special leasingprogram for high-tech electronic and “mecha-tronic” equipment, open to smaller manufacturers,

and also through selective tax breaks for high techinvestments (described below). There is evidencethat these targeted programs worked in Japan. Afterthey were offered, there was a surge in purchases ofNC equipment. (One Japanese manufacturer calledit ‘the NC-ization period. A possible drawback tosuch inducements is that they might encourage firmsto buy equipment that they really do not know howto use. They might even incite producers of theequipment to cash in by raising prices.

Another option is the one mentioned above: makeFederal financial aid conditional on the fro’sgetting a competent technical assessment and eitherfollowing its guidance or working out an alternativeplan with the advisor. The obvious difficulty withthis option is that adequate public technologyextension services don’t yet exist.

Tax Incentives

An option much used in Japan is to give compa-nies tax breaks-credits or accelerated depreciation—for investments in new production equipment. Espe-cially prominent are various tax incentives availableto small and medium-size enterprises (SMEs). Ineffect, these tax breaks are subsidies, paid forindirectly by the taxpayers. It has long been JapaneseGovernment policy to encourage business invest-ment with programs that keep the costs of capitallow, and this seems to be eminently acceptable to thepublic, who pay for it. In the United States, policiesfor this purpose have been less consistent and aremuch more controversial.

A general discussion of tax incentives as a way tostimulate investments in plant and equipment ap-pears in chapter 3 and an earlier section of thischapter (Financing Long-Term Investment). Dis-cussed there are the disagreements among analystson whether increases in investment due to taxincentives are significant or trivial; the fact thatmany special tax incentives were removed in the1986 tax reform act as a quid pro quo for loweringthe overall corporate income tax rate; the perverseeffect of this bargain, in rewarding old investments

s~e ent~e ~owt of d~t busjness Ioans and the guarantd portion of guaranteed business loans disbursed by the Small Business Administrationfrom fiscal years 1953 to 1989 was $50.5 billion, of which $3.9 billion had been charged off as losses by September 30, 1989. On this basis, the lossrate for SBA business loans and loan guarantees was 7,7 percent. (Information provided by the House Committee on Small Business.) However, the“net loss rate,” figured on the same basis that commercial banks use, is lower, For 1986-88, SBA’S net loss rate for guaranteed business loans was 3.60to 3.74 percent. This compares to net commercial and industrial chargeoffs by banks of 1.17 percent of commercial and industrial loans in 1987 (thelatest date available). Note that SBA takes greater risks than banks because its loans go to startups and other good prospects that need long-term loansbut have too literal equity or collateral to qualify for a bank loan. Allan S. Mandel, Assistant Deputy Administrator for Financial Assistance, U.S. SmallBusiness Administration, “The Role of SBA 7(a) Loan Guaranty Program in the U.S. Economy,” October 1989.


in productive equipment at the expense of newinvestments; the fact that tax incentives cost some-thing and worsen the budget deficit, unless revenueis found elsewhere to make up for them; and theurgency of weighing all reasonable options forimproving manufacturing technology. In view ofthese many complications and uncertainties, theconclusion was that Congress might wish to man-date a study, with an early delivery date, of theeffects of tax incentives as a stimulus to capitalinvestment in manufacturing. This could include aconsideration of special tax incentives for smallmanufacturers.

The broadest and most accessible of the Japanesetax incentives for capital investment by SMEs isaccelerated depreciation-14 percent in the firstyear, on top of normal depreciation-for any ma-chine an SME purchases.57 A measure more directlytargeted to high-tech equipment is the SME NewTechnology Investment Promotion Tax System(established in 1984) which offers SMEs twooptions for buying or leasing electronic and mecha-tronic technology: either a special first year depreci-ation of 30 percent, or a tax credit of 7 percent of thevalue of the machine, up to 20 percent of total taxes(in the case of leased equipment, 7 percent of 60percent of the total leasing expense).

Cooperative Networks of SmallManufacturing Firms

There is strength in numbers. Small firms thatband together to do cooperative research and devel-opment, get quantity discounts on new equipment,share equipment that no single owner can afford,find out about new technologies and new markets,share orders that are too big for any one firm tohandle by itself, and find work for members whenorders are scarce, can strengthen themselves andeach other without losing competitive drive. Coop-erative networks in textiles and metalworking grewand prospered in mid and northern Italy in the 1970sand early 1980s (but seemed to be undergoing somereversal in the late 1980s). Such networks haveproven stable in certain industries in Japan, and maybe growing in importance.

Both the national and prefectural governments inJapan are strongly supportive of cooperative associ-

ations. SME cooperatives can get the same taxbreaks and subsidized equipment leasing as individ-ual small firms, and are eligible for low-cost loansfrom some of the same government financial institu-tions. There are also special loan programs forcooperatives with low (sometimes zero) interestrates, as well as government support for joint R&Dby groups and cooperatives.

Nothing like this government support for coopera-tive networks of small manufacturing firms exists inthe United States. In fact, there is a certain deterrenceto cooperation among small firms from antitrust lawand enforcement—if not in demonstrable fact, atleast in widespread perception (see ch. 7 and thesection below on antitrust options.)

If Congress wishes to support the formation ofcooperative associations among small manufactur-ing firms, it might explicitly state that cooperativesare eligible for the technology extension servicesoffered by the Manufacturing Technology Centers.Cooperatives might also be eligible for small busi-ness loan guarantees, and if an equipment leasingprogram is established, for that as well. If Congresswishes to start in a modest way on a programspecifically targeted to cooperatives, it might beginwith a program of technical assistance on how toorganize cooperative activities, such as joint pur-chases of equipment at discount or shared use ofequipment.

Commercialization of Technology FromFederal Laboratories

Most R&D performed in Federal laboratories isnot directly applicable to civilian industry. Out of$21 billion spent per year, about $13 billion is fordefense, and much of the rest is for basic research.Some of this defense R&D and basic research can bemade useful to civilian industry, in two ways. First,the labs’ expertise and results can be transferred toindustry, which then performs further work tocommercialize the technology. Technology transfercan be accomplished in many ways, includingpersonnel exchange between labs and industry,private fins’ use of specialized lab facilities, andgranting licenses to firms for commercializing thelabs’ patented technology. Generally, effective tech-

syM~ri~ in tis ~tionon t~ incentives for SMEcapit,al investments is drawn mostly from D.H. Whittaker, “New Technology Acquisition in sm~lJapanese Enterprises: Government Assistance and Private Initiative,” contractor report to OTA, May 1989. This report also provides information onthe Japanese equipment leasing and financial aid programs.


nology transfer requires some person-to-person con-tact.

Second, there is cooperative R&D by the labs andindustry. Rather than simply transferring preexistingtechnology to industry, the labs cooperate withindustry to create new technology, which the firmsinvolved can then commercialize. Cooperative R&Dbuilds on the labs’ existing work but takes it in adirection useful to industry-helping to bridge thegap between the labs’ work and industrial applica-tions.

Cooperative R&D is a powerful tool. With theFederal labs sharing the expense and risk, industrycould be better able to take on large, long-termprojects with a highly uncertain payoff; and both laband industry researchers can benefit from sharingideas with each other. This approach implies thatFederal labs should make some of their R&Dchoices at least partly on the basis of their usefulnessto industry.

In some instances mechanisms for promotingcommercialization of lab technology have workedwell. For example, industry has benefited from usingspecialized facilities at the Department of Energy’s(DOE’s) multi-program national labs (e.g., BrookhavenNational Laboratory’s Synchrotrons Light Source,and Sandia National Laboratories’ Combustion Re-search Facility in Livermore, California). However,there is a consensus among industry, labs, andgovernment agencies that technology from Federallabs with defense or basic research missions is beingcommercialized much too slowly, despite the legis-lation that Congress passed throughout the 1980s toencourage such commercialization.

On consideration, this result is not surprising.These types of activities are difficult even when onlyindustry is involved. Firms with much in commonhave difficulty in agreeing on cooperative researchprojects, and it is even difficult to transfer technol-ogy from a firm’s central R&D facility to that firm’sown plants. Government-industry interaction is stillharder. It requires a fundamental reorientation onboth sides, since traditionally the Federal Gov-ernment and industry have opposed or ignored each

other. In particular, the Federal labs and their parentagencies must address many difficult issues involv-ing conflicts of interest, fairness to fins, nationalsecurity, and proprietary information. Labs also facethe formidable obstacle that U.S. firms are oftenslow to take advantage of new technologies devel-oped outside the firm (see ch. 6). When no firmexpresses an interest in a particular technology, it isdifficult for the government to identify those firmsthat could benefit--especially since the governmenttraditionally has not been skilled at marketing.Moreover, even if a lab finds a firm interested in itstechnology, negotiations can bog down because ofbureaucratic inertia and because government agen-cies often do not understand industry’s businessconstraints.

In the 1980s, Congress encouraged the labs toinclude technology transfer in their main missions .58Congress also authorized lab-industry cooperativeR&D,59 but made no special appropriations for it,apparently hoping that it could be supported withinexisting program budgets. This approach has oftenfoundered, for several reasons. Lab and agencypersonnel often consider the promotion of commer-cialization an improper distraction from the lab’sprimary mission. Agency security offices makeconservative rulings on what information can bereleased, general counsels are equally conservativeon which lab-industry arrangements are legallypermissible, and these rulings often actively inter-fere with the labs’ efforts to work with industry. Andin general, Federal labs and agencies face theinevitable problem of institutional inertia, a seriousbarrier to the new practices required for improvedlab-industry cooperation. Such a climate can stoplabs from working with industry unless there is astrong supporting voice within the agency.

Congress could provide stronger incentives forlab and agency personnel to help commercializetechnology. In practice, this probably means ear-marking money for promoting commercialization.Those who administer such money will want tospend it, and those who spend it will be evaluated onthe technology that was commercialized. Congresscould also remove some obstacles, including agency

SsFor exmple, in the Stevenson-Wydler Technology Innovation Act of 1980 Congress declared the @icy that ‘the Federal Government sh~l s~vewhere appropriate to transfer . . . technology . . . to the private sector. In the Federal Technology Transfer Act of 1986 Congress added that“[t]ednologytransfer, consistent with mission responsibilities, isaresponsibility of each laboratory science andengineeringprofessional.” 115 U.S.C.3710(a).

59FXSl Technology Transfer Act of 1986, 15 U.S.C. 3710~


red tape and legal problems with granting exclusiverights.

Earmarking Money for PromotingCommercialization

Most labs (or programs within labs) with missionsof either defense R&D or basic research do littlecooperative R&D with industry. Congress couldmandate that some part of the labs’ budgets be spentonly on cooperative projects with industry-perhapsrequiring equal matching funds from industry. Apossible model is DOE’s high-temperature super-conductivity pilot centers in three multi-programnational labs, which are collectively spending sev-eral million dollars only on R&D that industryproposes and cost-shares. Congress might start at afew percent of a lab’s total budget, and depending onexperience increase that amount to perhaps 10 to 20percent. Since cooperative R&D opportunities mustbe seized quickly, labs and agencies would need ageneral pool of money to apply as they saw fit tocooperative projects, without going through a budgetcycle to justify each project individually. Congresscould also provide stable multi-year funding to givefirms the confidence to enter into long-term projects.

Requiring certain money to be spent on collabora-tion with industry would change the labs’ missionssomewhat-or at least add to their missions acontribution to the commercial part of the economy.If the labs’ budgets were not increased, then theiroriginal missions might suffer. However, it mightnot damage a lab’s original mission to choose asmall fraction of its research projects on the basis ofrelevance to industry’s interests and needs; some ofthese projects might still be in some way useful forthe mission goals. In any case, Congress might deemit worthwhile to target some fraction of FederalR&D money to projects that have a good chance ofleading to commercialization.

Transfer of existing technology to industry alsorequires money. Activities include identifying ap-propriate technologies, patenting them as needed,marketing them, and in some cases giving startupfirms some support (e.g., office space, help in

writing a business plan, access to venture capital) toexploit lab technologies. Congress has directedagencies to set aside ‘sufficient funding, either as aseparate line item or from the agency’s research anddevelopment budget” to accomplish technologytransfer and to provide annual reports on past andplanned technology transfer activities.60 Congressmight wish to conduct oversight hearings to makesure that sufficient funds are being allocated. Alter-natively, Congress might mandate required fundinglevels. 6l

Congress could also increase the funding of theFederal Laboratory Consortium (FLC), currentlyabout $1 million per year. 62The FLC, with volunteerrepresentatives from over 300 labs and a smallcentral staff, functions for firms as a single point ofinquiry or entry into the Federal lab system. Addi-tional full-time staff would help the FLC meet itsgoal of matching an inquiry with an appropriate labresearcher within 1 day, and would also give theFLC more continuity. With its current reliance onvolunteers from the labs, the FLC inevitably suffersfrom high turnover of personnel. (Full-time staffmight be recruited from the labs’ ranks; they wouldthen be familiar with the labs.) Additional fundingwould also let the FLC pursue more projects todemonstrate new ways to facilitate commercializa-tion.

Congress might also designate funds specificallyfor facilitating personnel exchange. Currently, it isuncommon for industry researchers to take visitingpositions at Federal labs, and the reverse is quiterare. Subsidizing visiting positions from a specialfund would provide an extra incentive for the firm,the Federal lab, and/or the researcher. The fundcould at least be used to ensure that the researcher’spension benefits continue to accrue during his visit.

Removing Obstacles

Before undertaking either to commercialize exist-ing Federal lab technology or to perform cooperativeR&D with a Federal lab, firms often require exclu-sive rights to the technology; otherwise their invest-

~ational Competitiveness Technology Transfer Act of 1989, Public Law 101-189, Sec. 3133(e) (amending 15 U,S,C, 3710(b)).61Befo~ w p~~e of & National Competitiveness Technology Transfer Act of 1989, agencies were directed to set aside one-half percent of their

R&D budgets, though agency heads could waive this amount and some did. Stevenson-Wydler Tedmology Innovation Act of 1980, Public Law 96-480,sec. 11, amended and renumbered as sec. 10 by the Federal Technology Transfer Act of 1986, Public Law 99-502, sees. 3-5,9(e)(l), codifkd at 15 U.S.C.371O(I3).

~he FLC’S complex funding is set out at 15 U.S.C. 3710(e). Funding is set to expire aik FY 1991.


ment will not be worthwhile. Labs often face severalobstacles in granting these rights.

First, there is red tape while the labs’ parentagencies review the agreement. This is a seriousproblem, since delay can kill a deal. In 1986Congress permitted agencies to delegate to govern-ment-operated labs the power to make agreementsfor licensing and cooperative R&D (subject toagency veto within 30 days).63 In April 1987,President Reagan by Executive Order directed allagencies to do so,64 but it took many agencies untilwell into 1988 to comply and two (NASA and theNavy) still had not complied late in 1989. Congressmight wish to make the delegation mandatory andautomatic by statute. In December 1989 Congresspassed legislation permitting a similar delegation tocontractor-operated laboratories.65 Congress mightalso wish to make this delegation mandatory, and/orto conduct oversight hearings to determine whetherthe situation has improved for DOE’s contractor-operated labs, which have often experienced longdelays in getting approval for cooperative R&D.

Some of DOE’s labs have also been handicappedby having to negotiate with DOE for patent rightsbefore they can grant such rights to a firm. Currently,with certain exceptions, DOE’s labs run by non-profit contractors can automatically take title topatents from lab research;66 Congress may wish toextend that rule to include labs run by for-profitcontractors as well, and narrow the exemptions—allwith appropriate safeguards such as requiring royal-ties to be used within the lab.

Another legal problem concerns copyright. Underthe law, works created in whole or in part bygovernment employees cannot be copyrighted. Thisprohibition applies to software created at government-operated labs. Congress might wish to change thelaw to allow a copyright for such software, so thatfirms will have more incentive to commercializesoftware from these labs (commercializing it usuallyrequires substantial further development work) andto engage in cooperative R&D that will producesoftware. Congress might also wish to clarify thatDOE may maintain secrecy for software or otherdata developed cooperatively.

Lab-industry cooperation raises legal issues notonly about exclusive rights, but about many othersubjects as well, such as potential conflicts betweena researcher’s duty to the government and his desireto get personal gain from consulting, royalties, or acontemplated startup firm. To encourage generalcounsels to overcome their caution, Congress mightestablish an interagency legal task force for lab-industry interactions. If a general counsel feltuncertain about a proposed arrangement, he could ifhe wished submit the question to the task force,although the task force’s approval would not berequired.

University-Industry Collaborations

The National Science Foundation created Engi-neering Research Centers for several purposes: 1) tointegrate different engineering disciplines in R&Dprojects that are useful to industry and improve U.S.competitiveness; 2) to encourage cross-disciplinarytraining of engineers; 3) to improve relations be-tween university and industry researchers; and 4) togenerate strong participation from industry in re-search, education, and funding.

Early reports from this relatively new program(begun in 1984) indicate progress toward these goals(see the section on ERCs in ch. 7). In particular, theearly returns suggest considerable success in the keyobjective of educating engineers in several disci-plines. NSF is monitoring the centers closely to seethat their research is cross-disciplinary, is useful toindustry, and gives engineers a broad education.Under this scrutiny, 2 of 18 centers have lost theirNSF funding.

The two basic options with a program that seemsto be going well are to leave it alone or to expand it.In favor of leaving it alone is the argument theprogram is still experimental and all the results arenot yet in. In any case, the Federal Government isstrapped for funds. The strongest argument in favorof expansion is that a bigger program could producemore engineers with the kind of cross-disciplinarytraining that manufacturing needs. The vast majority

@ls USC. 3710a.6’$Ex~utive &&r 12591, l%cilitating Access to Science and Technology, Apr. 10, 19W, sec. 1. Pa. b(l).~~lic IAW 101-189, Sec. 3133(a).

W35 U.S.C. 202(a).

Chapter 2--Strategies To Improve U.S. Manufacturing Technology: Policy Issues and Options . 65

of U.S. engineering students take no part in theprogram. 67

As noted above, in the section on human re-sources, one way to increase support for manufactur-ing R&D and education in universities is to create aManufacturing Sciences Directorate in NSF. Inaddition, a much broader program of support formanufacturing R&D in universities might be one ofthe things a Civilian Technology Agency could do.(See the section below on Strategic TechnologyPolicy.)

Tapping Into Japanese Technology

Government-sponsored programs to encouragetransfer of technological research from Japan to theUnited States are of two main kinds: sendingresearchers to Japanese laboratories (people-to-people exchanges) and scanning the technical litera-ture. Federal programs of both kinds are quite newand still small; they have not yet come near theirpotential as a source of technological advances. Bothwould thrive better if more Americans learn to readand speak the Japanese language.

People-to-People Technology Transfer

NSF programs to promote long-term research byAmericans in Japanese labs were established byexecutive action. Congress has not enacted any lawsfor this purpose, other than including in the 1988trade act a direction to U.S. negotiators to ensuresymmetrical access to technological research.68 Asnoted in chapter 7, new government programs tosupport U.S. engineers and scientists doing long-term research in Japan, established in 1988 by theJapanese Government and the National ScienceFoundation, were not fully subscribed in 1989-90.There is reason to believe these programs will havemany more applicants within a few years, sinceprivately sponsored programs to send researchers toJapan have grown fast after a gestation period of afew years. Congress may wish to monitor theprogress of the Japanese government and NSFprograms, with an eye to supplementing them ifapplications multiply and, at some point, expansionis needed.

Meantime, another option would be to establish aCongressional U.S.-Japanese Fellowship Program,taking advantage of the prestige that the sponsorshipof Congress confers. Congress might also wish toencourage researchers working in Federal labs toundertake long-term projects in Japan. In oversighthearings, Congress might suggest that agenciesencourage sabbaticals for this purpose. For example,the three national labs that have pilot centersworking on lab-industry collaborations in high-temperature superconductivity might be able to sendsome of their people to the Japanese nationallaboratories, MITI facilities, or university labs thatare giving high priority to basic and applied researchin this field. A modest but useful initiative that NSFmight undertake would be to put together in oneplace information on all the programs, public andprivate, that offer U.S. researchers the chance towork in Japan.

In addition, Congress might consider establishinga program of post-doctoral or midcareer commercialfellowships in Japan, open to people other thanscientists and engineers, for example, economists,business administration graduates, and experiencedbusiness managers. The program might identifypositions in Japan that would enrich the fellows’understanding of Japanese management techniques,industry practice, and government-industry rela-tions. For example, positions might be found inJapanese Government agencies, in banks or securi-ties companies (whether Japanese or foreign-owned), or possibly in Japanese manufacturingcompanies. As with exchanges of scientists andengineers, any such program would have to startsmall and build gradually as U.S. candidates find outabout the program and learn enough Japanese toprofit from it.

Scanning Japanese Technical Literature

In the Japanese Technical Literature Act of 1986,Congress took steps to encourage the transfer oftechnology through the written word. The Office ofJapanese Technical Literature, set up under the actin the Department of Commerce, keeps up with newtechnical developments in Japan and publishesinformation about abstracts and translations ofJapanese technical literature. The office is small,

bTAt four ERCsex~~by OTA in visits and interviews, only about 1 percent of engineering undergraduates and4 to 11 IXXentof Wdwe ~ud~~took part inthe ERC program. Only 18 universities have ERCs (two of these are being discontinued but two were added in January 1990); this compareswith 280 colleges and universities in the United States that offer engineering education.

@Khnnibus Trade and Competitiveness Act of 1988, Public Law 100-418, Part II., SW. 5171.


operating with two people on an annual budget of$425,000.

If Congress wishes to take further steps to helpresearchers penetrate Japanese technical literature, itmight wish to increase the appropriation for theoffice. Possibly, ” the office could collaborate withprivate services that offer abstracts and evaluationsof Japanese technical literature and, on demand,translations. Because these services are expensivebut not very familiar to potential users, the Officemight consider offering users such as NSF granteesor industrial subscribers partial, temporary subsi-dies. This would get users started, and allow them tojudge the value of the services before they have tomake full payment.

Japanese Language Studies

The ability to read and speak the Japaneselanguage is fundamental to transferring technologyfrom Japan, both through people and throughpublications. The best way to learn languages is tostart young. Congress has already taken a steptoward getting Japanese language instruction in thepublic schools. The 1988 education act authorizedFederal grants of up to $20 million a year to helpfinance model foreign language programs.69 Theprogram supports instruction in “critical foreignlanguages,” as defined by the Secretary of Educa-tion. Congress might wish to oversee the programand evaluate whether it gives the study of Japaneseenough weight.

Congress might also wish to support an expansionof Japanese language programs at the college leveland beyond. The NSF language courses for scientistsand engineers-are getting an eager response, but arequite small-limited to 100 or so people a year-andare at the post-graduate (mostly post-doctoral) level.One option would be to fund a larger program of thiskind. Another would be to encourage the study ofJapanese at the undergraduate level, perhaps byproviding NSF fellowships for engineering under-graduates who want to study Japanese.

Antitrust Law

Antitrust law has a long and honorable history inthis country. It has been used to dismember monopo-lies (Standard Oil), induce dominant firms to yieldentry points to smaller firms (unbundling of IBMcomputer hardware and software), and open manyfields to innovative newcomers. In recent years,however, as international competitors have tight-ened the screws on domestic fins, some peoplehave questioned whether traditional tough enforce-ment of antitrust laws is still appropriate or wise.

In fact, antitrust law and enforcement have beenrelaxed in the past decade. Congress amended thelaw to make it easier for firms to get together forcooperative research or to form export tradingcompanies. The Reagan Administration was gener-ally considered less aggressive in antitrust enforce-ment than previous administrations. And the Federalcourts have interpreted the law in less stringentways.

Nevertheless, the antitrust laws may still detersome cooperation among firms that could help theircompetitive performance. Firms sometimes hesitateto undertake such things as joint R&D or manufactur-ing, cooperation to set voluntary industry standards,or simple sharing of information, for fear they willrun afoul of the antitrust laws. This is especially trueof cooperation among firms in the same business.Generally, the problem is not so much that thecooperation would actually violate the law, as thatthe law is unclear and penalties of misinterpreting itcan be severe. Thus, firms often shy away fromactivity that runs even a small risk of being deemeda violation.

To minimize these effects, Congress could bylegislation clarify and modify the legal standard forpermissible activities and change enforcement pro-cedures and penalties. It should be possible to draftsuch changes in the law without letting down ourguard against anti-competitive activity. Several billspending in Congress attempt to strike a properbalance by changing the law in certain limitedcontexts.70

@Au_F, [email protected]~ ‘Z S@ffgrd Elemm~ and Secondary School Improvement Amendments of 1988, Public Law 100-2fJ7, Tide IL PWB,

T%= bills inclu~ the Joint Manufacturing op~unities Act, H.R. 423; the National Cooperative Innovation md Commercialization A@ H.R.1024; the National Cooperative Reseamh and Reduction Amendments Act, H.R. 1025; the High Definition Television Competitiveness Act, H.R. 1267;the Cooperative Productivity and Competitiveness Act, H.R. 2264; the Advanced Television Competitiveness Act, H.R. 2287; the High DefinitionTelevision Development Act, S. 952; and the National Cooperative Researeh Aet Extension Act, S. 1006.


The Legal Standard

One uncertainty in antitrust law is whether anactivity will be judged using the rule of reason, underwhich activities are permissible if pro-competitiveoutweigh anti-competitive effects. Under the Na-tional Cooperative Research Act of 1984,71 jointR&D (as defined in the Act) is always judged underthis standard.

Joint manufacturing, cooperative manufacturingand marketing by small fins, and standard-setting,which in general are more likely to have anti-competitive effects than joint R&D, were notincluded in the 1984 Act. While the rule of reasonwould normally be applied to these activities as well,it is not clear that in all cases the pro-competitiveeffects will be fully considered. Congress couldclarify that the rule of reason applies in thesecontexts as well.72 This clarification would changethe existing legal rules (as interpreted by the courts)little if at all. It would remove doubt as to what therules are, and (especially if accompanied by congres-sional findings) would signal courts to take seriouslythe potential benefits of cooperation.

Congress “could also establish safe harbor marketshares, below which no violation would be found. Inpractice, antitrust violations are now rarely found ifthe firms involved have a combined market share ofunder 20 percent. Establishing a safe harbor at thatlevel would not change the law much, but wouldsimplify and clarify it. Firms with less than 20percent combined market share could proceed with-out fear; if sued they could get the lawsuit dismissedearly on. However, the measure would not applyautomatically to all firms claiming to fall below the20 percent limit; they might still be judged to havea greater combined market share, depending on howthe court defined the relevant market.

Antitrust law sometimes makes it difficult forU.S. firms to merge or form joint ventures to resiststrong actual or threatened foreign competition. U.S.firms do not get any special lenient treatment in thiscontext, because our antitrust law, as a matter ofprinciple, is nationality-blind (U.S. and foreignfirms are treated equally).

Congress might be reluctant to introduce nationalbias into our antitrust system. Yet even within anationality-blind framework, antitrust law could bemade more sympathetic to mergers or joint venturesof domestic firms under threat of foreign competi-tion. By law, Congress could instruct the Federalenforcement agencies and the courts to take along-term view and to listen seriously to factualarguments in particular cases that U.S. firms’ joiningforces will ultimately promote competition in theU.S. market.

For example, it might be argued that foreign firmscurrently having little share of some particular U.S.market will capture all of it in a few years, unlessU.S. firms in the same industry merge or form a jointventure to resist the foreign competition. Althoughthe merger would reduce the number of U.S.competitors in the short run, the number would begreater in the long run--e.g., one instead of none. Asa further example, it might be argued in a particularcase that competition in the U.S. market cannot beachieved without a healthy U.S. industry. Forexample, the exit of most U.S. firms from themerchant DRAM market in the mid- 1980s left U.S.computer firms exposed to high prices from foreignDRAM producers. Also, there is some evidence thatU.S. computer and semiconductor firms that dependon foreign, vertically integrated competitors forcritical components or equipment are last in line forthe latest technology.7 3 A j o i n t v e n t u r e o r m e r g e r

that has primarily anti-competitive effects in thenear term might be necessary in the long term tomaintain a healthy U.S. industry.

Both of these examples involve arguments thatU.S. firms in principle can make now in antitrustsuits. However, enforcement agencies and courts arelikely to reject such arguments as based too much onspeculation about the future. Congress could bolsterthe arguments by writing into legislation: 1) findingsthat scenarios like those described above can hap-pen, and 2) a direction that the law should be appliedto enhance competition in the long term.

71~bli~ ~w 98462, 15 U,S$C. 4301 -4305,”72HOR. 1025 ~~~dd~ ~ forjo~t ~an~actfing andm~keting; H. 1024 would do so forjointmanufacturing andmarketingto exploit R&D conducted

jointly or by one or more of the participants; H.R. 423 would do so for joint manufacturing and marketing by small businesses with at most 20 percentcombined market share; H.R. 2264 and S. 1006 would do so for joint manufacturing, but not joint marketing.

‘%3X Ch. 5.


Enforcement Procedures and Penalties

Federal antitrust law can be enforced both by thegovernment and by private parties. Successful pri-vate parties are awarded treble damages, plusreimbursement of reasonable attorney fees. Theseheavy awards in private suits increase the risks tofirms undertaking cooperative ventures; in particu-lar, these awards encourage private parties to filelawsuits even when they have weak cases, in thehope of extracting a payment to settle the case.

Some analysts believe that few private antitrustsuits are justified and have concluded that privateenforcement should be eliminated. However, thatwould leave enforcement of Federal antitrust lawtotally up to the Federal Government, which mightnot have the resources or the will to police the wholecountry effectively .74

A less extreme approach would be to award onlysingle damages in private antitrust suits. This is theprovision of Japanese and EC law.75 Even withsingle damages, Federal antitrust law would stillhave stronger enforcement provisions than mostother U.S. laws, as it includes both public and privateenforcement, attorney fee awards in private suits,and permission to States to sue on behalf of theircitizens.

Congress has taken some steps toward removingtreble damage provisions. Under the National Coop-erative Research Act of 1984, R&D projects (asdefined in the Act) registered for publication in theFederal Register are subject only to single damages.Congress is now also considering bills to allow onlysingle damages for registered cooperative manufac-turing ventures, registered cooperative manufactur-ing and marketing ventures, or registered coopera-tive manufacturing and marketing ventures by smallbusinesses with at most 20 percent market share.76

It might make sense to remove treble damagesonly for projects registered for public disclosure,because anti-competitive activities threaten compe-

tition less when they are disclosed to the public.(Treble damages might be needed to discouragefirms from secret, clearly anti-competitive activitiesthat might not be discovered. Disclosure enablesothers to quickly file suitor monitor the project.)

However, selective removal of treble damagesmight be only partially effective. Some companiesmight shun registration because it could give awaystrategic information, and it involves some extraexpense as well, including the need to amend theregistration if the project’s scope changes. If thereduction to single damages covers only certainactivities (e.g., as in the bills described above), firmsmight have trouble predicting whether certain activi-ties are covered. Adoption of single damages for allactivities would afford simplicity and certainty,although it could make the law less effective atdiscouraging some anti-competitive conduct.

A middle ground might be to adopt singledamages for certain registered activities and also inindividual cases where the accused firm can show itacted in good faith. Good faith might be shown, forexample, by an opinion from counsel, or by the factthat the firms had a reasonable (albeit losing)argument that their activity would pass muster underthe rule of reason. If treble damages were reservedfor the relatively rare egregious cases, the risks ofinter-fro cooperation would be less, and privateparties would have less incentive to file suit withweak cases .77

Another option, which could complement thesingle damages approach, is to let firms apply to thegovernment for advance certification that a proposedactivity is permitted. The Export Trading CompanyAct of 1982 followed this approach for exporttrading companies.

78 One bill before Congress takesthis approach for joint manufacturing and marketingthat exploits R&D results.79 So long as firms staywithin the scope of the certification, they could notbe sued for damages or penalties, either by the

T4s= for exwp~e, Report of the American Bar Assoc@ion Section of Antitrust Lxzw, Tmk Force on the Atiitmt Division of the U.S. Dep~tme~of Justice, July 1989, pp. 52-55 (finding that the Antitrust Division of the Department of Justice has inadequate resources and low morale).

Ts’rhom~J~de~d David T~e, “Innovation, Cooperation and Antitrust: Balancing Competition and Cooperation, ’ High IWvwlogyLawJowWvol. 4, No. 1, spring 1989, p. 56 and fmmote 157. EC antitrust law applies only in certain circumstances; in other cases, the member states’ own antitrustlaws apply.

715H.R. 2264 and S. 1006, H.R. 1025, and H.R. 423, respectively.77A ,similar rule exis~ for patent infringement. Treble damages may be awarded, but o~y in egregious c-.T~bfic IAIW 97-290, 15 U.S.C. 4001 et seq.

79H.R. 1024.


government or private parties. At most, they couldbe ordered to stop what they were doing.

Advance certification gives greater protection tofirms than just replacing treble with single damages,but could be costly and time-consuming. Presentprocedures for non-binding approvals from theJustice Department and the Federal Trade Commis-sion often take several months and require consider-able attorney time. Certification would be mostuseful if, at least in simple cases, a firm could applyfor one without assistance of counsel, and it could beissued within weeks, not months.

Innovation and Intellectual Property

Many concerned with our manufacturing compet-itiveness would put stronger intellectual propertyprotection worldwide for new technology (includingpatents, copyrights for software, and trade secretprotection) near the top of their list. Strongerprotection, it is argued, rewards invention, which isan American strength, and by encouraging R&Dwould make U.S. products more competitive. Also,it would discourage foreign firms from imitatingU.S. firms’ new products and processes—thusprotecting sales of U.S. firms, making them strongercompetitors in the present and better able to supportlong-term development for the future.

It is not clear, however, that stronger protectionalways encourages more R&D. And it is not clearhow much stronger protection would help increaseU.S. fins’ sales. There are limits, for example, tohow far we can push developing countries to goalong with stronger protection, since they do not seeit as to their advantage. From their point of view, itwould make their people pay more for foreign goodsand stop their firms from taking advantage of foreigntechnology. More fundamentally, patents and otherforms of protection for technology usually provideonly a temporary edge, until competitors find orinvent an alternative way to get the job done. A surerway to competitive success over the long run is toimprove the cost and quality of U.S. manufacturedgoods.

Nevertheless, some changes could improve theintellectual property environment. First, certainfeatures can be corrected in the United States-arelatively easy thing to do, since it can be doneunilaterally. These improvements at home mattersince the United States remains the most importantmarket for most U.S. firms today. Measures requir-

ing international negotiation can also be usefullypursued. These concern not only the substance oflegal rights but also the procedures for enforcingthem. If intellectual property law is poorly enforced,then even strong-sounding legal rights do notamount to much in practice.

Protection of Patent Rights in the United States

Prompt enforcement of patent rights is the mosturgent need for improvement of intellectual propertyprotection in the United States. Patent cases that goto trial take an average of over 21/2 years beforeending in a decision. During this time the firm withthe patent loses sales and must pay legal bills. Somefirms might not make it to the end of the trial. Evenif a firm survives and prevails at trial, compensationawarded by the court might not fully makeup for theharm caused by the infringer. (However, recent courtdecisions show particular concern to provide fullcompensation when possible, and also show willing-ness to find special circumstances justifying trebledamages or an award of attorney fees.)

One way to speed up patent infringement trialswould be to designate special judges for patentcases. At present, patent cases are normally heard byU.S. district court judges, who often have littleexpertise in patent law. Congress could encourage orrequire district courts to designate certain judges tohear all patent cases. They could be chosen for theirexpertise in patent law or build it up with experience.

This approach would conflict with the philosophythat Federal judges should be generalists. However,specialist Federal judges are not without precedent.Since 1982 the U.S. Court of Appeals for the FederalCircuit has handled all appeals in cases arising outof patent law and in certain other specialized areasof the law since 1982. That court is credited withbringing order and predictability to patent law.Because patent law is hard for the uninitiated tograsp, it seems a good area of the law for specialistjudges. (If the Federal Circuit is any guide, specialistjudges also tend to favor patent owners.)

Congress might also consider increasing thejudicial manpower devoted to hearing patent cases.One option might be to increase the number ofFederal district court judges across the board (withthe option of designating some of them patentjudges); alternatively, Congress might instruct thecourts to advance patent cases ahead of other cases.However, our Federal judicial system in general


suffers from delay, and Congress might not believethat patent cases need extra judges any more than,for example, cases against drug dealers do.

In evaluating whether patent cases deserve aspecial claim on limited judicial manpower, Con-gress might consider that, in effect, extra judges havealready been assigned to hear patent cases, and thosejudges’ ability to handle cases quickly and compe-tently has been hailed as a great strength of ourpatent enforcement system. These are the fouradministrative law judges at the U.S. InternationalTrade Commission. They are assigned to hear casesof “unfair imports” under Section 337 of the TariffAct of 1930, as amended,80 most of which concernpatent infringement. Under Section 337, U.S. firmscan apply for an order to be enforced by the CustomsService which stops infringing goods from enteringthe country. The law mandates that cases be decidedin 1 year (18 months in a minority of cases declared“more complicated’ ‘)—much faster than the aver-age time for trial in Federal district court.

However, the General Agreement on Tariffs andTrade (GATT) ruled in 1989 that Section 337enforcement proceedings violate U.S. obligationsunder the GATT treaty, by discriminating againstforeign goods. This decision put pressure on theUnited States to change Section 337 procedures.However, it is hard to satisfy the objections of theGATT panel while keeping the advantages of: 1) aquick decision, and 2) an order which can exclude allinfringing goods (or all infringing goods fromcertain manufacturers), no matter by what route andby whom they are imported. The Office of the U.S.Trade Representative has been considering variousoptions, including handling all patent infringementcases in a special court, or allowing the Commissionto issue temporary exclusion orders which wouldthen be reviewed by a court with a full trial. TheAdministration may propose a solution along theseor other lines for consideration by Congress.

Protection of Patent Rights Abroad

The United States is engaged in bilateral andmultilateral negotiations to strengthen intellectualproperty protection abroad. Two important goals arechanges in Japan’s patent system and a unified worldpatent system.

●

U.S. firms find Japanese patents not very effectivein stopping imitation by Japanese firms. Japan’ssystem is slower than ours in issuing and enforcingpatents, and it is strongly tilted toward licensing ofpatents (see ch. 7). Often, U.S. firms wish not tolicense patents to Japanese firms but rather toexclude them. The reason is fear of losing all theirsales in Japan, since Japanese customers stronglyfavor a Japanese supplier if one is available.Successful negotiations to change the Japanesepatent system could help some American firms holdon to sales in the rich and fast-growing Japanesemarket.

Besides the problems inherent in the Japanesepatent system, there is the added problem that manyU.S. firms are ignorant of how the system works.This ignorance sometimes extends to basic facts. Forexample, one firm did not know that after the initialapplication, a follow-up request must be made forthe Japanese patent office to examine the applica-tion. Congress might consider creating an office inthe Patent and Trademark Office to collect anddisseminate information about the Japanese patentsystem.

The second goal, creation of a unified worldpatent system, would help firms desiring patentprotection in more than one country. Currently, withsome exceptions, they must file separate applica-tions in each country. This is expensive, requiringlegal and translation services in each country. In aninternational patent system, one application wouldbe enough for a patent good in all participatingcountries.

A prelude to this long-term goal is the harmoniza-tion of different countries’ patent laws and applica-tion procedures. The United States has been negoti-ating to this end, especially with Japan and thecountries of the EC. Any agreement will probablyrequire substantial changes in our own patentsystem. For example, the United States now followsa first-to-invent system (in which the first person tomake an invention is entitled to a patent); we wouldprobably have to change to a first-to-file system (inwhich the first inventor to file an application isentitled to a patent), which almost all other countriesnow use. Also, the United States now keeps patentapplications secret; almost all other countries pub-lish applications after 18 months. In this too we

8019 U.s.c. 13370


would probably have to follow suit. While suchchanges might face strong political opposition inthis country, Congress may wish to consider themseriously if they are proposed by the Administrationas part of an overall treaty, containing importantconcessions from other countries.

STRATEGIC TECHNOLOGYPOLICY

In the past 40 years, and especially in the last 10,it has been an article of faith that governmentsupport of research and development should stick tobasic science, or else to the government’s ownneeds—mainly military security. Yet, governmentbacking for particular technologies seen as critical tothe nation’s economic progress is hardly unknown.The most obvious example is in agriculture. TheU.S. Government contributes well over $1 billion ayear to the Cooperative Extension Service foragricultural research and technology extension. TheService itself is 75 years old, and its origins go backstill further, to the foundation of the land-grantuniversities in the Merrill Act of 1862 and Federalfinding of State agricultural experimental stations,begun under the Hatch Act in 1887.

A venerable example from manufacturing is thecivilian aircraft industry. Established in 1915, theNational Committee on Aeronautics (NACA, laterthe National Aeronautics and Space Administration,or NASA) conducted or funded significant researchon airframe and propulsion technologies for years.NACA’s R&D typically went well past basicresearch, extending to pre-commercial proof ofconcept (tests of specific combinations of materialsand systems). The government’s decision in 1915 toback the aircraft industry with scientific and engi-neering R&D was grounded in the conviction thatthe entire nation had a stake in all phases of aviation,and that the country where powered flight wasinvented should be a leader in its continued develop-ment. The decision was made on patriotic, but notnarrow national security grounds.

After World War II, the idea took firm root thatonly defense needs justify government developmentof new technologies much beyond the basic researchstage. Although the government was the principalforce in the early development of computers andsemiconductors, both through R&D funding and

procurement, it did so in the name of defense.81

Sometimes the connection with defense was indi-rect. The Defense Advanced Research ProjectsAgency (DARPA), whose mission is to supportlong-term, risky research for national security needs,justified some of its computer R&Don the groundsthat, since the Department of Defense was a majoruser of computers, it would benefit in the end fromR&D that led to advancement of the technology inthe commercial sector.

A related argument was used recently to justifythe special government funding that semiconductorR&D is receiving. Alarm over the precipitous loss ofthe memory chip market to the Japanese led tourgent requests from U.S. semiconductor producersfor government R&D help. Congress respondedwith a contribution of $500 million over 5 years tothe Sematech consortium to improve the manufac-ture of DRAM chips, and put DARPA in charge ofthe government’s part in the project. The idea is thatmilitary security depends on a stable supply ofmemory chips from U.S. suppliers. Congress alsogave DARPA a total of $46 million in fiscal years1988-89 for R&D in materials, devices, and manu-facturing process technology for high-temperaturesuperconductivity.

The national security argument is wearing thin,however. As the military threat from the SovietUnion recedes, the economic challenges from Japan,the newly industrialized Asian countries, and aunified Europe loom larger than ever. In the publicdebates on government support for Sematech, high-temperature superconductivity, and lately on high--definition television (HDTV), the stakes in eco-nomic as well as military security got some frankrecognition. Not all parties agreed that our economicsecurity needs any bolstering from the government.But the stage was set for a new debate in which thegrounds for public support of technology advancecould shift.

Picking Winners

Government funding for R&D in semiconductortechnology, high-temperature superconductivity, andtechnologies for HDTV departs from usual U.S.policy since each of these projects concentratesmuch more on the applied than the basic end ofR&D. Indeed, the whole point of Sematech is to

glKenne~ F]-, “Gwcmment’s Role in Computers and Superconductors,” contractor report to the OTA, March 1988.


improve the manufacturing process for a particularproduct—the 16-megabit DRAM semiconductor.However, the recent cases are tentative and ad hoccompared to the steady long-term R&D support thatcivil aircraft manufacture has enjoyed, and thecombination of R&D and technology extension thathas been available to American agriculture sinceearly in the century, through the land grant collegesand the Cooperative Extension Service.

The widely accepted economic argument forselective but solid government support of commer-cially interesting technologies is that governmentshould share the risks of long-term, highly uncertainR&D projects in which the potential for benefits tosociety is great, but the payoff to individual firms islikely to be small and not worth the risk. In the U.S.financial environment, with its high cost of capitaland emphasis on short-term profit taking, the argu-ment for government’s sharing the risks of long-termR&D takes on special force.

The argument against giving selective support totechnologies that are vital to particular commercialindustries is mostly political. In brief, it runs asfollows: the American political system is pluralistic,disorderly, and open at so many places to influencefrom special interests that rational governmentdecisions on technology or industry policy are nextto impossible. The idea that government cannot“pick winners,’ and if it tries to will just bungle thejob, rests partly on this political argument and partlyon the simple claim that the market, for all itsfailures, is a better bet.

Politics probably interfere less in governmentsupport for R&D than in ventures more directlyconnected to commercial production, such as govern-ment-backed low-cost loans or purchase guarantees.Such ventures are likely to cost more than R&Dsupport, and are closer to the intensely political issueof jobs. Moreover, it is possible to erect safeguardsagainst ill-informed or political] y inspired choices oftechnologies for government R&D support. SharedR&D projects, in which industry takes part inselecting the subject and puts up at least half themoney, are one way for government to escapeblatant pressure from special interests and also to

enlist industry and market forces in the process ofpicking winners.

The record of the two industries that have receivedmost government support for technology advanceover the years belies the simple statement thatgovernment cannot pick winners. These industriescan hardly be described as failures. Until the recentchallenge from Airbus (which has had billions ofdollars in R&D and working capital support fromfour European governments) the U.S. air transportindustry was the undisputed world leader in technol-ogy, and it still produces a bigger trade surplus forthe United States than any other manufacturingindustry ($15.4 billion in 1988). Agriculture hascontributed trade surpluses for years ($16.4 billionin 1988) and is a technology leader as well. Laborproductivity on U.S. farms has increased more thanelevenfold in this century .82

The history of both industries suggests thatgovernment can not only pick winners but help tocreate them. (See box 2-A for a brief account ofgovernment support for the civilian aircraft indus-try.) Of course, there are failures too. For example,in 1980 Congress voted to create the SynfuelsCorporation that President Carter had proposed theprevious year, providing $20 billion in loan guaran-tees for plants making wood-based, coal-based, andshale-based substitutes for petroleum fuels, andprice guarantees for the output. Synfuels was one ofseveral initiatives designed to make the UnitedStates energy-independent, some of which stillcontinue today. But expectations that the SynfuelsCorporation would be able to produce fuels fromdomestically available feedstocks without addi-tional research and development were unrealistic, oilprices fell, and the Reagan Administration suc-ceeded in killing the program. Synfuels, it isgenerally conceded, was a failure.

Japanese industrial policies have missed the marktoo. Some examples of projects that did not achievetheir objectives include MITI’s effort to spur fastdevelopment of the biotechnology industry, the fifthgeneration computer project aimed at developingartificial intelligence, and the entry into the civilian

gzsme a~c~tu~ technologies develo- and disseminated by the Department of Agriculture, the land grant universities, and tie CooperativeExtension Service have raised labor productivity at serious cost to other values. For example, the overuse of broadscalepersistent insecticides in the 1950sand 1960s did much environmental damage, and in the end did not work because the target insects became resistant, secondary pests were released, andnatural predators were killed off. However, continuing R&D in the Federal-State agricultural research and extension system is working on saferapproaches to pest management.

Chapter 2-Strategies To Improve U.S. Manufacturing Technology: Policy Issues and Options . 73

aircraft industry with the YS-11 commercial trans-port.

Thus, there are examples of both success andfailure. The failures do not prove that government isinherently ineffective at fostering technologies ofinterest to particular industries. Said one DARPAemployee, “We defend our right to fail.” This is anessential right for anyone trying to develop some-thing new, whether it is new to the world, likeaircraft in the early 20th century, or new to a nation,like a commercial air transport industry was to Japanin the 1950s.

Another lesson may be learned from our limitedand uneven record of picking commercial winners;that is, if efforts are confined to crisis situations, theywill be more likely to fail than if a more proactive,strategic approach is adopted. Synfuels was con-ceived in 1979 when, for the second time in thedecade, oil deliveries from the Middle East weresharply curtailed for political reasons, prices shot up,shortages appeared, and anxiety over energy de-pendence was at a peak. Today, there is an air ofurgency over whether or how to support America’slate entry into the business of developing andproducing advanced television products. In a panicsituation, there is little time to construct or examineoptions or weed out the wilder ones.

Creating a Civilian Technology Agency

One option to help avoid the pitfalls of technologydevelopment by crisis is to establish a civiliantechnology agency. The last few years have broughtarising chorus of pleas by and on behalf of industriesthat are in danger, and it is likely there will be morein the future. If Congress wishes to respond to thosepleas in an organized fashion, it could benefit fromhaving an agency whose job would be to anticipatesuch developments, develop proactive options inresponse, avoid some crises, and improve thechances of responding well when they do arise. Thealternative is for Congress to continue responding adhoc—an option that some prefer, on grounds that

government support for commercial R&D should bethe exception, not the rule.

Congress has already established a small programthat might in time become a full-fledged civiliantechnology agency—NIST’s Advanced TechnologyProgram. Created in the 1988 trade act, the programgot its first funding, $10 million, in fiscal year 1990.The Program’s purpose, as stated in the law, is tohelp U.S. businesses apply research results to therapid commercialization of new scientific discover-ies, and to the refinement of manufacturing technol-ogies. The Program can assist joint R&D ventureswith technical advice or can take part in them—providing start-up funding or a minority share of thecost, or lending equipment, facilities, and people tothe venture.

In October 1989, the Senate passed a bill thatwould authorize the Advanced Technology Programto receive as much as $100 million funding per yearand gave quite specific directions on where to putthis R&D support.83 The bill directed the Program togive limited financial assistance to industry-led jointR&D ventures in “economically critical” areas oftechnology, and spelled out five areas that should getmost of the support: advanced imaging electronics,including advanced television; advanced manufac-turing; applications of high-temperature supercon-ducting materials; advanced ceramic and compositematerials; and semiconductor production equipmentfor the development of X-ray lithography.84

Other bills in the 100th and 101st Congresses,taking a broader but less directive approach for R&Dsupport of strategic commercial technologies, pro-posed to create an Advanced Civilian TechnologyAgency .85 It would be located in a new Departmentof Industry and Technology, replacing the Depart-ment of Commerce. The agency would make grantsto and cooperative agreements with R&D entities,with the government providing a minority share ofthe funding. The purpose would be to support highrisk projects with potentially great value to thecivilian economy that would otherwise lack ade-

S3S. 1191, entitled tie TdmoIogy Administration Authorization Act of 1989.

~ebill specified that $75 million of the $100 million should be available for these five areas, with individual projectstobe approved by the %cretatyof Commerce and the Directorof NIST; in reporting the bill, the Senate Committee on Commerce, Science and Transportation suggested specific amountsfor each of the five high technology areas. The bill also authorized $13 million for other technologies deemed of great economic importance by theSecretary and the Director; $10 million was reserved for small businesses with promising technologies; and $2 million was specified for programmanagement, analyses, and workshops.

8S@e of~e= bills, S. 1233 ~ the looth Cmgew, WaS reported out of the Senate Committee on Governmental Aff*, anached to tie 1998 W*act, and then dropped. Two similar bills, H.R. 3838 and S. 1978, were introduced in the IOlst Congress.


Box 2-A--Government Backing for the Civilian Aircraft IndustryAfter the Wright brothers flight at Kitty Hawk in 1903, the U.S. Government was slow to get behind

aeronautical research and development.l Twelve years went by before the creation of the National AdvisoryCommittee for Aeronautics (NACA), a U.S. Government institution whose purpose was to further the science andtechnology of aeronautics. Meanwhile, the Wrights (and some others, mainly Glenn Curtiss) had gone on buildingplanes and improving them, but with little research support. Most of the flying was left to barnstormers and stuntflyers, whose hijinks and appalling safety record did not help to commend aviation to serious research attention.The military services waited until 1907 to let their first contract for an airplane, and the first appropriation formilitary aircraft-$25,000 for the Navy-came in 1911.

At the same time, European governments were taking very seriously the possibilities opened up by the firstsuccessful powered flight. All over Europe, but particularly in France, Britain, and Germany, governments eitherestablished or contributed to aeronautical research centers. Advances came quickly. In July 1909 Louis Bleriot flewacross the English Channel. In the next couple of years, many new European planes emerged (Bleriots, Farmans,Antoinette), some demonstrating features such as allerons and monoplane design that were superior to the Wrights’designs,

Aviation enthusiasts in America were mortified. They “found it a national embarrassment-not to say adanger--that the country where aviation began should trail so far behind the Europeans.”2 By 1911, some of themstarted to campaign in earnest for a national aeronautical laboratory. They were not to succeed until 1915, whenTheodore Roosevelt endorsed the idea and the Congress looked on it with favor. Even so, the joint resolutioncreating NACA would have been lost in a close-of-session rush if it had not been backed by the powerful NavalAffairs Committee and tied to a navy appropriation bill.

NACA’s charge was to ‘‘supervise and direct the scientific study of the problems of flight, with a view to theirpractical solution, and to determine the problems which should be experimentally attacked.”3 By the 1920s, NACAwas an important contributor to R&D for the fledgling commercial industry. NACA pioneered in building and usinglarge wind tunnels, collaborated with both the civilian aircraft industry and the military on designing researchprojects, and made its test facilities and a stream of test results available to both throughout the 1920s and 1930s.

NACA boasted among its accomplishments the design, modeling, and testing of a family of airfoil shapes, sowell-characterized that designers could select wing sections for various purposes off the shelf. The famous NACAcowl, developed and tested in NACA’s propeller wind tunnel in the late 1920s, was credited with greatly reducingwind resistance in the then-standard air-cooled radial engine, cutting engine drag by 75 percent with hardly any lossin cooling. NACA research also helped to define optimal placement of the engine in the wing, thus contributing tomuch greater engine efficiencies and higher speeds, When airline cruising speeds rose from 120 to 180 miles perhour, overnight transcontinental runs became possible, and air travel boomed even in the midst of the depression.4

After World War II, NACA and its successor, the National Air and Space Agency (NASA) continuedaeronautical research and testing, but the aircraft companies were soon outspending them, and military R&Ddwarfed both.s However, the aircraft companies continued their close relations and collaborative research withNASA, and a liberal system of cross-licensing of patents (originally backed by NACA and continued under NASA)helped to diffuse technology advances throughout the industry.6 Technological spillover from military to civilianaircraft remained consequential at least through the 1960s. For example, the airframe design of the Boeing 707

l~ex Roland, Mo&l Research: The National Adviso~ Committee for Aeronautics, 1915-1958 (Washington, DC: U.S. @vernmentPrinting Office, 1985), vol. 1.

21bid., p. 4.s~blic Law 271, 63d Cong., 3d sess., Mar. 3, 1915, cited in Roland, Op. Cit., vOi. 2, p. 394.4Rol~nd, op. cit., vol. 1, pp. 92-94, 111-1 16; David C. Mowery and Nathan Rosenberg, ‘The Commercial Aircr* ~dus~~” Guver~@

and Technicul Progress, Richard R. Nelson, (cd,) (New York, NY: Pergarnon Press, 1982), pp. 128-129.5From 1945 t. 19w, tm~ R&D Spn&ng in the ~cr~t indus~, mili~ and civilian, w= $l@ billi~ (1W2 doll~), of which $81

billion was provided by the military, $18 billion by industry, and over $9 billion by non-military Federal agencies. David C. Mowery, “JointVentures in the Commercial Aircraft lndwstry,’’[nternutionul Collaborative Ventures in U.S. Manufacturing, David C. Mowery (cd.)(Cambridge, MA: Ballinger Publishing Co., 1988), p. 75, For a brief history of government R&D support for the civilian airmfl industry, seeDavid C, Mowery, “Collaborative Research: An Assessment of Its Potential Role in the Development of High Temperature Superemductivity,”comract report to the Office of Technology Assessment, January 1988.

%hma-licrxtsing was abandoned in 1975, due to the objections of the Antitrust Division of the Department of Justice,


passenger plane was such a clone of the KC-135 refueling tanker that Boeing made for the Air Force that the firstprototype 707 wheeled out of the Seattle plant had no windows in the fuselage.7 Boeing eventually made more than800 KC-135 tankers. Sharing development costs and moving down the learning curve together with its military twinbrought down costs for the 707 much faster than would have been possible otherwise.

The civil aircraft industry also benefited from other government policies besides NACA/NASA support forR&D. From 1930 to 1934, U.S. Government contracts with airlines to carry the mail included subsidies, and helpedto sustain demand for civilian aircraft during the depression. (Indeed, at that time, the major aircraft companies werevertically integrated with the airlines and with engine companies as well. The Air Mail Act of 1934 ended thesubsidies and forced dissolution of these vertically integrated firms.) In addition, regulation of airlines by the CivilAeronautics Board indirectly favored technology advance in aircraft manufacture. By ruling out price competition,the CAB encouraged the airlines to compete on performance instead, and thus indirectly supported the aircraftmanufacturers’ commitment to technological excellence.8

CAB regulation is now ended; the airlines are competing more on price and passing on competitive pressuresto aircraft manufacturers. And the civilian aircraft industry relies less than it did in the past on government R&D.The airframe companies-especially Boeing, which is far and away the biggest in the civil aircraft business---fundmost of their research and nearly all their development costs on the commercial side (engine companies still getsubstantial Defense Department funds for commercial projects that may have a military payoff).9 Also, spinoffs arefewer; civilian and military aircraft technology has increasingly diverged in the past 20 years or so, not only in theoverall product but to some degree in component technologies.

10 Nevertheless, NASA still spends a fair amounton generic aeronautical research and testing (about $350 million to $400 million a year), which complements theindustry’s private R&D and reduces its costs to this day.

7&@Wev and Ro~~krg, op. cit.> P. 131“

8Mowery, “Collaborative Research,” op. cit.9~ Coml=ion on industrial Productivity, “The U.S. Commercial Aircraft Industry and M Foreign Competitors,” The Working

Papers of the MIT Commission on lndu.mai Productivity (Cambridge, MA: The MIT Press, 1989), p. 16.

IOIbid., p. 17,

quate private support. The agency’s activities would It may be objected that DARPA is not appropri-be overseen by a 21 -member Board with at least 14from various industries and businesses, small andlarge, and the rest from State and local governments,academic institutions and nonprofit organizations.

A model that has sometimes been suggested for acivilian technology agency is DARPA. Establishedin 1958 (as ARPA-the D, for Defense, was addedlater), this small elite agency has gained a reputationfor flexible, impartial decisionmaking, and forintelligently placing its bets. It has of course, lostsome of its bets, and some have been a very longtime in paying off. For example, from its beginningDARPA has been a major supporter of research inartificial intelligence. Only in the early 1980s, after20 years of steady investment by DARPA, did thefirst commercial AI projects begin to emerge.86

ately compared to a civilian technology agency,since it has a military mission and can be heldaccountable to that mission. Yet, as noted above,DARPA has often interpreted its mission verybroadly. The Department of Defense buys on thecommercial market, and it benefits if that sectorexcels in technology, and suffers if it lags. And if thecommercial sector does lag, U.S. defense couldbecome too dependent on superior foreign produc-ers. The fact that commercial companies are sellingAI machines based on research that DARPA hasfunded for nearly 30 years illustrates how DARPA’sbroad interpretation of its mission can carry it wellinto the commercial side of the economy. (This is notalways the case; DARPA’s support for broad R&Dprojects with no obvious short-run military applica-

s~e f--st commerci~ AI machine was Xerox’s Interllsp work station, introduced in 1981. Although Xerox funded much of the developmentinternally, it also relied on DARPA projects and funds. By 1985, four U.S. firms were selling computers designed to program in the AI language LISP;all had direct ties to DARPA-funded research. Flamm, op. cit.


tions has waxed and waned, depending on budgetsand competing DoD demands. )87

The parallels between DARPA and a civiliantechnology agency go only so far. Choosing technol-ogies that must eventually prove their worth in themarket is tougher, even allowing for failures, thanchoosing ones for which there is some crediblemilitary use, so that at least one customer—thegovernment-is likely to materialize. Also, thechoice of technologies to support may lend itself topolitical pressure on the civilian more than on themilitary side (though decisions about military pro-curement are hardly free from the competing claimsof different regions and industries). A civilianagency would probably have to balance politicalpressures more deftly than DARPA is called upon todo, but the difference might be more a matter ofdegree than of kind.

A distinct difference is that a civilian technologyagency would need to interact much more closelywith industry than DARPA does in choosing tech-nologies to support, in the design of R&D, and injoint payment for R&D. Until very recently, withSematech and some small HDTV projects, DARPAhas not funded projects jointly with industry. AndDARPA staff members exercise a great deal ofindependent judgment about what technologies tofund.

Perhaps the biggest threat to the long term successof a civilian technology agency is exaggeratedexpectations. Technology push, even if planned anddirected intelligently, certainly does not guaranteesuccessful commercialization. One reason for thecontinuity and accomplishments of NACA/NASAsupport for the civilian aircraft industry is that it waslow-key and did not promise miracles. To restoreworld-class performance in U.S. manufacturingindustries will take much more than selectivegovernment support for technologies up to the pointof commercial production. Technology push is justone of the many things that must be done, byindustry and government alike.

Designing a Civilian Technology Agency

Any Civilian Technology Agency (CTA), whetherit develops from the NIST Advanced Technology

Program or is established more formally, wouldcertainly start small, and might remain so. DARPAhas a staff of 150, half of them in technologydevelopment and the other half in administration,and about $1.3 billion a year to spend on R&Dprojects. Too much smaller, and the agency wouldnot have a critical mass. Too much bigger, and itprobably could not operate in the anti-bureaucraticway DARPA does, which is to give each member ofthe technology staff almost total responsibility forthe areas he or she manages.

After a few years’ experience, a CTA might takeover some technology projects from other agencies,such as engineering projects of the National ScienceFoundation (e.g., the Engineering Research Cen-ters). But most of the big government technologyprograms now in existence are solidly ensconced intheir present homes (NASA, DOE labs, NationalInstitutes of Health). If a CTA were to grow, it wouldmore likely result from years of success andexpansion in its own line of work than fromreshuffling present programs. The bills in the 101stCongresses to establish an Advanced Civilian Tech-nology Agency in a new Department of Industry andTechnology propose a small agency, starting with astaff of 40 (primarily recruited from industry, ontemporary assignment) with a first year authoriza-tion of $100 million, rising to $240 million in thethird year.

A small agency funding technology R&D proba-bly works best if the staff members are not hemmedin by too many rules and guidelines, but can exercisetheir own good judgment. DARPA attracts itsexcellent staff by offering a combination of hardwork, low pay, great responsibility, and a chance todo something for one’s country. If a consensusdevelops that the foremost job for the Nation is tosecure our economic future, the chances would begood that a CTA could hold out similar attractions—with the difference that the staff would work muchmore cooperatively with industry. One caveat: thelow pay (relative to private jobs) that governmentcan offer to highly trained scientists and engineershas not stopped DARPA from getting good people,though they tend to leave when their children reach

STFor exmple, during the heyday of the Strategic Defense Initiative in the 1980s, DANA cut back on its support of broad advances in ComPutertechnology in favor of the Strategic Computing Program, which was a part of SDI. Funds were diverted horn universities responsible for the earlierprograms (and their eventual success) to military contractors. The emphasis changed from long-term open-ended resuhs to milestones and conc~tedeliverables.

Chapter 2--Strategies To Improve U.S. Manufacturing Technology: Policy Issues and Options . 77

college age. Low pay could be a greater handicap toa new agency just starting out.

Where in the government bureaucracy a CTA islocated may not matter much. Aside from thePresident’s own staff and the upper reaches of theOffice of Management and Budget, power in theexecutive branch of the Federal government is fairlydispersed. It is probably an advantage to the NationalScience Foundation to be independent of anydepartment. Yet DARPA, a tiny appendage to thebiggest and most hierarchical of all the Federalagencies, still makes its voice heard through sheercompetence and dedication.

Defining Goals, Choosing Projects

Desirable as it may be to give the CTA manage-ment and staff freedom from red tape in workingwith industry and choosing technologies for support,some explicit overall goals should serve as aframework for the choices. If Congress wishes toestablish a CTA, it might give the agency the duty ofdeveloping a set of goals, based on a more generalmission defined by Congress. For example, propos-als before the 101st Congress for a CTA defined itsmission as contributing to U.S. competitiveness by‘‘supporting generic research and development pro-jects . . . that range from idea exploration to proto-type development and address long-term, high riskareas . . . that are not otherwise being adequatelydeveloped by the private sector, but are likely toyield important benefits to the nation.”88 Similarly,S. 1191, the bill passed by the Senate in 1989 thataimed to beef up NIST’s Advanced TechnologyProgram, referred to “research that no one companyis likely to undertake but which will create newgeneric technologies that will benefit an entireindustry and the welfare of the Nation. In definingthe mission, it would be unwise to limit the supportonly to long-term, high-risk technologies with su-pernova potential. This could rule out catch-upprojects like Sematech, or projects for incrementalimprovements in technologies that are alreadywell-known, such as a next-generation controller formachine tools.

The “visions” that Japan’s Ministry of Interna-tional Trade and Industry develops in consultationwith industry for Japan’s economic developmentoffer an example of goals that a CTA might advance.

MITI’s current vision is for a knowledge-intensiveeconomy. This means support not only for technolo-gies important to Japanese industries that are obvi-ously knowledge-intensive themselves (e.g., com-puters) but also projects that deepen knowledgeintensiveness in traditional industries (e.g., theAutomated Sewing System, a 7-year $90 millionMITI project that brought together 28 textile,apparel, and textile-apparel machinery manufactur-ers in a cooperative R&D effort).

How the government’s fund should be dividedamong various broad areas of technology is the mostfundamental of the choices to be made. Howmuch—if any—should go to high-temperature su-perconductivity? How does high-temperature super-conductivity compare with competing claims fortechnologies important to computers, or advancedtelevision, or industrial robots, or advanced automo-bile engineering? S. 1191 was specific in directingNIST to support technologies in five particular hightechnology areas. The bills aiming to create a CTAwas less directive, leaving it to the agency to makethese choices, with the guidance of its advisoryboard. Thus, the CTA would have to pick winners;that would be the nature of business. While theagency would have the final responsibility fordeciding how government money should be spent ontechnology R&D, it would rarely choose to supporta technology that did not also have strong industrybacking, including a financial commitment.

Another point is that a CTA would need toconsider whole technological systems rather thanisolated bits of systems. For example, if (as is quitelikely) it should select semiconductor technologies,it would have to be mindful of R&D needs through-out the system, starting with improved materials forthe silicon crystals that are made into wafers, andcontinuing through such things as X-ray lithographyfor etching circuits on the wafers (including thewhole paraphernalia of a source for the X-rays,lithographic equipment, photochemicals, masks andsubstrates); automated techniques for packagingchips; advanced methods for placing chips on aboard and interconnecting them; and so on. As partof its strategic approach, the CTA should also lookfor technologies that are central to more than oneapplication. Examples are advanced displays and thetechnologies for manufacturing them, applicable toboth HDTV and computers; high-temperature super-

8SH.R. 3838 and S. 1978, part B—Advanced Civilian Technology Agency, SeC. 212(a).


conducting magnets, which could be important forseveral steps in semiconductor manufacture (e.g.,compact synchrotrons as a source of X-rays forlithography) as well as for such futuristic things asmagnetically levitated trains.

Once a technology is selected for support, thechoice should be given a fair chance. Just as the CTAcould provide a way to look ahead and makestrategic choices rather than react to the technologycrisis of the day, it could also impart steadiness.Continuity-a long-term, multi-year commitment—may be the most important benefit government hasto bestow on a risky undertaking.

Government-Industry Collaboration

The main reason for government to put moneyinto technology R&D of commercial interest is thatthe risks are too great for individual companies tobear. But if private companies are not interestedenough to take some of the risk and do some of thework, then the commercial potential may be veryremote. It might make sense for a CTA to reserve asmall portion of its funds for projects that are solong-term and chancy that they do not attract muchindustry support. But for the most part, if industry isnot willing to pay a hefty portion-usually at least40 to 50 percent—the projects are probably notworth pursuing. Other requirements for membercompanies could be willingness to put well-qualified employees on the project, carry on comple-mentary research, and make a fairly long-termcommitment, say 3 years.89

In the few collaborative projects that the U.S.Government has recently proposed or undertaken,industry participation has been no problem. All havehad enthusiastic takers. In the case of Sematech, itwas the semiconductor industry that did the propos-ing; the industry lobbied hard for the program.Member companies pledged to contribute 1 percentof their revenues, and they are paying about half ofthe costs. The three national laboratories with pilot

programs for R&D leading to commercialization ofhigh-temperature superconductivity have coopera-tive agreements with two dozen companies, allpaying half the costs of their projects, and still morecompanies want to join if the labs can find enoughmatching funds. When DARPA proposed to put up$30 million for collaborative R&D projects inHDTV 87 companies wanted in.

Sematech has its own facilities, but some government-industry collaborative R&D could take place in thecompany labs. Much more could be done in Federallabs, especially the Department of Energy’s well-endowed national labs. (See the discussion in anearlier section of this chapter on how DOE’s labs canbe made more hospitable to collaborative R&D.)

Since government money is involved and thepurpose is to bolster U.S. competitiveness, it maymake sense generally to limit membership in thesejoint government-private R&D projects to U.S.companies. Once again, the definition of a U.S.company would have to be settled, and conditionsfor foreign participation defined.90 European experi-ence may shed some light here, since government-industry collaboration on R&D is increasinglycommon in Europe. The European Community isspending over $1 billion a year on its Frameworkprogram (R&D collaborations with industry anduniversities). Also, 19 European countries plus theEC Commission collaborate with industry on applica-tions-oriented R&D under the umbrella organizationEUREKA. In both the Framework and EUREKAprojects, foreign-owned companies can often takepart provided they have an ‘integrated presence”--that is, research, production, and marketing-inEurope. Not always, however. Foreign-owned com-panies with an integrated presence have been ex-cluded from some of these R&D consortia (e.g., Fordand General Motors of Europe are excluded fromPROMETHEUS, a consortium working on ad-vanced transportation technologies). In some cases,the determining factor seems to be whether Euro-

89H.I?. 3838, S. 1978, and S. 1191 would all require R&D entities receiving funds from the Advanced Civilian Tdmology Agency ortie AdvancdTechnology Program to put in more money than the government contributes. Also, see ch. 7 for a discussion of the factors that make for success in R&Dconsortia, and favorable conditions forgovernment-industry collaborations. See also U.S. Congress, Office of Technology Assessment, CommercializingHigh-Temperature Superconductivity, OTA-ITE-388 (Springfield, VA: National Technical Information Service, 1988), pp. 133-37.

~,R. 3838 and S, 1978 (lOlst Congress) provide that “noproject which contains aforeigncompany orentity orasubsidiary thereof” shall reeligiblefor government financial support, unless the foreign company makes material conrnbutions to the projects; the foreign company makes a substantialcommitment to manufacture products arising from the projects’s R&Din the United States and to buy from North American suppliers; the home countryof the foreign company affords reciprocal treatment to U.S. companies; and the Secretary of the Department of Industry and Technology certifies (afterconsulting with North American participants in the project and the advisory board of the Advanced Civilian Technology Agency) that the foreigncompany’s participation is in the interest of the United States. S. 1191 contains similar provisions.


pean members want the foreign-owned companies inor out.

Beyond Technology Policy

Technology policy, even a strategic one, carriesgovernment involvement only so far-to the brinkof commercialization. After that it is up to industry.Of course, many governments, including our own,have gone farther than that in support of particularindustries seen as having a special importance to thenation. Among the industrialized countries, Japanhas probably gone farthest down this road. Twonewly industrializing countries, Korea and Taiwan,observing Japan’s success, have employed elementsof the same strategy, sometimes carrying it farther.

Japan and other Asian countries have combinednumerous policy tools besides long-term govern-ment support for technology R&D to promoteselected industries: preferential loans from govern-ment banks or banks that follow the government’slead; guaranteed purchases by governmental bodiesfor home-grown products (e.g., semiconductors forNippon Telephone & Telegraph, supercomputers forgovernment agencies); government-subsidized leas-ing companies making guaranteed purchases ofadvanced equipment and leasing them at preferentialrates (e.g., robots, CNC machine tools); formal orinformal barriers against imports, removed (or partlyremoved) only after the domestic industry hasbecome a world-class competitor; strict limits onforeign investment in manufacturing; governmentnegotiations for technology licenses on behalf ofindustry; government guidance (not always fol-lowed) to rationalize industries, scrap overcapacity,and encourage companies to get economies of scaleby specializing in certain parts of an industry (e.g.,machine tools).

This is industry cum trade policy on a comprehen-sive scale. Other nations have used some of theconstituent policies with greater or lesser success.For example, several European countries favor theirnational champion computer and semiconductorcompanies almost exclusively in government pur-chases. The members of the Airbus Industrie consor-tium get low-cost loans from their governments(France, West Germany, the United Kingdom, andSpain) and can wait to pay it back from revenues.

This is an enormous advantage in an industry whereit takes 10 to 14 years and at least 500 unit sales tobreak even on a new transport plane. The BuyAmerican act in the United States gives a priceadvantage to domestic producers. U.S. Governmentpurchases of semiconductors and computers werecritical to the success of those industries in theirinfancies (though it cannot be said that thesepurchases deliberately favored domestic producers,since there were hardly any other producers at thetime).

The next, and final report in OTA’s assessment ofTechnology, Innovation, and U.S. Trade will con-sider trade and industrial policies of Europe, Asiannations, and the United States in depth. This report,which focuses on technology, touches only lightlyon these matters, but it is relevant here to considerhow strategic technology policy relates to industrialand trade policy. The justification for government’sspending money on technology R&D—potentiallygreat benefits for society, coinciding with returns toindividual firms that are too small or remote tooutweigh the risk-could apply, in some situations,to commercial production. This is part of theargument for protection and support of infantindustries, especially ones where capital require-ments are extremely high or the manufacturingtechnology is complex and demanding, so that ittakes a long time to learn how to do it right and getcosts down. Both conditions apply, for example, tocivilian aircraft manufacture. According to the MITCommission on Industrial Productivity, “no avia-tion company has ever succeeded without govern-ment help,” though the form, degree, and timing ofhelp has differed.91

This kind of thinking has led to calls for govern-ment help to get U.S. companies into the business ofmaking consumer electronics items such as highdefinition TV that use advanced digital integratedcircuit semiconductors and have many core technol-ogies in common with computers (see the discussionof advanced television at the end of this chapter).One proposal is to set up a private corporation,backed by “pledges of support’ from Federal, Stateand local governments, to provide “low-cost, verypatient capital” to U.S. companies making ad-

91 MIT Commission on Industrial Roductivity, “The U.S. Commercial Aircraft Industry and Its Foreign Competitors,” The Working Papers of theMIT Commission on Industrial Productiviq (Cambridge, MA: MIT Ress, 1989, vol. 1, p. 16.


vanced consumer electronics products.92 Govern-ment backing of this kind would tilt the odds in favorof investing in consumer electronics. It is onepart-but a small part-of the package that adds upto industrial policy. A comprehensive public policyaimed at building up an industry for nationaleconomic security reasons would involve muchmore, and would probably include some aspects oftrade policy, such as domestic content requirementsor government negotiations on behalf of industry forforeign technologies. Opposition to such policies isbased on the idea that if government actions overridemarket signals, the result will be economic ineffi-ciency and high prices, the extreme case beingcentral control of the economy with shortages ofeverything people want, as in Poland.

A look around the world, however, shows thatsome governments have selectively helped indus-tries they consider crucially important to the nation,using a full panoply of technology, financial andtrade policies while still leaving the economy opento market signals. It is not an easy trick, and it iscertainly not cost-free. Japanese consumers, forexample, pay higher prices for some of the goodsthat Japanese industry excels in producing (e.g.,compact cars, color television sets) than do Ameri-can consumers for imports of the same products, andthis difference has something to do with governmentpolicy. Yet those same Japanese consumers areworlds better off than they were 20 or 30 yearsago-and this has something to with governmentpolicy too.

The last report in this assessment will take on thequestion of how industrial and trade policies in othernations have helped-or failed to help-their indus-trial advance, and which if any of these policiesmight be useful for the United States to try. In thisreport, we can say that, based on its limited use inthis country and more extensive application abroad,strategic technology policy offers some attractiveoptions for Congress to consider. This is the leastintrusive and least expensive of public policies toimprove the performance of industries seen ascritical to the nation’s economy, yet it has never yetreceived a broad trial in the United States. The

traditional U.S. science and technology policy,which shunned government support of commercialtechnologies, served well enough in the postwaryears when the United States was king of themountain. Now, with U.S. manufacturing in obviouscompetitive difficulties, it may bean opportune timeto try other approaches.

One Example of Technology Policy:The Case of Advanced Television

HDTV is an improved form of television, with alarger screen, more detail, and better color thanconventional TV. If that were all it is—a bigger,more alluring form of television for home entertain-ment—HDTV might not have become the front pagenews item and center of political controversy that itwas in 1989. But it is something more. Its require-ments could drive a range of technologies that haveimportant applications in other parts of the electron-ics industry—in particular, computers and telecom-munications. 93

There are two key reasons why technologicalspillovers from HDTV are likely. First HDTV’s coretechnologies-for production, storage, transmis-sion, processing and display of information-are inthe same family as those used in computers andtelecommunication devices. They are based ondigital electronics. Conventional TV and many otherconsumer electronics items depend mainly on ana-log electronics technology. (Box 2-B outlines thedifferences between digital and analog electronics.)

In some digital electronic technologies, HDTV isahead of computers. For example, one of HDTV’srequirements is the ability to process and displayhuge amounts of picture data very rapidly. Becauseof this, HDTV must advance the state of the art indisplay technology (also in fast processing, althoughthe chips and hardware being developed for thispurpose for HDTV are specialized). Computersdon’t yet need such advanced display technologybecause their general-purpose hardware is slower atgenerating data. As computers’ speed of operationincreases, they will be able to take advantage ofHDTV’s display technology, using it in such activi-ties as weather forecasting and computer-aided

‘%4 Strategic lti~ at Risk, a rep to the President and the Congress fmm the National Advisory Committee on Semiconductors (WttSh@On,DC: The Committee, 1989), p. 20.

Q3Much of tie maten~ ~ ws ~ti~ is ~a~ from a foficoming OTA report, The Big Picture: High-De finitwn Television and High Resola”onSystems, which provides a comprehensive account of HDTV’S history, technology linkages to other electronics industries, and relation to the U.S.communications infrastructure.

Chapter 2--Strategies To Improve U.S. Manufacturing Technology: Policy issues and Options ● 8 1

Box 2-B--Digital and Analog Data: Television Transmission

In electronics, information can take two forms, digital and analog. In the digital form, numbers representinformation; the numbers are generally written in the binary system, which has only two numerals, zero or one. (Thefamiliar decimal system has ten numerals, zero through nine,) Modern digital computers represent each binary digit,or bit, as a switch; if the switch is on, the bit is one, and if off zero. In computer calculations, numbers are simplynumbers, written in binary; e.g., 8 is 1000. Other data, such as letters of the alphabet, are converted to numbersaccording to a code, Letters usually take up eight bits; for example, the capital letter “A” is often denoted as thesequence 01000001.

In the analog form, information is represented by physical characteristics (e.g., distance or voltage) which varycontinuously. Traditional sound recordings are analog. Grooves in the record have tiny physical patterns that varycontinuously and correspond to the original sound. The needle of the phonograph arm rides over small bumps inthe grooves, which apply pressure to a crystal (or other pickup system) in the cartridge, which in turn generates avoltage that varies with the degree of pressure applied by the needle. The electronic signal thus generated is thenconverted back into sound. Compact disk recordings, in contrast, are digital. Sounds are recorded on an optical diskas small pits, representing zeros or ones, which denote various characteristics-frequency, volume, and soon-according to a prearranged code. In the disk player, a solid state laser detects the pits (or their absence), andthat digital signal is then converted into the corresponding sound.

When continuously varying quantities are represented in digital form, the original quantities are onlyapproximated. For example, frequencies and volume vary continuously in music, but only certain discrete levelsof frequency and volume can be represented on an optical disk. It might therefore seem that digital representationis inferior. However, the problem is handled by allowing for a great many finely spaced choices of frequency,volume, etc. The more choices allowed, the greater number of bits the system must use to represent the information.The cost of storing and manipulating great amounts of digital data continues to decline, so that a very goodapproximation can be quite affordable-the compact disk is one such example.

The digital form has some important advantages. Even though the initial representation in digital form is anapproximation, it can be held to its original form without subsequent errors. Each copy of a digital recordingreproduces exactly the sound pattern of the master, because it copies the master’s pattern of ones and zeros. Intraditional analog sound recording, the copying of masters introduces some distortion-which generally differsfrom one record to another. Distortion shows up even more in electronic transmissions. For example, when a cabletelevision program is transmitted to a home, the signal typically passes through about 25 amplifiers along the wayto keep the signal strong. Each amplifier introduces some distortion, and the distortions are compounded in the finalsignal received in the home. If the picture were represented in digital form, at the end of each leg the pattern of onesand zeros could be sensed and a fresh, distortion-free signal sent along the next leg of the trip. So long as the signalis good enough at the end of each leg to tell which bits have value zero and which bits have value one, the finalpicture can be received error free.l

Another advantage of the digital form is that information is easier to manipulate. For example, splicing filmsegments or creating special visual effects (e.g., superimposing two images) is much easier to do if the picture isstored in digital form: it is easier to rearrange data inside a machine (essentially, a special-purpose computer) thanto cut or otherwise manipulate film. For another example, filtering ghost images out of television is practicable onlyif the picture is represented in digital form. Still another advantage is that digital data can be compressed, allowingmore information to be conveyed over a given TV channel (see the discussion below). Its intrinsic advantages andsharply declining costs have made the digital form increasingly popular in recent years. Sound recording is oneexample. Television promises to be the next.

Conventional television uses predominantly analog information, while high definition television (HDTV)relies much more on digital information. This difference is at the heart of what is new and important about HDTV.In conventional analog TV, the picture is recorded in the studio as a series of frames (30 per second) on film or tape.Each frame shows continuous gradations in color and brightness, corresponding to the original scene. Fortransmission, each frame is broken down into hundreds of horizontal bands, called lines. A scanner sweeps

Isome ~wlY ~OmPUter~ ~ePreWnt~ ~tir~ in analog fo~ and had tie same problems of increasing distofion. Numbers wodd berepresented, for example, as voltage differences. But tie voltages could not be set wff~tly accurateh) so quantities rePresent~ ~side hemachine had some error. As these quantities were added, multiplied, etc., the error increased; moreover, the errors were somewhat random, sothat the same calculation might yield different results. For these reasons analog computers were rejected in favor of digitat computers.


continuously across each line in turn, sensing the color and brightness of each part of the picture as it goes. Thesecontinuously varying characteristics are encoded into a continuously varying electromagnetic wave (the carrierwave) which represents the visual signals through variations (modulations) in its amplitude (strength). Informationcan also be encoded by modulation of the wave’s frequency (number of wave cycles per second), or phase (whenthe cycle begins); the TV sound signal is encoded by frequency changes. Any of these modulations has the effectof changing slightly the observed frequency of the carrier wave. The range over which the frequency may vary iscalled the bandwidth. The carrier wave, sent over the air or over cable, is picked up by a television receiver tunedto the wave’s frequency band. The receiver senses the modulations in the wave, and decodes them to reconstructthe original, continuously varying, pattern of color and brightness for each line. Because of noise in transmission,the received signal has slight errors, causing some distortion in the picture displayed.

HDTV, in contrast, is a largely digital system. In some proposed systems, transmission will be entirely digital;others include an analog component for compatibility with existing receivers. While HDTV systems might bedeveloped in ways that vary somewhat, for simplicity one example is chosen for discussion here.2 For HDTV, thescreen is divided into about 1 million or more equal rectangular or square segments, known as pixels, In any oneframe, each pixel is treated as having uniform color and brightness. These characteristics are recorded in the studioas numbers on magnetic tape.3 The color and brightness of each pixel are represented together as a sum of the threeprimary colors in appropriate brightnesses. For each primary color, 256 different brightnesses are possible(including the dimmest, no light at all); this requires eight bits to represent each brightness, or 24 bits to representall three. Color and brightness do not vary continuously because only certain discrete combinations of primarycolors are allowed. However, so many variations of color and brightness are available that each pixel can come veryclose to the original. Also, the size of the pixel limits the physical detail that can be shown, but with 1 million ormore pixels, that is fine detail. These slight imperfections are less than those caused by noise in conventional TV.

Each television frame is recorded as a string of numbers that represent the color and brightness of each of the1 million or so pixels. To record the 30 frames which comprise one second of television requires about 1 billionbits. This large amount of data must be recorded very quickly to produce HDTV programs, and it must also bemanipulated quickly for transmission, reception in the home, and display on the screen,

The numerical data are encoded into an electromagnetic carrier wave, modulating its amplitude, frequency, andphase. (As with analog television, the result is to vary slightly the observed frequency of the carrier wave.) Whilefor conventional analog television the wave’s amplitude and frequency vary continuously, for HDTV they vary inonly a limited number of steps, corresponding to the numerical patterns being encoded. The television receiversenses the discrete but swiftly changing variations in the incoming wave’s amplitude, frequency, and phase, andthen reconstructs the original pattern of bits for each frame. Based on the information for each frame, the displaymust be quickly updated.

For both conventional television and HDTV, the television carrier wave is allowed to vary only within a certainrange of frequencies, or bandwidth; other frequency bands over the air are used for other television charnels, or forother uses such as radio and cellular telephones. Generally, the more bandwidth is available, the more informationcan be sent per second, As noted, the frames in 1 second of HDTV are represented by about 1 billion bits. To sendthat much information per second would require much more bandwidth than is available for television channelsbroadcast from terrestrial towers; while more bandwidth might be available by cable or satellite, even that amountwould probably be insufficient. This is not surprising, since it takes much more information to transmit the finerresolution HDTV image than that to transmit the image for conventional television programs.

The solution to this shortage of bandwidth will probably involve a combination of techniques. First, the numberof bits actually transmitted can be reduced or compressed, primarily by getting rid of redundant or otherwiseunnecessary information, For example, if a blue sky background does not change for several seconds, it does notneed to be rebroadcast in every frame. (Analog data, used in conventional TV, cannot be similarly compressed.)Also, since the eye cannot perceive fine details of fast moving objects, those objects could be sent in less detail. Thecalculations that do this compression before transmission, and then decompress the information on reception, aredone by digital signal processor (DSP) chips, a kind of integrated circuit. HDTV will require advances incompression techniques.

2The ~l~tion of @is ex~ple does not imply that any particular system of design specifications iS Superior to any Otim.3~ ~me C*S, a pqram is first recorded in analog form and later converted to digital form.


Even with compression, however, HDTV will probably also require an improvement over current technology inthe amount of information that can& transmitted per second in a given bandwidth. Improved equipment will beneeded to encode the bits into modulation of the carrier wave and to decode the modulation on reception. HDTVwiil also require developing DSP chips in the receiver to perform calculations to reduce or eliminate ghost images,flicker, snow, and other picture imperfections.

Actually, conventional television and HDTV are merely points on a continuum. Intermediate versions oftelevision improved Definition Television (IDTV) and Enhanced Definition Television (EDTV), offer a finerresolution picture than conventional television, but not as fine as HDTV, For IDTV and EDTV, analog picture datais sent over the air (or over cable) but upon reception the picture is converted to digital form,

IDTV and EDTV have the advantage of being compatible with existing television systems. IDTV receiversare designed to receive current television transmissions, are being sold commercially, and are already in use in somehomes. EDTV receivers require some change in the transmitted signal, but the new signal would still work withconventional receivers, HDTV transmissions that are composed of encoded compressed digital data would makeno sense to conventional television receivers, which are designed to receive transmissions with analog data encoded.

IDTV and EDTV receivers perform some digital data handling similar to that needed for HDTV. For example,DSP chips reduce or remove ghost images and other flaws in the picture; also, each frame must be displayed quicklyas for HDTV. However, IDTV and EDTV break the screen into fewer pixels, so that not as much data has to bemanipulated each second. In sum, IDTV and EDTV are technological stepping stones to HDTV, and some of thistechnology is already in commercial use.

design. Other business applications, e.g., medical floor. Once advanced manufacturing techniques areimaging, education, and publishing, might also usethe advanced display technology developed forHDTV--indeed some early versions are already inuse.

Manufacturing processes under development forHDTV might find still wider application. Forexample, in the long run, the most promisingmedium for displaying the fine-grained HDTVpicture is the flat panel liquid crystal screen. Thetechniques needed to make these screens can beapplied to methods for interconnecting chips onboards (a process that is common to almost allconsumer electronics products and computers), andto other electronics products and processes as well.

mastered for making electronic components forHDTV, those same techniques can be applied tolower volume business products.

Cost reductions through mass production can bedramatic. For example, in the early 1970s, PlesseyLtd., a British semiconductor firm, developed ahigh-speed digital device able to count about 1billion events per second. These counters, made forlow-volume military and business applications,were expensive and required care to ensure properperformance. RCA, then a leader in the manufactureof television sets, saw the counters’ potential appli-cation to TV tuning systems. Within about 3 years,RCA had made its own circuits, with similar

This spillover to a variety of manufacturing performance characteristics but more robust, and

processes in electronics brings up a second major was mass-producing them for about $1.50 to $3.00

point. To succeed in mass markets for consumer apiece--one-fiftieth of their former cost.94

electronics products and their components, manu- The technological importance of consumer elec-facturers must meet some exacting demands: high- tronics is sometimes underestimated, but the fact isvolume production, low costs and profit margins, that some aspects of the industry---especially manu-and high product reliability. HDTV is interesting not facturing processes—are at the leading edge. Notjust because it demands new microelectronic com- infrequently in the past, manufacturing technologyponents, but because it is a potentially large market developed for consumer electronics has been appliedthat will also push advances in manufacturing to good effect in business products, and this kind ofprocesses, These advances come both from labora- transfer is increasing as the consumer electronicstory R&D (e.g., designing for manufacturability) products converge with business products in the useand from continuous improvements on the shop- of digital technology (box 2-C). U.S.-owned firms

94JohII Henderson, Head, systems Technology Research, David Sarnoff Research Center, personal COmInIJniCatlOn, Jan. 5, Iw.


Box 2-C—Technology Spillovers From Consumer Electronics

Technology developments in consumer electronics have often paved the way for advances in other familiesof electronics products, such as computers. For example, automatic insertion of components into a printed circuitboard was first developed for car radios and other consumer products, and was refined for television. That processhas since been used to build computers and many other products. Another example: mass production of cathoderay tube (CRT) screens for television brought down their price enough that it was attractive to use them in personalcomputers.

Technological spillovers from consumer electronics to computers and other business applications are gainingimportance, because the technologies are converging, For many years, business applications used mostly digitalcircuits, while consumer products relied more heavily on analog circuits. Recently, consumer goods have used moreand more digital circuitry; and HDTV, with its huge appetite for digital circuits, some of them quite advanced indesign, promises to accelerate the trend.

Already, some digital technologies that first appeared in consumer electronics are finding applications incomputers. For example, the digital magnetic tape Sony developed for its 8-millimeter portable camcorder, and thedigital audio tapes developed by Sony and others, are now used in computer systems to store backup data---at aboutone-twentieth the cost of tapes previously available.1 Also, the digital optical disks developed for compact disksound recordings are now used for permanent data storage for personal computers (they are known as CD-ROMs,or compact disk read-only memories in the computer world).

The spillover of technologies honed for high-volume consumer goods to other electronics sectors is uncommonin U.S. companies today. Only one major U.S.-owned company (Zenith) is still in the television business. Butforeign firms-especially the Japanese-continue to use their consumer electronics technology to improve theirposition in computers and other business products. While Japanese firms have had other advantages as well, thistransfer of technology within the firm was often a significant factor. For example, firms in Japan, Korea, and Taiwanadapted the superior CRTS they developed for television to computers, and took a 1arge share of that CRT market.Seiko and Casio exploited their liquid crystal display technology, first developed for watches, to move up to pocketcomputers (used for such things as computerized address books) and then to laptop computers which they sell inJapan. Canon used its expertise in optics, developed in producing consumer cameras, to help in gaining its presenteminence in photocopiers. Perhaps most important, Japanese firms producing consumer products such as VCRsgained experience with automated production lines which they are now applying to the manufacture of computers.2

IRofe~r ~vid Me~rsctiitt, Department of Electrical Engineering and Computer Science, University of California W Be*eley,personal cmnmunication, Dee, 7, 1989.

2Th~ ex=ple~ ~gre #ven by ~k w~n, ~r~t~r, International and Associated Eograms, k’fkrOt?kXtiOnicS & Comput~rT~~~I~sCorp., personal communication, Dec. 13, 1989 and Dec. 28, 1989.

have largely retreated from the consumer electronics and Extended Definition Television (EDTV); to-field; this has sometimes put U.S. firms makingbusiness electronics products, such as computerCRT displays, at a disadvantage (box 2-C). HDTV,which could be one of the premier next-generationconsumer electronics products, might either reverseor accelerate this trend, depending on whether U.S.firms get into HDTV production in a significantway.

At this point, some questions are in order. First, aswith all new products, projections of the eventualmarket for HDTV are uncertain. One question iswhether consumers might settle for intermediateimprovements that go partway towards HDTV.These are Improved Definition Television (IDTV)

gether with HDTV, they are known collectively asadvanced television (ATV). EDTV and IDTV han-dle less data than HDTV; however, all the ATVsystems rely on digital electronics (HDTV being thefarthest along this path) and all require advances inmanufacturing processes. In any case, both Japanand the European Community are pouring substan-tial government as well as private resources intomaking HDTV a reality. This dedication of re-sources into a new technology itself affects themarket’s growth, since it helps to drive down prices,Moreover, the Japanese Government and industryare whetting the consumers’ appetites. The 1988Seoul Olympics were broadcast in HDTV to televi-sion sets at 81 public sites in Japan; daily l-hour


HDTV broadcasts by satellite began in 1989; andNHK (the Japanese national broadcasting company)was planning to broadcast 6 or 7 hours of HDTVprograms every day by 1991.

Another question is whether semiconductors,computers, telecommunications, and other electron-ics fields in which American firms are strongcompetitors might not do as well as HDTV inadvancing technologies with important spillovers toelectronics sectors other than their own. The forego-ing discussion suggests that HDTV itself is not sosignificant a technology driver as are the underlyingsystems for data processing, transmission, anddisplay, and the process technologies for manufac-turing these systems. Two answers suggest them-selves. First, HDTV is pretty clearly ahead in a fewof the core technologies. But second, it is oftenimpossible to be certain which application is ahead,or will remain ahead, as the driver of many of theseimportant core technologies-and this uncertaintydoes not really matter. HDTV, computers, commu-nications, and other electronics fields are all devel-oping on separate but related tracks. So long as manyof their core technologies are fundamentally similar,then advances in any or all of them are synergistic.The same research can be used to advance differentindustries. Each helps the others along.

This kind of synergism is less available toU.S.-owned electronics companies than to Japaneseand European, because few U.S. firms are in theconsumer electronics business in a major way. TheJapanese Government and electronics industry arewell aware of the synergisms and do their best toexploit them.95 The same is increasingly true in theEuropean Community.

Advanced Television as Technology Driver

Some of the core technologies being developedfor HDTV, and to a lesser extent for other forms ofATV, look to be pathbreaking, and could havesignificant spillovers to other electronics appli-

cations. 96 Others are based on technologies that werealready well developed for other uses; furtherdevelopment for ATV probably will not create majorbreakthroughs, but might offer incremental im-provements useful elsewhere. Still others that areneeded for ATV may be developed first for otheruses. While some of the following examples oftechnologies in which ATV seems to have the leadmay turn out to be mistaken, others, in hindsight,probably could be found to take their place.

Flat Panel Liquid Crystal Displays-Display ishigh on the list of technologies likely to driven byHDTV—indeed by all forms of ATV. Not only willthe displays themselves be adaptable to other uses,the manufacturing processes for making them couldalso be widely applied.

Looking ahead to the year 2000, the best candi-date for displaying the HDTV picture (and probablyany ATV picture) appears to be flat panel liquidcrystal displays. This form of display has theadvantages of low power consumption, good colorrange, and compact size.

97 The display contains aglass screen with elements made of a liquid crystal,which change the way they pass or reflect light whenthey are subjected to a small polarizing voltage.Electrical circuits are put right on the glass to controleach of the liquid crystal elements to produce thedesired picture.

Liquid crystal displays have long been in use, e.g.,in digital watches. The challenge is to make them inthe large size and with the fast response and greatdetail (millions of display elements) needed forHDTV--all at a cost that consumers can afford.Making liquid crystal displays for HDTV will pushsome areas of manufacturing technology that havewide application in other electronics sectors. (Thesame is true of IDTV and EDTV, although to aslightly lesser degree, because they require fewerpixels for display than HDTV and the screen might

g5GngoV T=y, c ‘S~tW~ Chage and com~tit;veness: The U.S. Semiconductor Industry, TechnofogicalForec@ing a~soci~c~nge) vol.38,1990 (forthcoming); Barry Whalen, Senior Vice President for Plans and Programs, and Mark Eaton, Director, International and Associated Programs,Microelectronics & Computer Technology Corp., letter to John Glem, Chairman, Senate Committee on Governmental Affairs, July 31,1989, reprintedkProspectsforDevebpmnt o~a U.S. HDTV Industry, hearings before the Senate Committee on Governmental Affairs, Aug. 1,1989, [S, Hrg.] 101-226,pp. 522,524-25 (letter discusses Japan’s Giant Electronics project, and includes translation of two pages of project’s plan); Lansing Felker, Director,Industrial Technology Partnership Program, U.S. Department of Commerce, personal communication, Nov. 21, 1989.

%Fm a more de~~ discussion of llnkages ~tween HDTV and o~er el~tronics indus~~, ~ OTA’S forthcoming report, The Big ~lCtUW, op. Cit.,

ch. 5.gTThe ~T displays currently used for television consume much more power. They are also bulky-nearly as deep as the screen is wide, ~ todaY’s

models-and breakable. Unless greatly slimmed down, with the large screen required to show off HDTV to advantage (40-inch diagonal or more), theywould weigh several hundred pounds and would scarcely fit in the door of most houses.


be somewhat smaller. See box 2-B for a definition ofpixels.)

Some of the advances in manufacturing requiredfor making liquid crystal displays for ATV are: theability to make extremely flat glass panels of largesize (the area of the display screen); precise etchingof electric circuit patterns over the entire screen area;deposition of thin films of material over this areawith uniform thickness; and new techniques forattaching electrical leads and testing finished cir-cuits. Japan’s Ministry of International Trade andIndustry expects that Japanese R&D for flat panelliquid crystal displays, all told, will have applica-tions in a great many areas. Some examples areultra-high density optical recording systems, ultra-thin photocopying systems, solar cells, opticalengraving, large flat light sources, high-precisionelectronic components, and a better method forinterconnecting semiconductor chips.

This last application is particularly significant.The requirement for interconnection of integratedcircuits (chips) is ubiquitous in consumer electronicsand computer applications. The traditional practiceis to put each chip in a plastic or ceramic packagewith metal electrical leads, then mount the packageson a printed circuit board (a pattern of circuitsconsisting of copper foil laminated to sheets offiberglass reinforced epoxy), and then connect thechip’s leads to the board’s circuits. The method isexpensive and somewhat unreliable (connectionsoccasionally come loose), and it limits how denselycircuits can be packed. The less dense the packing,the longer the path the electrical signals must take;longer paths slow down computations, and thus limitthe speed of computers based on this technology forinterconnections.

The emerging ‘chip on glass” technology allowsgreater density and reliability. In this system, thebare, unpackaged chips are mounted directly ontoglass (or another insulating substrate), and the chips’own tiny leads are connected to a fine pattern ofcircuits etched on the glass. The technology de-mands high precision over a large area both inetching the circuits and in film deposition. Largearea lithography-a technique to do these steps atlow cost for mass production of chips on glass-will

probably be developed frost (at least in part) formanufacturing HDTV displays.

Another requirement for the chip on glass technol-ogy is a method of connecting the chip’s minuteleads to the precision etched circuit on the glass. Onesuch technique is tape automated bonding (TAB), inwhich adhesive tape with electrical leads connectsthe chips to the circuit board-and in television witha liquid crystal display, to the display as well.Japanese firms are already using TAB to makeminiature televisions with liquid crystal displays; infact, the Sony Watchman miniature television usesmore demanding TAB than the NEC SX-2 super-computer. 98 In developing HDTV, Japanese firmsare pushing TAB technology still further. U.S.electronics firms have lagged behind in TAB tech-nology, even though it was invented in the UnitedStates.

As manufacturing of liquid crystal displays forATV improves, the displays will become cheaperand more reliable, and will probably find manyapplications in business products-specially com-puters. Liquid crystal displays for ATV and forcomputers are essentially similar, although ATVdisplays require more choices of color and bright-ness and computer displays require more closelyspaced pixels. Lap-top personal computers alreadyuse flat panel displays. More powerful computerswill probably follow.

Digital Signal Processor Chips and ComputerSimulation--The amount of information in a real-time, high-definition, full color HDTV signal ishuge—as much as 1.2 billion bits per second in somesystems. HDTV is driving state-of-the-art technol-ogy in processing so much information at highspeed. The chips that process the information flowsfor HDTV are tailored to its specific needs but mightbe adapted to other signal processing applications,such as compressing speech for transmission. Moregenerally, some of the technologies needed to handleHDTV’s complex, high-speed chips could haveimportant spillovers--e.g., high-performance cir-cuit boards made of new, cheaper materials. Anotherspillover could come from the methods used todesign chips for HDTV.

HDTV picture data are so voluminous that theydemand more bandwidth than is available in most

98N~i~~ReXh~~cil, Commission on Engin~fig and T~hnic-al Systems, Man~acttingStudies Board, The Future ofElectrow”csAssernbly:Report of the Panel on Strategic Electronics Manufacturing Technologies (Washington, DC: National Academy Press, 1988), p. 55.


transmission systems (certainly in broadcasts fromterrestrial towers), and therefore have to be com-pressed before transmission. This compression ofdata, and decompression upon reception, are done byspecialized integrated circuits, digital signal proces-sor (DSP) chips. DSP chips are also used in all ATVto reduce or eliminate ghost images, flicker, snow,and other picture imperfections. The design of thecomplex calculations to be performed by those chipsis made much faster and cheaper by computersimulation. Operations the chip would perform withhardware are first tried out by software on thecomputer, since the computer can readily be repro-grammed to experiment with different designsbefore a real chip is ever made.

Because the DSP chips for HDTV must performcalculations with many billions of steps per second,computer simulations of their operation are difficult.Normally, computers running simulation programsperform much more slowly than the hardware beingsimulated. It is a major challenge to get computersto simulate DSP calculations fast enough to generatevideo images at the normal viewing speed. (Viewingat normal speed is necessary to assess the picturequality.)

A working prototype computer to perform suchsimulations has been built by the David SarnoffResearch Center in the United States, under contractto Thomson Consumer Electronics, a U.S. subsidi-ary of the French firm (partly owned by the FrenchGovernment) Thomson SA.99 In late 1989, Thomsonbegan to use this machine as a design testbed todevelop IDTV receivers; Thomson expects to use itto help develop DSP chips for all future advancedtelevision systems.100 Japanese firms have beendeveloping similar testbed computers.

To achieve simulation at actual viewing speeds,the firms involved have chosen a parallel processingapproach, in which many processors (essentially,many individual computers) all work on the problemat the same time. Parallel processing-especiallywhen it uses many hundreds of processors—is a

cutting-edge area of computer technology, useful forsolving a great many problems from aircraft designto weather forecasting. Massively parallel machineswill take an increasing share of the supercomputermarket because they provide great computing powerat relatively low cost. The firms that use parallelprocessing computer testbeds to design DSP calcu-lations are gaining experience in hardware andsoftware for parallel processing generally. Thishelps Japanese firms’ efforts to catch up to U.S.firms in parallel processing.

Digital Filters--The digital filter, a kind of DSPchip, has many uses in electronics products, includ-ing selecting frequencies and reducing noise. ’In TVreception, for example, the home set may receive notonly the direct television signal but also a weaker,delayed version of the signal reflected off a building.This causes a ghost image, which digital filters canreduce or remove when the picture is representedinside the TV receiver in digital form. Digital filtersare also used in other systems--e. g., in telephonenetworks, to reduce noise from reflections within thesystem; in military radios, to select frequencies andreduce noise; and in compact disk players, to selectfrequencies. Despite much past R&D, digital filtersare still hard to design. As part of its HDTVdevelopment work, Thomson Consumer Electronicshas engaged the David Sarnoff Research Center forwork on making the design easier. This research willpermit easier design of digital filters for otherapplications as well.101

Digital Modulation Techniques—HDTV willrequire new transmission and reception systems, toallow the transmission of more information in agiven bandwidth than is needed for conventionalcolor television today. Among other things, thesesystems will use new, more efficient ways ofencoding digital data into variations in the ampli-tude, frequency, and phase of an electromagneticwave. This encoding is called modulation. Oncemore efficient modulation techniques are developedfor HDTV, they might be used generally to enhancethe information-carrying capacity of other digital

-Omson Consumer Electronics consists of the old RCA consumer products group, which General Electric bought and then sold to Thomson SA.loo~, D. Joseph I)o@w, Senior Vim President, Technology and Business Development, Thomson Consumer Electronics, PrSOnd co~unication,

Jan. 2, 1990; see also Danny Chin, Joseph Passe, et al., lle Princeton Engine: A Real-Time Video System Simulator,’ IEEE Transactions on ConsumerElectronics, vol. 34, No. 2, 1988, p, 285.

IOIJOh.11 H~erson, Hd, systems Technology Research, David Sarnoff Research Center, personal communication, Dec. 7, 1989. while digitifaltering can be done by soflware, that would be too slow for television applicadons. ‘he digital filters used for television are hardware devices. ‘fheyare adaptive filters, meaning that they can adjust their operation to a changing delay between the original signal and its reflection. The filter senses thedelay using a special calibrating signal transmitted at regular intervals.


communications systems, such as microwave phonelinks and digital satellite transmissions.

Fiber Optic Communications--HDTV mightprovide the first demand for fiber optic communica-tions to the home. If a large proportion of U.S. homesare connected to a fiber optic network, the network’selectronic components residing in the homes wouldbe manufactured in very large quantities, and wouldjustify R&D to reduce manufacturing costs. Theelectronics needed to connect each home have agreat deal in common with electronics neededelsewhere in the network. For example, require-ments for wiring up the home would include: 1)electronic components for receiving and amplifyinglight signals;1°22) digital signal processing adaptedto the available bandwidth (greater than that availa-ble for over-the-air broadcasts); and 3) fiber opticcable which is easy for a service technician to installand repair. All of these features are also needed atother points in fiber optic networks.103 Companiesthat cut costs for mass wiring of homes would realizecost advantages generally in building a fiber opticnetwork. They would have the advantage in provid-ing other fiber optic services, such as data transferbetween computers.

Government Policy and ATV

Although the technological spillovers amongdifferent branches of the electronics industry cannotbe pinned down or forecast with precision, theexamples given above suggest the breadth of thesynergism between the advancing, increasinglydigitalized consumer electronics branch-with HDTVin the lead—and the computer and telecommunica-tions branches. Because of these interactions, someof the technologies that have to be developed foradvanced television systems look like strong candi-dates for government support. There are strongcandidates as well in fields other than advancedtelevision. As of now, however, no agency of theU.S. Government has the mandate to select fromamong these possibilities, or the money to givestrong R&D support to civilian technologies thathave the potential for large, long-term benefits to

society, but are too risky to attract adequate privateinvestment.

The U.S. microelectronics industry is at a doubledisadvantage in creating and exploiting advancedtechnologies that are common to consumer and otherelectronics sectors. First, the consumer electronicsindustry in this country is limited. In television, onlyone major company (Zenith) is U.S.-owned; all therest are foreign-owned. This is not an insuperablebarrier to development of new technologies impor-tant to ATV within the United States—witness thefact that Thomson Consumer Electronics (French-owned) engaged the Sarnoff lab (American-ownedand staffed) to build a computer that could helpdesign DSP chips for ATV. It is a handicap,however, that most U.S. electronics companies arenot in the TV business. Second, government isplaying a critical role in developing HDTV technol-ogies in both Japan and Europe. 104 This kind of helpis almost entirely lacking in the United States.

The Japanese Government has worked with in-dustry for over 25 years on developing HDTV andits components, putting HDTV in the wider contextof technology development for a knowledge-intensive economy. NHK, Japan’s quasi-public na-tional television and radio broadcast company, hasinvested about $150 million in R&D related toHDTV since the mid- 1960s, financing its contribu-tions from household TV subscription fees. NHKalso organized and parceled out some of the R&Ddone by private companies. (Private investment overthe years is estimated at $700 million to $1.3billion.) MITI and the Ministry of Posts & Telecom-munications (MPT) added support for R&D,105

while the government’s low-cost loan programshave encouraged private investment in productionfacilities. NHK and private companies have alsoconcentrated on developing programs for HDTV,and the government-supported space program launchesthe satellites for broadcasting.

European countries got a later start but, accordingto those close to the scene, were only a couple ofyears behind the Japanese by 1990. First, at ameeting on international telecommunications stand-

lU21f ~Nice~ ~qufing tw~way ~ommmlcatlon, such ~ te]e.shopping, are provid~, Components tO ~ansmit optical signals would dso k needed.I(LiThe= exmples ~em @ven by J~es Bellisio, Manager, Video Systems Twhnology Research Division, BellCore, personal communication, Jan. 2,

1990.104For a mom det~l~ di~cuwlm of fomlw goverment~’ suppofi of HDTV development, w OTA, ~~ Big Picfure, Op. cit., ch. 2.105M1T1, for exmple, orgai~ and p~i~ly supwfis tie Giant El~~onics ~oj~t, a ‘7.year effo~ to develop co~ twbologies relevant to a40-inch

flat panel display and many other applications by 1996.


ards in May 1986, the European countries refused toaccept the Japanese HDTV production standard, ongrounds that it was incompatible with Europeansystems, but also because of the threat to EuropeanTV manufacturers. A month later, the Europeansformed the joint venture EUREKA Project 95 todevelop their own version of HDTV; the consortiumnow includes two dozen organizations from nineEuropean countries. When the first phase of theproject ended in December 1989, the members hadspent $318 million, of which 40 percent wascontributed by governments and the rest by privatecompanies, and the consortium was ready to beginsatellite transmission tests. It expects to makefull-scale HDTV broadcasts by 1994. Meanwhile,the European Community has adopted local originrequirements for electronics, in which EC goods aredefined as those where the “most substantial trans-formation’ took place in Europe. Non-EC goods aresubject to tariffs, quotas, discrimination in publicprocurement and, when dumping is claimed, anti-dumping actions (which the EC is vigorouslypursuing).

The U.S. Government, by contrast, has been verylittle involved with HDTV. Indeed, the U.S. StateDepartment originally supported the Japanese stan-dard for producing HDTV program material; notuntil May 1989 was this position reversed. The onepositive government action to support HDTV tech-nology came from the Department of Defense. InDecember 1988, the Defense Advanced ResearchProjects Agency (DARPA) invited industry partici-pation in a 3-year $30 million program of R&D forhigh-resolution displays and supporting electronics.Within a few months, 87 companies applied tocollaborate with DARPA, in proposals totaling $200million. By the end of 1989, DARPA had selectedfive contractors, with more to come.

Despite its technical savvy and fine record,DARPA is not the ideal agency to support technolo-gies of great importance to the civilian economy. Itscentral mission, after all, is to fund long-range R&Dthat supports military security. Although it hassometimes interpreted that mission broadly enoughto encompass technologies on the commercial side,since they have military as well as civilian uses, ithas also, on occasion, had to narrow its focus and putstrictly military needs frost. A civilian technologyagency could be given the job of weighing the claimsof various commercial technologies in a systematicand proactive way. Guided by industry’s counseland industry’s willingness to put up its own money,the agency would have to consider what technolo-gies are likely to fortify the long-range economic aswell as military security of the Nation, whethergovernment R&D support is needed, and if so, whereit would count most.

Government support for R&D is clearly noguarantee of success in developing new commercialtechnologies-especially when it comes to a con-sumer product like advanced television. Recall thatsome of the most important linkages betweenconsumer electronics, on the one hand, and suchthings as computers and telecommunications, on theother, are in the manufacturing process. Both thecondition and dividend of success in the demandingmass consumer electronics market is excellence inmanufacturing. And excellence in manufacturingcomes from interaction between the R&D thatgenerates better equipment and processes, andpractice on the shop floor. Thus, government R&Dsupport for the core technologies of importance toseveral electronic sectors is only one ingredient inthe synergism that nurtures all of them.

Chapter 3.

Financing Long-Term Investments

CONTENTSPage

INTERNATIONAL CAPITAL COSTS ...... +... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94The Japanese Financial Market: Sharing the Risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

THE AMERICAN FINANCIAL MARKET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105The Decline in Savings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , ...+........,+,, . . . . . 106Mergers and Acquisitions . . . . . . . . . . . . . . . . . . . .. . . . . . ...,...+. ., . .. . .. +.,,+,... . . 107

FiguresFigure Page3-1. Capital Input Prices, United States and Japan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 953-2. Comparative Capital Costs: Equipment and Machinery, 20-Year Life . . . . . . . . . . . . . 953-3. Comparative Capital Costs: R&D, 10-Year Payoff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 953-4. Comparative Capital Costs: Factory, 40-Year Life . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 953-5. Net Savings, Percentage of Gross Domestic Product ........, .. .. .. .... +., . . . . . . . 973-6. Capital Formation in the United States, Japan, and OECD Europe . ........,..+,., 97

Chapter 3

Financing Long-Term Investments

Developing improved technology requires long-term investment. This is true of all the activitiesinvolved in technological advance-research, de-velopment, commercialization, and acquisition ofnew capital equipment. All these undertakings havea better chance of success when there is a steadycommitment of money, often for several yearsbefore the investment begins to pay off.

Much has been said about the short planninghorizons of American business managers comparedwith the longer term view taken by foreign competi-tors, especially the Japanese. Because Japan’s eco-nomic success shows most clearly what long-terminvestment can accomplish, this section concen-trates mostly on Japan, although examples fromother countries (e.g., Germany and South Korea)would be equally appropriate.

Several explanations have been offered for theJapanese propensity to take the long-term view, andfor the American focus on shorter term returns. Oneis, simply, national culture and, by extension,business culture. But this is less an explanation thanan observation. A factor with more explanatorypower is the remarkable growth of the Japaneseeconomy since World War II, and the comparativelysluggish growth, on average, of the post-1960sAmerican economy. American firms, doing most oftheir business domestically, faced potential growthrates whose mean was close to overall economicgrowth-3 percent per year or so, in real terms.Japanese manufacturers, however, were also lookingoutward, and had not only their own rapidly growingmarket to expand into, but the U.S. market as well.When markets are expanding at a rapid clip,investment for greater market share over the longterm can reap more rewards than playing forshort-term gains. Conversely, economic stagnation,recession, or even sluggish growth can work to thedetriment of long-term investors and make winnersout of short-term profit takers.

Japan’s rapid economic growth in the postwarperiod and its government’s effectiveness in promot-ing swift recovery from the oil shocks and recessionsof the 1970s and 1980s partially explain the pench-ant of Japanese managers to focus on the long term.Likewise, sluggish growth explains some, but notall, of America’s managerial myopia. Another deter-

mining factor is the financial environment. If a focuson short-term returns and profits is hurting Americanfirms in competition with Japanese and Germanfins-and this is widely accepted as true-then itfollows either that U.S. managers persist in ill-judged strategies in the face of evidence to thecontrary, or that there is something about suchstrategies that is rational, viewed from the perspec-tive of the managers. To achieve any long-lastingchanges in the strategic behavior of American fins,it is necessary to understand how the Americanfinancial environment fosters short-term strategies,and how the Japanese financial environment resistssuch pressures.

A major part of the answer lies in the terms onwhich capital is provided, which includes, but is notlimited to, its cost. By common consent, Japanesefirms have deep pockets and patient capital. Patientcapital is, almost by definition, low-cost capital, orit behaves like low-cost capital. And there issubstantial evidence that Japanese businesses haveenjoyed lower cost capital than American firms overmost of the postwar period. Moreover, the financialclimate has encouraged relatively heavy investmentin things like R&D and fixed capital to an evengreater extent than differences in simple cost ofcapital suggest. The question is why.

Today, when Japanese national income per capitais among the world’s highest and Japanese corpora-tions are swimming in profits, it may be hard toremember that, not so long ago, capital was rela-tively scarce in Japan. The Japanese personalsavings rate has been extraordinarily high through-out the postwar period. But initially, incomes werelow, so the total amount saved was not very great.On the other side of the ledger, demands for capitalwere high, mainly to feed the appetite for investmentcapital of a rapidly industrializing economy but alsoto finance frequent deficits in the national govern-ment budget. The workings of free capital marketsdo not explain the low cost of capital to Japanesefirms during those years. The wide gap betweenAmerican and Japanese capital costs, through themid-1970s at least, was a result of governmentregulation of the Japanese financial market.

Today, after years of deregulation, Japanesefinancial markets have become more open, and real

-93-


interest rates, many suggest, have converged some-what with American ones. Yet even if interest rateswere the same, the risks to business in makinglong-term investments might still be lower in Japan.That is, in large part, because both debt and equityfinancing are provided on a less risky, more long-term basis in Japan (and Germany) than in the UnitedStates, in effect lowering the cost of capital toJapanese firms even if the cost of funds (interest ratepaid on debt capital, for example) were the same asAmerica’s.

INTERNATIONAL CAPITAL COSTSAn often-repeated argument holds that if money

flows freely between nations there should be nodifference in the cost of capital based on the nationalidentity of fins. Investment capital, regardless of itsorigin, will seek investments that are expected toyield the highest return, and investors will seek thebest terms from creditors. If there are enough of both(that is, if no investor or creditor has inordinatemarket power), capital flows should be sensitiveonly to risk. This argument presumes, logically, thatthere is no difference in risk based on nationality.And indeed, one study concludes that there is nopersistent difference in real short-term interest ratesbetween the United States and Japan (the nationmost often alleged to enjoy favorable terms oncapital provision).1

There are many flaws in this kind of argument.Short-term interest rates are not a very relevant basisfor comparison, and comparisons of other real ratesdo show a difference between Japan and the UnitedStates. For instance, the real lending rate in theUnited States in the 1980s was higher than that ofJapan by 1.1 to 4.8 percentage points, averaging 2.6

percentage points.2 But a more fundamental flaw isthe failure to take into account the differencebetween cost of funds—interest rates or the cost ofequity-and the cost of capital, which is influencedby corporate tax rates, the economic depreciation ofthe investment and its tax treatment, and other fiscalincentives for investment.3 Numerous studies havedocumented the gap-sometimes several percentagepoints—between Japanese and American capitalcosts over the past two or more decades.4 Jorgensonand Kuroda, for example, estimate that Japan’slower capital costs have been a very importantcontributor to the increasing international competi-tiveness of Japanese firms over the postwar period,excepting the years 1973, 1978, and 1989 (figure3-1).5

The most thorough study, comparing capital costsof the United States, Japan, West Germany, and theUnited Kingdom, calculated capital costs for varioustypes of investment, including research and develop-ment, new plants, and machinery and equipment.The study concluded that American and Britishcapital costs for all types of investment weresubstantially higher than those of Japan and WestGermany over the period 1977 to 1988 (figures 3-2to 3-4). Specifically, each year from 1977 to 1988,the cost of capital in America averaged 3.4 percent-age points higher than the cost of capital in Japan forinvestments in machinery and equipment with aphysical life of 20 years; 4.9 percentage pointshigher for a factory with a physical life of 40 years;and 8 percentage points higher for a research anddevelopment project with a 10-year payoff lag.6

The impact of differences this great is profound.Even small disparities can be important and havelong-lasting effects. A 1-percentage-point difference

IN~on~ Science Fo~dation, The semico~uc~r[~~, Report of a Federal Interagency Staff Working Group (Washington, DC: NOV. 16. 1987).p. 36. This point is quite debatable, even on short-term rates. The NSF study does not mention which short-term rates were compared, and other studieshave concluded that there are substantial differences in short-term interest rates.

z~e p~e lending ra~ in tie United States, and tie lending rate in Japan, according to International Financial St~iSticS. The rates were deflat~using GDP deflators, from the Organization for Economic Cooperation and Development.

3Rc)befl N. McCauley and Steven A Zimmer, “Explaining International Differences in the Cost of Capital,’ Federal Reserve Bank of New YorkQuarterly Review, summer 1989, pp. 7-28.

qFor example, w ‘‘U.S. and Japanese Semiconductor Industries: A Financial Comparison, ’ Chase Financial Policy for the Semiconductor IndustryAssociation, June 9, 1980; George N. Ha~sopoulos and Stephen H. Brooks, ‘The Gap in the Cost of Capital: Causes, Effects, and Remedies,’ TechnologyundEconomic Policy, Ralph Landau and Dale Jorgenson (eds,) (Cambridge, MA: Ballinger Publishing Co., 1986); Albert Ando and Alan J, Auerbach,‘‘The Cost of Capital in the U.S. and Japan: A Comparison,’ Working Paper No. 2286, National Bureau of Economic Research, Inc., June 1987; andDale W. Jorgenson and Masahiro Kuroda, ‘Productivity and International Competitiveness in Japan and the United States, 1960- 1985,’ paper presentedat the Social Science Research Council Conference on International Roductivity and Competitiveness, Stanford, CA, Oct. 28-30, 1988.

SD~e W. Jorgenson and Masahiro Kuroda, “Productivity and International Competitiveness in Japan and the United States, 1960-1985,” paperpresented at the Social Science Research Council Conference on International Productivity and Competitiveness, Stanford, CA, Oct. 28-30, 1988.

GMcCa~ey and Zimmer, op. cit., p. 16.

Chapter 3--Financing Long-Term Investments ● 95

Figure 3-1--Capital Input Prices, United States and Japan

Prlce index

3~~

1 -+

0.6 -

Ot 1 1 t 1 I 1 1 I 1 I 1 1 1 1 1 # , 1 1 I 1 1 i 1 I11960 1962 1964 1966 1966 1970 1972 1974 1976 1978 198019821984

— J a p a n 1 + Japan 2 -++ Us.

SOURCE: Dale W. Jorgenson and Masahlro Kuroda, “Productivity andInternational Competitiveness in Japan and the United States,1960 -85,” paper presented at the Social Science ResearchCouncil Conference on International Productivity, Stanford, CA,CM. 28-30, 1988.

Figure 3-2-Comparative Capital Costs: Equipmentand Machinery, 20-Year Life

1s1

12 -

10 -[ 1

F

0’ 1 1 1 1 1 1 1 1 1 I I

1 9 7 7 1 9 7 6 1 9 7 9 1 9 8 0 1 9 8 1 1 9 8 2 1 9 8 3 1 9 8 4 1 9 8 5 1 9 8 6 1 9 6 7 1 9 8 8

— Uni ted S ta tes + J a p a n + Germany -% United Kingdom

SOURCE: Robert N. McCauley and Steven A. Zimmer, “ExplainingInternational Differences in the Cost of Capital,” FederalReserve Bank of New Yotk Quarter/y Review, summer 1989,table 2.

in the after-tax cost of capital can result in differ-ences in capital stock of 7 to 13 percent in the longrun. 7 Even if American and Japanese capital costswere the same today —which they are not—markedly lower costs in previous decades in Japanwould still favor the Japanese firms.

Figure 3-3-Comparative Capital Costs:R&D, 10-Year Payoff

30 -

25 -

15 -

10 -

5 ‘

L I 1 1 , I 1 1 1 I I I1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988

‘– United States + J a p a n 4 Germany * United Kingdom

SOURCE: Robert N. fvfeCauley and Steven A. Zimmer, “ExplainingInternational Differences in the Cost of Capital,” FederalReserve Bank of New York Quarterly Review, summer 1989,table 2.

Figure 3-4-Comparative Capital Costs: Factory,40-Year Life

1 4

I ,’ I12 -

10[

8 -

6 -

\

0 1 1 1 1 1 I 1 I 1 , I1 9 7 7 1 9 7 8 1 9 7 9 1 9 8 0 1 9 8 1 1 9 8 2 1 9 8 3 1 9 8 4 1 9 8 5 1 9 8 6 1 9 8 7 1 9 8 8

— Uni ted S ta tes + J a p a n ‘+ Germany * United Kingdom

SOURCE: Robert N. McCauley and Steven A. Zimmer, “ExplainingInternational Differences in the Cost of Capital,” FederalReserve Bank of ~w York Quarter/y Review, summer 1989,table 2.

Sustained differences in capital costs of themagnitudes shown by McCauley and Zimmer arenot likely under free market conditions in interna-tional finance.8 Based on evidence of capital-costdifferences alone, we would conclude that thefinancial market of either the United States or Japan

TM. Fukao ~d M. Hmaz~, “Internation~izati~ of Fin~ci~ Markets: Some Implications for Macroeconomic Policy and fOr the Allocation ofCapital,” OECD Working Paper, No. 3, November 1986.

81t is q~~ ~sible, however, mat sm~ler differences could be sustained simply by different cdcdat.ions of ~vestment ri* b~ on c~ncYfluctuations, even if capital moves across national borders without restriction. A Japanese investor, for example, might insist on a higher return on aforeign investment than on a comparable domestic one simply to cover the risk of losses induced solely by changes in currency value.


is not free to seek its own equilibrium. Since theAmerican financial market is known to be relativelyopen internationally, and interest rates are higherhere, the hypothesis is that the Japanese financialmarket has been controlled. That is in fact the case.

Moreover, regulated financial markets are not theonly influence on capital investment or formation.Tax incentives and exemptions are widely used topromote capital investment in Japan, often for quitespecific purposes. The Japanese main-bank systemhas also played a crucial part in lowering capitalcosts and reducing the risk of investment in Japan.9

So, too, has the Japanese network of stable share-holding, designed to help managers resist pressurefrom equity owners to concentrate on short-termprofits and dividends at the expense of market share.

The American financial environment is markedlydissimilar. Not only are there fewer provisions,public and private, to promote investment, but thegovernment gives less effort to maintaining overallmacroeconomic stability, shareholders demand muchgreater accountability, and relationships betweenbanks and companies they lend to are more distant.Moreover, the pressure exerted by the financialenvironment to focus on short-term payoff, orsimply to invest less compared with Japan, isgrowing.

The Japanese Financial Market:Sharing the Risk

Capital costs are based on risk. Riskier invest-ments must promise higher returns to induce inves-tors to provide capital. There is evidence based onthe likely future earnings potential of American andJapanese firms in 1989 that the international Japa-nese manufacturing firms could now be better betsthan the American ones. While they were oftensatisfied with lower profits in the past, many

international Japanese firms are earning handsomeprofits now; their reputations are sounder, and theircapital spending plans are lavish. A 26.3 percent realincrease is anticipated in Japanese capital spendingin manufacturing in fiscal year 1989, and 11.8percent overall,10 compared with a 12.1 percentincrease planned expenditures on new plant andequipment on the part of U.S. manufacturers.11 Astable prosperous future for Japanese manufacturersis a recent development, at least in the eyes ofinternational investors. In the 1960s and even in the1970s, large, long-term investments by Japanesecompanies in markets dominated by European andAmerican corporate giants must have been viewedwith much more skepticism than comparable largeinvestments in Japan now. Yet this higher degree ofrisk was not perceived in the same way in Japan, norwas it reflected in the costs of capital for largeJapanese manufacturing concerns.

The regulation of many facets of the financialsystem of Japan made it possible for these compa-nies to get low-cost capital. According to Abegglenand Stalk, "[t]he policy of the Japanese governmentis, and long has been, to hold interest rates toindustry at as low a level as prudent monetary policymanagement allows. ’ ’12 Until the 1980s, Japan’sfinancial market was effectively closed to outsiders,and Japanese investors had few options for invest-ment outside Japan.13 Moreover, Japan’s financialsystem spread the risks of long-term investments inindustrial development widely among banks, savers,consumers, and corporations. This was done throughcontrolled interest rates; tax policies that limitedconsumer spending, encouraged saving and trans-ferred household savings to businesses on veryfavorable terms; and a variety of tax incentives thatreduced the cost of investment. In America, muchmore of the risk of long-term investment is borne by

9y. KUOWIW~ op. cit.l~e Jqm ~vclqment B~nk, ‘ ‘me Japan ~velopment Bank Reports on capi~ Spending: Survey for Fiscal yew 1988 -90,’ mimeo, September

1989, pp. 2-3. Mr. Nobuyuki Arai, Deputy Manager and Economist of the Economic and Industrial Research Department of JDB expects these plannedtargets to be met. Personal communication with Mr. Arai, November 1988.

1 Iu.s. ~partment of Commerce, Bureau of fionomic Analysis, “Plant and Equipment Expenditures, the Four Quarters of 1988, ’ Survey ofCurrentBwines.s, September 1988, p. 19.

12J~~s c. A@@ and George Stw, Jr., Kais~, the ~~anese co~or~ion (New York, NY: Basic BOOkS, hlc,, 1985), p. 178.13The following dlsc~sion ~aws heavily from tie fol]ow~g ~Uces: M. ‘f’hemse Flaherty and H.iro~ Ita,mi, ‘ ‘Finance,’ Competitive Edge: The

Serniconductor[ndustry in the U.S. and Japan, Okimoto, Sugano and Weinstein (eds.) (Stanford, CA: Stanford University Press, 1984), pp. 135-76.Philip A. Wellons, “competitiveness in the World Economy: The Role of the U.S. Financial System,” U.S. Competitiveness in the World Economy,Bruce R. Scott and George C, In@ (eds.) (Boston, MA: Harvard Business School Ress, 1985), pp. 357-394.


the corporation itself.14 In addition, Japan’s high rateof savings and rapidly rising income levels haveprovided an increasingly generous pool of capital forinvestment. Since World War II, net savings as apercent of GNP averaged well above 20 percent inJapan through the late 1970s, and have declined onlymodestly since. Net savings as a percent of GNPhave rarely approached as much as 15 percent inother advanced industrial democracies. 15 America isthe worst performer among the most advancedOECD nations; net saving hovered at just below 10percent of GNP through the end of the 1970s, andthen plummeted, reaching a low of 2.4 percent in1987, and then recovered slightly (figure 3-5).Capital formation, as a percent of GDP, has alsobeen higher in Japan than in the United States orOECD Europe (figure 3-6). Finally, Japanese lend-ers—stockholders and large city banks-tend tohave much closer and more influential relationshipswith their corporate debtors than is the case in theUnited States.l6

Although some of the conditions described aboveare slowly changing as the Japanese financial systemis deregulated, their combined influence over thepostwar period was to give Japanese firms substan-tially more freedom to make riskier, long-terminvestments at lower cost than American (or proba-bly European) firms enjoyed. From this perspective,Japan’s much-touted long-term vision—and corre-spondingly, the much remarked myopia of Ameri-can managers—becomes understandable. Rationalmanagers, operating under the rules and conditionsof financing in both countries, could be expected tobehave quite differently. This view is persuasiveeven if the numerical difference in interest rates—aslow as 1 to 3 percentage points, according to someanalyses--is modest.

The sharing of risk in Japan is not the result of anysingle action or actor, but rather of a variety ofinstitutions and laws. Moreover, the risk-sharingthat lowers the cost of capital to corporations doesnot apply to consumers. The factors that spread therisk of business investment include closed or con-trolled financial markets, channeling of funds to

Figure 3-5-Net Savings, Percentage of GrossDomestic Product

15

t I)

10 -[

5 -

I 1 I , I 1 1 1 1 1 I1977 1978 1979 1980 1981 1982 1983 1984 1986 1988 1987

— Uni ted S ta tes + J a p a n +7 Germany ~ Great Britain

SOURCE: Organization for Economic Cooperation and Development,Historical Statistics 1960-87 (Paris, France: 1989), table 6.16.

Figure 3-6--CapitaI Formation in the United States,Japan, and OECD Europe

Percent of GDP40% [ I

1‘+\ + -+~+—./ —+--+30% -+%+—+—

I3 G . . . *

-...*. .- .*. - . . + - - - * - - - . * - - -*-. -.

20% - 6

10%

I0% ‘ I I 1 1 1 1 1 1 , , 1

1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985

— Us. + Japan - *- OECD Europe

SOURCE: Organization for Economic Cooperation and Development,National AccourIts 1960-1964 (Paris, France: 1986).

businesses and away from consumer loans, a largepool of savings for investment, and close relation-ships between companies and capital providers(banks, affiliated financial institutions, governmentinstitutions, and stockholders). For targeted indus-tries-those viewed as having most promise fordevelopment —there are other mechanisms as well,

lq~e @@p~ ~onomies of western Europe, except West Germany, more closely approximate the American model than the Japanese, at 1east interms of capital costs, according to available evidence. See, for example, Y. Suzuki, Money and Banking in Contemporary Japan (New Haven, CT: YaleUniversity Press, 1980).

ISFlahe~ and Itsmi, op. cit., p. 137.16For Cxmple, cor~tt m~es he Point hat Japane= ba~S probably monitor the companies they lend heavily to more actively than is the Case in

other countries. See Jemy Corbett, “International Perspectives on Financing: Evidence from Japan, ” O#ord Review of Economic Policy, vol. 3, No.4, 1987, p. 45.


some of them explicit (subsidies for R&D andcapital investment, for example) and some implicitor consensual, such as protection from the threat ofhostile takeovers.17

Controlled Financial Markets—The history ofthe Japanese financial system is a study in controland fragmentation. Although recent market-openingmoves have gained widespread attention. it is onlyin the 1980s, under intense internal and externalpressure, that real liberalization has occurred, andeven so, Japan’s financial market remains one of theworld’s more controlled.18 Between World War IIand the early 1980s, a dominant purpose of theJapanese financial system was to revive and strengthenJapanese industry, often at the expense of consum-ers. Guidance of the financial system had two aims,subsumed under the single purpose of promotingJapan’s reconstruction and economic development.First, the system was designed to favor businessinvestment instead of current consumption, or, in thewords of an official of the Ministry of Finance, ‘‘toprepare the ground for industry to walk on.”19

Second, the government selectively promoted heav-ier investment in certain sectors as a part of Japan’sindustrial policy, and also helped non-targetedindustries cope with the costs of adjustment.

Preparing the Ground-Japan was a poor coun-try after World War II. Its needs for capital wereenormous. Much of its industry had been devastatedby or dismantled after the war, and the zaibatsu,family-controlled bank-holding companies that weremajor providers of capital pre-war, were dismantledduring the occupation.

20 To rebuild industrial pro-duction--and then, beginning in the 1950s and1960s, to accelerate development of targeted indus-

tries like machinery, motor vehicles, and electronics--required what capital there was in Japan to bepreferentially provided to utilities and manufactur-ing. Several things made this transfusion possible.

Japan’s financial institutions were compartmen-talized and fragmented, each with its own rathernarrow purpose and with many proscriptions on itsbehavior. Briefly, the institutions worked together toincrease savings rates (generating capital for invest-ment) and pass them on to industrial users withouthigh costs. They also worked to reduce the riskassociated with financial downturns and the costs offinancial distress to the firms.2] The institutions thatpromoted high savings rates in Japan included alump-sum payment at retirement (rather than alifetime annuity) and a marginal system of socialsecurity (though this is changing to become moregenerous); large required downpayments on houses;the absence of scholarships at universities; a systemof postal savings banks authorized to pay interestrates higher than rates available elsewhere ondeposits, and tax exemptions on interest on postalsavings up to a certain level (14 million yen in theearly 1980s); a bonus-pay system of compensationin Japanese corporations; and very high interest rates(with no tax deductibility of interest paid) onconsumer loans .22

Together, these measures discouraged consump-tion and encouraged saving. In addition to providinga large pool of capital, the system also controlled thecost of raising it. Households were paid low rates ofinterest on the savings they put into banks,23 butrewarded by the tax benefits, or ‘‘maruyu,’ fordoing so. Securities markets were tightly controlledso as to concentrate household savings in postal

ITpC~~ ~O=miC~tim ~i~ROn~d ~re, ~wri~ college, University of London; and Edward J. Lincoln, The Brookings ktitution, ~ch 1989.lg~on V&r, Jqanese Financial Markets (Homewood, IL: DOW Jones-Win, 1988).l~erW~ ~m~c~on, OTA st~f witi W. Kit~~a, Financial Bureau, h4inistry of Finance, Tokyo, Japan. Mm. 13>1989.

~efollowingdiscussionof Japan’s financial system depends heavily on the following sources: Viner,op. cit.; Andreas R. Prindl, JqpaneseFinance:A Guide to Banking in Japan (New York, NY: John Wiley & Sons, 1981); Philip A. Wellons, ‘‘Competitiveness in the World Economy: The Role ofthe U.S. Financial System,’ in Bruce R. Scott and George C. Imdge, U.S. Competitiveness in the World Economy (Boston, MA: Harvard Business SchoolPress, 1984).

21 Wellms, op. cit., p. S(jl. ~o~er ~t of ~stltutlons, ~u~ly impfiant, gave J~anese fiis preferenti~ access to the domestic market, helping toassure a demand for the products of Japanese industry without ruinouscompetition from (at that time) abler foreign competitors. Japan’s trade policiesand their relation to industry policy will be discussed in the final report of this assessment of Technology, Innovation, and U.S. Trade.

22Tobe specific, a change in the rules governing consumer finance companies-known as sarakin—in 1985 reduced the maximum rateon consumerloans from 109.5 percent per annum to 73 percent, and set a maximum of 10 percent of amual salary of 500,000 yen to the amount one customer couldborrow. Source: Viner, op. cit., p. 339. For an explanation of how the bonus-pay system promotes savings, see Abegglen and StaJk, op. cit., p. 1%.

mB~s did not pay ~ hi@ intere~ rates ~ ~st~ ~vings, but tie upper limit on be ~o~t of any one ps~ savings account, the trouble of keepingseveral accounts, and the fact that company employees are often encouraged to use the company’s main bank or an affiliate, kept some household savingsaccounts in banks.


savings and in banks, so that banks, with theircontrolled interest rates, did not have to compete forsavings by paying high rates of interest to deposi-tors, and thus narrow their profit margins. Interbanktransfers of funds were also handled so as tominimize the eventual interest rate that industrypaid. The result of all this control was that moneywas channeled from households through severalbanks to corporations, at rates that greatly favoredindustrial investment and expansion at the expenseof consumption. The extent of the transfer was huge.According to one estimate, if these measures low-ered the interest rate to business by 2 percentagepoints in 1971,800 billion yen was transferred fromhouseholds to businesses in that year-money that,under free market conditions, would not have goneto the corporate sector.24

Both commercial and governmental banks lendmoney to Japanese corporations, but the distinctionbetween them is rather more blurred than is the caseinmost other industrialized nations. The commercialbanks include the large city banks, which specializedin lending to large, blue chip corporations during thehigh growth period;25 regional banks, which tend tolend to small and medium-sized companies; theBank of Tokyo, technically a city bank, but the onlyone that could make foreign exchange transactionsuntil World War II, and is still a specialist in foreigntrade financing and foreign exchange; trust banks,which specialize in managing pension funds; spe-cialized banks; the postal savings system; andlong-term credit banks created in the 1950s and1960s by government to make long-term fundsavailable for industrialization. These last (whichinclude the Industrial Bank of Japan and the LongTerm Credit Bank) were able to provide funding tocompanies even when there were severe liquidityshortages, thus reducing the vulnerability of Japa-

nese firms to ordinary fluctuations in economicconditions.

The government exercises control over andthrough the banks in many ways. First, interest rateshave been tightly regulated since 1947, when theTemporary Interest Rate Adjustment Law waspassed. 26 By 1986, after 2 years of steps towardderegulation, about 80 percent of deposits in Japanstill came under fixed interest rate regulations.27

Interest rates have historically been negotiated bythe Ministry of Finance, the Bank of Japan, andlong-term credit banks, the financial institutionsmost concerned with the competitiveness of Japan’sindustry. Equity-to-asset ratios have also been ex-tremely low by international standards; they aver-aged 2.19 percent for the city banks as of March1986, compared with 5 to 6 percent for U.S. banks.28

This allows Japanese banks to make low-interestloans both domestically and (lately) abroad.

There are informal controls as well. The Ministryof Finance exercises enormous (though waning)control over all aspects of Japanese finance. Much ofthis is through so-called administrative guidance,which takes a variety of forms, and can affectbehavior at the level of the individual firm or bank.MoF’s instructions and desires are not often ignored,even when they are not backed by force of law. Itsstaff are “the most gifted graduates of the bestuniversities. 29 Like many other powerful Japaneseinstitutions, MoF operates through frequent contactand consensus building; it holds regular meetingswith the management of main Japanese banks,influencing the actions of Japanese branch banks inforeign nations as well as at home. When its seniorstaff retire,30 many of them accept positions at thelong-term credit banks, which were privatized dec-ades ago. According to Viner, “. . . it is neitheraccurate nor meaningful to describe the threelong-term credit banks as private institutions. Their

24y+ K~o~wa, q). Cit., p. 13.

2SBo~ km~auon and tie fmaci~ SWcess of tie lwge corporations of Japan have encouraged the city btis to look for new kinds of b@ne~-Now, with many large businesses financed mainly by bonds, depreciation, and retained earnings, the city banks are turning increasingly to medium-simdbusinesses for customers. Personal communication with Mr. Tatsuo Takahashi, Manager, Public Relations Division, Japan Development Bank, March1989.-e word “temporary” is misleading; the law is still in effect.zTViner, op. cit., pp. X16-3W.Zgviner, OP. cit., p. 20*. ~s low ~uity.to.~wt ratio is typic~, despite the fact that the 1954 Banking Act required a ratio of 10 percent. wording

to Viner, “this level was considered absurdly high by banks and was ignored. ”2~~, op. cit., p. 9.

q~e tem for this is um@duri, or ‘descent from heaven’—which by itself connotes a status of civil servants that is very different from Americanexperience.


ties with the government are so close that in manyrespects they resemble auxiliary components of theMinistry of Finance. ”

Industrial Policy--Formal and informal controlscan be used both systemwide—to advance capitalrelatively cheaply to firms and away from personalconsumption, for example-and in pursuit of moreindustry-specific goals. The government acts both asa direct lender and as a bellwether for other privatesector lenders. Its direct role is small-in 1980, only5.6 percent of all funds placed in financial institu-tions in Japan reached business directly from gov-ernmental institutions,31 and long-term credit banksprovided another 5.2 percent. But this governmentalrole is more powerful than its modest funding wouldsuggest. According to Wellons, “few dispute thatprivate lenders in Japan treat this lending as a signthat the firm or project has government support,which would reduce the risk of the credit. ” ManyJapanese sources agree. According to Kurosawa,

The government also helped to reduce risk; MITIestablished specific goals and initiated investmentfor companies, and when necessary, adjusted theorder [of] which group of companies should investfirst and which next (Rinban Toshi).32

One way the Japanese Government primes theprivate lending pump is through the Japan Develop-ment Bank (JDB). When motor vehicles werechosen for rapid development in the 1950s, andelectronics in the 1960s and 1970s, the Japanesecompanies were generally far behind American andEuropean companies in technology, and financialreturns from heavy investments in those industrieswere therefore quite uncertain. City banks, with

much of the lendable capital, might have been waryof making heavy investments in such industries, butwere reassured by JDB’s lending. Throughout thepostwar period, JDB loans have been among themost important sources of funds for new equipmentacquisition in manufacturing. In fact, even in the1980s, long after the end of any real capital scarcityin Japan, about one-fourth of JDB’s funds still go tomanufacturing. 33 Where JDB lends is, in turn,decided by a variety of government departments,with strong participation from MITI, and its lendingis meant to help major strategic industries directly.%

Financial support for both industry as a whole andstrategic industries in particular has been a crucialelement of Japanese industrial policy, but it is by nomeans the only one. Government support takes avariety of forms, including preferential access to theJapanese market,35 support for research and devel-opment, market segmentation among domestic firms,and control of foreign investment. With such apanoply of tools at hand, and the demonstratedwillingness to use them to support development ofindustries, government can pack a powerful punchwith a relatively modest direct financial role.36 Also,the variety of available tools helps to make up forweaknesses in the use of any one. For example,pump priming alone would not have induced Japa-nese banks to invest in certain sectors where theexpected returns were especially low; it was deci-sive, however, where both expected returns and riskswere high.37

The government’s control over the financialmarkets is lessening. Many Japanese financial insti-tutions see narrowing opportunities for growth

31’’fh= fi~tutionS ~c]u~e the J~p~~ Development Ba~, the Japan Export.~pofl B*, and agencies to finance smd and MdkM-Skd business.Source: Wellons, op. cit., p. 380.

32y- K~=wa, op. Cit., p. 16.XIR~fi J. B~lon @ ~wao ~mita, The Fi~n~~l Behavior of Japanese corpor~~~ (Tokyo: Kodm~ International, 1988), p. 37.

sdperm~ cornm~c~on with W. Kitamura, Ministry of Finance, op. cit., and Ballon and Tomita. op. cit.35This is ~t ~t~ Mwket pro~tion, ~ is ~met~e~ cl~m~; however, access to Japan’s markets in Wgeted tid~tries is c~fully controlled and

limited, as are opportunities for direct foreign investment and direct investment abroad. Preferential access allows Japanese producers to sell goods inJapan at higher prices or of lower quality than they could if foreign products were allowed unlimited access. Barriers to foreign competition are usuallyphased out once the Japanese indusrnes have grown to be formidable competitors. However, we are now beginning to see Japan resorting to voluntaryrestraint agreements in industries that are under pressure with the rise of the yen and the growing competence of other Asian competitors. A more completediscussion of these mechanisms will appear in the next and f-real, report in this OTA assessment.

36Althou@ ~e~ nu~r is ~lin~g, there we expe~s who @pute the dqy~ to w~ch Japan’s indu~d policies have been responsible for thepostwar success of her industries. Clearly, other nations have used tools similar to Japan’s without the same results, and Japan herself has demonstratedremarkable ability to develop industries in eadier periods when policies were quite different, as in the decades following the Meiji Restoration in thelate 19th century. Thus, more than industrial policy is responsible for Japan’s reeent performance. However, industrial policy has been and remains acritical factor in Japan’s development, as will be explained more fully in the next and fuial report in this OTA assessment.

3TS~ibWaEi~e, Ro&t Feldmm, ad Yum H~~a, The Japanese Fina~~ System in Compar@”ve perspective, study Pmpa for the U= Ofthe Joint Economic Committee (Washington, DC: U.S. Government Printing Office, 1982).

Chapter 3-Financing Long-Term Investments ● 101

domestically, as prosperous Japanese firms areincreasingly able to finance themselves, or havemore freedom to choose among domestic andforeign financing options. International pressure hasalso been a factor forcing liberalization of Japanesefinancial markets. However, it would be a mistake toregard Japan’s financial market as open—the dereg-ulation is proceeding deliberately, so as to avoidmajor shocks--or to discount the advantage thattight controls gave to Japanese industry during thepostwar period through the early 1980s. Without thedeliberate channeling of capital away from personalconsumption and towards industry-particularlythose that were targeted-it is unlikely that so manyJapanese industries would be so prominent on theinternational scene as they are now. It is also prudentto assume that, if Japanese manufacturing comesunder increasing international pressure, the financialsystem is capable of mobilizing quickly in response.

Corporate Finance-It is well established thatJapanese firms rely more heavily on externalfinancing-both debt and equity—than Americanfirms, and that the reliance was greater in the pastthan it is now. Debt financing in particular hasplayed a greater part in corporate finance in Japanthan in the United States (until very recently) andother western industrialized nations, and it still doesso today, even though the percentage of equityfinancing is growing in Japan.

Precise figures are somewhat deceptive, as manycritics have pointed out. The gearing ratios38 re-ported are based on the book value of companies’assets, which are reported at historic cost. Inflation,especially the run-up in the value of property andland in Japan, tends to understate asset value andthus overstate gearing ratios. However, even whenthe figures are corrected to reflect more realisticmeasures of Japanese (and American) fins’ assetvalues, gearing ratios in Japan were still roughlytwice as high as those in the United States only a fewyears ago. In 1981, for example, Japanese gearingratios were estimated at 0.56 to 0.62; American at0.28 to 0.30.39 Japanese dependence on bank financ-ing is also high compared with that of European

nations. American companies have depended muchmore heavily on retained earnings (internal financ-ing) and equity. This remains true even with modestmoves away from debt as a source of new funds inJapan and increases in debt in America,40 the latterresulting mostly from takeovers and leveragedbuyouts to defend against the possibility of take-overs.

Japanese reliance on bank financing, particularlywhen capital was much less available there than it isnow, underlines the importance of low interest ratesin Japan. It also means that fins’ relationships withbanks are more important than their relationshipswith shareholders, compared with the United States(and much of Europe). As long as Japanese banks_ aresympathetic to the need to make long-term invest-ments with little immediate return, firms are morelikely to make such investments. This would be trueeven if Japanese fins’ relationships with theirshareholders were the same as those of Americanfins; however, Japanese shareholders are also moresympathetic to the long-term interests and perform-ance of Japanese firms than in short-term financialgains, compared with American shareholders.41 Inshort, while the structure and regulation of Japanesefinance would alone lead to the conclusion thatJapanese firms are better able to make long-term,relatively heavy investments than American firms,the nature of the relationships between capitalproviders and firms supports this conclusion as well.

Japanese banks-including both commercial bankslike city and regional banks, and governmentinstitutions like the Japan Development Bank-aremore involved with their clients than are Americanbanks. This is true at every step of the process, fromscreening to monitoring of firm performance.42 Tobegin with, Japanese firms usually have a specialrelationship with one bank, a system known as themain-bank system, and this relationship is animportant part of the risk-sharing that allows Japa-nese firms to enjoy or act like they have lower capitalcosts. Kurosawa characterizes the main bank systemthis way:

s~e~ng r~o is defined as the sum of short- and long-term liabilities divided by total assets.s~lWes ~P~ in Jenny ~~~, ‘‘l,n~rnation~ Perspectives on Financing; Evidence from Japan,’ O#Ord Revi6’w of Ecown”c po/icY~ vol~ 3!

No. 4., p. 34.

Wen Bemanke, “Testimony on corporate debt,” mimeo, May 25, 1989.AIThis is l~gely due to tie institution of stable shareholding, as is explained later in this chapter.4~s coWlwim, ad much of tie following disc~sion about banks’ relationships with firms, depends heavily on CorbeV oP. cit., Passim.


The main bank almost always has the largest sharein such business relationships as lending, sharehold-ing, trusteeship of bonds, deposits, and so on. It givesspecial priority to the client firms in credit rationing,and in the case of a severe slump or bankruptcycrisis, coordinates the responses of other lendingfinancial institutions and acts as a mediator andsupporter for the clients’ survival. Consequently, itis essential for the main bank to monitor the firm, andfor the other banks the actions of the main bank actas a signal. If the actions of the main bank remainunchanged, there are no problems in the fire-t. Themain bank’s additional loans in effect guarantee thesecurity of the other banks’ loans.43

Differences begin with the way they screenpotential borrowers. For example, city banks are lessconcerned about debt/equity ratios and are moresensitive to the firm as a going concern (rather thanas a default risk) than are non-Japanese banks. Thescreening is extensive, so when a city bank takes ona client it is generally considered a good credit riskby others. Part of the screening is done by the citybanks, but they are also able to rely on extensivescreening by the Japan Development Bank and theIndustrial Bank of Japan (IBJ).44 There is somegenial disagreement between these two institutionsas to which developed the screening procedures bothemploy—both lay claim to it—but in any case, it isthorough. According to IBJ, the screen consists ofincreasingly smaller sieves. First, the IndustrialResearch Department (IRD) develops informationon specific industries, examining in detail possibili-ties for growth and international competition. TheIRD also examines new sectors and technologies,such as biotechnology and superconductivity, fortheir eventual commercial possibilities. Once indus-try prospects are understood, the Credit Departmentscreens individual companies. If IBJ accepts acompany, that is a powerful signal to other financialinstitutions of the company’s creditworthiness, anda pattern of heavy lending to any particular industryor sector is also a bellwether.

There are several reasons why the close tiesbetween main banks and their corporate customerscould lead to a longer term outlook on the part ofbusinesses, and possibly even to better decisionmak-

ing than in countries like the United States orEngland, where ties between banks and the compa-nies they lend to are more frequently arm’s-length.As noted above, the close relationships between citybanks and their customers are based on massiveamounts of information, always a good basis forsound advice and decisionmaking. The city banks,along with other major Japanese financial institu-tions like JDB, have become powerful informationbrokers, and their ability to gather and processinformation about businesses and business condi-tions in a variety of industries around the globeprobably exceeds that of all but the very largestcorporations. Banks can therefore serve as importantsources of information for strategic and operatingdecisions for their closest customers. This assistanceon the part of banks is influential in encouragingcompanies to focus on longer term goals in Japanand Germany.

Another difference between Japanese and Ameri-can bank lending is that loans from city banks aremuch more likely to be long term. According to theBank of Japan, about 40 percent of Japanesecorporate borrowing had a maturity of more than ayear, compared with only 19 percent in the UnitedStates, as of 1985. However, the longer maturities ofmany Japanese loans are not exceptional comparedwith France and the United Kingdom (where about40 percent of loans are classified as being long ormedium term) or Germany (where about 60 percentof corporate loans are long term) .45

Finally, it is well established that the conditions ofloans are changed when economic conditionschange in Japan. Although this practice is alsocommon in western industrialized nations, the kindsof changes made are different. Corbett points outthat a shortening of the term of a loan would beexpected if a firm gets into trouble; yet in Japan loanmaturities have lengthened at the same time thatbankruptcies increased. With heavy investments ofboth capital and prestige in the successor failure oftheir clients, Japanese (and also German) banks arefar more likely, in a crisis, to extend additionalfinancing and assistance before pulling the plug than

43Y. K~~Wa, op. cit., p. 18.

44~e ~~~ B~of Japmis me of J~an’s~ ~ong.~~ c~it ba~, ~d it is~s~ly descri~ u tie most prestigious of d Japanese @Vatebanks. Its purpose is to provide long-term capital to private corporations, witb priority given to industries that are part of the government’s industrialpolicy.

d5Bti of JqM& ECO?WW St@stics Att?u@ VtiOUS ye~; ~d COrb@t, Op. cit.. P. 42.

Chapter 3-Financing Long-Term Investments ● 103

an American or a British bank.46 Japanese banksoften forgive payments on debt principal duringtough economic times, or restructure debt in order toallow firms additional options to overcome theirproblems. 47 While some firms do eventually gobankrupt or are forced to restructure severely, banksexplore many other options with their clients (oftenat great cost to themselves) before declaring loans indefault. Prindl tells the story of Ataka, the fourthlargest Japanese trading company in the early1970s. 48 It got into trouble over excessive creditextended to a refinery in Canada, and eventually hadto merge with another firm, C. Itoh. However, $370million in uncollectable receivables were absorbedby its house banks, Sumitomo and Kyowa. This waspossible, in part, because of the widespread beliefthat no large bank would be allowed to fail. Indeed,in 1986, Japan had its first bank failure since WorldWar H, and that was a result of ‘massive, long-termcorruption. This situation is changing, like somuch of Japanese business. According to Viner,“banks have been informed that they can no longerexpect central bank rescue in the event of a liquiditycrisis.”49 So far, this new policy has not been tested.

Even the promise of government support does notseem adequate to explain why Japanese banks aremore willing to go the distance with their clients, aslong as there is some chance of maintaining thecompany in business. In part, it is because the mainbank’s relationship with a client company goes farbeyond a loan. Companies generally encourage theiremployees to deposit their savings in their mainbank, and deal with the main bank or its affiliates forlife insurance and managing the pension fund. Inaddition, the main banks, in return for bearing someof the risk of the company’s long-term investment,are privy to a great deal of information about thecompany, and are allowed to take part in itsmanagement should it get into trouble. Main banksoften accept deferment of payment on principal and

interest if a client gets into trouble,50 and willcoordinate rescue funds from other banks. In addi-tion, however, they investigate whether the com-pany can be restructured to get it out of trouble, andoften draw up the restructuring plan.51 Corbettpoints out that exchanges of personnel at both seniorand junior levels between banks and large firms (andgovernment ministries) are common.52 Banks some-time suggest changes in strategy when evaluating acustomer’s request for a loan, and make moreforceful suggestions of strategic changes when afirm gets into trouble.

The kind of involvement that large banks main-tain with their customers resembles that of preferredstockholders more than creditors, according toKurosawa. Preferred stock may have a fixed divi-dend, but if profits are insufficient to support it, therate will be reduced and carried over.53

But what about actual equity holders? Here, too,there are different relationships in Japan. Most largeJapanese firms belong to groups known as keiretsu,which translates as “group arranged in order. ”These are companies that have primarily beenassociated with one city bank, and hold relativelylarge amounts of each other’s stock—1 to 3 percent,typically, of the stock of each other member of thegroup. The result is that a majority of shares of allmembers are held by other members of the samekeiretsu.54 Japanese city banks also typically holdstock in the companies they provide credit to, withthe maximum amount now limited to 5 percent.Finally, although intra-keiretsu shareholding is de-creasing, a majority of stock in Japanese corpora-tions is still typically held by corporate and otherinstitutional investors, rather than by individualshareholders. As of 1988, 69 percent of all shareslisted on the Tokyo exchange were held by domesticinstitutional investors—19 percent by banks, 13percent by life insurance companies, and 26 percentby other corporations—while 25 percent was held by

‘WOrbett, op. cit.a~erW~ com~c~on wi~ David HI,@ Whittier, 1988; Flaherty and Itami, op. cit., p. 144; Corktt, pp. 46-51 Passim.4~~, op. cit., p. 64.49Viner, op. cit., p. 196.

5TM5 should not be reg~d as a distant possibility, Ballon and ~rnita point out th~, “more often than not, [the] bank at some point in time hashad to stage a rescue operation for its mqjor clients with the cooperation of other parties concerned,” Ballon and ‘fbmita, op. cit., p, 60,

51Y. Kuosaw~ op. cit., pp. 19-20.5~a~ti, op. cit., p. 45.

53Y. Kwsawa, op. Cit., p, 20.54Viner, op. cit., p. 2.


individual Japanese stockholders and 6 percent bynonresidents. 55 In contrast, 57 percent of U.S.equities were held by individuals as of mid- 1989.56

More important than the pattern itself is thecharacter of equityholding in Japan. Until the early1970s, it was virtually impossible for more than atiny trickle of foreign capital to find away into Japanwithout the express permission-indeed, sponsor-ship--of government. In 1971, the door was openeda crack through revision of the Securities ExchangeLaw, and along with the liberalization came mount-ing concern that foreign companies would take overJapanese corporations. To prevent that, Japanesecompanies—at the urging of government—resortedto a system known as stable shareholding.

Stable shareholders are Japanese nationals whocan be counted on to keep their shares, no matterwhat happens to their price. It is a primary duty offinancial officers of corporations to find stableshareholders. According to Ballon and Tomita,

When a capital increase is planned, financialexecutives usually visit the major shareholders whomight be willing to subscribe to new shares andrequest their cooperation in purchasing the newshares at par while retaining both old and new shares.However, a request for further subscription of sharesfrequently implies a favor in return. . . the firm mayat this time confirm its friendly relationship with thebank by promising (albeit unwillingly) to buy morebank shares.57

Stable shareholding has had the direct result ofpermitting companies to keep a longer term view intheir capital investment. Stable shareholders preferretaining earnings to receiving high dividends,permitting the company that issued the stock toreinvest its earnings, This reinvestment, in turn, isviewed as directly contributing to higher shareprices. Since stocks are carried on their owners’books at purchase price, rather than market value,the rapid increase in share value has allowedJapanese banks and corporations to carry substantialhidden reserves. These hidden reserves are the utility

infielders of Japanese accounting: they can be usedto manipulate the reported levels of profit, andthereby, taxes and dividends. For example, if thecompany has a loss and needs to show a small profit,it can sell a portion of its investment securities,whose book value is usually significantly underre-ported. Often, it sells these to an affiliate or anotherstable shareholder, and expects in its turn to pay thesame consideration to its affiliates when needed.58

The amount of hidden reserves is staggering: at theend of March 1988, the hidden reserves of securitiesof the 13 city banks alone totaled $229 billion.59

Stable shareholding has served the needs of theJapanese economy admirably. It permitted long-term investment at a time when Japan’s companieswere much more vulnerable to foreign competitionthan they are now. It has helped Japanese companiesto continue expansion and market share-buildingduring the various economic upheavals that para-lyzed their competitors—through energy shocks ofthe 1970s, the recessions of 1974 and 1982, andthrough endaka in 1985-86. Most observers expectstable shareholding to continue for the foreseeablefuture, although it will face increasing challenges inthe years ahead. Financial liberalization in Japan andthe expansion of Japan’s business and financial tiesaround the world have made it more vulnerable tooutside economic uncertainties. While its recoveryduring the postwar period has been robust, this newinternational exposure could well reduce its powerin the future. The high yen, too, has put the wholeeconomy on a more precarious footing. Some of theadvantages Japanese firms receive have narrowed ordisappeared, and strong competition from a new setof industrializing nations has left Japanese manufac-turers with less ability to ride out a prolongeddownturn. In a downturn, stable shareholding mightstart to unravel, as companies in trouble draw downtheir hidden reserves. The demise of this institutionis unlikely without a major recession, and not certaineven with one; however, if it does happen, thesystem is likely to come apart rapidly.60 That,according to Ballon and Tomita, “would have

55H1dW Ishihwa, “Jap~’s Compliant Shareholders, “ The Asian Wall Street Journul Weekly, June 13, 1988, p. 17.56sW~tia ~dwq ASWciation ~ta, ~mpi]~ from Fl~ of F~ A~~ou~s, F~er~ Re~~e Bo~d. Thi.s total is down from 65 percent in 1985

and 85 percent in 1%5.sTB~lon and Tbmita, op. cit., p. 52.ssB~lon and Tomita, op. cit., p. 202.59y. K~o~wa, Op. Cit., p. 20.

@Personal communication, OTA staff with Kimihide Takano, Senior Analyst, Corporate Division, The Nikko Research Center, Ltd., Tbkyo, Mar,22, 1989.

Chapter 3---Financing Long-Term Investments . 105

profound repercussions on the stock market and theJapanese economy as a whole.”6l It would tend toshorten the perspectives of Japanese managers andfirms, making them more like American fins.However, given the pervasive effect of administra-tive guidance from the Ministry of Finance, it seemsunlikely that the Japanese financial market willbehave a great deal like that of the United Statesanytime soon.

In sum, a network of policies, practices, andrelationships acts to support heavy investment inlong-term performance in Japanese industry byspreading risk. In contrast, American firms mustcarry more of the risk of such investments bythemselves. While changes are occurring in theJapanese financial market, the backlog of more thanthree decades of such advantages has been highlyeffective in putting Japanese firms in the securepositions they now hold, relative to American andEuropean competitor. Even if the changes weredramatic and rapid (which they are not) theseadvantages would not disappear quickly. It may wellbe that alterations in the way American managers aretaught to think about business could foster a morepositive attitude toward long-term investment, par-ticularly in improved technology. But it is the rulesunder which they must operate rather than theireducation that is the principal influence on how U.S.managers view long-term investment.

Even with changes in the rules, however, therewill be outliers. High capital costs have hobbled butnot crippled American firms in international compe-tition; some firms are able to make substantialinvestments in technology development for manyyears. If a firm exploits its R&D effectively, suchinvestments are rewarded, not penalized, by equityholders. But now, with increasing competition, morefirms are forced to choose between supporting profitmargins or stock prices and postponable expendi-tures like R&D.

Some long-term investments pay off, and somedon’t. We should not expect that risk-sharing willnecessarily result in longer term investment acrossthe board in America, or that every long-terminvestment will be successful. However, withoutsome changes in the financial rules of the game,

American companies will continue to focus mostlyon short-term profit, to their detriment in interna-tional competition.

THE AMERICAN FINANCIALMARKET

The problem for America is not only that Japan’scapital costs are lower than those of the UnitedStates, or that Japan’s providers of debt and equitycapital are content to take more of their rewards ascapital gains rather than as cash payments. Amongthe developed nations, Japan goes unusually fardown these paths. America is, for the most part, atthe other end of the scale. Our capital costs are highnot only relative to Japan’s, but relative to those ofmany European countries as well, and they are highin real terms, compared to what they were in the1960s and 1970s. Institutional investors are, ifanything, more insistent on receiving short-termfinancial gains than they have been, and they havepowerful tools to use if their interests are notaddressed. Rather than mobilizing its resources tosupport American manufacturing during its difficul-ties, the United States often seems indifferent to orcontemptuous of the nation’s manufacturers. Theproblems of manufacturers, we often say, are self-generated; manufacturing is badly managed, andbadly managed firms ought to fail, or change hands.The contrasts with Japan, and with Europe as well,are great.

Some—not all--of what we attribute to badmanagement is simply a matter of intelligent peopleplaying by the rules. If our interest rates are such thatAmerican managers can prudently invest $0.37 inreturn for $1.00 in 6 years, while a Japanese managercould invest $0.66 for the same return,62 we wouldexpect to see about half as much long-term invest-ment in America as in Japan. If stockholdersevaluate a company’s performance on the basis ofquarterly or half-yearly reports of profit, we wouldexpect managers to emphasize short-term profits,even when it raises possible conflicts with longerterm investment. And if showing a profit forshareholders is one of the most important factors inthe survival of a business, we should expect to seefinancial specialists wielding more power in compa-

61B~lon and lbtnita, op. Cit., p. 53.6~e= fiWes reflWt tie actu~ co5taf<api~ difference of Japan and Americ& according to one calculation. See J~es M. po~rbas ‘me cow of

Capital Consequences of Curbing Corporate Borrowing, ’ Testimony before the committee on Ways and Means, U.S. House of Representatives, May16, 1989.


nies than in nations where share price is a lesspressing daily concern to company managers. Thepreoccupation with finance and short-term shareprice performance was reinforced by the wave ofmergers and acquisitions American business experi-enced in the 1980s. Rather than moving toward anenvironment more conducive to long-term invest-ment in the development and use of outstandingtechnology, the U.S. system raised the hurdles.

Another complicating factor is instability in thefinancial environment. Federal decisions affectingthe value of the dollar and interest rates take businesscompetitiveness into account only tangentially, if atall; yet such changes can have profound effects onthe ability of businesses to make prudent long-terminvestments. Again, Japanese policies contrast sharply.U.S. Government support for long-term research,development and investment has also been some-what shaky, leaving businesses that invest in suchprojects vulnerable. For example, the Administra-tion sent confusing signals about its support fortechnology development in semiconductors andhigh definition television in 1989. Even if themodest support for R&D in these areas is continued,the unreliability of Federal commitment to suchprograms could make industry wary of such ven-tures. 63 Another example of the inconstancy ofFederal efforts to promote technology developmentand diffusion is the impermanence of tax measuresthat favor capital spending or R&D.

In short, America’s financial environment isgenerally unfavorable to long-term investments intechnology development and diffusion, and govern-ment actions that mitigate the effects of this unfavor-able environment have lacked commitment.

The Decline in Savings

Nations must continuously invest in productiveassets-plant and equipment, people, and technol-ogy development—to sustain investment and livingstandards. Investment funds come from saving,domestic and foreign. In the 1980s, an increasingproportion of U.S. investment has come from

foreign saving, because U.S. savings rates havefallen.

In the 1970s, net national saving (the percent ofnational income saved by business, government, andhouseholds) averaged 7.9 percent. Of this, 96percent was invested domestically, and 4 percentwas invested abroad. In the 1980s, savings ratesdropped, and by the middle of the decade-1985 to1987—net national saving dipped to 2.1 percentbefore rising to just above 5 percent in 1989. Netdomestic investment (the percent of national incomeinvested) dropped to 5.7 percent, lower than in the1970s but greater than the amount of investablecapital provided domestically. The United Statesmade up the difference by becoming a net importerof investment funds, borrowing $417 billion fromabroad over the 1985-87 period.64 To attract savingsfrom abroad, the United States has had to raiseinterest rates, or the return to investors. Importingcapital allowed the United States to invest more thanits own savings would permit, but it also raised theprice of domestic investment. This means thatimproving and replacing productive assets andtechnology for U.S. firms became more expensive inthe 1980s. A nation trying to keep pace withwell-financed and technologically sophisticated com-petition can ill afford this.

The decline in savings occurred across the board.The sharpest change in the 1980s was a decline ingovernment saving, manifested by budget deficits atthe federal level. Falling household and businesssavings contributed to the decline as well. TheFederal budget deficit resulted from a tax cut, whichslowed the growth of revenue, and from increasedoutlays, principally for defense.

The reasons behind falling household savings areless obvious. Many explanations have been ad-vanced for this drop-and conversely, the rise inconsumption as a percent of national income—butthere is little consensus on which are most signifi-cant. Some analyses attribute part of the decline tohigh interest rates, which made it possible forcorporations to decrease contributions to pension

631n ]~e 1989, -Or$ of an A&ninis~ationpropos~ t. kill f~ding for Sematwh in tie fisc~ ye~ 1991 budget s~fac~. The nunor woseconcurrentlywith Administration proposals to shut down the Defense Manufacturing Board, and an OMB proposal to reduce DARPA funding for HDTV, While theAdministration eventually denied any plan to kill funding, the rumor was widely believed and taken seriously by much of the electronics industry. See‘‘Administration Charged With Seeking Funding Cuts for Sematech, Other Projects,’ lnternatwnal Trade Reporter, Nov. 15,1989, pp. 1481-1482; andLucy Reilly, “Death Knell for Sematech?” New Technology Week, Nov. 6, 1989, p. 1.

@George N. Hatsopoulos, Paul R. Krugman, and James M. Poterba, Overconsutnptwn: The Challenge to U.S. Econom”c Policy (New York, NY andWashington, DC: American Business Conference and Thermo Electron Corp., 1989), pp. 6-7.

Chapter 3---Financing Long-Term Investments ● 107

funds (these are included in household savings). Thejury is out on the effect of demographics. Some thinkthe baby boom was a major factor in increasingconsumption rates: since young people typicallysave less than the middle-aged, they expect personalsavings rates to rise as the baby boomers mature.Others dismiss demographics as having little ex-planatory power. Another often-cited argument isthat gains in wealth in the 1980s--capital gains oncorporate equities and homes----encouraged con-sumption. If people feel richer because their assetsare increasing, goes the argument, they feel less needto save. On the other hand, since real wages andsalaries dropped during the 1980s, falling savingsmay reflect attempts to keep up consumption pat-terns in the face of (for most families) decliningincomes.65 Another theory is that the propensity toconsume may have been fueled by the easy availabil-ity of consumer credit.66

The enormous increase in Federal Governmentdebt and the fall in household savings rates wereenough by themselves to force a curtailment ofcapital formation, or a switch to capital imports, orboth. The decline in business saving has been lessremarked, but is important for two reasons. Betweenthe mid-1960s and the late 1970s, business saving—measured in national accounts by the retainedearnings of corporations-fell from 4.5 percent ofGNP to 2.75 percent. By the mid-1980s, businesssaving fell still further, to 1 percent of GNP.67 Unlikethe ballooning Federal deficit and falling householdsavings, the decline of business savings is long-standing, and cannot be fully understood in terms ofthe events of the 1980s alone. Nonetheless, thedepression of business savings to the lows of the1980s is part of another change in the financialenvironment-that is, mergers and acquisitions—that limits the willingness of American companies tomake long-term investments.

Mergers and Acquisitions

Mergers and acquisitions are a normal feature ofthe U.S. financial landscape, and ordinarily not acontroversial one. Occasionally, though, merger andacquisition (M&A) activity heats up, as it did in the1980s, provoking debate and examination. M&Aactivity has raised many questions including those ofbasic efficacy (are mergers and acquisitions really aneffective managerial disciplinary force, for example)and effect (do mergers and acquisitions generallyimprove long-term productivity, or produce out-comes as desirable from society standpoint as fromtarget shareholders’?). None of the questions areresolved. Even questions that are somewhat periph-eral to the whole debate—such as the effect onmanagers’ willingness to undertake longer terminvestments in technology development and diffu-sion—are hotly debated. While there is a growingbody of research and empirical evidence on thecauses and consequences of M&A, there are fewpoints of consensus in the argument. But it is clearthat the takeover wave of the 1980s is a specialfeature of the American financial environment,much more prominent here than in any other nation.The length of the following discussion is not meantto imply that M&A is the only, or even the major,factor that causes American managers to focusstrongly on short-term profit, but M&A does inten-sify the pressures of the American financial environ-ment, characterized by high interest rates and capitalcosts and macroeconomic instability.68

Briefly, the argument goes as follows. One pointof view-often articulated by businessmen—is thatcorporate raiders have forced a preoccupation withshort run performance that has disrupted businessplanning. With access to new capital instruments(junk bonds), acquirers can afford to pay inflatedprices to get controlling interest in their targets. Thefirst defense against potential raiders, therefore, is tokeep the stock price high enough to fend them off.Since stock prices can fall significantly on disap-pointing quarterly profit performance, business man-

GsKatherine Gillrnan and Joy -erley, “Is the Middle Class Shrinking?” Furures, April 1988.~he following sources discuss reasons for falling savings rates: Barry P. Bosworth, ‘ ‘There’s No Simple Explanation for the Collapse in Saving,’

Challenge, July-August 1989, pp. 27-32; George N, Hatsopoulos, Paul R. Krugman, and James M. Poterba, Overconsumption: The Challenge to U.S.Economic Policy (Washington, DC: American Business Conference, 1989); David E. Bloom and Todd P. Steen, “Living on Credit,” AmericanDemographics, October 1987, pp. 22-29; and William D. Nordhaus, ‘‘What Wrong With a Declirting National Saving Rate?’ Chullenge, July-August1989, pp. 22-26.

67Nordhaus, Op. Cit., p. 23.

6s The u~t~ States is not ~W~le compm~ t. most Comtnes, but the American financial environment for business is less stable than that of eitherJapan and West Germany, our premier international competitors.


agers must focus on keeping short term profits atacceptable levels. This, in turn, exaggerates thealready short-term planning horizons of Americanbusiness. 69

In some cases, more drastic steps maybe taken tofend off a potential takeover, such as taking thecompany private by means of a leveraged buyout(LBO), or implementing some kind of “poison pill”defense. While these strategies can keep the com-pany from changing hands, the effects on planninghorizons can, ironically, be no friendlier to long-term investment and planning. In the case of adefensive LBO, the company exchanges equity fordebt, making it safer from raiders but harder pressedto maintain cash flows. Debt payments must bemade, while dividends can be postponed during thintimes. Cash flows that could have been invested inresearch and development, plant and equipment, orother long-term projects must be at least partlydedicated to paying interest and debt retirement; socompanies may defer long-term projects in favor ofmeeting their short-run obligations.70

Current concern is spurred by the fact that theavailability of high-risk, high-return bonds hassubjected many more companies to the threat of atakeover than in the past. Junk bond financing canturn even relatively small operators into potentialraiders, and even large companies are not immunefrom the possibility of a takeover. Any company thatappears undervalued may be fair game.71 Moreover,a company’s value to a raider can seem inordinatelyhigh to many business managers;72 company manag-ers feel pressed to keep their stock price above eveninflated asset value.

The foregoing argument raises two questions.First, it is difficult to accept at face value thecontention that a price can be too high if a willingbuyer agrees to pay it. The difference between

managers’ estimation of the real value of theircompanies and that of potential acquirers maytherefore be that outsiders can see higher yieldingopportunities for managing companies’ assets thanmanagers do. Experts hold divided opinions onwhether acquisition prices are too high.

The concern implicit in the arguments of manybusinessmen is that equity markets consistentlyundervalue long-term investments. If the resultingstock prices do not fully reflect the companies’investments in future output, then perhaps acquisi-tion prices are not too high, but represent a morerealistic appraisal of long-term company value.Here, too, there is no consensus of expert opinion,but it should be pointed out that there is no necessaryinconsistency here: while ordinary stock prices maybe too low, acquisition prices may be too high.73

The opponents in the debate view debt verydifferently. Those who see takeovers and mergers asa necessary disciplinary force on management seethe higher debt levels that result from much of thecurrent takeover activity as keeping managers fromsquandering corporate assets on less productiveventures. 74 Others regard the high debt that oftenresults from a hostile takeover, or a defense againstone, as a ball and chain hampering companies’abilities to invest, particularly in long-term ventureslike R&D. The pressure of high debt load is expectedcause many defaults or bankruptcies in a recession.Even without a recession, however, the junk bondmarket is troubled; in 1989, corporate bond defaultswere up 136 percent over 1987, largely due todefaults on junk bonds.75

Most of the evidence indicates that the directeffect of all kinds of M&A activity on R&Dexpenditures or intensity (R&D as a percent of sales)is small or negligible. Bronwyn Hall, examiningapproximately 250 manufacturing acquisitions be-

@John C. Coffw, Jr., ~uis ~we~~in, and Susan Rose-Ackerman, Knighrs, Rai&rs and Targets (New York, NY: Oxford University ~ess. 1988),pp. 34.

7Wor a briefsummary of the arguments on both sides of the controversy, see Robert R. Miller, ‘‘The Impact of Merger and Acquisition Activity onResearch and Development in U.S.-Based Companies,’ contractor report to OTA, November 1989. The report is a summary of interviews with R&Ddirectors of 19 fms with a variety of M&A experiences. Some had undergone friendly mergers, some hostile takeovers, some leveraged buyouts, anda couple had no resent experience with M&A.

71 Miller, op. Cit., p. 3.

TzW~en E. Buffett, Mich~l D. Dingman, and Harry J. Gray, with Louis Lowenstein, Moderator, ‘‘Hostile Takeovers and Junk Bond Financing: APanel Discussion,” in Coffee, et al., op. cit., pp. 10-27.

T3Coff=, et al., op. cit., P. 4.TqMiller, op. cit., p. 6.75Richmd D. Hylton, “Corporate Bond Defaults Up Sharply in ‘89, ” The New York Times, Jan. 11, 1990.

Chapter 3--Financing Long-Term Investments . 109

tween 1977 and 1986, concludes that the post-acquisition R&D intensity of the firms was about thesame as pre-acquisition; moreover, the R&D inten-sity of the post-acquisition firms was not differentfrom the R&D intensity of all manufacturing firmsduring the same period.76 In addition, there is abroadconsensus that R&D-intensive firms are unlikely tobe attractive takeover targets, and that the majorityof M&A happens in firms that do relatively littleresearch and development.77

Some use this kind of evidence to dismiss thepossibility that M&A is having corrosive effects onR&D in particular or long-term investment inparticular.78 Yet there is reason for skepticism. First,while much of the evidence supports the contentionthat the effect of M&A on R&D is small, it is notunanimous. The National Science Foundation exam-ined the R&D spending and intensity of the 200largest industrial R&D performing companies in1984-86. 79 These companies account for almost 90percent of all U.S. industrial research and develop-ment. Among the 200 firms were 24 firms that hadeither merged or undergone an LBO during theperiod; these 24 accounted for nearly 20 percent ofthe R&D spending of the entire group of 200 in1987. The firms that did not undergo restructuringincreased real spending on R&D by 5.4 percent,while the 24 firms that were restructured throughM&A reduced their R&D spending by 8.3 percent inreal (deflated) terms from 1986 to 1987. Theseoverall findings were consistent with comparisonsof restructured and unrestructured firms at theindustry level as well.80 The NSF data should beinterpreted cautiously-the study spans only 3years, and some of the reductions in R&D might beelimination of redundant programs in newly merged

companies—but they indicate a need for equalcaution towards studies that show negligible impactsof restructuring.

One possible reason for inconsistencies betweenthe studies cited above is that not all restructuringsare alike. One of the few points of consensus in thedebate is that M&A in the 1980s is unlike earlierwaves of M&A activity, and is certainly differentfrom the background level of restructuring. Differentkinds of restructuring-friendly mergers, hostiletakeovers, defensive LBOs, and other managementbuyouts, for example-would be expected to havedifferent effects on managers’ abilities and incen-tives to invest in R&D and other activities consid-ered discretionary in the short run.

The last wave of M&A activity, which occurred inthe 1960s, was characterized by diversification andagglomeration. The 1980s, in contrast, are character-ized by so-called bustup takeovers of diversifiedcompanies with subsequent selloffs of the compo-nents.81 Hall’s study includes many mergers fromwhat could be considered another era--the late1970s--which may blur the effects observed by theNSF study which focused on the mid-1980s. Highdebt is closely associated with the bustup takeover.Friendly mergers often have little or no effect onoverall corporate debt levels, while hostile takeoversand defensive LBOs, in particular, often leave veryhighly leveraged companies in their wake. One ofthe striking effects of the 1980s wave of M&A is thesubstantial increase in corporate debt attributed to it.According to one estimate, the corporate debtburden was 20 percent higher in 1988 than it wouldhave been without the effects of corporate restructur-ing. 82

7~ew ~eS~tS ~ Sm~~ ~ c ‘TeStimony of Bronwyn H~l in He~ngs on corporate Res~c~g and its EffWts on R&D Before the Science,Research, and Technology Subcommittee of the House Committee on Science, Space and Technology, July 13, 1989;” and Bronwyn Hall, “Effwt ofTakeover Activity on Corporate Research and Development,’ Alan J. Auerbach (cd.), Corporate Tdeovers: Causes and Consequences (Chicago, IL:University of Chicago Press, 1988), pp. 69-96.

77sW, for exmple, Lawrence SummerS> “LBO Debt and Taxes,” Across the Board, April 1989; Hall, op. cit.; and Abbie Smith, “CorporateOwnership Structure and Performance: The Case of Management Buyouts,’ Leveraged Buyouts and Corporare Debt, Hearing Before the Committeeon Finance, United States Senate, Jan. 24, 1989.

TWW exmple, w Jo=ph A. Grundfest~ ‘‘M&A and R&D: In Corporate Restructuring Stifling Research and Development?” Address to NationalAcademies of Sciences and Engineering, Academy Industry Program of the National Research Council, Oct. 11, 1989.

79The tem “~dm~~” refers to companies in mining, cons~ction, and manufacturing. The vmt majority are in mmUfaCtUring.8~est~ony of ~. Willim L. Stewm, Nation~ Science Foundation, ~fom tie Committee on Science SpaCe and TtxhnoIogy, Subcommiu* on

Science+ Research and Technology, House of Representatives, July 13, 1989.81Lym E. Browne and Eric S. Rosengren, “The Merger Boom: An Overview, “NW Enghd Economic Review, March/April 1988. P. 23.s~ol~an Sachs, Fi~ncia/MarketF’ersPectives, December 1988, quot~ in Lawrence SummerS~ ‘‘Taxation and Corporate Debt,’ in U.S. Congress,

House of Representatives, tiveraged Buyouts and Cmporate Debt, Hearing Before the Committee on Finance, U.S. Senate, Jan. 25, 1989. TheGoldman-Sachs analysis shows the outstanding debt of nonfhancial corporations as a percent of the gross domestic product of those corporations at 66percent in 1988, compared with an estimated 55 percent without restructuring.


It is quite possible that high-debt restructuring hasa greater impact than friendly mergers on R&D. Thisproved to be the case in OTA’s interviews with 19manufacturing companies representing a variety ofdifferent restructuring experiences. Although thesample was not a statistically valid sample of M&Aas a whole, the firms that had increased debt as theresult of a takeover or as a defense against a takeoverconsistently reduced R&D following the event. Thereductions may not prove permanent-companiesmay rebuild R&D as they pay down their debt—butmost of the R&D managers of the firms that had cutback also believed their firms’ future ability tocompete was compromised as a result.83 Halldownplays the overall importance of R&D cutbacksfollowing LBOs (which invariably results in muchhigher leverage), citing evidence that most firms thatundergo LBOs do no R&D. Also, Hall points outthat in her sample of 200 manufacturing acquisi-tions, 30 were LBOs. Those 30 had very low R&Dintensity-on average, 0.4 percent of sales-andaccounted for only 1 percent of the R&D done in theprivate sector in the years 1984-86.84

What all this seemingly conflicting evidence maymean is that LBOs as a whole have not directlyaffected R&D overall by a measurable amount, butthat LBOs in large manufacturing firms have re-sulted in reduced R&D, at least in the short run,because of the pressures of high debt. Indirectsupport for this conclusion comes from anotherstudy. Abbie Smith found that R&D intensitydeclined in firms that reported R&D expendituresbefore their LBO, and that sold assets after the LBO.Smith warns against any conclusory interpretationof this result, however, because so few of the firmsin the population of LBOs studied reported anyR&D at all.85

Another complicating factor is firm size. Mostservice firms and small manufacturing firms per-form very little or no R&D. The fact that NSF’s top200 R&D spenders accounted for 90 percent of allindustrial R&D is telling. Summers points out thatmany LBOs occur when the owner-manager of a

small establishment approaches retirement, and thatthese are “almost certainly benign.”86 In anothercommon LBO situation, a company finds that acertain line of business no longer fits into its overallstrategy, and makes amicable arrangements with themanagers of a division for the sale. Again, thesebuyouts could be expected to have little or no effecton R&D, either because many of the firms involveddo little or none, or because amicable transfer ofownership of a division to its current managers canoften be accomplished without the high acquisitionprices often associated with LBOs.

Analysts have concentrated more on the effects ofM&A on research and development than on itseffects on other discretionary expenditures. ButR&D isn’t the only kind of discretionary expendi-ture that affects a fro’s technology; the other iscapital expenditure. There are no clear and consis-tent answers to questions about the effects ofcorporate restructuring. Capital expenditure is nec-essary if firms are to keep up with and advancetechnology, but like R&D, capital expenditure maybe postponed for a short time without long-termmaterial damage to a fro’s technological base. Theduration and depth of sustainable cuts varies byindustry and by firm, but even so, available evidencegives some cause for concern. Smith reports asubstantial and significant reduction in capital ex-penditures as a percentage of sales that occurred in58 management buyouts between 1977 and 1986.87

This finding is consistent with anecdotal evidence.For example, consider Houdaille, a machine-toolmaker that underwent an LBO in 1979. Pressured byforeign competition and (later) the effects of the1982 recession as well as its high debt burden,Houdaille cut capital spending as a percent ofrevenues in half following its post-buyout restruc-turing. 88 One owner of a machine-tool makingbusiness states, “When we hear LBO, we knowthey’re not going to be buying anything.”89

Most analyses of the consequences of M&A havebeen confined to measurable direct effects—spending on various activities or overall perform-

83Milkr, op. cit., p. 14.

‘Hall, op. cit., p. 3.8SSmj@ Qp, Cit. PI 71o

%hunmers, op. cit., p. 187.S7S~~, op. Cit., p. 47-

88M= Holl~d, “HOW to Kill a Company, “ The Wmhington Post, Apr. 23, 1989.8gHowad G~is, ~sident, Kinefac, personal communication, NOV. 16, 1989.


ance of companies that have undergone restructur-ing. Two others should also be considered. First,there are qualitative effects, not readily measurable,on R&D or firm activities. Again, we would expect(and find, according to the limited evidence) thatdifferent kinds of restructuring have different quali-tative effects. In OTA’s interviews, firms thatmounted successful defenses against hostile take-overs (leaving the companies with high debt)long-term R&D had invariably been significantlycut back in favor of projects with promise of quickerpayoff. 90 Some analysts interpret this kind of cut-back as making R&D more efficient, and this isindeed possible in the short run. R&D is by its naturea long-term process, and firms can cut back on newlong-term projects without impairing their ability toexploit the results of projects undertaken in the past.So a shift in emphasis toward shorter term projectswould be unlikely to show up as detrimental for atleast a year or two. But in the long run, it seemsunlikely that increasing the focus on short-termprojects on the part of American firms will permitthem to maintain even their current level of compet-itiveness.

Friendly mergers, on the other hand, had eitherlittle impact on R&D, or effects that would begenerally accepted as positive. One example is thepurchase of Celanese Corp. by the West Germanchemical firm Hoechst. Hoechst was interested inexpanding its U.S. operations through the purchaseof an American firm with strong R&D, and after theacquisition increased Celanese’s R&D expendituresby 10 percent annually. Significantly, the newGerman managers were also more willing to commitsubstantial resources to long-term projects with lesscertain payoffs.91 A similar story was told by thepresident of Materials Research Corp., a semicon-ductor equipment and materials company recentlyacquired by Sony. After the deal was completed, thepresident was told by Akio Morita, the president ofSony, that he had “essentially unlimited capital,”and was no longer obliged to concern himself withquarterly profits. “I can think of projects that taketwo years, ” said Dr. Sheldon Weinig, the president.“It’s a wonderful way to live.”92

It is difficult to make a few cases add up to astrong finding, but the anecdotes about the qualita-tive effects on manufacturing R&D of differentkinds of M&A activity are consistent with quantita-tive evidence, if the focus is adjusted correctly. Inother words, both the qualitative and quantitativeevidence suggest the following: in manufacturingfirms that have appreciable amounts of R&D,restructurings that result in high debt levels depressR&D spending or intensity, or both, and oftenshorten the allowable time for completion of R&Dprojects. Because such restructurings are not common—most happen in firms that do little R&D, and manyof them are in service fins-the overall directeffects of M&A on overall national R&D are not yetlarge, and may never be, particularly as hostiletakeover/LBO activity seems to be winding downfor now. This does not justify complacency aboutM&A. NSF’s data are disturbing, and will be moreso if the highly leveraged companies continue to lagin R&D spending or long-term planning. Additionaldepression of discretionary expenditures on capitalequipment or R&D could well occur in the event ofa recession, or perhaps even when growth is less thanrobust. Such cutbacks, normal in recessions, aremore likely when companies are highly leveraged.

Finally, the indirect effects of M&A must beconsidered. The 1980s added a new wrinkle to thetakeover enterprise: the expansion of the pool ofpotential raiders. In the past, in most takeovers, largefirms acquired smaller ones. In the 1980s, junkbonds made it possible for “individuals, smallerentities, and investment banking fins” to takepart. 93 In another contrast to past takeover waves(and ordinary M&A activity), these new playersoften intended to dismantle the acquired companyrather than to assimilate it. Both factors-theincrease in number of raiders, and the consequencesof a successful takeover-have apparently increasedmanagers’ fears of takeovers markedly, and mayalso have depressed discretionary expenditures.Managers, feeling that an unwelcome takeover bidmight come at any time, might take steps thatapproximate what they would do to defend againsta real hostile takeover bid, with the same effects onspending for R&D and capital equipment. In mid-

~iller, op. cit., p. 18.glMiller, op. cit., p. 31.WAII*W Pollack, ‘Cne Challenge of Keeping U.S. Technology At Home, ” Z% New York Times, k. 10, 1989.

93John C. Coffee, Jr,, “Shareholders Versus Managers: The Strain in the Corporate Web,” in Coffee, et al,, op. cit., p. 77.


1989, for example, Honeywell acted to discouragepotential raiders by cutting out certain lines ofbusiness (reducing the breakup value of its assets),eliminating 4,000 jobs, repurchasing up to 10million shares of its own stock, and increasing itsannual dividend to shareholders by 31 percent.94

There had been speculation that Honeywell might bea takeover target, but no actual bid.

Few companies make moves as dramatic asHoneywell’s, but many members of corporateboards and senior managers report that hostiletakeovers came to dominate corporate board meet-ings and decisionmaking to an unprecedented extentin the 1980s. The effect on overall business plan-ning, almost certainly, was to increase the emphasison distributing profits to shareholders in preferenceto reinvesting in the company.

Hostile takeover activity seems to be windingdown, although not crashing; the number of dealscompleted in the first 9 months of 1989 was smaller,according to a preliminary estimate, than the number

in the first 9 months of any of the preceding 3 years.The first three quarters of 1989 saw 2,298 completedacquisitions, compared to 2,790 in 1988, 2,851 in1987, and 2,707 in 1986. However, the value of thesedeals in 1989 was $144 billion, just below the peakof $144.7 billion in 1988. The story is different forLBOs: there were slightly fewer completed in thefirst 9 months of 1989 (214) than in a similar periodof 1988 (221), but the total value of those LBOs in1989—$47 billion—was quite a bit higher than theprevious high of $29.1 billion in 1988.95 T h enumbers aren’t the only story. There is a widespreadperception that the market has grown pickier aboutthe kind of deals that can be approved, and there hasbeen a flight from junk bonds.96 Acquisitionscontinue, but many believe that the wave of highlyleveraged, bustup takeovers is on the wane. If this istrue, it could provide time to examine how much ofthe negative effects of M&A is associated with thisparticular type of financial activity, and time forpolicymakers to evaluate how to tailor possibleregulation to the real problems.

%Tony Kennedy, “Honeywe]l ACE Agtist PotentiaJ Raiders,” The Washington poM j~Y 2571989,95jUdi~ H. Dchyzinslci, “Deals, Yes. Maniac Deals, No,” Business Week, Ox. 30, 1989.~hristopher Farrell, with I-ah J. Nathans, “The Bills Are Coming Due, ” Business Week, (M. 30, 1989.

Chapter 4

Human Resources

CONTENTS

EDUCATION: PREPARATION FOR COMPETITIVENESSTHE MANUFACTURING CONNECTION . . . . . . . . . . . . . . .

Page. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117THE TECHNICAL AND ENGINEERING WORK FORCE . . . . . . . . . . . . . . . ● . . . . . . . . . . . . . . + , 120

The Engineering and Scientific Work Force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120Numbers and Distribution of Scientists and Engineers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . +... 121The Functions of Engineers in Japan and America . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124

FiguresFigure Page

4-l. Twelfth Grade Achievement Scores in Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1164-2. Twelfth Grade Achievement Scores in Advanced Algebra . . . . .+ . . . . . . . . . . . . . . . . . . . . . . . . 1174-3. Scientists and Engineers per 10,000 Labor Force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1224-4. Employment of Scientists and Engineers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1224-5. Trends in Science and Engineering Labor, 1976-88 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1234-6. Salaries of Engineers and Laborers of Large Establishments in Japan, 1987-88 +.. . . . . . . . . 124

Chapter 4

Human Resources

Manufacturing, like the rest of the economy,depends on the competence and ingenuity of work-ers, from the shopfloor to the executive suite.Sophisticated technology demands able people. Justas powerful machines can enhance the productiveabilities of people, it takes well-trained people to getthe best out of the machines.

The need for highly qualified people is notconfined to an elite; the most productive technolo-gies are those that exploit the talents of skilledpeople at all levels. This has been a cherishedprinciple of American development, manifested inmany ways. One is the commitment of the UnitedStates to universal education, probably the mostimportant investment a nation makes in its people.During most of the 19th and 20th centuries, theUnited States enrolled a larger percentage of itspopulation in school than did European countries.1

Even now, although there are many serious prob-lems with educational quality, American enrollmentin primary and secondary education is among thehighest in the world, and in postsecondary educationthe United States ranks much higher than any othernation. Fifty-seven percent of the relevant age groupwas enrolled in postsecondary education in theUnited States in 1987, compared to a weightedaverage of 38 percent in all other industrial marketeconomies and lower averages for developing andless developed nations.2 Nathan Rosenberg, describ-ing the factors that led to the rapid rate of technolog-ical innovation in 19th-century America, writes,

Not only did American society devote a largeproportion of its resources to inventive activities; itis also apparent that the human resources of thecountry were well-equipped through formal educa-tion with the skills which might raise their productiv-ity both as inventors and as successful borrowers andmodifiers of technologies developed elsewhere.3

Kazuo Koike, writing about contemporary Japa-nese manufacturing and skills, puts it this way:

The essence of the contribution of high morale is. . . in devising better work methods and production,which in turn demand technological knowledge byworkers for maintenance . . . This kind of wide-ranging skill contains such knowledge and promotesthe ability of workers to determine the causes ofproblems on the shopfloor and thus to contribute toproductivity. 4

Rosenberg and Koike both stress technologicalknowledge, and that is no accident. All fast develop-ing and developed nations put heavy emphasis oneducation-both on high-quality education and onbroad participation by all ranks of citizens. Amongthe developed countries, those best known for theirheavy investments in education are either the richest(West Germany, Sweden) or the fastest growing(Japan).

Many leading-edge companies that have beenmost successful in applying advanced automation inmanufacturing put a particularly high premium onthe cognitive skills of workers. By replacing humanlabor in the more routine tasks, they create a greaterconcentration of tasks that require judgment andcomplex knowledge. The best preparation for aworklife that puts increasing emphasis on judgmentand knowledge is a good education. Providing thispreparation is now a grave challenge for America. Itis the wellspring of competitive ability in Japan,several Asian developing nations, and many Euro-pean nations.

EDUCATION: PREPARATION FORCOMPETITIVENESS

During much of the 20th century, the UnitedStates had the best educated work force in the world,and American manufacturing was the world’s mostdynamic and competitive. There is a causal connec-tion between these two, although it is not perfect. Atthe turn of the century new forms of industrial andwork organization, known now as Taylorism and

lllichmd A. ~terlin, ‘A Note on the Evidence of History, ’ Education and Econon”c Development, C. Arnold Anderson and MMY Jean Bowrnm(eda.) (Chicago, IL: Aldine Publishing Co., 1965). The figures are reproduced in Nathan Rosenberg, Techno/ogyandArnerican Economic Growth (NewYork, NY: Harper & Row Publishers, 1972), p. 38.

z~e World Bank, Worfd DeveJOp~ntRepOr-t 1987 (New York, NY: Oxford University hss, 1987), Pp. 262-263.qNfi Ro~nberg, ~ec~~fl mtdAmerkan Economic Growth (New York, NY: Harper& ROW Publishers, 1972), P. 35.

qKazuo Koike, “Human Resource Development and Labor-Management Relations, ” The Political Economy of Japan, Volume 1: The DomesticTransformation, Kozo Yamamura and Yasukichi Yasuba (eds.) (Stanford, CA: Stanford University Press, 1987), p. 327.


Fordism, tried to reduce jobs to their simplestcomponents, which sometimes also had the effect ofreducing the educational demands made on workers.However, many ordinary workers continued to bringingenuity and creativity to their jobs, and it was thisfact, as much as the efficiency of the assembly-linemethod, that impressed foreign observers aboutAmerican manufacturing.s It is not a coincidencethat America is now slipping on both counts,educational performance and manufacturing com-petitiveness.

American students perform poorly on standard-ized tests compared with their counterparts in manynations of Asia and Europe. Since the 1970s, theyhave compared unfavorably with their predecessorsin American schools as well. In the mid-1980sAmerican junior high school students ranked 10th inarithmetic, 12th in algebra, and 16th in geometry ina survey of mathematics competence in 20 coun-tries. 6 Twelfth graders, compared with students from14 other nations, ranked 12th in geometry and 14thin advanced algebra, according 1981-82 survey7

(figures 4-1 and 4-2). American students scoredbelow students in Canada (Ontario), Scotland,Finland, Sweden, Japan, New Zealand, Belgium,England and Wales, and Israel in functions andcalculus. Of the students tested, only those inHungary and the Canadian province of BritishColumbia performed worse. Moreover, the surveyshowed that the performance of American studentshad worsened in the past two decades. At the time ofthe frost international mathematics study in the early1960s, the top 5 percent of American students wereperforming as well as the top 5 percent anywhere inthe world. By the 1981-82 survey, the top 5 percentof American students had sunk to the bottom quarterof the scores of the top 5 percent in other nations.8

The results are similarly dismal in science. Also,compared with students in many other developednations, American students are less likely to learnforeign languages.

The deterioration in the performance of Americanstudents since the 1960s is just as disturbing as their

Figure 4-l-Twelfth Grade Achievement

Hong Kong

Japan

England\Wales

Sweden

Finland

New Zealand

Belgium(Flemish)

Scotland

Canada(Ont)

Balgium(French)

Israel

Us.

Hungary

Canada(B.C.)

Thailand

Scores in Geometry

o 10 20 30 40 SO 60 70 80Mean score

SOURCE: International Association for the Evaluation of EducationalAchievement, The Underachieving Curriculum: Assessing U.S.School Mathematics From an International Perspective (Cham-paign, IL: Stipes Publishing Co., January 19S7).

poor showing in international comparisons. Formany decades, American students scored higheryear by year on standardized tests such as theScholastic Aptitude Test and the Iowa Test ofEducational Development. This progress was all themore impressive considering the fact that the Amer-ican educational system was at the same timereaching more and more people. From 1890 to 1960,time spent in school, daily attendance, and thenumber of years of schooling completed all in-creased. For instance, the scores of 12th graders onthe Iowa Test of Educational Development roserobustly between 1942 and the mid-1960s, with adramatic spike in test scores after Sputnik’s launch.9

During about the same period (1941-68), highschool graduation in Iowa, where the test wasadministered, increased from 65 to 88 percent of therelevant population. In the late 1960s, the gainsstopped. Scores on many standardized tests began adecline that lasted for over a dozen years. The upturnin test scores in the early 1980s has only partiallyoffset the decline. Young adults who entered the

5Je~.Jqu~ seman-sc~ei~r, The American Challenge (New York, NY: Atheneum, 1%9).q-Iarold W, Stevenw, “America’s Math Problems,” Educational Leadership, October 1987; and International Association fOr the Evaluation of

Educational Achievement, The Underachieving Curriculum: Assessing U.S. School Mathematics From an Internatwnal Perspective (Champaign, IL:Stipes Publishing Co, January 1987).

T~ternatio~ Association for the Evaluation of Education Achievement, op. cit.g~len Hoffmm, “The ‘Education Deficit’, ” The Nationaf Journal, Ma. 14, 1987.

gJohn H. Bishop, “Is the Test Score Decline Responsible for the Productivity Growth Decline?” The American Econorru”c Review, March 1989.

Chapter 4--Human Resources . 117

Figure 4-2-Twelfth Grade Achievement Scoresin Advanced Algebra

Hong Kong

JapanFinland

England\ Wales

Belgium(Flemish)

Israel

Sweden

Canada(Ont)New Zealand

Belgium(French)

Scotland

Canada(BC)

Hungary

Us.Thailand

I

1 1 1 1

0 20 40 60 80 100Mean score

SOURCE: International Association for the Evaluation of EducationalAchievement, The Undwachkwing Curriwlum: Assessing U.S.School Mathemati From an International Perqmctive (Champ-aign, IL: Stipes Publishing Co., January 19S7).

work force in the 1970s were less well preparedacademically than their predecessors, an unprece-dented occurrence in America.10

THE MANUFACTURINGCONNECTION

The strength of a nation’s scientific and engineer-ing work force is connected to manufacturinginnovation and competitiveness in immediate andobvious ways. The academic accomplishments ofshopfloor workers are not so obviously related tocompetitiveness. When we consider the nature ofmuch factory work—short-cycle repetitive tasks—the relevance of performance in science and mathe-matics may seem slight.

Yet manufacturing work is changing with ad-vances in technology, and the changes often demandskills that are more in line with academic compe-tence than those required in earlier generations ofmass-production factory work. Automated produc-tion makes each worker responsible for a largershare of the production process, and creates a greaterneed for each worker in the system to understand

other parts of the system. Emphasis on productquality, often formalized into statistical processcontrol (SPC) procedures, requires workers to havebasic skills in reading and math. For example, at aFujitsu Microelectronics semiconductor plant in SanDiego, California, most production jobs requiregood arithmetic skills, including proficiency withfractions and decimals, to cope with the demands ofSPC. ll

Automated production also requires sound judg-ment and skill in problem solving. An account ofwork in a silicon wafer plant in North Carolinastates,

At DNS, the silicon log in its raw state is worthbetween $2,000 and $5,000. This fact and the costand expense of the machines employed in sawingmake “down time” far more acceptable than scrap.Although the only direct control an operator mayhave over his or her process is an on/off switch,timely and judicious use of that switch is becominga high skill.12

Programmable automation and/or flexible manu-facturing systems require multiple skills, many ofthem new for production workers. Programmableequipment enables one machine or group of ma-chines to make a much wider range of parts orproducts than dedicated machines. In the past,workers could learn in a few days or weeks, bywatching and working with an experienced worker,how to operate a particular machine. Now, workersmust identify more closely with products than withprocesses or machines, and they are less likely to bebuffered from other machines and workers by largestocks of parts and loose schedules. As a result, theymust be more familiar with the whole productionprocess and able to operate multifunctional ma-chines. In such a system, operators can no longer relyon learning by example, but instead must be able toread and understand manuals and specifications.13

These skills are hard to translate into grade-levelequivalents, but training directors of firms that haveconfronted difficulties with problem-solving abilityrecognize that a basic proficiency in reading andmathematics is both a good foundation for and anindicator of problem-solving ability. Motorola, for

lqbid., p. 193.llpa~ V. ~l~r, “Wo~er Training: A Study of Nine Companies, ” contract report to OTA, September 1988.lzIbidO

lsLarry Hirschhom, “Training and Technology in Context: A Study of Four Companies, ” contractor report to OTA, September 1987.


example, determined that workers in its Factories ofthe Future-fully automated semiconductor produc-tion facilities-needed at least sixth grade math andseventh grade reading to cope with demands formastering different jobs in a rotation system, assum-ing responsibility for quality control, and participat-ing in problem-solving work teams.

While these requirements are modest, manyworkers do not possess them. Of a group of 278Motorola production workers who volunteered fortesting, 85 percent were in need of some remedialinstruction in order to meet the standard of sixthgrade math and seventh grade reading. Most of thepeople who failed to meet the standard in bothreading and math were workers whose nativelanguage was not English. Fujitsu’s San Diego plant,producing integrated circuits and semiconductors,had the same problem: lack of the basic skills neededfor effective participation in quality circles (workgroups focused on problem solving). Here, too, thetrouble stemmed largely from the fact that manyemployees, including the Japanese plant managers,were not native English speakers.

This does not mean that basic skills deficienciesin American manufacturing are confined to immi-grant populations. Many companies have found thatpoor basic skills among native workers limit theirability to adopt new technologies. Their experiencesare confirmed by results of the National Assessmentof Educational Progress’ survey of literacy profi-ciency among young adults aged 21 to 25. Althoughthe NAEP findings show that nearly all young adultsare literate in a rudimentary sense, 20 percent ofyoung American adults read no better than a typicaleighth grader and 6 percent do no better than theaverage fourth grader.14 Moreover, very few youngadults were proficient in tasks requiring even amoderate level of complexity. For example, only 9.5percent of the group, given typical grocery storeprice information on a unit-cost basis, could selectthe least expensive of two brands of peanut butter.l5

While it is not focused on the basic skills require-ments for work, the NAEP study makes clear thatlarge numbers of young American workers do notcome into the workplace with the basic academic

skills that employers could expect from their yearsof formal schooling. Such problems are not confinedto new entrants, products of an educational systemwith slipping standards. They are found also amongmidcareer and older workers, people whose basicproficiencies were perhaps not strong to begin with,or whose skills have rusted with little use.

With the quality of American academic achieve-ment only now showing signs of rebounding, theprospect is that things will get worse, not better. Thegrowth rate of the labor force is slowing, and a highproportion of the new entrants over the next decadewill be from demographic groups (blacks, Hispan-ics, and immigrants) that traditionally have beeneducationally disadvantaged. Faced with a decliningpool of qualified applicants, employers may not beable to be as selective in their hiring as in the past.Even if educational quality rebounds strongly in theprimary and secondary schools, the generation ofpeople that entered the work force in the 1970s, andinto the early 1980s, could still depress overallAmerican productivity growth well into the nextcentury, unless employers and public programs takestrong measures to help large numbers of workerslearn to read, calculate, and communicate better. l6

Well-designed programs can help workers withrusty basic skills improve enough to handle suchchallenging tasks as statistical quality control anddaily maintenance of sophisticated equipment.17

In some countries-West Germany is a primeexample-a nationwide system for teaching youngpeople technical skills adds a further advantage tothat provided by a sound basic education. Abouttwo-thirds of Germany’s young people go through a3-year work apprenticeship after finishing compul-sory academic schooling at age 16. The vocationaltraining combines classroom studies 1 day a weekwith organized work the other 4 days, either in aworkshop or a regular workplace. To qualify as acraftsman, the trainee has to pass practical tests anda 4-hour written exam. There is evidence that thiscentury-old system (it started with Bismarck) paysoff handsomely in productivity, quality, and flexibil-ity in manufacturing.

ldIIWiII S, Kirsch d Ann Jungeblut, Literacy: Profiles of America’s Young Adufts (Princeton, NJ: Educational Testing Serviw. 1986), P. ~.

IsIbid., p. 34.160TA is conduc~g ~ as~ssmcnt of ‘Worker Training: Implications for U.S. tim~t.itiv~=s,” to be completed in 1990. Preliminary results of

this assessment indicate that the lack of basic skills among manufacturing workers is a solvable problem, but does require effort and expense.17~~, op. cit., paati.

Chapter 4--Human Resources ● 119

A mid- 1980s series of studies comparing matchedBritish and German manufacturing plants-in metal-working, kitchen cabinet manufacture, and garmentmaking-found that the German plants had laborproductivity advantages of 60 to 130 percent.18 Ineach case, the studies concluded that a major reasonfor the German advantage was the country’s bettertrained, more highly skilled shopfloor workers.(Technical training of foremen and higher managerswas found to be at least as important, in some casesmore so.) For example, in the kitchen cabinet plants,nine-tenths of all the German workers on theshopfloor had had 3-year apprenticeships followedby qualifying examinations. At best, one-tenth ofproduction workers in the British plants were soqualified, and several British plants had no workerswith similar training. One result: the Germanworkers were adept at using computerized wood-working machinery and a linked system for feeding,unloading, and stacking materials. Fully linkedmachine lines were hardly to be found in the Britishplants, one main reason being fear that one of thelinked machines would “go wrong” and stop thewhole line.

Breakdowns of all kinds of machinery were farmore frequent in the British plants-another sign ofinsufficient worker training. The German operativesroutinely clean and maintain their machines, whereasthis kind of planned maintenance is virtually un-known in the British plants, according to the study .19Similarly, in metalworking, breakdowns of machin-ery-especially of advanced computer, numericallycontrolled machinery-were a serious, continuingproblem in British plants, while the German plantsreported only startup problems, never continuouslongstanding difficulties.20

Apprenticeship training was also credited withhelping German shopfloor workers adapt easily tochanging requirements. This adaptability is essential

to the strategy of the German clothing industry,which concentrates on short runs of high-pricedquality products and pays relatively high wages—atleast 50 percent higher than wages in the Britishindustry. In the German plants visited for the study,80 percent of sewing machine operators had com-pleted a full 2-year apprenticeship; no British firmhad a single machinist with equivalent training.21

The German machinists needed only 2 days to reachtop-speed production on a new style, and most wereable to work on new operations directly fromtechnical sketches. The British machinists typicallytook several weeks to master a new style, and fewcould work from technical sketches. Also, qualitywas apparently much better in the German plants,since the number of quality controllers (passers) wasonly 1 for 23 machinists, compared to 1 for 7 inBritain. Undoing of faulty work was often observedin the British plants visited, but not once in theGerman.

It is not the apprenticeship training alone thatserves German manufacturing so well. The level ofmath competence of the average school leaver (age15 to 16) is substantially higher in Germany than inBritain, and the relative advantage is especiallymarked for the less academically ambitious students(those most likely to take up operative work) .22 Noris a public system of vocational training the onlyway to give production workers the technical skillsthey need for advanced manufacturing. In Japan, forexample, immensely successful international firmssuch as Toyota or Mitsubishi hire high-schoolgraduates with no special technical training and givethem company training. Japan’s publicly funded,vocational training institutes typically serve theneeds of smaller companies. Many American man-agers also think they can train production workersadequately, if the workers know how to read, figure,and communicate adequately and have good workhabits. The sine qua non is good basic skills.

l%Wtivi~ wss fi~ on he basis of physical units of production for similar items. The studies were: A. Daly, D.M.W.N. Hitchens, and K.Wagner, “Productivity, Machinexy and Skills in a Sample of British and German Manufacturing Plants,’ National Institute Economic Review, February1985; Hilary Steedman and Karin Wagner, “A Second Imok at Roductivity, Machineg and Skills in Britain and Germany,’ Nutiontdlmtitute EconomicReview, November 1987; Hilaty Steedrnan and Karin Wagner, “Productivity, Machinery and Skills: Clothing Manufacture in Britain and Germany, ”National Institute Economic Review, May 1989. S= also these papers by S.J. Rais and Karin Wagner in the National institute Economic Review “SomePractical Aspects of Human Capital Investment: Training Standards in Five Occupations in Britain and Germany, ‘‘ August 1983; ‘ ‘Schooling Standardsin England and Germany: Some Summary Comparisons Bearing on Economic Performance,’ May 1985; “Productivity and Management: The Trainingof Foremen in Britain and Germany,” February 1988.

lgs~m and Wagner (1987), op. cit., p. 89.

~aly et al., op. cit., p. 55.zls~mm and Wagner (1989), op. cit., p. 49.

22Prais and Wagner (1985) and (1988), op. cit.


THE TECHNICAL ANDENGINEERING WORK FORCEAlthough good basic skills throughout the work

force are fundamental for good manufacturingperformance, the defects of ordinary Americaneducation and the lack of a robust vocational trainingsystem may be more damaging to the nation’stechnical operatives than to its blue collar workers.Assuming that production workers are competent inreading and simple math, or need no more thanbrush-up courses, they can be trained for manyshopfloor jobs in a matter of weeks. Training oftechnicians-those who do nonroutine maintenance,programming, and repair of equipment—takesmonths to years, on top of decent reading and mathskills.

One conclusion of the comparative studies ofGerman and British manufacturing plants was thatthe superior training of foremen in Germany was akey advantage to manufacturers. The German fore-man combines technical and managerial skills. He orshe supervises workers in the routine care andmaintenance of machinery, adapts standard ma-chines to specialized needs, and works with suppli-ers in developing new machines. The foreman is alsoresponsible for scheduling work (often using com-puters for the purpose) and ensuring delivery ontime.

Most foremen are qualified as Meister, or ad-vanced mechanic. Candidates for the Meister quali-fication must first have at least 3 years’ full-timework experience following their apprenticeship andqualification as craftsman. Then they take a pre-scribed set of courses in technical topics, businessorganization, and training responsibilities, eitherpart-time over 2 or 3 years or full-time for about 9months. The courses are free but candidates takethem on their own time. The written examinations atthe end of the course typically take about 17 hours,spread over 3 days. Advanced mechanics in textiles,for example, must pass an exam covering thefollowing subjects:

1. origins and qualities of raw materials andtextile products;

2. yarn and thread production;3. yarn and thread construction;

4. the organizational structure of the firm;5. the rights and duties of workers;6. safety rules and first aid;7. adjustment and operation of fiber preparation

machines;8. adjustment and operation of spinning ma-

chines;9. ability to determine the quality of yarns and

threads;10. maintenance of tools, machines, and equip-

ment;11. machine parts;12. electronics;13. fundamental metalworking; and14. installation and repair of machines.23

This rigorous training and accreditation systemfor technicians or foremen is routine in Germany,but practically unknown in America. Yet, particu-larly in automated manufacturing systems, the needis increasing for numbers of people who have thekind of broad mastery described above, people whounderstand the entire production system and keep itrunning. Remedying a shortage of these skills ismade considerably more difficult when the workforce is populated by men and women whose basiceducational preparation is poor.

The Engineering and Scientific Work Force

The problem of poor preparation in public schoolsmay turn out to be more acute in the engineering andscientific work force. It takes at least 4 years toproduce an engineer, assuming the student has hada solid secondary education. It takes longer toproduce most scientists. If because of inadequatebasic education the United States cannot keep ahealthy flow of scientists into research and develop-ment and engineers into R&D and industry, Ameri-can manufacturing industries will find it increas-ingly difficult to keep up with, not to mentionoutperform, industries in other nations.

Several trends are worrisome. First is the numberof scientists and engineers in the work force,particularly those employed by industry. The pro-portion of scientists and engineers in America’swork force has remained fairly constant through thelast two decades, while in Japan it has risen steadily.Now, Japan has about as many scientists andengineers employed per thousand workers as Amer-

Zswape Br~ke Nelson, I~roving Competitiveness in Mature Industries: Lessons From the West German Textile Industry, M&$ter’s ~esis,Massachusetts Institute of Technology, October 1987.


ica, but will soon have significantly more, unless thetrends change. Second, it will be hard to spur growthin the number of engineering graduates in Americabecause of three impediments: the poor performanceof average American students in science and mathe-matics in secondary schools; the increasing propor-tion in the population of America’s young people ofminorities, who have traditionally done poorly inscience and math; and the increased efforts byforeign governments to attract home their ownnationals who are graduates of American engineer-ing and science programs.

There are also some more specific problems.Improving productivity and quality in manufactur-ing means attracting more engineers to manufactur-ing, and not just to the lucrative electronics indus-tries. Manufacturing engineering has enjoyed muchlower status than other engineering specialties, andthere are few signs of change. Also, many engineersand scientists are diverted from civilian industries towork on defense technology; it is estimated that 20percent of U.S. engineers are in defense work.24 Thedebate over how much of the engineering andscientific knowledge generated by the DoD spillsover into civilian sectors will not be resolved here.However, defense work provides few benefits tomost manufacturing industries (aerospace and, to alesser extent, electronics are where DoD technologyhas most of its civilian application).

Finally, there are qualitative differences in howJapanese and American engineers spend their days.Japanese companies are structured to do what theyare renowned for: make things better, and faster, andless expensively. Accordingly, their use of engineersis well adapted to continual incremental improve-ment of products and especially manufacturingprocess. They are not particularly known for comingup with a steady stream of larger technologicalbreakthroughs. American companies, on the otherhand, are better known for the stimulation ofengineers’ creative abilities, but are less effective inday-to-day improvement or in meshing engineers’design with shopfloor production. While both coun-

tries are making efforts to reproduce each other’sstrengths, there is little doubt that the Japanesesystem has served manufacturing competitivenessbetter than the American system has in the past fewdecades.

Numbers and Distribution ofScientists and Engineers

The concentration of scientists and engineers in anation’s work force says much about its capacity forinnovation and improved productivity.

Among five industrialized nations-France, WestGermany, Japan, the United Kingdom, and theUnited States—the United States ranks first in thenumber of scientists and engineers per thousandpeople in the work force by a small margin (figure4-3). When it comes to the engineering work force,the United States, with 175 engineers per 10,000workers in 1984, has a slightly lower concentrationthan Japan (187 per 10,000 in 1985) or WestGermany (194 per 10,000 in 1985), and a higherconcentration than the United Kingdom (144 per10,000 in 1981) or France (105 per 10,000 in1982). 25

The number of people entering or graduating fromscience and engineering programs in this countryhas responded readily to market forces in the past.The boom in industrial demand for computer scien-tists, for example, has made computer science thefastest growing field of science at all degree levels.26

Patricia Flynn, analyzing the shift in industrialcomposition of the Lowell, Massachusetts areabetween 1970 and 1982, found that:

The occupational education network was highlyresponsive to overall occupational trends in the areaand to the particular needs of the high-technologyindustries. Three-quarters of the occupational educa-tion programs, accounting for 85 percent of all of thetrained graduates, were “on target” or “reasonablyaligned” with occupational employment changes inthe Lowell area during the 1970s.27

Specifically, Flynn showed how local educationalinstitutions shifted to meet the change in local

zqNtioti Actimy of Sciences, The impact of Defense Speti”ng on Nondefense Engineering Labor Markets (Washington, DC: Nation~ ActiemyPress, 1986), p. 74.

~National Science Foundat,i~, National Science Board, Science and Engineering indicators-1987, NSB 87-1 (Washington, DC: U.S. GovementRinting Offke, Nov. 30, 1987), p. 226, appendix table 3-15.

~U,S. con-, Office of TechnoloW Assessment, Ehcating Scientists andEngineers: Grade School to Grad School, OTA-SET-377 (WAin@n,DC: U.S. Government Printing (Mce, June 1988).

?-~a~ciaM. Flynn, Facilitating Techno~gicalC~nge: The Human Resource Chalienge (Cambridge, MA: Ballingermbli*ing CO.? 1988), P. 101.


Figure 4-3-Scientists and Engineersper 10,000 Labor Force

250

I200

t150

100I

-,France West Germany Japan United Kingdom United States

= Scientists m Engineers

SOURCE: National Science Foundation, National Science Board, ScienceatndEngineering h?dicators-1987, NSB87-1 (Washington, DC:U.S. Government Printing Office, Nov. 30, 1987), appendixtable 3-15.

employment patterns and industrial growth. Tradi-tional manufacturing in Lowell was marked bydeclining average annual employment of 4.0 percentin textiles, 4.9 percent in apparel, and 8.4 percent inleather between 1976 and 1982. At the same time,employment in high-technology sectors took off:annual employment growth in nonelectrical machin-ery (including computers) was 43.3 percent; ininstruments, 23.6 percent; in transportation equip-ment (mostly aerospace), 7.2 percent; and in electri-cal and electronic equipment, 7.2 percent.28 Low-ell’s educational institutions responded, and thenumbers of graduates from high-technology pro-grams grew more than twice as rapidly as the numberof graduates from all the other occupational pro-grams.

More generally, engineers and scientists seem tobe in adequate supply in the United States-so far.During the past decade there has been healthygrowth in the nation’s scientific and engineeringwork force (figure 4-4). Both market forces andgovernment policies have proven effective at draw-ing people into engineering and science schools, andat attracting people who are qualified to work inengineering from other fields. Federal funding ofgraduate fellowships has encouraged enrollment in

Figure M-Employment of Scientists and Engineers

Millions

6~

3

2

1

01976 1978 1980 1982 1964 1986 1966

Y e a r

- Total, Sci. & Eng. = Engineering

*Estimate.

SOURCE: National Soience Foundation, “U.S. Scientists and Engineers:1988,” NSF 88-322, 1988, table 1.

science and engineering and caused some students toshift their postdoctoral plans.29 Federal science andtechnology initiatives, such as NASA programs andthose at the National Institutes of Health, have alsohelped to create a healthy job market for graduates.Finally, the boom in microelectronics and computerindustries in the 1970s and 1980s also drew manypeople into science and engineering curricula, espe-cially electronic engineering specialties and com-puter science. Between 1976 and 1986, for instance,the work force increased just over 2 percent per year,while the number of computer scientists increasednearly 17 percent per year, and the number ofelectrical engineers increased 7 percent per year(figure 4-5).30

But the trend is a bit bleaker. In the past, engineersand scientists were typically white males. They nowmake up a shrinking proportion of the pre-collegepopulation, which is itself growing smaller. Thegreatest growth is in the Hispanic population, witha more slowly rising proportion of black people. Bythe year 2000, 25 percent of the college agepopulation will be black or Hispanic. These twogroups, which are more likely to live in poverty,perform less well in school and have had higherdropout rates than white or Asian ethnic groups. Itwill take greater efforts to prepare and recruit them

‘gIbid., p, 81.2gIbid., p. 17.~.S. Department of Labor, Employment and Eurnings, any issue; and National Science Foundation, U.S. Scientists and Engineers: 1986, NSF

87-322 (Washington, DC: U.S. Government Printing Office, 1987).


Figure 4-5-Trends in Science and Engineering Labor,1976-88

Scientists, engineers (thousands) Total work force (millions)3000, ~ 140

2500120

1002000

80

1600

60

100040

500I20

0 0

1976 1978 1980 1982 1984 1986 1988

Y e a r

9 Computer scientists ~ Electrical engineers m AI I scientists

= All engineers * U . S . w o r k f o r c e

● Estimates, except for total work force.

SOURCES: National Seienoe Foundation, “U.S. Scientists and Engineers:1988,” NSF 88-322, 1988, table 1; and U.S. Department ofLabor, Bureau of Labor Statistics, “Employment and Earn-ings,” vol. 36, No. 12, Deeember 1989, table Al.

into the ranks of scientists and engineers. If feweryoung people enter engineering and science pro-grams, salaries will be bid up, and employers mightface rising costs of securing technical talent. Evenwith manufacturing employment shrinking, the de-mand for engineers and scientists might not declineor might even rise, as it takes increasing numbers ofscientists and engineers to keep manufacturingcompetitive. If salaries rise, it will be more expen-sive to solve technical problems in manufacturing,develop technology, run and adapt equipment.While large companies and high-technology com-panies will continue to employ engineers andscientists, more small companies will find it hard toafford even one engineer.

To guarantee a steady stream of qualified entrantsinto college engineering programs, many actionswill be needed. One is investment in primary andsecondary school programs designed to improveperformance in math and science. Actions to attractand retain larger numbers of students into engineer-ing and science would require a substantial commit-ment of resources, and take many years to yieldsignificant results .31

In the meantime, the Japanese system is alreadyprimed to prepare, recruit, and educate engineers.

Currently, the concentration of engineers in theJapanese work force is only modestly higher than inthe U.S. work force (187 per 10,000 workers inJapan v. 175 per 10,000 in the United States) andtheir concentration of scientists is much lower (65per 10,000 in Japan, compared with 101 per 10,000in the United States) .32 But the educational systemof Japan is effectively geared to produce newengineers of a high caliber, while the Americansystem needs substantial improvement before thefeed rate into engineering curricula can be steppedup, or even maintained. Over 4 percent of 22-year-old university graduates in Japan hold degrees inengineering, compared with less than 2 percent of22-year-old college graduates in America. While theabsolute numbers are roughly comparable-71,400new engineering graduates in Japan in 1985, and77,900 in America-the emphasis of the Japanesesystem is clear, considering that Japan’s populationand GNP are about half that of the United States.

Despite its current favorable position, Japan facesits share of problems in engineering. Maintainingstrength in manufacturing may prove a bit moredifficult than Japan’s impressive record wouldindicate. Endaka, or high yen, squeezed Japanesemanufacturing, and while industry responded admi-rably to the challenge, the constraints of being ahigh-cost nation are beginning to have effects thatconcern many Japanese observers. Specifically, withthe pressure to increase productivity and hold downwages, many newly graduated engineers are optingfor careers that offer greater financial rewards thanmanufacturing. Currently, beginning engineers inmanufacturing earn only a bit more than workerswith no more than a high school education. In1987-88, average earnings for male systems engi-neers 20 to 24 years old were 150,000 yen per month($1,071 at 140 yen to the dollar); their earningspeaked at 401,400 yen per month ($2,867) for 45 to49 year-olds (figure 4-6). Prospects for graduatingengineers are much more lucrative in Japanesefinance, at least for now. The salary of a midcareer(35-year-old) employee in a Japanese bank is theequivalent of $70,000 to $80,000 per year, about

31s= U.S. ConweSS, Office of T~~~@y Aswwment, &f~ati~g scie~~ts ad Engineers: Gr&e SC/WOi to Grad SChOOf, OTA-SET-377

(Washington, DC: U.S. Government Printing Office, June 1988) for a detailed discussion of these policy options.szNatio~ Science Fo~dtiion, National Science Board, Science and Engineering indicators-1987, op. cit.

21-700 0- 90 - 5


Figure 4-8-Salaries of Engineers and Laborers ofLarge Establishment= in Japan, 1987-88

Thousand yen per month600

500 -

4 0 0 -

300 -

200 -

100

tO L 1 I I 1 1 1 1

20 25 30 35 45 50 55 60

A v e r a g e a g e

— P r o d u c t I o n w o r k e r . + College graduates + System engineers

SOURCE: Japan Productivity Center, Practica/ Handbook of Productivityand Labour Statisifcs ‘87-’88 (Tokyo, Japan: Japan ProductivityCenter, 1988), tables 11 and 13.

double the salary of a midcareer manufacturingprofessional. 33 Little wonder, then, that new engi-neering graduates should opt for other sectors as theyleave school.

The salary differentials-and possibly, the sink-ing prestige of a career in manufacturing comparedwith other opportunities-are taking a toll. While 60percent of graduates from all Japan’s engineeringuniversities still entering manufacturing (roughlythe same proportion as in the past), the sector islosing its appeal for engineers graduating from thethree most prestigious universities (Tokyo Univer-sity, Tokyo Institute of Technology, and WasedaUniversity). About 80 percent of the engineers fromthose institutions chose to enter manufacturing in1982. The proportion has been declining ever since,dropping under 60 percent in 1988. Many of thesegraduating engineers are being lured into banks andsecurities companies, where the jobs pay more andthe opportunities are regarded as more exciting. Arecent survey of electrical engineers showed thatyounger engineers feel more strongly than otheryoung workers in Japan that they are unable to fullyuse their talents, and that they cannot do what they’reinterested in. In addition, like other young workersin Japan, they feel underpaid.34

Thus, Japan is not free of difficulties in attractingengineers into manufacturing. However, the supe-rior educational preparation of Japanese students

may make Japan’s problems easier to solve thanours. Japan’s large pool of people who are able toenter science or engineering could be an importantsafety valve as it enters its own version of unchartedwaters. Just as the United States is trying to copewith international competition on an unaccustomedscale, Japan is trying to improve its ability togenerate breakthrough advances in science andtechnology while maintaining its strength in manu-facturing process. The new emphasis on innovationprobably means that Japan will need many morescientists than it has, and that it will have to spendmore on basic research both in industry and inuniversities-which, compared to American uni-versities, contribute much less to the national streamof technological development and innovation. Inaddition, some departures from the traditional,seniority-based career paths of Japanese scientistsand engineers may be needed.

So far, it is hard to make any case that Americadoesn’t have enough engineers, particularly inmanufacturing. There are nearly as many engineersin manufacturing in the United States as in Japan andGermany, and more scientists; there is no artificiallycreated scarcity. The number of people entering orgraduating from science and engineering programsseems to respond readily to market forces or at leasthas done so in the past. The boom in industrialdemand for computer scientists, for example, hasmade computer science the fastest growing field ofscience at all degree levels. The principal worry forthe near future, so far as supply is concerned, is thetrend in demographics.

The Functions of Engineers inJapan and America

Japanese and German manufacturing, both re-nowned for their attention to precision and quality,employ about the same number of engineers perworker as American manufacturing, which no longerhas the same reputation. Obviously, it is not just thenumber of engineers in manufacturing that countsbut also how they spend their time.

The Japanese have consistently surpassed theirU.S. competitors in manufacturing things reliably,with high precision, and at reasonable costs. In otherwords, they have devoted more effort than Ameri-

ssBob John~one, “A Tmhnical Hitch,” Far E@ern Economic Review, Feb. 16, 1989, p. 49.~~wo Kojirna, Yoshio Nishimura, ~ Tom SUZ~I)“ “The Changing Role of Japan’ sEEs,” Electronic Engineering Times, Dec. 5, 1989.


cans to ironing out the large and small problems ofmanufacturing. In comparison, American firms havetended to put more emphasis on innovation. Jobassignments differ for engineers in America andJapan, as does the relation between design andproduction engineers.

The careers of Japanese and American engineersin industry differ starting from the time theycomplete their schooling and join a manufacturingfirm. In sharp contrast to American engineers,Japanese engineers are likely to stay with the firm until retirement, and to progress along a fairlypredictable path through the hierarchy of the com-pany. Few leave firms and move to another inmidcareer. They are more likely to be transferred bytheir company to an area outside their specialty. Theobjective is to broaden their job skills and broadentheir knowledge of other functions. American engi-neers are likely to become managers earlier thantheir Japanese counterparts, and to broaden theirknowledge by transferring between companies ratherthan within them.35

About one-third of American engineers work inresearch and development (940,000 out of a total of2.8 million, as of 1988).36 In addition, some 275,000engineers are involved in the management of R&D.Only about 17 percent of American engineers(495,000 people) work on the shopfloor, in produc-tion and inspection.37 The same pattern, in a moreextreme form, prevails in West Germany, where 50percent of engineers work in R&D, and only 12percent in manufacturing production and repair.38

While comparable data are not available for Japan,there are strong indications that the Japanese firmdeploys its engineers differently. Japanese engineersare much more likely than their American counter-parts to have at least one assignment in a new areato broaden their skills: 62 percent of Japaneseengineers report at least one job rotation assignment,compared with only 35 percent of American engi-neers. Thirty-five percent of Japanese engineerswere assigned at some point to production, com-pared with only 14 percent of American engineers,

and 50 percent of Japanese engineers have servedone outside assignment in research, design, anddevelopment activities, compared with only 14percent of American engineers.39

These standard job rotations afford Japaneseengineers the opportunity to acquire a firsthandknowledge of and sensitivity to the problems andconstraints of manufacturing. Most observers agreethat this understanding explains much of the abilityof Japanese manufacturers to bridge design andengineering functions effectively. American engi-neers, who rotate functions less frequently butchange firms far more often, may acquire someunderstanding of both manufacturing and design,but the record of Japanese and American manufac-turing suggests that it is relatively unusual. In Japan,the transfer of research or development to manufac-turing is accomplished by transferring people di-rectly, while in the United States one manager ismore commonly assigned the responsibility fortransferring the knowledge from design teams toproduction people.40 The fact that American firmsgenerally make much less effort than Japanese tosmooth the differences between product and manu-facturing process design and startup shows up indesigns that are harder to manufacture, longerstartup times, and lower process efficiency.

Japanese engineers are more likely than theirAmerican counterparts to take responsibility formaking sure their designs are manufacturable, a factsupported by considerable anecdotal evidence. A

—also typical of the kinds of storiesgood exampletold about interactions of design and manufacturingengineers--comes from an engineer now at Sema-tech, the U.S. semiconductor manufacturing devel-opment consortium. The engineer once worked fora major U.S. semiconductor manufacturer producing1 megabit DRAMs, and then for Siemens on theMega Project, the European program to design andmanufacture 1M and 4M DRAMs. He recounted thetale of the U.S. fro’s unsuccessful attempts tomanufacture 1M DRAMs efficiently (e.g., with highyields and low cost). After developing the process

35s= ~nwd I-I. Lynn, Henry R. Piehler, and W. pad fiaY. “Engineering Careers in Japan and the United States: Some Early Findings From anl%pirical Study,” rnimeo, n.d.; and D. Eleanor Westney and Kiyonori %lmkibara, “Designing the Designers,” Tech~lou ~evi~, APril 1986”

36Nation~ Science F~~dation, U.S. scie~ists ad Engineers: ]988, NSF 88-322, 1988.371bide38Nat10@ science F~mdati~n, Scientiso a~Engi~ers in ~~wtria~ized cou~~es (Washington, DC: CIR Sttif Paper, November 1986), p. 25.3~ym, et ~+, op. cit. The= ~rcen~ges describe only the job rotation experiences of engineers, not heir c~nt Positions.

WVestney and Sakakibara, op. cit., p. 28.


and prototypes in the laboratory, the company turnedthe design over to the factory, where manufacturingengineers were unable to get chip yields up tocompetitive levels. The manufacturing engineersprotested that the process had no margin for error,but they did not themselves have the resources orknowledge to do paper analysis and make improve-ments. The designers, on the other hand, insisted thatthey had developed a robust and manufacturableprocess, and shied away from correcting the prob-lems. At Toshiba, where the 1M DRAM processquickly resulted in very high yields, the engineersand scientists who developed the 1M DRAMprocess and presented the results to the scientificcommunity 4l were also responsible for yield im-provement activities.42

Case studies also indicate that Japanese firmsoften have more engineers on the shopfloor than doU.S. firms. In a study of flexible manufacturingsystems (FMSs) in the United States and Japan,Jaikumar concluded that the Japanese companiesused the systems far more effectively than theAmerican fins. They got their systems up andrunning in much shorter time and made many morekinds of parts. Further, their machines had far lessdown time. Much of the difference arose from theways in which the two countries used their engi-

neers. U.S. managers treated their FMSs inflexibly,like hard-wired equipment, while the Japanesecontinued to tinker and make incremental improve-ments.

The adjustments needed to exploit the flexibilityof programmable machinery can generally only bedone by engineers. In Japanese firms using FMSs, 40percent of the staff were college-educated engineers,and all the workers were specially trained in the useof computer numerically controlled (CNC) ma-chines. In the U.S. companies, only 8 percent of theworkers operating the FMSs were engineers, andfewer than 25 percent of all workers had been trainedon CNC machines. In the U.S. firms, the projectteam of engineers and software specialists whodesigned the system disbanded and left after theyhad it debugged and running. In Japan, the engineerswho designed the system remained to operate it,m a k i n g c o n t i n u a l programming changes, writingnew programs, and staying with it until theyachieved untended operation at least 90 percent ofthe time. In a fully automated FMS metal-cuttingoperation, Jaikumar found, engineers would out-number production workers three to one, but thesystem would require less than half the number ofengineers needed in a conventional U.S. system.43

41 Syuso Fujii et al., “A 50 [mu]A Standby IMxl/256Kx4 SMOS DRAM With High Speed Sense Amplifier,” IEEE J. Solid-State Circuits, vol. SC21, October 1986, pp. 643-647.

Q~erW~ comm~cation, D. Robyn, S. Baldwin, and A. Buyrn of OTA with Peter Nunan, Semattxh, May 10-12, 1989.dsR~c~ndr~ Jailn.unar, “Postindustrial Manufacturing, “ Harvard Business Review, November-December 1986.

●

5Chapter

Links Between Firmsand Industries

CONTENTSPage

LINKS BETWEEN MAJOR MANUFACTURERS AND SUPPLIERS . . . . . . . . . . . . . . 129LINKS BETWEEN SEGMENTS OF AN INDUSTRY COMPLEX . . . . . . . . . . . . . + . . . 135LINKS BETWEEN MAJOR MANUFACTURERS AND CAPITAL

EQUIPMENT PRODUCERS . . . . . . . ........ + . . . . . . . . . . . “ ● . ● -. ● + . . ● . ● . + . ● ‘ ‘ . “ . . . 137INTERNAL LINKS: VERTICAL INTEGRATION, PRODUCT DIVERSITY, AND

LARGE SIZE . . . . . . . . . . . . . . . . . . . . . . . . . . “ ~ð . . . . . . . ● . . . . . ● ● + . . . ● ● . . ● . ● ● ● . ● . * ..- ● 143Links to Markets, Financial Stability . . . . . . . . . . . . . . +... ”.. !..””.”..”””+ +.. w.. ” +..143Links With Consumer Electronics . . . . . . . . . . . . . . . . . . . . +.. ● ● ”.. ”... +++. i.• ● +. w.. ” + 147Technology Links . . . . . . . . . . . . . . . . . . . . . . . . . . . .. ”. .. .. .” S. .. ..+. .” S.. .+O.. ...9.-0

FiguresFigure

Page

5-l. Supplier Network for Fuji-Xerox . . . . . . . . . . .. .. . .. .. .. ... .. .......+. . . . . . . . . . . 1315-2. Shift in Market Shares for Wafer Steppers . . . . . . . . . . . . . . . . . . . . . . . . . . ... ....,..+ 1395-3. U.S. Market Shares of Selected Semiconductor Equipment .. .. .. .. . ..+. ......+. 1395-4. World Semiconductor Equipment Sales ........””.o.0+0.’.’.+.* t+.’.+.”...””” 1435-5. World Semiconductor Sales...*..............*. .....o.J.c...+0..+o’o +..”+....

TablesTable Page

5-1. Main Reasons for Subcontracting, Japanese Firms,1986 .. .. .. .. .. ... ... .......+ 1335-2. Top Ten Semiconductor Equipment Suppliers, World Sales .. ... ... +...... . . . . . . 141

Chapter 5

Links Between Firms and Industries

Industries do not standalone. They are linked withsuppliers of machinery and materials in one direc-tion and with a chain of customers in the other. Howa firm or an industry handles these relations has agood deal to do with its competitive performance. Inall kinds of manufacturing industries, close links andstable relations between suppliers and purchasersseem to be important factors in boosting overallperformance, including productivity, quality, andinnovation.

U.S. industries, on the whole, have not beenstrong on collaborative vertical links. The traditionalrelation between supplier and customer has insteadbeen distant, even adversarial, and based mainly onprice. 1 But there are signs of a trend toward morecollaboration. U.S. auto companies are trying toform closer, longer-term relations with parts suppli-ers. Sematech, the industry-government consortiumdedicated to improving manufacturing technologyin semiconductors, began by strengthening tiesbetween chipmakers and producers of the materialsand equipment used to make chips. Textile compa-nies are forging stronger links backward to fibersuppliers, forward to apparel makers, and beyond toretailers. Individual firms that have made a come-back against foreign competition use close supplierlinks as part of their strategy, a leading examplebeing Xerox.

The trend toward closer links is certainly notuniversal. Nor is it likely that American manufactur-ers will ever replicate the distinctively Japanesestyle of close, mutually obligating bonds betweenparent and subsidiary companies (even in Japan thebonds are weakening somewhat). But the advan-tages of collaborative links, throughout an industrycomplex and between related industries, are increas-ingly appreciated.

LINKS BETWEEN MAJORMANUFACTURERS AND

SUPPLIERSTraditionally, U.S. manufacturers have either

supplied their own materials and parts (in verticallyintegrated companies) or, when dealing with outsidesuppliers, have kept them at arm’s length. Acommon strategy has been to pit one supplier againstthe other and drive the hardest possible bargain onprice. In offering their own goods to the nextproducer down the chain, the main selling point hasalso been price, with quality, service, and respon-siveness to customers’ needs taking a lesser place.This approach is not confined to the United States,but is typical in many market-oriented industrialcountries.

A different pattern is common in Japan. In theworld-class industries that have led Japan’s strongtrade performance, manufacturers generally main-tain long-term, collaborative relations with theiroutside suppliers. They are demanding on price andequally demanding on quality and just-in-timedelivery, but they also give their suppliers technicalhelp--occasionally financial help as well—in meet-ing these demands. Suppliers who show they areable to satisfy the manufacturer’s demands can befairly confident of keeping the business, rather thanlosing out to a price-cutting competitor. This patternis part of the overall Japanese approach of carefulattention to all aspects of manufacturing, includingthe quality of components and supplies.

The manufacture of motor vehicles offers anexceptionally clear picture of these alternate ways ofhandling links with suppliers. Organization of sup-ply is a central feature of the auto industry, since theaverage car or truck contains some 15,000 parts.Historically, U.S. automakers have chosen one ofthe two opposite approaches: either vertical integra-tion (as practiced by General Motors, which is 70

l~e pat~rn is not inv~able, For ex~ple, major ~r11ne5 have ]ofig had close, cooperative ties with the manufact~ers of aircritft, with tirlineengineers taking a leading part in design and purchase decisions. However, with deregulation of the industry, the ties are loosening; airlines are cuttingtheir engineering staffs andmakingpurchase decisions more strictly on the basis of price. See the case study of the commercial aircraft industry in MichaelL. Dertouzos, Richard K. Lester, Robert M. Solow, and the MIT Commission on Productivity, Made in America: Regaining the Competitive Edge(Cambridge, MA: The MIT Press, 1989).

-129-


percent integrated), or arm’s-length purchase fromsuppliers bidding against each other (Chrysler, 25percent integrated). Vertical integration is supposedto have the advantage of lowering barriers betweensupplier and main company (reducing transactioncosts)--e.g., by assuring that suppliers’ interests arethe same as the company’s, or by making it easier totransfer new technology to the supplier. The arm’s-length bidding system is supposed to do a better jobof keeping suppliers’ prices low.

Japan’s highly successful auto manufacturingindustry uses the third approach, a middle way thatis sometimes termed the supplier group system.2 Itconsists of a pyramid, topped by the final assembler,who deals with a group of first-tier companies justbelow that are responsible for major components.The first-tier suppliers manage relations with thesecond tier, who supply individual parts and them-selves often deal with third-tier groups, which mayin turn reach down as far as a fourth tier of tiny firmsspecializing in very narrow tasks. Some of thesesupplier groups are tightly bound. This is especiallythe case with the Toyota group, composed of 225companies that own each other’s shares and lendstaff and equipment from purchaser to supplier,starting with the assembler and reaching downthrough the various tiers. Other companies, such asHonda, have a looser structure, relying more onindependent suppliers who also serve other majorassemblers. But here too the relationships are closeand long-term.3

A leading virtue of the Japanese system is that itis easier to manage than the older U.S. systems. Astudy for the International Motor Vehicles Programcomparing General Motors procurement with Toy-ota’s found that, despite GM’s 70 percent verticalintegration, and despite stringent efforts to cut backits purchasing departments, GM still had 6,000

buyers of outside components and supplies in 1987.The Toyota Motor Co., only 20 percent integratedand producing about 40 percent as many vehicles,might be expected to need as many buyers as GM butreportedly had 337.4 These figures very likely drawan exaggerated picture of the differences, becauseToyota often uses engineers as purchasing agents sothat the number of its buyers is probably understated.But the disparity is so large that some of it is boundto be real, not definitional.

The answer to the seeming paradox is that, in theToyota system and others like it, purchasing isdelegated down the line. So are other responsibili-ties. The final assembler makes the car bodies,engines and drive trains, and integrates the system.But the first-tier suppliers are assigned the tasks ofdesigning, engineering, and testing components, aswell as producing them. Often, the supplier deliversto the assembler pre-packaged subassemblies thatcontain many parts (e.g., instrument panels orsuspension systems). The suppliers moreover havethe burdens of assuring quality and managingjust-in-time delivery. What they get in return is areliable purchaser for their particular componentsfor the life of the vehicle model, and often beyond—subject to the understanding that they will continu-ally reduce the component’s cost while maintainingits quality. At the same time, to keep competitionkeen, assemblers often do business with more thanone supplier of the same component.

Industries other than automating are just aswedded to the supplier group system-e. g., themanufacture of cameras (e.g., Canon), office copiers(Fuji-Xerox), personal computers and printers (NECand Epson).5 Figure 5-1 illustrates the suppliernetwork for Fuji-Xerox. A rough indication of theextent of the system is that the share of Japanese

2The Japanese gOUp system has ~n described by many authors; a comprehensive treatment of the system as practiced in the auto indusv is inMichael Cusumano, The Japanese Automobile Industry: Technology and Management at Toyota and Nissan (Cambridge, MA: Harvard University Press,1985), see esp. pp. 241-61. A succinct description is in James P. Womack and Daniel Roos, “Case Study: The Automouve Industry,” contract reportto the Office of Technology Assessment, Sept. 15, 1988; some of the material in this section on the motor vehicle industry is drawn from this report.

3A recent Japanese suwey found that 68 percent of subcontractors had never changed tieir ‘‘ Pment, ’ and that 53 percent had been doing businesswith the same parent for 15 years or more. Chusho kigyo cho cd., Chusho kigyo hakusho (Small and Medium Size Enterprise White Paper) (Tokyo:Okurasho instasu kyoku, 1988), p. 61, cited in D,H. Whittaker, ‘‘New Technology in Small Japanese Enterprises: Government Assistance and PrivateInitiative,” contract report to the Office of Technology Assessment, May 1989.

dToshihiro Nishiguchi, ‘‘ Competing Systems of Automotive Components Supply: An Examination of the Japanese ‘Clustered Control’ Model andthe ‘Alps’ Structure,” Massachusetts Institute of Technology, International Motor Velucles Program Working Paper, May 1987, p. 15.

sKen-lchl ~~, ~ujlro Nonaka, and Hirotaka Takeuchl,“Managing the New Product Development Process: How Japanese Companies Learn andUnlearn, ‘‘ in Kim B. Clark, Robert H. Hayes, and Christopher L,orenz (eds.), The Uneasy Alliance. Managing the Productivity-Technology Dilemma(Boston, MA: Harvard Business School Press, 1985).

Chapter 5--Links Between Firms and Industries ● 131

Figure 5-1-Supplier Network for Fuji-Xerox

Fuji-Xerox a

I

Toritsu-Kogyo b

(primary subcontractor)

(secondary subcontractor)

v

Number of secondary subcontractors

a Has other primary subcontractors.b Serves as subcontractor for other manufacturers.

SOURCE: Ken-lchi Imai, Ikujiro Nonaka, and Hirotaka Takeuchi, “Manag-ing The New Product Development Process: How JapaneseCompantes barn and Unlearn,” in Kim B. Clark, Robert H.Hayes, and Christopher Lorenz (eds.), The Uneasy Alliance:Managing The Productivity Technology Dilemma (Boston, MA:Harvard Business School Press, 1985), p. 364.

manufacturing companies using subcontracting rosefrom 32.5 to 37 percent from 1976 to 1981; in theelectrical machinery industry, the share rose from 55to 58 percent.6 And subcontracting in Japan usuallyinvolves long-term relations and mutual obligations—what the Japanese call the oyakigyo-kogaisha rela-tionship (literally, parent business-child company,but with connotations extending to many forms ofsuperior-subordinate relationships).7

The supplier group system is doubly advanta-geous to the lead manufacturers. They get many ofthe benefits of both arm’s-length subcontracting(control over costs) and of vertical integration(responsiveness to the lead company’s needs).Moreover, the requirement of uniformly high qualityfrom suppliers is part of the system of building inquality throughout the manufacturing process, ratherthan inspecting for defects at the end of the line.With this system quality need not cost extra, since itsaves the cost of keeping large inventories of partsand requires less re-work.

Close interactions between the major manufac-turer and its suppliers also helps the lead companyfield new models quickly, by dividing the labor ofproduct development among many small firms withspecialized skills. Shaving time off development cangive a firm a crucial headstart. Firms that are first torespond to market changes and to adopt newtechnologies in their products open a lead that is hardfor competitors to close.

In a study of the world’s motor vehicle assem-blers, a Harvard Business School team found thatJapanese automakers take about 3.5 years to producea new car design, compared to 5 years for Americanand European producers, and that the Japanese do itwith half the engineering effort.8 This takes intoaccount the different amounts of engineering effortcontributed by components suppliers in Europe, theUnited States, and Japan. The advantage, the studysaid, ‘‘appears to lie in the strength of the Japanesesupply base, and the way projects are organized andmanaged. Within the lead company, the mainadvantage lies in simultaneous rather than sequentialengineering, made possible by a continuing informaldialog between people at different stages of thedesign process, with give-and-take in both direc-tions. But suppliers contribute to this interactiveprocess too. Often they take part in collectiveengineering and analysis of key new components 2years before manufacture of a new model. About 1year ahead of time, first- and second-tier suppliers

6Ro&fl J. B~lon and 1wao ‘fomi~, ~~ F1~n~i~/ B~~~lor ~fJ~pa~se co~oratio~ (Tokyo and New York: K~ansha International Ltd., 1988),p. 45, citing the Ministry of international Trade and Industry, White Paper on Small and Mealum Enterprises in Japan, 1987 (Tokyo: MITI, 1987).

l’~ld., ~h. 3. Many o~er ~u~ors have alW descn~ ~is interm~iate system, ~tween ~’s-leng~ con~acting and Vexlical integration, in a V~etY

of Japanese industries. For recent examples, see Nishiguchi, op. cit.; and Mari Sake, ‘Neither Markets nor Hierarchies: A Comparative Study of InformalNetworks in the Printed Circuit Board Industry, ’ paper prepared for The First Conference of the Project ‘Comparing Capitalist Economies: Variationsin the Governance of Sectors,” Wingspread, Wisconsin, May 1988.

s~m B. Clwk, W. Bruce Chew, and Takahiro Fujirnoto, ‘‘Product Development in the World Auto Industry: Strategy, Organization, Performance,’paper presented to the Brookings Institution Macroeconomics Conference, Dec. 3, 1987 (available from Graduate School of Business Administration,Harvard University).


may be brought in to run an assembly or subassem-bly line, to solve startup problems before actualmanufacture. 9

The supplier group system is also credited with animportant role in the Japanese strategy of manufac-turing a greater variety of products at lower volumethan U.S.-style mass production has done, while stillkeeping costs competitive. This ability is particu-larly striking in the auto industry. When it started outin the early postwar years, Japanese auto productionwas, perforce, in small batches and great variety. TheJapanese domestic market was small, exports werevirtually nonexistent, and producers were numerous(they still are). The answer to the fragmented marketand extreme competition was to develop a flexibleproduction system within the factory. This includedmulti-skill training of workers and efficient layout ofthe factory and organization of work. It also includedthe supplier group system, with its collaborativeengineering, just-in-time delivery, and assurance ofhigh-quality parts and components. The result wasan industry that initially succeeded in the rich U.S.market with a niche product (the well-made, econom-ical small car), and has continued to broaden its salesappeal with frost-class entries into specialized mar-kets (e.g., sports and luxury cars).

Today, the average annual production per modelof the Japanese automakers is about 120,000, halfthat of U.S. producers. Since they introduce newmodels more often and more quickly, the lifetimeproduction for the average Japanese model is about500,000 units-less than one-quarter of the 2.1million units for U.S. producers and well below thelifetime 800,000 units per model for the high-pricedEuropean specialists (BMW, Mercedes, Porsche,Jaguar, Volvo and Saab), The group supplier systemis only one of the factors that make this flexibilitypossible, but it is a considerable one.

As for the suppliers, they also get multiplebenefits from the system. Besides gaining reliablemarkets for their products, they often get loans ofup-to-date equipment and sometimes financial help

in buying it; assistance from borrowed engineers ortechnicians in learning how to use the equipment ororganize work more efficiently; and in general aflow of advanced technology that has helped tomake many first-tier suppliers first-rate industryleaders.l0 This technology transfer is not confined tothe first-tier companies but frequently extends to thelevel of tiny family-run metalworking firms.11

Table 5-1 lists advantages of the subcontractingsystem from the participants’ points of view, asreported by Japan’s Small and Medium EnterpriseAgency. At the top of the list, for suppliers, is a‘‘steady amount of orders. ’ This stability some-times extends to a change in product line. Forexample, one Japanese subcontractor who hadworked with an electronics manufacturer for manyyears reported that he had changed from supplyingpaint and sheet metal to supplying printed circuitboards, at the customer fro’s request.12

On the down side, the system has a high level ofstress. Lead companies demand continual pricereductions as well as high quality, and if a supplierfails to meet the demands, he may find his share ofsales cut back (or even cut off eventually) in favor ofa more compliant supplier. As noted, lead compa-nies often have two or more firms supplying thesame item, and the competition is tough. While anexisting supplier may be safe from sudden shifts toa new competitor offering drastically lower prices(e.g., one electronics producer stuck with his sup-plier of printed circuit boards despite an offer froma newcomer of a 40 percent lower price), frequent‘‘requests’ by the lead firm for price cuts can narrowthe difference fairly quickly .13 Moreover, in arecession, the supplier is expected to make do withsmaller orders, cut prices to the bone, and forgoprofits. In Japan’s economic downturn of 1986,profit margins for the printed circuit board industryfell from 2.5 percent of sales to 0.3 percent.14

However, the lead company has the obligation totighten its belt too; suppliers trust that their largecustomers will not squeeze them into bankruptcy.

$“1’’oshihiro Nishiguchi, op. cit., p. 10.IOReputable suppliers may get ind~~t financi~ benefits as well. Major manufacturers generally belong to a group that includes a lmge ba~; 1oam

on favorable terms from that bank are often made to a supplier on the lead manufacturer’s recommendation.llSee ch. 6.

12Mari Sake, op. cit.13Mari Sake, op. cit.141bid.

Chapter 5---Links Between Firms and Industries . 133

Table 5-l—Main Reasons for Subcontracting, Japanese Firms, 1966

Subcontractor Parent company

Reasons Percentage Reasons Percentage

Steady amount of orders . . . . . . . . . . . . . . . . . . . . . . . . 50.1 Know-how of contractor not held by oneself . . . . . . . . 57.6Product design and development difficult by oneself . . 45.8 Efforts concentrated into best suited work . . . . . . . . . . 48.2Efforts concentrated on production activities . . . . . . . . 38.7 Past business relations with and reliability ofNo worries about default or debts . . . . . . . . . . . . . . . . . 27.7 subcontractor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.5Improved reputation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.2 Increased flexibility through size of orders . . . . . . . . . . 37.1Supply of raw materials, etc. . . . . . . . . . . . . . . . . . . . . . 21.7 Lower personnel costs and lower unit costsTechnical assistance provided . . . . . . . . . . . . . . . . . . . 14.7 of products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.5

Small lot sizes and thus greater efficiencythrough production by small enterprises. . . . . . . . . . . . 30.6

, Overly large size of own company wouldreduce operating rate . . . . . . . . . . . . . . . . . . . . . . . . . . 9.4Competition among subcontractors ensureshigh quality and lower unit price . . . . . . . . . . . . . . . . . . 8.8

SOURCE: Small and Medium Enterprise Agency, Survey In Division of Labor in Manufacturing Industries (Tokyo: SMEA, 1966), pp. 24-25.

Suppliers are also expected to push themselves tothe limit to meet urgent needs of important custom-ers. For example, when Fuji-Xerox changed thedesign of a part midway through development of anew copier, it made an “utterly insane’ request forearly delivery of a newly designed part which thesubcontractor was able to meet only by workingthrough the nights. The subcontractor was laterrewarded with a generous payment. 15 But a moreimportant motive for such sacrifice is the fact thatthe subcontractor’s own future depends on thesuccess of the lead manufacturer.

Finally, wages among subcontractors, especiallyin the lower tiers, are at least 25 percent lower thanwages of the privileged lifetime employees of majormanufacturing firms.l6 Indeed, low wages for the‘‘mom-and-pop’ suppliers at the bottom of thepyramid has long been considered a competitiveadvantage of Japanese producers. An integratedcompany like GM could credibly claim this as ahandicap--although GM presumably found advan-tages in vertical integration to compensate, since itcompeted successfully for years against Chrysler,which had a substantial discrepancy between the payof its own employees and that of its suppliers. Recentresearch suggests that disparities in incomes be-tween small and large firms are not as great asdisparities in wages. The published data cover the

workers’ wages in small family-run companies, butnot the income of the owner, who gets profits as wellas wages. 17 Many of these small entrepreneurs makea good living. One investigator of subcontractingfirms in the Japanese auto industry reported thatowners of small firms made about 10 million yen ayear ($71,000) on average, compared to 5 millionyen for people of the same age and same high schooleducation who work for big companies. In inter-views with over 100 of these small subcontractors,the author found them “remarkably confident andsatisfied despite their seemingly unstable position inthe industrial economy.”18

In any case, many American managers now seempersuaded that the system of buying from autono-mous, but closely linked suppliers, offers benefitsquite apart from wage differentials. The big threeautomakers are making moves toward adopting thegroup supplier system, or parts of it. The GM-Toyotajoint venture, New United Motor Manufacturing,Inc. (NUMMI), has adopted the system successfully,largely with North American suppliers. It took time.At first, NUMMI found three times as many defectsin the parts supplied by North American companiesas in those coming from Japanese companies. ButToyota and NUMMI engineers worked with the 70North American suppliers, and 4 years after the 1984

ls]m~, Non&a, and Takeuchi, op. cit., p. 371.lbNom~~ly, wages in e~ablishments wi~ 5 to 29 workers are only 57 percent of wages in firms with 500 or more workers. Controlling fOr differences

in occupational employment eliminates about 20 percentage points of the 43 percent difference. The discrepancy has been growing; wages in the smallestestablishments were 63 percent of those in the largest in 1965, but dropped to 57 percent in 1983. (OTA interview with officials of the National Instituteof Education and Vocational Research, Tokyo, Mar. 15, 1989.)

17Tm~~ro Nishiguchi, op. cit.

lgIbid., p. 21.


startup these suppliers were as good in cost, quality,and delivery times as their Japanese counterparts.19

For U.S. companies in general, both lead compa-nies and suppliers, the changes involved in movingto the supplier group system are great and conse-quently slow. It means going from l-year contractswith specifications and drawings (sometimes diesand tooling as well) provided by the assembler tomulti-year, less formal arrangements, in whichsuppliers are expected to help design and developparts, continuously improve them, and respondquickly to requested changes during the model run.20

It also means requiring suppliers to deliver just theright number of defect-free parts precisely when theassembler needs them. The just-in-time deliverysystem depends on getting high-quality parts, sincethere are no stacks of backup parts to replacedefective ones. When the system works, it savescosts in storage, handling, end-of-the-line inspec-tion, rework, and repair after sale, and the qualitybuilt in at every stage of supply up the pyramid leadsto a reliable product and customer satisfaction. Butthe system also requires high competence on the partof suppliers and a good working relationship be-tween assembler and suppliers. These attributes arenot easy to develop overnight.

According to the General Accounting Office(GAO), Japanese auto assemblers operating in theUnited States impose on suppliers the rigorousexpectations described above. (GAO reports thatU.S. firms are also beginning to expect the samekind of quality, prompt delivery, and engineeringcapabilities from their suppliers.) A good many U.S.suppliers are having trouble meeting the expecta-tions. Japanese supplier firms, accustomed to work-ing in this way and also benefiting from longtimerelationships with Toyota, Nissan, Honda, or Mazdain Japan, often have the advantage. The number ofJapanese suppliers in America (some of them in jointventures with U.S. fins) is growing fast. Of 104Japanese-affiliated suppliers operating in the UnitedStates in August 1987, 102 answered queries byGAO. Of these, 60 had opened up for business in

America since January 1981; 23 were establishedfrom 1970 to 1980, and 19 before 1970.

Some U.S. suppliers have succeeded with theJapanese transplant automakers. Of 30 representa-tive firms GAO selected for interviews, 15 had donebusiness with at least one of the Japanese assemblers—some in joint ventures with Japanese supplier firms.Most of these U.S. firms found big differences in theway the Japanese assemblers operated, comparedwith their American counterparts. The Japanesecompanies not only gave the suppliers added respon-sibilities but, several said, also kept in closer contact.Where the U.S. assemblers would send a few peopleon an occasional courtesy visit, the Japanese turnedup often, bringing a wide range of staff to give thesuppliers’ operations a complete evaluation. Onetrim and body parts supplier said the Japaneseassembler he deals with calls every day to consult ondefects. A steelmaker said the Japanese companyvisits were “preventative” where the Americancompany’s were ‘‘reactive. ’ Most of the U.S.suppliers doing business with the Japanese trans-plant automakers rated the results positively. Theycited benefits of greater efficiency, better qualitycontrol, and more attention to process and productimprovements. Some said the experience made themmore competitive, and that they were now demand-ing more from their own suppliers. And some notedthat U.S. automakers are adopting more and more ofthe Japanese practices.

These positive comments came from the firmsthat had succeeded in supplying the Japanesecompanies. From less successful firms came com-ments that it is hard to overcome the longstandingties between Japanese assemblers and suppliers, andthat U.S. firms are at a disadvantage in culture andlanguage. These companies feared growing compe-tition from Japanese-affiliated suppliers now locat-ing in the United States. Although the Japaneseautomakers have stated that they intend to increasethe U.S. content of their cars and trucks from about50 percent in 1987 to about 70 percent by the early1990s, it is not clear that “U.S. content” means theproducts of U.S.-owned firms.21

19jo~ Fe fiafclk, “A New Diet for U.S. Manufacturing, “ Technology Review, Jan. 28, 1989, pp. 31-32.me following discussion of U.S. and Japanese firms supplying automakers in the United States (both U. S.- and Japanese-owned) is based mostly

on U.S. General Accounting Office, Foreign investment: Growing Japanese Presence in the U.S. Auto Industry, GAO/NSIAD-88-l 1, March 1988.21A~cording t. GAO, u-s. automakers repo~~ ~a[ tie domestic content of heir cars and ~cks was 86 to over 99 percent, depending on the model,

in 1986; the average for the industry was about 90 percent. These figures applied to auios made in North America, including Canada, and did not includeforeign-made cars with a U.S. nameplate (“captives” such as the Dodge Colt, which is Mitsubishi-made). U.S. automakers were expected to increasethe foreign content of their cars to about 17 percent by 1990, GAO said.

Chapter 5--Links Between Firms and Industries . 135

A fundamental change in outlook would have toevolve if Japanese-style supplier relations were tobecome the norm rather than the exception in U.S.manufacturing. It is longstanding custom for Ameri-can manufacturers to discourage-even forbid—design engineers from developing close relationswith suppliers. Direct approaches to suppliers areknown as “going around the purchasing depart-ment,’ and are against company rules. Purchasingagents themselves are frequently reassigned todifferent types of supplies, so they won’t developoverly cozy relations with suppliers. The ideasbehind all this are, frost, that maintaining arm’s-length, impersonal, strictly contract-based relationswith suppliers is the best way to get a good price andkeep costs down; and second, that it is unfair to giveany supplier a privileged position and deny theothers an equal chance. Some company officialseven believe they might be subject to lawsuits ifsuppliers were deprived of the chance to bid forcontracts.

For suppliers themselves, the Japanese-style sys-tem has distinct drawbacks as well as strong points.While some may welcome the demands for con-stantly improving performance combined with helpin achieving it, others find the system entirely toostressful. Moreover, the American tradition of rug-ged individualism exerts a pull against close bondswith customer fins. Some small companies thinkthat if their quality and delivery times improve, theyshould be rewarded with higher prices-not with along-term tie to a demanding customer. Some seesuch ties as threatening to their independence. Theywould prefer to take their chances in the biddingbattle rather than find themselves beholden to toofew major customers. The Japanese system doesmake for heavier dependence on a few customers——only tolerable, perhaps, in a situation where manysuppliers trade with their major customers for 15 or20 years .22

A Japanese engineer who has observed relationsbetween large and small companies in both Japanand the United States put it this way. In Japan, smallcompanies making parts for computers or copiers orfacsimile machines are very conscious that they arein the office automation business. They carefullymonitor the price they have to stay under so that theircustomers, the companies that assembles the ma-chines, can be competitive. In the United States,small companies are not so conscious of being partof a whole.

Dependence may be lessening even in Japan; aseconomic growth has slowed, some lead companieshave actively encouraged their suppliers to seekother customers. The bonds of long-term relationsare still strong however. It must be remembered thatthe system has roots in the centuries-old tradition ofmutual obligation, and that it developed over dec-ades in the postwar period when it suited the needsof all parties quite well. The major manufacturerswere growing too fast to do all their own work; thesmaller companies were eager to take part in thegrowth, and also to get access to modern technologyat a time when foreign currency was scarce andgovernment restrictions allowed only a few firms toimport the latest machinery from Europe and Amer-ica. Today, the parties to the bargain still seemsatisfied, on the whole, that it is working to theadvantage of all.23

LINKS BETWEEN SEGMENTS OFAN INDUSTRY COMPLEX

A variant of the strategy of close relationsbetween major manufacturers and their suppliers isclose links between different segments of an indus-try complex--e.g., between the manufacturers ofchemical fibers, textile producers, apparel makers,and retail clothing businesses. There is more than ashade of difference in this variant. A chain of moreor less independent industries selling to and buying

22A 1983 survey of 1,54.() Japane~ subconwactors in the metal/machining industry found that, on average, these firms relied on one large customer(parent firm) for 60 to 65 percent of their business. (D.H. Whittaker, op. cit.) Mari Sako found in her study of printed circuit board suppliers in Japanand the United Kingdom (where customer-supplier practices are similar to those in the United States) that the Japanese suppliers depended much moreheavily on fewer customers. Comparing companies of similar size, Sako found that in Britain orders from the largest customer made up 6 to 25 percentof suppliers’ total sales. In Japan, the largest customer accounted for 15 to 85 percent of the supplier’s total sales. (Mari Sake, op. cit.)

z3Korea and T~wan, ~~ch me followlng tie Japanew model of expofl-1~ gro~h in many Ways, have not emulated the Wppher gK)l,Ip SySteII1.

Korea’s chuebol are industrial empires, typically doing business in a few related sectors, under the ownership and management of a founding father andhis heirs. They do not rely on long-term, stable relations with small subcontractors but rather buy or start up new firms to meet their needs. In Taiwan,business groups are much less prominent than in Japan or Korea. The groups that do exist arc made up of rather small firms in different economic sectors,with horizontal rather than vertical relations; the same people or their relatives hold management positions in the different fins. Relations withsubcontractors are not particularly close or long-lasting. (Gary G. Hamilton, Marco Orru, and Nicole Woolsey Biggart, ‘‘Enterprise Groups in East Asia:An Organizational Analysis, ’ Shoken Keizai, September 1987.)


from each other differs considerably from thesuperior-subordinate relation of a lead manufacturerand its network of suppliers. Nevertheless, evenamong the nominally independent members of anindustry complex, a purchaser who has attractivealternatives for material supplies wields more powerthan the supplier. In the case of the fiber-textile-apparel complex, it is designers and retailers whohold this power. They can buy anywhere in theworld. Increasingly, in the past quarter-century, theyhave done so. Imports account for well over one-third of what Americans spend on apparel.24

In nearly all high-wage countries, the textile andapparel industries face a tough challenge from poorcountries. Apparel manufacture is labor-intensive.Modern textile production is less so, but the capitalrequirements are relatively modest—well within themeans of newly industrializing countries (e.g.,Taiwan, Korea) and not out of reach for some poorerones (China). The textile-apparel industries thatseem to do best in high-wage countries are thosewith close ties between industry segments, wherefirms in the supplier industry focus their efforts onresponding to customers’ needs.

In the United States, textile producers and apparelmakers have traditionally had standoffish or evenhostile relations.25 The main concern of textileproducers was to mass-produce with high-speedequipment, rather than deal individually with cus-tomers’ needs. Apparel makers, if they were bigenough, treated their textile suppliers as inter-changeable and disposable, bargaining with numer-ous firms to drive the price down. This situation hasbegun to change. Industry leaders are realizing thatcloser links, from fiber production through retailing,

can lower costs, lend stability to all parts of thecomplex, and give an edge to domestic producers.

The Quick Response system was devised by U.S.industry leaders to foster these tighter links andcapitalize on the advantage of being close geograph-ically to the big American retail market,26 Imports(most of which are from low-wage countries) havethe attraction of lower prices;27 but there are alsoextra costs in doing business with importers. Besidesthe obvious ones—transportation, travel overseas,advance letters of credit, and extra paperwork-thelong leadtimes usually involved in overseas pur-chases also mean extra cost. When retailers order ayear ahead of time they pay carrying costs for largeinventories; they lose profits when they have to markdown unsold goods at the end of the season; and theypay still more in lost sales when items the customerswant are out of stock. One industry expert estimatesthat these costs add up to 25 percent of the value ofnet retail sales.28

The Quick Response system uses just-in-timeprinciples to reduce these costs. It allows the retailerto start the season with a wide but shallow selection,and when stocks get low, to re-order and get fastdelivery. About 80 percent of retail apparel businessis in items that have a shelf life of only 10 to 20weeks, either because they are ‘‘fashion” items instyles that are quickly changed or because they areseasonal. Quick Response is most obviously a usefulstrategy in these lines. However, some producers oftextiles for non-seasonal products, such as beddingor men’s underwear, are finding that close, stableties with their customers make it possible to cutinventories nearly to zero by just-in-time manage-ment, and thus to save costs.

z40TA’~c~timateof fiprt ~ne~atjon ~ ~ppwel is 36 ~rcen~ for 1987. It is b~ on dollarval~e, and includes freight, insurance, and import duties;shoes are not included. Other dollar value estimates of impons, which include the costs of transponation within the United States and other extras, putthe import penetration ratio for clothing at 57.5 percent. See “Import Penetration in the Apparel Industry: A Technical Study,’ prepared for the Fiber,Fabric and Apparel Coalition for Trade, September 1988. Import penetration is less for textiles, about 9 percent. (The apparently low figure for textileimports are misleading. Over the past 30 years, many foreign producers have switched from textile to apparel exports, because apparel has more valueadded. The quotas limiting imports combine textiles and apparel; textiles embodied in the apparel are not counted separately.) The combined importpenetration ratio for textiles and apparel was 25 percent in 1987, according to OTA’s estimate.

~This is ~u~ly not~e of textile producers and industri~ cons~em, such as auto Compafies buying seat cover fabrics, or hotel chains buying carpet.~pically, U.S. textile producers keep close ties with these industrial customers, and are very responsive to their needs. This is probably one reason forthe greater success of the industrial fabrics sector, compared with the apparel fabrics sector, in fending off imports.

26A conmlt~g fjm, Km s~mon Associates helped to devise the plan; the DuPont chemical company and Roger ~lliken of tie ~lliken textilecompany have been leading champions. DuPont is an important producer of textile fibers.

27’rhe top foW ~xtile ad ~p~el ex~ers t. tie Unitti states~~a, TAwan, Korea, and Hong Kong have textile wages ranging from about 2to 23 percent of U.S. wages; the next two-Japan and Italy-now have textile wages 30 to 40 percent above those of United States, since the fall of thedollar.

Z8KW Salmon Associat~, The KSA Perspective (New York, NY: January 1986).

Chapter 5-Links Between Firms and Industries ● 137

To make Quick Response work, each of theindustry segments upstream not only has to cutresponse time in its own operations but also needs tocooperate with the purchaser or supplier next in line.The example of Greenwood Mills, a large textilefirm specializing in denim, is illustrative. Until a fewyears ago, Greenwood bought fiber from fourdifferent suppliers, shopping around to drive theprice down. Greenwood’s biggest customers fol-lowed the same tactics, shifting orders among eightor nine suppliers on the basis of lowest price.

Greenwood and its suppliers and customers havesince adopted a more collaborative way of savingcosts. Greenwood now buys fiber from just twosuppliers, who offer quality, service, and guaranteeddelivery times in return for assurance of a long-termrelationship. Using this system, Greenwood has cutinventories from 3 weeks to a fraction of a week andis able to hold $40 million less in stock. In the sameway, Greenwood’s two biggest customers are now,by mutual agreement, reliable long-term purchasers.Greenwood takes the responsibility of loadingdenim into the trucks in sequence so that colorsalways match, marking the cuts electronically, anddelivering so quickly and reliably that one jeansmaker cut inventory from 4 weeks to 3 days and theother got rid of its warehouse. The denim isdelivered directly to the sewing room.29

The heart of Quick Response is responsivenessand interaction with customer firms. Another exam-ple comes from the Dan River textile mill, acompany that concentrates on making high-qualityapparel fabrics and emphasizes close customerrelations. Individual looms at Dan River are markedfor production for specified customers. And a DanRiver representative was on the floor at one shirt-maker’s plant so often that he was mistaken for anew employee.

A major achievement in the Quick Responseprogram was inter-industry adoption of a commonbar code standard. This allows electronic communi-cation between retailers and producers all the wayback through the supply chain. When and if thesystem is widely adopted, a textile mill, say, couldstart preparing anew order to send to apparel makerson the basis of electronic data passed back automati-cally from department store sales.

The close, responsive inter-industry links justnow being developed in the United States have longbeen a feature of the Italian, Japanese, and Germantextile-apparel industries-the three high-wage coun-tries that are usually considered the most successfulin these industries. This is not to say that closelinkages are a guarantor of success. All theseindustries have other features in their favor. Forexample, the Italian and Japanese industries benefitfrom a dense network of technical, organizationaland financial support, private and public. TheGerman industry has the advantage of an excellentcentury-old vocational education system. All threetextile industries (and the U.S. industry as well) arewell-equipped with modern machinery. None ofthese industries, even the best, is invulnerable tocompetition from low-wage countries. But it seemsclear that suppliers’ ability to respond quickly to theneeds of their customers and purchasers’ willingnessto form stable, cooperative relations with theirsuppliers are part of the mix that makes theseindustries more competitive, and helps them tosurvive without constantly escalating trade protec-tion.

LINKS BETWEEN MAJORMANUFACTURERS AND CAPITAL

EQUIPMENT PRODUCERSA special case of linkage with suppliers is the

relation between lead manufacturing companies andthe firms that make production equipment for theindustry. Perhaps even more than suppliers of partsand components, makers of capital equipment de-pend for their success on close relations with themanufacturers down the line who are their custom-ers. In the semiconductor industry, for example,customer firms (the chipmakers) were the source oftwo-thirds of the ideas for advances in productionequipment in the last few years.30

Customer firms, in turn, benefit from easy andcontinuing exchanges with the makers of theirproduction machinery. Sometimes they can achievethis with foreign manufacturers, as seems to be thecase in textile manufacture. The virtual disappear-ance of U.S. firms from production of the mostimportant kinds of textile machinery is apparentlynot crippling to textile producers. But in a rapidly

Z90TA intemiew wi~ Thomas 0’Gorrnan, President, Greenwood Mills, Dec. 11, 1987.Swric Vm Hip@, The ~o~ce~ of ]nnovation (New York, NY: Oxford University Press, 1988), p. 4, table 1-1.


advancing high-technology industry, close links canbe crucial. Already, U.S. semiconductor manufac-turers are at something of a disadvantage becauseU.S. equipment makers have lost out to Japaneserivals, and the handicap could become greater.

The story of the U.S. textile machinery industryillustrates the dependence of equipment makers onclose ties with their customers, The industry’sprecipitous decline was due largely to its failure torespond to customers’ needs. In 1960, Americanmakers of spinning, weaving, and knitting machin-ery dominated the U.S. market, accounting for 93percent of sales. By 1986, their share was 42 percent,most of it in spare parts and ancillary machinery.Several leading firms in the industry were organizedin the traditional Ford manner for mass production,with semi-skilled workers on the assembly lineturning out long runs of limited kinds of machinery.The attitude of these companies toward their cus-tomers was, “This is what we make; how many doyou want?”31

The merger mania of the 1960s also played a partin the industry’s decline. During that decade, all theBig Five U.S. textile machinery firms and manysmaller ones sold out to conglomerates. RockwellInternational, for example, not only bought Draper,the leading U.S. manufacturer of looms, but alsosmaller companies such as the Textile MachineWorks of Reading, PA, which made knitting ma-chinery. Some of these companies had built theirbusinesses on a solid tradition of close relations andgood service to their customers. But the newconglomerate owners lacked both technical knowl-edge of the business and interest in serving individ-ual customers.

Scanty spending on research and developmentwas another major cause of the deline, with U.S.producers lagging well behind the R&D spending ofcompetitors in Europe and Japan. When the Ameri-can textile machinery industry was seriously chal-lenged in the 1970s by innovative, responsiveEuropean and Japanese manufacturers, willing and

able to make a wide range of sophisticated machines,it lost.

According to people in the textile industry, theretreat of U.S. textile machinery makers from thebiggest part of the field (spinning, weaving, andknitting equipment) is not a serious technicalhandicap. They say that their German, Swiss, Italian,and Japanese suppliers keep improving equipmentin response to their needs, and that service (espe-cially from the Japanese) is outstanding. The mainproblem in dealing with foreign suppliers, as of thelate 1980s, was the fall of the dollar, which madenew equipment and parts suddenly much moreexpensive.

Semiconductor producers are faring worse. Japa-nese firms are now the world leaders in making theequipment that is most critical to chip production.According to industry sources, Nikon was notselling its leading edge model of this equipment toU.S. chipmakers in 1989, though the model wasalready widely used in Japan.32

As recently as 1979, U.S. firms dominated themarket for semiconductor production equipment,accounting for 79 percent of world sales. By 1989,the U.S. share was down to 47 percent and stilldropping 33 (figures 5-2,5-3, and 5-4). A central partof chipmaking is the fabrication of wafers, the 2- to8-inch silicon disks on which dozens to hundreds ofindividual chips are made. The most vital piece ofwafer fabrication equipment is the step and repeataligner, or stepper, which uses ultraviolet light toproject an outline of the chips’ circuit on the wafer;the circuit is then etched in an acid bath or reactivegas. An American firm, GCA, was first to commer-cialize a stepper, and it dominated the field until theearly 1980s. Nikon first pulled ahead in 1983.Today, GCA (which was bought by General Signalin 1988) is out of the Japanese market, has about 5percent market share in Europe and 20 percent in theUnited States. Nikon now occupies a commandingposition (table 5-2). It was Nikon’s latest and beststepper, the G-body, that was unavailable to U.S.firms in 1989.

Slchmles F. sa~l, Gary Herrigel, Richard Kazis, and Richard D&g, “How To Keep Mature Industries Innovative,” Technology Review, Apr. 28,1987.

gz~cip~ ~Wces for the following section are OTA’S review of the literature and interviews with leaders in the semiconductor and allied industries,and with officials of the Sematech consortium; other sources include Industry and Trade Strategies, ‘‘The U.S. Electronic Industry Complex,’ contractorreport to the Office of Technology Assessment, October 1988; William F. Finan and Jeffrey Frey, ‘‘Study of the Management of Microelectronics-RelatedResearch and Development in Japan,’ contractor report to the Office of Technology Assessment, November 1988.

33 VJ-,SI Rese~h, Inc., personal communication, Jan. 5, 1990.


Figure 5-2--Shift in Market Shares for Wafer Steppers

80

60

Ewl-l&L

40 - Japan

20 \

o I I I I I I I’83 ’84 ’85 ’86 ’87 ’88 ’89

est

NOTE: The wafer stepper is a device central to manufacturingsemiconductors.

SOURCE: VLSI Researeh, Inc.

Of the several reasons why Japanese firms havebested U.S. equipment makers, a leading one is thatU.S. chipmakers were themselves losing out to theJapanese competition.34 Japanese firms began tospend more on capital equipment than their U.S.rivals in 1983, and continued to outspend Americanfirms throughout the industry’s worldwide slump in1985 -86.35 Increasingly during this build-up, Japa-nese chipmakers bought Japanese-made productionequipment—in the case of steppers, overwhelm-ingly from Nikon. GCA, which had geared up toproduce 500 to 600 steppers (at $1 million apiece) in1985, sold barely 100 for the year, and wound uplosing $94 million. Financially weakened, sufferingdelays in delivery of lenses from the German firmCarl Zeiss (Nikon made its own lenses), and making

Figure 5-3--U.S. Market Shares of SelectedSemiconductor Equipment

100

U.S. resist

processing80 \ .\ . .-.

/’ y‘:K

\ \.\\

60- - U.S. integrated \ - --a ‘-’p- \circuit testers -~. —. - - -

‘x.A’ ‘\ “\

40 1 P I

/1’\\ \.U.S. stepping ‘~ .,aligners ‘x

20

1979 1980 1981 1982 1983 1984 1985 1986 1987 1988Year


a stepper that was no longer clearly the world’s best,GCA never recovered.36

How Nikon caught up with GCA technologicallyis another part of the story. Close relations betweenthe maker of production equipment and the customerusing it played an important role. The Nikon stepperwas an outgrowth of the very large-scale integration(VLSI) project which MITI directed from 1976 to1979. The goal of this cooperative industry-government project was to help Japanese companiesmaster the technology for making the newest genera-tion of semiconductors and, more broadly, to en-courage the national move toward more knowledge-intensive industries.

The emphasis of the VLSI project was on themanufacturing process. One-third to one-half of thebudget went for purchase of equipment (including aGCA stepper), and the five chipmakers who were the

3.11n 1981, Us. merchant ~ompafie~ (ho= hat pr~du~e chips for the open mmket, not just for their own intern~ u=) ~ared the big important marketfor dynamic random access memory chips (DRAMs) equally with Japanese firms. By 1988, U.S. firms had 8 percent of the world merchant DRAMmarket (this excludes chips made by integrated firms such as IBM for their own internal use), Japanese fms had 87 percent, and most of the rest wasdivided between West Germany and Korea.

3sDatWuest fiWes, ~ show ~ T~ Sem”con&ctor ~~~~, rew~ of a F~er~ ~teragency Staff working Group (Washington, DC: NationalScience Foundation, 1987), p. 28, chart 24. The data cover U.S. merchant (but not captive) producers. The rate of capital spending by Japanese companies(i.e., spending on plant and equipment as a percent of revenues in the integrated circuit business) has been higher than the U.S. rate since 1982. For yearsbefore 1981, the data on rates of capital spending are in conflict. OTA’s data for 11 U.S. merchant producers and 11 or 12 Japanese producers, showthat the Japanese rate was higher from 1973 to 1980, and nearly the same in 1981; see U.S. Congress, Office of Technology Assessment, lnrernutionufCompetitiveness in Electrom”cs, OTA-ISC-200 (Springfield, VA: National Technical Information Service, November 1983), p. 274. According toDstaquest, the Japanese rate was about equal to the U.S. rate from 1977 through 1981 but in 1982 and thereafter was higher; see The SemiconductorIndustry, op. cit.

36~cording t. Us. ~du~ ~~ces, (he GCA stepwr has better focus and more pr~i~ alignment ~an tie NikOn—but only when WlgiIleel’S Set lt

up. The Nikon stepper is more robust and requires far less set-up time. h can run well day after day with little adjustment, and therefore is much superiorin throughput (an important consideration for mass production of commodity chips).


Figure 5-4--World Semiconductor Equipment Sales

Billions of dollars

‘~

4

3

2

1

0LIL1 9 8 4 1985 1 9 8 6 1987 1 9 8 8 1989 1 9 9 0

- United States = Japan D Rest of world ~ Joint ventures

-Forecast


main participants from industry worked hand inglove with equipment makers. The project managersselected Nikon to develop a made-in-Japan waferstepper—a logical choice since Nikon already had afine reputation as a maker of precision opticalequipment, cameras and lenses, and also precisionmechanical measuring instruments. Toshiba, thefirst Japanese company to have used GCA stepperson a production line, was chosen to work with Nikonon behalf of all the member companies.

The collaboration was extraordinarily close. Ac-cording to a GCA engineer familiar with the effort,Toshiba set performance specifications but did notprovide a design. Instead, Toshiba engineers re-viewed all details of development, manufacture, andtesting; provided technical help in design concepts,electronics, and materials and components selection;and in the process visited Nikon several times aweek. 37 The result was a stepper which, though notradically different from GCA’s, gained the reputa-tion of being more reliable.

The close relation between vendors and users wasnot unique to the VLSI project. It is characteristic ofthe Japanese semiconductor industry, and remains apotent factor in the industry’s success. The sameengineers who oversee supplier companies’ devel-

opment of a new piece of equipment are thenresponsible for putting it to work on the productionline, where their familiarity with it pays off in rapidachievement of high productivity and quality. Thiskind of collaboration is largely missing in the U.S.semiconductor industry. According to officials ofSematech, the U.S. industry-government consor-tium that is working on generic improvements in thesemiconductor manufacturing process, the lack ofclose relations between equipment producers andchipmakers is a serious handicap. The consortiumhas given top priority to improving those relations,and to developing a full range of high-quality,reliable, affordable equipment and materials for theU.S. semiconductor industry.

Some firms-notably big ones like IBM andAT&T—have worked closely with equipment pro-ducers. But the merchant firms (those that sell chipson the open market rather than producing chipslargely for their own use) have typically hadarm’s-length relations with their equipment suppli-ers. Sometimes the relations are downright distrust-ful; new equipment firms are often started byexecutives defecting from companies that manufac-ture chips or from other equipment firms. TheJapanese firms’ habit of collaboration extends totheir American as well as their Japanese suppliers.Spokesmen for GCA noted that their Japanesecustomers were more demanding than Americanfins, asking for more fine-tuning and changes in theequipment they bought. But they were also morehelpful in making suggestions for improving theequipment.38

It is worth repeating that vendor-user relationswere not the only factor in Nikon’s (and laterCanon’s) emergence as world leaders in photo-lithographic equipment.39 The nearly instant pref-erence Japanese semiconductor firms gave to theNikon stepper, combined with the large investmentsin new equipment that these firms made through themid-1980s, were critically important. In 1981, GCAhad 95 percent of the Japanese market. The nextyear, it had 40 percent. Today it has next to nothing.Toshiba took the Nikon stepper as soon as it was out,in April 1981. NEC followed in early 1982 when

3~~m ~d FRY, op. cit., citing JotI slgurd~n, “Industry and State Partnership in Japan: The VLSI Project,” Discussion Paper No. 168 (Lund,Sweden: Researeh Policy Institute, 1986), p. 48.

SSOTA interview with GCA.39when Nikon fwst brought out its stepWr, it Wm tie only Japane= pr~ucer; Croon stuck to m~ng the older process ~igner. Later, 8S GCA

weakened, Canon entered the stepper market.

Chapter 5----Link-s Between Firms and Industries ● 141

Table 5-2—Top Ten Semiconductor Equipment Suppliers, World Sales(millions of dollars)

1982 1988

Perkin-Elmer . . . . . . . . . . . . . . . . . . . . . . ..$162 Nikon . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..$521Varian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 Tokyo Electron (TEL) . . . . . . . . . . . . . . . . . 508Schlumberger . . . . . . . . . . . . . . . . . . . . . . . 96 Advantest . . . . . . . . . . . . . . . . . . . . . . . . . . . 385Takeda Riken(Advantest) . . . . . . . . . . . . . 84 Applied Materials.. . . . . . . . . . . . . . . . . . . . 382Applied Materials . . . . . . . . . . . . . . . . . . . . . 84 General Signal... . . . . . . . . . . . . . . . . . . . . 375Eaton . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 Canon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290Teradyne . . . . . . . . . . . . . . . . . . . . . . . . . . 79 Varian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211Canon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 Perkin-Elmer . . . . . . . . . . . . . . . . . . . . . . . . 205General Signal . . . . . . . . . . . . . . . . . . . . . . . 77 Teradyne . . . . . . . . . . . . . . . . . . . . . . . . . . 190Nikon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 LTX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180

(Japanese Firms Italicized)

SOURCE: VLSIReaearch,lnc.

prices offered by the two rivals, after intensecompetition, were equal. NEC said its decision infavor of Nikon was based on technical superiority,availability of local service, and early delivery.Then, when semiconductor sales nosedived in 1985,U.S. chipmakers canceled their orders for GCAsteppers—a near mortal blow to a company that hadjust invested heavily to expand capacity.

The troubled condition of Perkin-Elmer’s semi-conductor equipment division, a major U.S. supplierof photo l i thographic equipment , underscores thepoint that other factors besides relations with cus-tomers are important to success in the semiconduc-tor equipment business. For over 20 years, IBMworked closely with Perkin-Elmer on various kindsof equipment (though not the stepper, which Perkin-Elmer effectively ceded to GCA). Recently, withIBM’s financial and technical help, the companydeveloped an advanced step-and-scan machine, theMicraScan,that is said to be a technological wonder,w i t h t h e p o t e n t i a l t o v a u l t o v e r t h e J a p a n e s ecompet i t ion . @ Yet despite Perkin-Elmer’s technicalabilities, and despite its close-working relationshipwith IBM, its semiconductor equipment divisionwas a financial loser in the 3 years 1987-89 (the mainpar t of the company’s bus iness i s in sc ient i f icins t ruments) .41 In April 1989, Perkin-Elmer offeredits semiconductor equipment division for sale.

One reason for Perkin-Elmer’s decision to bowout is the heavy spending for technology develop-ment that the fast-moving semiconductor businessdemands; new generations of both chips and equip-ment appear about every 3 years. Perkin-Elmer(andIBM) spent $100 million in 4 years to develop theMicraScan, and faced costs of $50 to $l00 millionmore to refine and update the equipment. The highcost of capital and pressures for short-term profits inthe United States add to the burden of makingcontinuing high investments in advancing technol-o g y .42 Nikon and Canon, Perkin-Elmer’s Japanesecompetitors, have the advantage of easier access tolow-cost capital and less pressure to show short-termprofit;and both these firms excel in engineering andmanufacture.

IBM declined to buy Perkin-Elmer’s semiconduc-tor equipment division, on grounds that the expertisefor running a toolmaking business was outside itsarea of competence. No other U.S. buyers had comeforward by the end of the year. Nikon; which hasboth the technical and financial resources to run thecompany, was the leading suitor but then backed off,apparently because of U.S. political objections to thesale.

The erosion of leadership in production equip-ment is already a handicap for the U.S. semiconduc-

40~-.ording t. ml R~searCh, ~ ~om~ting fim that sp~l~izes in semiconductor equipment, perkin-Eher’s new machine has a 3- to A-year leadon ail the competition, including Nikon and Canon. Alan Cane, ‘‘Chips Are Down for Perkin-E1mer,’ Financial Times, Dee. 7, 1989, p. 21; see alsoAndrew Pollack, “The Challenge of Keeping U.S. Technology at Home,” The New York Times, Dec. 10, 1989, p. 1.

QIThedivit&onhad revenues of$l~ mil]ion and operating losses and charges against earnings of $z~ million h tiethrm years, according to an aIldySt

with Shearson IAtrnan Hutton, Inc. (Pollack, op. cit.) The Perkin-Ehner company as a whole lost money in 1987 but made a profit in 1988.42sW ch. 3 for a dl~cussion of tie us. financial environment and its eff~t on t~hno]ogy development.43KoNm pr~ucem, wi~ ~elr Iow.co~ labor, me a bigger heat to Japan in l-megabit memory chips than U.S. manufacturers, and tie Koreans are

reported to be worried that their access to Japanese equipment may be restricted. (David E. Sanger, ‘‘South Korea’s High Tech Miracle,’ The New YorkTimes, Dec. 9, 1988, p. D1.)


tor industry, at least in wafer fabrication.43 (Ameri-can companies are apparently able to buy some otherkinds of Japanese-made production equipment, suchas automatic assembly and testing equipment, onfairly equal terms with Japanese chipmakers.) Thesituation could deteriorate. Microelectronics is oneof the world’s most dynamic industries. Chipmakersin the nation where critical new technology in waferfabrication is first developed will almost certainly bethe frost to use it, and thus will gain a vital advantage.X-ray lithography is a strong candidate for the nextstep, and in this emerging technology, the Japaneseare well ahead.

Photolithographic steppers, used to etch today’sl-megabit chips, can go up as far as 16-megabits, butthe chips of the late 1990s and the early 21st century(64 megabits and beyond) will have circuitry withlines too fine for etching even by ultraviolet light.X-rays, with their shorter wave length, have muchgreater potential. Ultimately, chips made with X-raytechnology might be able to store 1,000 times asmuch data as the current l-megabit chip.

Development of X-ray lithography is expensive.The Sematech consortium, for example, had to ruleout extensive work on the technology because itsfunds-about $200 million a year over 5 years—would not stretch that far. In Japan, the half dozenleading chipmakers, several supplier firms, and theofficially privatized (but still mostly government-owned) Nippon Telegraph & Telephone (NTT) wereall involved in a MITI-led program to developvarious aspects of the X-ray technology. In the late1980s, the Japanese effort was outspending Ameri-can efforts at least five to one, and several firms wereengaged in developing a compact synchrotrons togenerate X-rays at the right wave-length and inten-sity, at commercially acceptable costs.44 In Japan, in

1989, ten compact synchrotrons were under con-struction or already at work on experimental proj-ects, and five more were on the drawing boards.Development of a compact synchrotrons was also faralong in Germany; government funds have helpedsupport the Siemens company in its development. Inthe United States, only IBM was constructing acommercial-type synchrotrons, and the cost wasstraining even its resources. IBM invited other U.S.chipmakers to participate in the effort, and Motorolasigned on in late 1989.

Commercial use of X-ray technology may verywell come about in the 1990s.45 If it does, the firstcommercial use will most likely be in Japan, givingthat country’s semiconductor producers a big lead ina new round of world competition. As the Japanesesemiconductor industry itself has shown, it ispossible to catchup even when one is far behind. TheUnited States has yet to demonstrate this ability,although projects such as Sematech are a move inthis direction.

INTERNAL LINKS: VERTICALINTEGRATION, PRODUCT

DIVERSITY, ANDLARGE SIZE

Japanese firms are the world’s leading producersof semiconductors, with 45 percent of the worldmarket in 1989. U.S. companies, which held 53percent in 1984, were down to 42 percent anddeclining 46 (see figure 5-5). Six of the world’s top 10companies in sales of semiconductors on the openmarket are Japanese, and all of them are large, stable,integrated electronics firms that make everythingfrom chips to computers and consumer electronics.They make more chips than they need and sell the

43 Korean pr~ucers, wi~ their low-cost labor, arc a bigger threat to Japan in l-megabit memory chips than U.S. manufacturers, and the Koreans arereported to be worried that their access to Japanese equipment may be restricted. (David E. Sanger, *‘South Korea’s High Tech Miracle,’ The New YorkTimes, Dec. 9, 1988, p. D1.)

~~e J~pane~ Prowm had spnt $XM to $750 million by late 1988, and planned to spend $200 million more; comparable fig~es for tie ufit~States were $50 to $100 million already spent, and $100 million planned. (John Markoff, “Experts Warn of U.S. Lag in Vital Chip Technology, ” TheNew York Times, Dw. 12, 1988, p. 1.) A technology that generates X-rays by pulsed laser sources is a possible alternative to the synchrotrons, slowerbut perhaps more practical. Japanese R&D is also pursuing this possibility.

4@pucal (u]&a-violet) ]i~o~aphy may lwt a while longer, however; Sematech is betting that it will, and is putting much of its effon into s~etchlngthe technology to its farthest limits. The history of the semiconductor industry shows that tedmologies sometimes last longer than expected, For example,several Japanese companies got a headstart on U.S. firms in manufacturing the 64K dynamic random access memoty chip because they used an older,conventional technology while the U.S. companies were trying to get the bugs out of a newer one.

4~e= figwes me from VLSI Re=uch Inc., and are for all semiconductor production, including ifltra.company captive production aS Wel] ilS

merchant production for the open market. Figures from Dataquest are for merchant semiconductor production only; they show U.S. producers holding37 percent of the world market in 1988 (down from 61 percent in 1980), and Japanese producers holding 50 percent; see National Advisory Committeeon Semiconductors, A Strategic Industry ar Risk, a report to the President and the Congress (Arlington, VA: The Committee, 1989).


Figure 5-5—World Semiconductor Sales

Billions of dollars35 I

3 0

25

I20

t

10

5

L. . .:::

o . . .1 9 8 3 1 9 8 4 1 9 8 5 1 9 8 6 1987 1 9 8 8 1989”

- United States = Japan ~ Europe D Rest of world

-Forecast

SOURCE: VLSI Researeh. Inc.

rest. Most U.S. companies that sell on the market aresmaller, less stable, and less integrated. Some makevirtually nothing but chips. Although there are largeintegrated U.S. companies making chips—IBM,AT&T, the Delco division of GM—they typicallyuse most of the chips they make and buy morebesides.

One explanation for the explosive success of theJapanese is the structure of their industry. It has beenstrongly argued that the vertical integration, largesize, and product diversity of Japanese firms givethem an advantage of staying power that is almostunbeatable---even if the U.S. industry succeeds in itsstrenuous efforts to catch up to the Japanese inmanufacturing excellence.47 Moreover, the big inte-grated firms can use their most advanced chips toimprove their own end products (computers, workstations, robots) well before they sell the chips on theopen market to competitors.

It is true that Japanese firms are using theirstructural features to advantage (American firmswith much the same features have not been sosuccessful, however). Possibly other arrangements—close collaborative relations between suppliers andcustomer firms, say--could give U.S. companiesmany of the same benefits that the integratedJapanese firms enjoy. These arrangements wouldnot, however, provide the kind of financial strength

that helped the Japanese firms weather the steepsemiconductor recession in 1985-86. (Volatile de-mand, independent of the business cycle in theeconomy as a whole, is typical of the semiconductorindustry, although the 1985-86 downturn was deeperthan usual.) But it is well to remember that theJapanese industry was not always so well-heeled asit is today. One must look to other factors to explainJapanese staying power before the plush era of thelater 1980s. Government support, financial andotherwise, had much to do with it. So did thewell-known ability of Japanese managers to take along-run strategic view, rather than going for short-term profits.

The supposed advantages of integration and largesize are most relevant to the semiconductor industry.(Lesser integration is often proposed as a remedy forother industries--eg., the supplier group system aseasier to manage and more conducive to innovationthan GM-style integration in autos; mini-mills asmore flexible, responsive, and efficient than theintegrated behemoths of the steel industry.) But thesemiconductor industry is well worth considerationon its own, for it is at the heart of technologicaladvance in every sector of the economy, from autosto computers to banks to defense.

Links to Markets, Financial Stability

All the leading Japanese semiconductor produc-ers belong to big, vertically integrated firms. All sellchips on the open market, but some 50 to 70 percentof the chips they make are used internally or sold toan affiliated firm in their industry group.48 Facingcompetition in the open market probably strengthenstheir performance, and having a large, reliabledemand lessens the risk in investing the $300million or more that building a state-of-the-artsemiconductor plant now requires. In addition, a big,diversified company can see its semiconductordivision through periodic downturns in demand, asin 1985-86, when Japanese producers are estimatedto have lost $3 to $5 billion, and U.S. producerssome $2 billion in memory chips. While the demandfor computers, and consequently for memory chips,plunged, the Japanese companies’ sales of otherelectronics products such as VCRs and compact disk

47For ~ exmple, ~ Michael L. mt-IOUZOS, Richard K. Lester, Robert M. Solow and the MIT Commission on Industrial Roductivity, Made inAmerica: Regaining the Pro&cave Edge (Cambridge, MA: The MIT Press, 1989), pp. 248-262.

dgMichael G. BOITUS, C’omperingfor Contro/ Arnerlca>s Stake in Microelectronics (Cambridge, MA: Ballinger Publishing CO., 1988), P. 111> table54.


players held up. The top six Japanese semiconductorproducers (NEC, Hitachi, Toshiba, Fujitsu, Mat-sushita, and Mitsubishi) are divisions of integratedelectronics firms that had sales of $10 to $23 billionin 1987. Semiconductors accounted for only 8 to 17percent of their sales.

The top U.S. firms are much more various. Ofeight leaders (companies ranking in the top 20 forworld market share), two, IBM and Hewlett-Packard, produce chips almost solely for their ownuse. The rest range from about 20 percent outsidesales (AT&T, which recently began to push externalsales) to more than 90 percent. The leading U.S.merchant producers—those that sell on the openmarket-range in size from medium to modestcompared to the Japanese electronic giants. The twolargest are Motorola and Texas Instruments, bothdiversified electronics firms of medium size, withtotal company sales of around $5 to $7 billion in1988. Much of their chip production is for their ownuse but they are also big producers for the market;semiconductors count for one-third or more of theirsales. The other leading merchant firms (Intel,Advanced Micro Devices, National Semiconductor,Fairchild) are primarily in the chip business.49 Theysell mostly to outside firms, and therefore lack theassurance of a large internal market. They areconsiderably smaller than Motorola and TexasInstruments (not to mention the Japanese electronicscompanies), with sales that run from about $1 billionto $2.5 billion. (IBM is primarily a computer andelectronic systems company but is also the biggestof all the semiconductor producers; its total sales inall product lines were $54 billion in 1988.)

Large size, diversity of product, and verticalintegration can have their down sides too; forexample, bureaucratic clumsiness and top manage-ment that does not understand the semiconductorbusiness. Indeed, in the United States, the moderatesize and flexibility of entrepreneurial semiconductorfirms have been hailed as the source of creativity andinnovation. And, in this country at least, some highlydiversified and vertically integrated companies havetried the semiconductor business with only limitedsuccess (e.g., RCA and Westinghouse). AT&T, avery large company ($34 billion in sales in 1988)

with a big internal market in telecommunicationsequipment, recently abandoned production of DRAMs.

It seems that large size and a high degree ofintegration are no guarantee of success in thesemiconductor business. But are they necessaryeven if not sufficient? (The question applies to majorplayers in the game, not to small niche producers.)And if large, diverse, integrated firms have a built-inadvantage, why hasn’t the U.S. industry taken thatdirection? The answers to these questions are notsimple or obvious. The U.S. industry developed astructure that was well-suited to an earlier period ofthe microelectronics business but does not fit as wellwith the requirements of a more mature industry.(However, other factors besides industry structurehave also favored Japanese semiconductor produc-ers as the industry matured; see the discussionbelow.)

The pioneering era of the business, from 1950s upto the mid- 1970s, was one of repeated technologicalupheaval as products were rapidly introduced andthen just as quickly superseded. The germaniumtransistor gave way to the silicon transistor; inte-grated circuits ousted the single transistor for mostuses; MOS (metal oxide semiconductors) succeededand largely replaced bipolar logic in highly inte-grated systems, as in the memory chips used incomputers. This environment of turmoil and fre-quent change was favorable to startups of new,creative companies bankrolled by venture capital.High turnover—20 percent a year on average,including top ranking professionals and managers aswell as production workers—has been the hallmarkof Silicon Valley since its early days. Engineers andscientists repeatedly peeled off and spawned newgenerations of highly innovative, but often short-lived, firms with a strong focus on new products.Along with the new products came substantialchanges in the manufacturing process.

In about the last dozen years, microelectronics hassettled down. Important changes are still occurring,but they have become more incremental than revolu-tionary. The 1-megabit memory chip of the late1980s is a fairly direct descendant of the 16-kilobitchip of 10 or 12 years earlier. It is made byessentially the same methods. But making chipswith ever finer lines and greater density requires ever

d~airchild Semiconductw cow., fo~erly a subsidiary of the international conglomerate Schlumberger, Ltd., was recently acquired by Nation~Semiconductor Corp. The sales figures cited here are from the 1989 edition of Standard and Poor’s Register of Corporations, Directors and Executives(New York, NY: McGraw-Hill, 1989); they do not reflect the acquisition.


more complex machinery, more exacting conditionsof manufacture, and greater capital investments. The$300 million or more that it takes to build and equipa semiconductor plant for memory chips todaycompares with an entry cost of $5 million or so in theearly 1970s.

Three characteristics of Japanese industry, apartfrom structure, are advantageous in the present stageof the microelectronics business. One is the demon-strated excellence of the Japanese in manufacturing.As an industry matures and incremental improve-ment takes the place of radical innovation, whatcounts most is the ability to shorten the cycle ofproduct development, to get the latest version of aproduct to market quickly, and to make the productto high standards of quality and reliability, at acompetitive price. This is just what the Japanese did,beginning with the sudden conquest of half the U.S.market for 16K random access memory chips in1980 and continuing with its successors, up throughthe l-megabit chip. Many well-known aspects of theJapanese system of manufacturing--collaborationbetween design and manufacturing engineers, scru-pulous attention to every detail of manufacturing,the team system for shop floor workers and theirclose involvement in quality and productivity, closecooperative links with suppliers, and a long-termview on the part of managers-contributed to thisoutcome.

Another factor is the relatively low cost of capitalin Japan and the favorable conditions banks andshareholders have long offered to major manufac-turers—a factor that grows in importance as capitalcosts rise.50 Related to this is the long-term viewcharacteristic of Japanese managers. The lowercapital costs are, the longer a company can reasona-bly wait for payoffs on its investment. Also, thelifetime employment offered by large Japanesecompanies, and the fact that employees typicallystay with one company for their entire careers,contributes to a strategy of counting on marketgrowth for prosperity, rather than taking instantprofits. The highly unstable attachment to compa-nies in Silicon Valley pushes in the oppositedirection. Still another factor, not discussed here butto be considered in a following report, is the

contribution of the Japanese Government’s indus-trial and trade policies to the success of industriesconsidered critical to the nation’s economic future—government support of R&D and assurance of aplentiful supply of low-cost capital, combined withexport promotion, tight restriction of foreign invest-ment, and protection of the domestic market.

With this perspective, it may be seen that themore-or-less assured markets for semiconductorsthat a vertically integrated electronics firm can offer,the stability furnished by product diversity, and thegreater power to make capital investments thatcomes with large size are great assets for the bigJapanese electronics companies, but are not bythemselves the decisive assets. A recent examplefrom Japan underscores the point that verticalintegration is not prerequisite to success. NMBTechnologies Inc. is a subsidiary of a prosperous butnot very large Japanese company, Minebea, whichhas a $1.5 billion yearly business in precision ballbearings. Entering the semiconductor business in1983, NMB invested $250 million in a world-classfabrication plant, and started producing superfastDRAMs in volume in 1985. Despite the worldrecession in chips, NMB hung on, and was readywith suitable fast memory chips when Intel andMotorola introduced their 32-bit microprocessorsfor top-grade personal computers in 1986. Granted,fast DRAMs are something of a niche market; yet theinvestment required to get into the business was farfrom trivial. NMB may later fall victim to a biggercompany deciding to compete in fast DRAMs, but in1989, only 4 years after starting production, it had 90percent of the world market in fast DRAMs, andexpected to double its 1988 sales of $250 million.51

For stand-alone semiconductor firms there maybealternatives to the internal markets that integratedcompanies provide. Long-term, stable relationshipsbetween chipmakers and chip users (i.e., builders ofcomputers, work stations, telecommunications equip-ment, industrial machinery, automobiles, consumerelectronic goods) might offer similar benefits. Anexample is the close ties between IBM and Intel,which makes a microprocessor for IBM computers.NMB owed much of its success to cultivation ofclose links with users such as Compaq Computer

5~hc~~=t~~S@~ of the J~~~~~se ~]w~ofics Companies in the late Iggos (and inde~ of he entire Japanese Uonomy) has reduced the lnlpOllitllCe

of bank loans and outside equity holders; many of these companies today are capable of meeting most of their own financial needs. See ch. 3 for discussionof this issue.

51 Larry Wailer, “How NMB Took Over the Fast-DRAM Market,” Eiecfronics, November 1988.


Corp. (its biggest customer) and Lockheed AircraftCorp.

Compared with their performance in manufactur-ing standard commodity memory chips, U.S. pro-ducers do better in the kind of chips where individualproduct design and attention to customers’ needs areparamount and price is secondary (e.g., micropro-cessors and application-specific integrated circuits,or ASICs). However, two can play at this game. Thetop three firms in ASIC sales are Japanese. Not onlydo they sell ASICs at home, they operate designcenters in Boston and Silicon Valley, send thedesigns back to Japan by satellite communication,and deliver the custom (or semi-custom) chips byair. The greatest remaining U.S. advantage is inmicroprocessors, where the creative talent of design-ers (a U.S. strong point) is of paramount importance;also, users of microprocessors tend to form long-term ties with producers because they invest insoftware that fits their particular microprocessor andits progeny.

For mass-production chips, however, investmentsin semiconductor plants have become so huge andthe sales needed to justify them so large that it maybe a good deal harder for an independent. undiversi-fied company to prosper now than it was in the past.The plenitude of capital that the large, integratedelectronics companies of Japan possess may be acritical asset. It is sobering to reflect that among U.S.firms, only IBM, Texas Instruments and the muchsmaller Micron Technology, Inc. stuck with DRAMproduction through the late 1980s, and most ofTexas Instrument’s production was in its Japanesefacilities. (Motorola was getting back into produc-tion of DRAMs in 1989, after making an agreementwith Toshiba to swap a license to produce Motor-ola’s microprocessor in return for access to Toshiba’sl-megabit DRAM technology.) As recently as 1980,there were 11 U.S. companies making DRAMs. Thismass-production chip is essential to computers,telecommunications, and many other kinds of equip-

ment, and has been a favorite technology driver forthe industry .52

The purpose of Sematech, the government-industry R&D consortium in semiconductor manu-facturing technology, is to help U.S. producersregain competitiveness in DRAMs and other mem-ory chips. Sematech is a novel venture for the UnitedStates; not only has it put together industry andgovernment funding on a large scale, it is formingstronger vertical links than have existed before in theU.S. microelectronics industry and is creating un-precedented horizontal links between competitors.Sematech is confined to R&D, stopping short ofmanufacture. A more radical approach was theproposal by several U.S. computer and semiconduc-tor companies, announced in June 1989, to form aconsortium and produce DRAMs commercially.The project failed to attract enough computer firmsas participants, however, and was abandoned inJanuary 1990.53

Links With Consumer Electronics

Another question about linkage in microelectron-ics is whether the loss of the U.S. consumerelectronics industry has deprived American chip-makers of an essential market. For Japanese chip-makers it is a huge market, taking 40 percent ofproduction; this compares to 7 percent for U.S.producers. The decline in U.S. producers’ share ofthe world semiconductor market does track to someextent the decline in U.S. market share of consumerelectronic goods; in other words, other purchasershave not fully made up for the lack of sales to makersof television and radio sets, VCRs, compact diskplayers, and the like.

Up to this point, the loss of sales in consumerelectronics has hurt U.S. chipmakers more finan-cially than technologically. This is because most ofthe chips used in consumer electronics differ basi-cally from those used in computers, telecommunica-tions equipment, and other high-technology prod-ucts. Analog devices, which receive an analog signal

52Te-.~olog dnve.S ~e ~~pS ~how mmufacture provides learning experience mat can hen ~ app]ied to o~er kinds of chips or later generationsof the same chip. DRAMs are good technology drivers because: 1) they are produced in large enough volume to supply data quickly for statisticalanalyses; 2) they are high-density integrated circuits that push the limits of current lithography technology; 3) they have a simple repetitive design, whichmakes it easy to test them for design or production defects; and 4) the manufacturing equipment and process technology required for DRAM productionis similar to that required for other chips.

ssch~ermem~rs were tiee computer manufacturers (IBM, Digital Equipment Corp., and Hewlett-Packard) and four chipmakers (Intel, Advanc~Micro Devices, Nationat Semiconductor, and LSI Logic). Both Apple Computer and Sun Microsystems decided not to participate, A spokeswoman forSun cited its “global purchasing strategy” and “existing long-term”agreements with other chipmakers as reason for refusal to join. ‘‘ElectronicsNewsletter,” Electronics, December 1989, p. 17.

Chapter 5-Links Between Firms and Industries ● 147

and amplify it, are much used in consumer electron-ics. Computers and the like use digital chips. At onetime (until the late 1970s) this divergence wassomething of a handicap to Japanese semiconductorproducers. While they excelled in manufacture ofanalog devices for their booming consumer electron-ics business, much of that experience did not carryover into the production of large-scale integratedcircuits based on digital electronics (with the VLSIproject, Japanese manufacturers got over this tech-nological hump).%

Today, consumer electronic goods are changingcourse toward much greater consumption of digitalchips. Compact disk players already use them. Newgenerations of television sets and related productswill use far more. There are about 160 milliontelevision sets in the United States, which suggeststhat the potential market for digital chips fortelevision alone could be large. Semiconductorproducers who fail to get into this market could findthemselves at a disadvantage-but not just in the TVmarket. More importantly, they could fall behind inthe know-how required for making successive gen-erations of computers and their applications. This isbecause the core technologies for consumer elec-tronics on the one hand, and computers plus manyother advanced business products, on the other hand,are converging.

All digital chips are in the same family—i.e., theyare made with similar kinds of equipment andmanufacturing processes. Anyone who can meet theexacting requirements for mass-producing digitalchips for consumer electronics items—high volume,low cost, high reliability—gains valuable learningexperience in making similar kinds of chips, welland cheaply, for computers and other businessproducts. The same goes for other components thatcomputers, telecommunications, and other businessproducts have in common with consumer electronicsitems. Moreover, advanced television could be theapplication where certain leading edge technologies—e.g., advanced displays and the new manufacturingtechnologies needed to make them—will be neededfirst.55

The Japanese are making rapid progress towardcommercializing high definition television (HDTV),and some companies are already poised to sell anadvanced version of conventional television that hasmuch improved definition. The United States is farbehind. Zenith, the last remaining U.S.-ownedproducer of television sets, has not yet brought tomarket an improved definition TV (IDTV), and is alate and uncertain entrant in the HDTV race.(Foreign-owned firms with production facilities inthe United States are pursuing advanced televisionsystems, however. The Dutch-owned Philips has anIDTV on the market, and the French-owned Thom-son Consumer Electronics has demonstrated anextended definition TV.) After the Defense Ad-vanced Research Projects Agency (DANA) setaside $30 million to encourage U.S. producers tomake HDTV’s, Zenith proposed to collaborate withAT&T on such a venture; a number of computer andwork station manufacturers are also interested. Butso far, HDTV activity by U.S.-owned companies isconfined to research and the earlier states ofdevelopment, with commercial production still yearsaway.

Technology Links

Another way that a vertically integrated companymay get ahead of the competition is to develop itsown advanced technology, and keep it for itself. Forexample, both Hitachi and IBM are said to havedeveloped some superior production equipment thatthey never sold or licensed to anyone else.

A similar kind of technology link is part of therising threat from Japan to the U.S. lead in super-computers, the fastest and most powerful of comput-ing machines. Three Japanese electronics companies(NEC, Hitachi, Fujitsu) are narrowing the U.S. lead.These large, integrated companies make their ownhigh-speed components. Currently the world leader,and the only U.S. company making supercomputers,is the comparatively small, stand-alone firm Cray

MU.S. Cmgess, Office of TtxhnoIon Assessment, internutwrud Competitiveness in Electronics, OTA-ISC-200 (Springfield, VA: Nation~Technical Information Service, November 1983), pp. 196-198.

SSFor det~ls, see tie section on advanced television in ch. 2, and OTA’s forthcoming report, The Big picture.


Research Inc.56 Cray does not make the high-speedcomponents needed for supercomputers, and theyare hard to get from other U.S. companies. This isone of the main reasons why the U.S. lead isevaporating, according to a report by a panel ofcomputer science experts to the Institute of Electri-cal and Electronics Engineers, Inc. (IEEE) .57 Thereport said:

The highest performance memory and bipolarlogic components useful for supercomputers . . . areavailable only from Japan. The managements ofCray and ETA have been quoted in the press atvarious times as stating that these Japanese compo-nents are “not yet available for export’ from Japanto Cray or ETA as devices-but they are available toend users in the Japanese supercomputer systems.Those systems are definitely available for export.

A senior NEC manager, Akihiro Iwaya, under-scored the point in an interview with The New YorkTimes. He said: “We have our own chip divisions.They can custom-make the high-speed chips weneed. Cray can’t. They have to buy them fromJapan. ’58

Officially, Cray managers have no complaintabout their Japanese suppliers. And indeed it isunlikely that Japanese producers would cut offsupplies to Cray, partly because that would cause

political troubles for Japan, and partly because saleof the chips is highly profitable. What is more likelyis delay in providing the latest and best chips toCray. According to one informed observer, both theJapanese firms are delaying up to a year in providingtheir latest chips to their American competitor.

Cray is under challenge from larger, more inte-grated companies in another way as well. Craygained its leading position in supercomputers by itsexcellence in what the industry calls packaging, thatis contriving to arrange chips in close quarters forspeedy operation, while draining away the heat theygenerate. While Cray still has the reputation foroutstanding engineering, it is facing very toughcompetition from bigger companies that make a fullline of computers, from mainframes down throughpersonal computers. Such companies can afford todevote a lot of engineering talent to solving packag-ing problems, since the results can eventually beapplied not just to supercomputers but to the fullline, and the costs recovered from this broad rangeof products. The same consideration may apply toother kinds of R&D spending as well, Not all of thisparticular advantage resides in Japan, however. IBMtoo makes a range of computers, and is supportingthe efforts of a former Cray engineer (Steven Chen)to build a new improved supercomputer.

%e still smaller ETA, a subsidiary of Control Data Corp., dropped out of supercomputerproduction in April 1989. Also, in May 1989, Cray Researchspun off anew company, Cray Computer, to be run by Seymour Cray, the founder and chief designer of Cray Research. Funded by Cray Research with$100 million over 2 yeim, the new company was setup to pioneer a promising but risky technology based on gallium arsenide chips. Reportedly, thereascmfor the move was to free Seymour Cray from the short-term pressures of Wall Street. (AIan Kane and Imuise Kehoe, ‘Challenge to the U.S. BrainsTrust,” Financial Times, May 18, 1989.)

571EE~SAB Committee on Communications and hformation pO1iCy, “U.S. Supercomputer Vulnerability, ‘‘report to the Institute of Electrical andElectronics Engineers, Inc., prepared by the Scientific Supercomputer Subcommittee, Committee on Communications and Information Policy, UnitedStates Activities Board (Washington, DC, August 1988).

SsDavid E. Sanger, ‘‘A High-Tech Lead in Danger, ’ The New York Times, Dec. 18, 1988, sec. 3, p. 1.

Chapter 6

Technology Transfer and Diffusion:Some International Comparisons

CONTENTSPage

DIFFUSION OF ADVANCED MANUFACTURING EQUIPMENT . . . . . . . . . . . . . . . . . . . . . . . : 152LOOKING OUTSIDE THE FIRM FOR NEW TECHNOLOGIES . . . . . . . . . . . . . . . . . . . . .,..., 156TECHNOLOGY DIFFUSION TO SMALL FIRMS . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

Major Companies and Their Suppliers +...... . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . .

159Japanese Government Programs for Small and Medium-Size Firms ~‘, : : 161Horizontal Links Between Small Firms . . . . . . . . . . . . . . . *...*........*..........,,,.*..,,, 167

BoxesPage

6-A. A Small Plant in Yokohama,. . . . . . . . . . . . . . . . . . . . . . . . . , . . . . . . . 1646-B. A Plastic Mold Equipment Cooperative in Japan . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . 170

TablesPage

6-1. Adoption Rates of NC Machine Tools in Five Major Industries,United States (1988) and West Germany (1986) . . . . . . . . ., . . . . . . . . . . . ., .,, ,., ,+ . . . 153

6-2. Penetration Rates of NC Machine Tools in Manufacturing Industries,. . . . .

United States and Japan, 1987 . . . . +... . . . . . . . . . . . . . . . ... ...0 ... *., ,++, .,** +*, ++6. *,++6 1546-3. Defense production and Use of NC Machines in U.S. Manufacturing

Establishments, 1988 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-4. Japanese Government Equipment Modernization Loan and Equipment

Leasing Systems for Small and Medium Enterprises . . . . . . . . . . . . . . . . . .

Chapter 6

Technology Transfer and Diffusion:Some International Comparisons

Compared to our strongest competitors, the UnitedStates is lacking in institutions to diffuse technologyto manufacturing companies. This is true in both thepublic and private sectors, and it applies especiallyto small companies. For example, scattered Federaland State efforts to help small U.S. firms raise theirtechnological level are no match for the densenationwide program of financial and technical assis-tance to smaller manufacturers in Japan. Not manymajor U.S. manufacturers give their suppliers tech-nical help, as Japanese firms customarily do. Nor isthere anything in this country to compare with theapprenticeship training taken by half the youngpeople of Germany and Sweden and credited withproducing a high level of technical skills in the workforce. In those countries, good worker skills are akey factor in the diffusion of manufacturing technol-ogy.

Large companies as well as small ones suffer fromfailures in technology transfer. With their typicallystandoffish relation to suppliers, large U.S. manu-facturing firms rarely get the benefit of collaborationwith suppliers on developing and applying newtechnologies —a common practice in Japan (see ch.5). Moreover, many U.S. firms are not as good astheir foreign competitors at scanning the outsideworld for new technologies that would improve theircompany’s products or manufacturing techniques.Often, the company culture is inimical to anythingNot Invented Here—the NIH syndrome.

There is one kind of technology transfer in whichAmerican companies do have an excellent trackrecord. That is in taking fundamentally new ideasout of the laboratory and using them as the basis fornew families of products. Whole industries havebeen founded on science-based inventions. Forexample, the transistor, an invention that dependedon accumulated knowledge in quantum mechanics

and solid-state physics, was the progenitor of thecomplex of microelectronic industries, includingsemiconductors and computers. In the same way,commercial biotechnology has risen on the founda-tion of scientific advances in molecular biology.

But U.S. firms are weaker at the more ordinarykind of technological advance in which improve-ments are added bit by bit to existing products andmanufacturing processes. Over the past quartercentury, Japanese manufacturers have repeatedlybeaten American producers with incremental prod-uct and process improvements-first in transistor-ized radios and TVs, then autos, now semiconduc-tors. 1 Some companies in Europe also excel at thiskind of evolutionary advance. For example, theGermans, with their mastery of mechanical engi-neering and metalworking, are leaders in makinghigh-quality industrial machinery.

The strengths and weaknesses of American firmsin adopting new technologies reflect our institu-tional biases. U.S. Government science and technol-ogy policy is light on technology diffusion andheavy on the traditional government missions ofdefense, health, and basic research.2 In the privatesector, there is plenty of venture capital to supportattempts to commercialize science-based innova-tions coming out of research labs, and there areplenty of footloose managers and engineers ready toshift to promising new ventures. Thus, public policysupports the kind of R&D that sometimes leads totechnological breakthroughs, and private institu-tions are suited to exploiting them commercially.3

What is lacking is a web of institutions to spreadthroughout manufacturing, to small as well as largefirms, the more mundane and more gradual improve-ments in technology that spell success in the laterphases of a product’s lifecycle.

IJapae~e success is not confimed t. improvements of fami]iar products. For example, although the video cassette recorder was a descend~t of tieU, S.-made Ampex commercial video tape recorder, it embodied so many new engineering ideas that it might be regarded as a new invention. And moreand more, the Japanese are putting efforts into scientific work as the basis for new technologies, m in high-temperature superconductivity.

ZFor a &scussion ~mp~ng “mission. oriented’ [echnology plicy (~ practiced in the united Slates, Britain, and France), ‘‘diffusion- oriented”tedmology policy (Germany, Sweden and Switzerland), and a combination of the two (Japan), see Henry Ergas, “Does Technology Policy Matter?’in Bruce R. Guile and Harvey Brooks (eds, ) Technology and Global Industry, Companies and Nations in the World Economy (Washington, DC: Nationalkademy Press, 1987).

sAt lemt, tie ~~itutions ~ suited t. supw~ing st~.up firms. However, high-tech st~-ups often f~ter in the transition to large-scale production.

–151–


Government technology policies in other coun-tries are much more strongly directed toward tech-nology diffusion than are U.S. policies. Japan’slong-established programs of general financial assis-tance to small firms and special measures toencourage small manufacturers to adopt moderntechnologies are of particular interest. Althoughthere are many differences between small manufac-turers in Japan and the United States, some featuresof the government programs that have worked wellthere might be translated into American terms.

Small firms have received special attention fromthe Japanese Government for several reasons. First,they are numerous. The Japanese economy as awhole is heavily weighted to small fins, and this istrue of manufacturing as well. Small and mediumsize firms account for three quarters of manufactur-ing employment in Japan, compared to a bit overone-third in the United States. Also, small Japanesefirms have often been technologically backward,paid low wages, and operated under primitiveworking conditions. Despite these disabilities, manysmall Japanese manufacturers have turned in re-markable performances-especially those that aresuppliers for Japan’s world champion industries(e.g., electronics, automobiles). Technology assis-tance given by the major customer firms has helpedthe performance of these supplier companies, but thegovernment’s technical and financial assistanceprograms get much of the credit too.

Many of Japan’s large firms have now entered theranks of the richest and most successful in the worldand no longer need much of the governmentassistance that helped them get established. More ofthe nation’s resources, public and private, areavailable to smaller fins. This chapter describes atsome length the extensive technical and financialprograms available to small Japanese firms today,keeping in mind their possible relevance to U.S.policy. The relatively sparse Federal and Statetechnical and financial assistance available to smallmanufacturers in the United States is described inchapter 2 and chapter 7.

DIFFUSION OF ADVANCEDMANUFACTURING EQUIPMENTOne measure of technological sophistication in

manufacturing is the presence of advanced equipment—such things as computerized machine tools, robots,flexible manufacturing cells. This is only one kind ofmeasure, and by no means a complete one. Otherfactors, especially the so-called soft technologiesinvolving organization of work and use of people,are at least as important as hardware to manufactur-ing performance. Nevertheless, an industry that fallsbehind the international competition in installingadvanced machinery will very likely find itselffalling behind in the cost, quality, and variety of itsproducts.

In the use of robots-defined as programmable,multifunctional manipulators—U.S. industries arefar behind the foreign competition, especially theJapanese.4 Although the invention and first use ofindustrial robotics was in the United States, it is nomore than a minor factor in American manufacturingtoday. Even in Japan, where robots have beenadopted far more aggressively, they are mostlyconfined to special uses in a few industries (mainlyautos and electronics). A much more broadly usedtechnology is numerically controlled and computernumerically controlled (NC and CNC) machines(also invented here). These machines are the kind ofcomputerized production equipment most com-monly found on manufacturing shop floors in theUnited States, West Germany, and Japan (and inother industrialized countries as well).

American manufacturers are closer to their topforeign competitors in the use of NC machines thanin robotics.5 However, Germany leads by a fairmargin, and the margin is wider if U.S. militaryproduction is omitted. The Japanese, who startedlater than American firms in adopting NC machinetools, were nearly even by the late 1980s and wereon a faster track. In a few years, unless thingschange, NC machine tools will be more common inJapanese factories than in American ones.

In 1988,41 percent of U.S. manufacturing estab-lishments with 20 or more employees in five major

gK~~e~ Fkmtm, ‘The Changing Pattern of Industrial Robot Use, ’ in Richard M. Cyert and David C. Mowery (eds.), The im@Ct o~TechnologicalChange on Err@ymettt and Econom”c Growth (Cambridge, MA: Balhnger Publishing Co., 1988); Edwin Mansfield, Department of Economics,University of Pennsylvania, “Technological Change in Robotics: Japan and the United States, ” Managerial and Decision Economics, Special Issue,spring 1989, Pp. 3-12.

s~ tie following discussion, the term NC includes CNC.

Chapter 6--Technology Transfer and Diffusion: Some International Comparisons ● 153

industry groups were using one or more NC ma-chines, according to a survey by the Bureau of theCensus. 6 The figure for German plants of the samesize in a similar group of industries is 48 percent.7

The Japanese data are shown on a different basis,that is, NC machine tools as a percent of all themachine tools in the shop. In 1987, 12.2 percent ofmachine tools in Japanese establishments with 50 ormore employees in metal machining industries wereNC. The comparable figure for U.S. establishmentsof the same size, at the same time, was 13.1 percent8

(tables 6-1 and 6-2). The seeming parity of U.S. andJapanese metalworking plants in ownership of NCmachine tools may be misleading, however, sincethe Japanese firms are acquiring the machinery at afaster rate. U.S. metalworking firms increased theirinstalled computerized automation (mostly machinetools) at an estimated rate of nearly 16 percent a yearfrom 1983 to 1988;9 the Japanese added NC machinetools at a rate of 24 percent per year from 1981 to1987.10

Another complicating factor in making thesecomparisons is U.S. military procurement. TheCensus Bureau’s survey of U.S. metalworkingestablishments found that plants producing for themilitary are more likely than the general run ofplants to use NC machines. Of all the plants in thesurvey, 41 percent used this automated machinery.

Table 6-l-Adoption Rates of NC Machine Tools inFive Major industries, United States (1988)

and West Germany (1986)

Size of establishment West Germany United States(number of employees) (percent) (percent)

Under 20 . . . . . . . . . . . . . . . . . . . . 15.820-29 . . . . . . . . . . . . . . . . . . . . . . . 36.0 35.9100-499 . . . . . . . . . . . . . . . . . . . . 55.9 50.0500 and over . . . . . . . . . . . . . . . . 87.3 69.8

20 and over . . . . . . . . . . . . . . . . . 47.7 41.4

NOTES: “Adoption hate” means the percentage of surveyed firms report-ing installation of at least one NC machine tool. “NC machinetools” include CNC machine tools.

The five major industry groups are SIC 34-36 (Fabricated MetalProducts, Industrial Machinery and Equipment, Electronic andOther Electric Machinery, Transportation Equipment, and instru-ments and Other Related Products) for the United States. TheGerman industry groups are similar although they may not beidentical.

n.a. = not available.

SOURCES: 14baf Germany Hans-Jurgen Evans, Carsten Becker, andMichael Fritsch, “The Effects of Computer-Aided Technologyin Industrial Enterprises: It’s the Content that Counts,” inRonald Schettkat and Michael Wagner (ads.), TechnkalChange and Ernp/oyn?ertt (New YoM, NY: de Gruyter, inpress), and Michael Fritsch, Technische Universitat Berlin,personal communication. United States: U.S. Department ofCommerce, Bureau of the Census, Manufacturing Technology1988, SMT (88)-1 (Washington, DC: Department of Com-merce, 1968), table 6-B.

For plants making products to military specifica-tions, the figure was 58 percent; for those making nomil spec products, only 36 percent reported usingNC machines. Similar discrepancies were reported

6u.s. ~p~ment of Commeme, Manufacturing Techno/ofl 1988, Current Industrial Reports, SMT (88)-1 (Washington, DC: U.S. GovernmentPrinting Office, 1989). The survey covered 10,526 establishments, selected to represent a total universe of 39,556 manufacturing establishments in SICMajor Groups 34, fabricated metal products; 35, industrial machinery and equipment; 36, electronic and other electric equipment; 37, transportationequipment; and 38, instruments and related products.

THam.J~gen Ewers, Carsten Becker, and Michael F~tsch> “The Effects of the Use of Computer-Aided Technology in Industrial Enterprises: It’sthe Context That Counts, ’ in Ronald Schettkat and Michael Wagner (eds.), Technical Change and Ernpioyrnenz (New York, NY: deGruyter, in press),and personal communication, Michael Frit.sch, Sept. 21, 1989.

SFor tie Unitd Shtes, daw me from tie 1987 National Survey Data about Machine Tool Use in Manufacturing Plants in Maryellen R. Kelley andHarvey Brooks, Modernizing U.S. A4an@acturing (Cambridge, MA: MIT Press, forthcoming), and personal communication, Maryellen Kelley, Sept.20,1989. The survey covered a representative sample of establishments of all sizes, including 1,368 metalworking plantsin21 industries. “Computerizedautomation’ in the study was defined to include programmable numerically controlled (NC) machine tools, which are controlled by tape and have beencommercially available for more than 20 years; computer numerically controlled (CNC) machine tools, which include a microprocessor and a keyboardat the machine, so that programs can be written and edited at the machine; and flexible manufacturing systems (FMSS), which consist of a number ofprogrammable machines (either NC or CNC) connected by automatic materials handling devices (e.g., conveyors or robots). At the time of the survey,38 percent of computerized machine tools in use were the older NC type.

For Japan, data are drawn from a survey covering establishments of 50 or more employees in metal machining industries, conducted every 6 yearsby the Ministry of International Trade and Industry (MITI). The MITI survey, like the Kelley-Brooks study, combines NC and CNC machine tools. Datafrom the two surveys are only roughly comparable, because the industries covered differ somewhat. The source for the data in English is D.H. Whittaker,“NC/CNC Penetration in Japanese Factories,” Appendix 1 to “New Technology in Small Japanese Enterprises: Government Assistance and PrivateInitiative,” contract report to the Office of Technology Assessment, May 1989. In Japanese, the source is Tsusansho, Showa 62 nen duinanakui kosu.kukikaisetsubito tokeichosu hokokusho (Report of the 7th Sumey on Machine Tool Installation) (Tokyo: Tsusan todei kyokai), Appendix 1, pp. 282-284.

gMqe]len R. Kelley and H~ey Brooks, The state of co~~erized Autom#ion in U.S. Mawfacturing, H~~d University, John F. KemedySchool of Government, October 1988, p. I-6. The average annual rate of adoption from 1968 to 1983 was 13.7 percent, with a slowdown in the years1973-78 (8.4 percent per year) and a speedup in 1978-83 (18,6 percent per year). Anderson Ashbum, “The Machine Tool Industxy: The CrumblingFoundation,” in Donald A. Hicks (cd.), /s New Technology Enough (Washington, DC: American Enterprise Institute for Public Policy Research, 1988),p. 55. Sources of the data are the loth through 13th American Machinist Inventories.

IOMI~ smeys feud mat Japane~ plans in met~ machining indus~es had 4,861 NQcNc machine t~]s in 19’73, 19,549 in 1981, and 70,465 in1987. Whittaker, op. cit. and D.H. Whittaker, ‘‘Machine Tool and NC Development in Japan, ” rnimco, n.d.

154 Making Things Better: Competing in Manufacturing

Table &2—Penetration Rates of NC MachineTools in Manufacturing Industries,

United States and Japan, 1987

Size of establishment United States(number of employees)

Japan(Percent) (Percent)

Under 50 . . . . . . . . . . . . . . . . . . . . 8.150-99 . . . . . . . . . . . . . . . . . . . . . . . 12.6 10.7100-299 . . . . . . . . . . . . . . . . . . . . 13.7 11.2300-499 . . . . . . . . . . . . . . . . . . . . 12.7 12.6500-999 . . . . . . . . . . . . . . . . . . . . 12.3 13.5Over 1,000 . . . . . . . . . . . . . . . . . . 13.8 12.8

Total over 50 . . . . . . . . . . . . . . . . 13.1 12.2

NOTES: ’’Penetration rate” means the ratio of NC machine tools to the totalnumber of machine tools installed in the establishments surveyed.“NC machine tools” includes CNC machine tools. The metalwork-ing industries surveyed are similar but not exactly the same in theUnited States and Japan.

For Japan, the category “other machine tools” was excluded Inthis table, because it was not included in the U.S. survey.

n.a. = not available.

SOURCES: Japan: Ministry of international Trade and Industry, Report ofthe 7th Survey on Machine Tool Installation (Showa 62 nendainanakai kosaku kikai setsubito tokei chosa hokokusho)(Tokyo: Tsusan todei kyokai), pp. 282-84; The source inEnglish is D.H. Whittaker, “NC/CNC Penetration in JapaneseFactories,” Appendix 1 to “New Technology in Small Japa-nese Enterprises: Government Assistance and Private initia-tive,” contractor report to the Office of Technology Assess-ment, May 1989.Url/ted Statea: 1987 National Survey Data about MachineTool Use in Manufacturing Plants; Maryellen R. Kelly andHarvey Brooks, Modernizing U.S. Manufacturing (Cambridge,MA: MIT Press, forthcoming), and Maryellen R. Kelly, Carnegie-Mellon University, personal commumeation.

by prime defense contractors and subcontractors fortheir military and non-military products (table 6-3).This means that NC machines are used in Americanplants much more for producing military goods thanfor commercial goods, and thus contribute less thanit might appear to the Nation’s trade performanceand competitiveness.

The differences in NC machine tool use in theUnited States, Germany, and Japan reflect differ-ences in government policy. The policy with mosteffect in the United States is satisfaction of militaryneeds. Numerical controls for machine tools wereinvented here in the 1940s, and MIT developed ahighly sophisticated version for the Air Force in the1950s. NC machining offered the great precisionthat was needed for making integrally stiffened wingskins for aircraft. The first substantial use of NCmachining, in the late 1950s, was in five-axis millingmachines that could hollow out the wing, leaving

stiffeners in place, and contour the outside skin to theairfoil shape—all in one piece from a solid thickplate of metal (an advance from the old method ofriveting the skin to ribs and stringers). The Air Forcebought the first 100 of these machines (after theaircraft industry refused to invest in them) and putthem in its contractors’ factories. 11 Around the sametime, other machine tool builders were developingsimpler, cheaper, more flexible machines, takingadvantage of the progress in NC controls.

Just as defense contracts were critical in develop-ing NC machining, military requirements have hada continuing effect on its diffusion. The U.S.Government has given little attention to specificpolicies that would promote adoption of NC technol-ogy outside the military-industrial complex. Anexception, perhaps, was the investment tax credit, ineffect off and on from 1962 to 1986, that allowedfirms to deduct from their income tax 7 to 10 percentof the price of any productive capital equipment,including machine tools. There is some evidencethat the investment tax credit may have encouragedorders for NC machine tools. ’z

Many people expected NC machine tools tosweep U.S. metalworking shops soon after theirinvention. They did not. Nevertheless, diffusion ofthese machines has not been slow by historicalstandards.13 Says Ashburn Anderson, an expert onthe machine tool industry, “It is not so much thattechnology diffuses more slowly in the United Statesthan in the pastas that it now diffuses more rapidlyin Japan." 14

Early on, the Japanese licensed NC technologyand within 10 years had adapted the Americaninvention into simple, cheap, and robust machines oftheir own design. Computerized controls (also a U.S.invention) were added in the 1970s, and Japanesefirms became the world’s premier producers ofsturdy, relatively inexpensive workhorse CNC ma-chine tools. The Japanese Government supportedthese efforts, contributing generous amounts toresearch and development consortia, and encourag-ing the thousands of small firms making machinetools to coalesce and specialize in different segments

11A. Anderson, Op. Cit., pp. 44~7.

12A. hder~n, op. cit., pp. 69-71.Issm, for exmple, ~win Mansfield> ‘‘The Diffusion of Industrial Robots in Japan and the United States, ” mimeo, n.d., which found that it took

the relatively short time of 5 years for half the major potential users to adopt NC machine tools.14A. Anderson, op. cit., p. 79.


of the market to achieve economies of scale. (Thisadvice was not always heeded; firms tended to staysmall, but they did specialize more. )15

At the same time, Japanese government policyactively supported widespread adoption of NCmachine tools. The government’s equipment leasingsystems bought machine tools and leased them atlow rates to small and medium-size manufacturers,thus providing both a stable market for machine toolbuilders and subsidies for machine tool users. Thegovernment also provided low-cost capital to aquasi-public leasing company that bought machinetools and leased them to companies of any size.Japan’s nationwide technology extension services(discussed below) have helped small firms learn touse the equipment effectively. In addition, Japanesetax law was changed in 1984 to allow very rapiddepreciation of investments in high-technologyequipment (including NC machine tools) by smalland medium- size firms. This seems to have set offa flurry of buying; one Japanese manufacturer callsit the “NC-ization period. ”

In Germany, emphasis in many industries onmedium batch production rather than mass produc-tion may account in part for high adoption rates ofNC machine tools (hard-wired automation is oftenmore efficient in mass production) but basically,both the production and use of NC machine toolsreflects Germany’s tradition, more than a centuryold, of excellence in vocational and technicaltraining. The German training system is supportedby both government and industry; it includes 3-yearapprenticeships from ages 16 to 19 for operators andfurther rigorous training, practical and theoretical,for the master craftsmen who become foremen andoften middle managers.

Production machinery is an important export forGermany, and that includes CNC machine tools atthe high end of the range. Germany’s dominance inproducing these complex and costly machines is duein large part to the quality of its workers. Thetraining system also pays off in the use of NCmachine tools. A study of matched metalworkingplants in Germany and Britain (described in Chapter4: Human Resources) found productivity two-thirdshigher in the German plants, with most of thedifference credited to training, especially of fore-

Table 6-3-Defense Production and Use of NCMachines in U.S. Manufacturing Establishments, 1988

Number of Percent usingestablishments NC machines

All establishments . . . . . . . . . .Products made to military

specificationsYes . . . . . . . . . . . . . . . . . . . . . . . .No . . . . . . . . . . . . . . . . . . . . . . . .Don’t knowa . . . . . . . . . . . . . . . . .Not specified . . . . . . . . . . . . . . . .

Prima defense contractorYes: percent of products

shipped to defense:1 to 25 percent . . . . . . . . . . . .26 to 75 percent . . . . . . . . . . .Over 75 percent . . . . . . . . . . . .Don’t knowb . . . . . . . . . . . . . . .

No . . . . . . . . . . . . . . . . . . . . . . . .Don’t knowc . . . . . . . . . . . . . . . . .Not specified . . . . . . . . . . . . . . . .

Subcontractor to defenseYes: percent of products

shipped to prime defensecontractor1 to 25 percent . . . . . . . . . . . .26 to 75 percent . . . . . . . . . . .Over 75 percent . . . . . . . . . . . .Don’t knowb . . . . . . . . . . . . . . .

No . . . . . . . . . . . . . . . . . . . . . . . .Don’t knowc . . . . . . . . . . . . . . . . .Not specified . . . . . . . . . . . . . . . .

39.556

14,58819,439

2,1413.388

10,0101,012

683601

22,8741,0283,349

11,5332,738

8801,83

12,9016,0703,605

41.4

58.136.138.4

1.5

51.762.561.237.341.242.0

2.0

53.767.167.444.332.942.0

4.1

NOTE: “NC machine tools” includes CNC machine tools.a “Don’t know" means the respondent didn’t know what percentage ofproducts are made to military specifications.

b “Don’t know” means the respondent didn’t know what percentage ofproducts in the plant are shipped to Federal defense agencies or to primecontractors of defense agencies.

C “Don’t know” means the respondent didn’t know whether any of the plant’sproducts are shipped to Federal defense agencies or to prime contractorsof defense agencies.

SOURCE: U.S. Department of Commerce, Bureau of the Census, Manu-facturing Technology 1988, Current Industrial Reports, SMT(88)-1 (Washington, DC: US Government Printing Office, 1989).

men. Computerized machinery worked far moresmoothly in the German plant, with little downtime.

To summarize: the U.S. Government policy withmost effect on both the invention and diffusion ofCNC machine tools has been concern to meetmilitary requirements. In Japan, the governmentsupported efforts by machine tool builders to makeincremental improvements in the known NC tech-nology, and it underwrote diffusion of the technol-ogy to machine tool users through subsidizedleasing, tax breaks, and technology extension serv-ices to smaller firms. In Germany, training was the

ISFOr a detailed accomt of the development of NC controllers and machine tools in Japan, see Ezxa Vogel, Comebuck (New York, NY: Simon ASchuster, 1985).

21-700 0 = 90 - 6


most important contribution the government madeto the production, diffusion, and effective use ofcomputerized equipment.

None of this means that government policies werethe only or most important factor in either thedevelopment or diffusion of NC machine tools inthese countries. A great deal depended on the privateactions and decisions of the companies and peopleinvolved. For example, Fanuc, under the direction ofDr. Seiuemon Inaba, has from the start combinedexcellence of product with exemplary manufactur-ing practice, in which the latest automated equip-ment is used to make reliable, inexpensive control-lers. (In Fanuc’s factory near Mount Fuji most of themachining and some of the assembly is done withoutoperators.) American NC machine tool buildershave been much slower to install the very kind ofequipment they make-a case of the shoemaker’schild, according to Anderson. Most important,Japanese designers applied microprocessors (anAmerican invention) to CNC controls in 1976, a full4 years before U.S. companies followed suit. That4-year lead was probably decisive in giving Japa-nese NC machines first place in the U.S. market. l6 In1988, half the NC machines sold here were made inJapan and, according to preliminary estimates, theJapanese share of the U.S. market rose to two-thirdsin 1989.

Finally, the point that hardware is only one part ofmanufacturing success bears repeating. For exam-ple, studies of auto assembly plants in Japan, NorthAmerica, and Europe for the International MotorVehicles Program found that automation and a“lean’ Japanese-style management system are each,separately, important factors in manufacturing per-formance. l7 But they contribute most to high pro-ductivity and high quality when they occur together.The best performing, world class companies (mostlyJapanese) first established a lean management sys-tem, and then improved their performance withhigher levels of automation. U.S. and Europeancompanies that automated first and then tried to

improve their management of people and organiza-tion of work had a harder time reaching topperformance.

LOOKING OUTSIDE THE FIRMFOR NEW TECHNOLOGIES

Incremental improvement of an existing productis part of the “cyclic development process” inmanufacturing. 18 It is engineering-dominated, com-pared to the science-dominated process of makingcommercial products from radically new technolo-gies bred in the laboratory. Despite its less dramaticcharacter, cyclic development is no less significantthan radical breakthroughs, for its cumulative effectscan be profound. For example, just 20 years ago,memory chips held 1,000 bits. The newest genera-tion of commercial chips are capable of holding 4million bits.

If U.S. manufacturing firms have fallen behindforeign competitors in pursuing cyclic development,one reason is their backwardness in exploitingtechnological advances that originate outside thecompany (the NIH syndrome). A well-known studycomparing R&D in a random sample of major firms,50 Japanese and 75 American, found that theJapanese companies spent less time and money thantheir U.S. counterparts in developing new productsand processes. 19 But the Japanese advantage layentirely in innovations based on external technol-ogy. For innovations based on technology developedinternally, U.S. companies performed as well as theJapanese. (The study did not attempt to assess whatopportunities these large U.S. firms might havemissed altogether because of their weakness inexploiting external technologies.)

The timing demands of the product developmentcycle suggest a possible reason for this seeminglyimpervious attitude. Ralph Gomory, former chiefscientist for IBM, explains it this way:

If you want to get new ideas into the developmentand manufacturing cycle from outside, timing is

16A. ~de~n, op. cit., p. 58.

ITJohn F. fiticik and JOhII Paul MacDuff’te, ‘Explaining High Performance Manufacturing: The International Automotive Assembly plant Study,’paper presented to the IMVP International Policy Forum, May 1989, available from International Motor Vehicles Program, Massachusetts Institute ofTechnology, Cambridge, MA. The authors described the lean production system as one ‘ ‘that nms ‘lean’ in its avoidance of pr~blem-hiding buffers andstays ‘fragile’ in its willingness to rely on a skilled, flexible, motivated workforce for problem-solving and continuous improvement. ”

l~his term and much of the following discussion is drawn from Ralph E. Gomory, ‘‘Reduction to Practice: The Development and ManufacturingCycle,” in National Academy of Sciences, National Academy of Engineering, Institute of Medicine, and National Research Council, industrial R&Dand U.S. Technological Leaders@ (Washington, DC: National Academy Press, 1988).

l~win Mansfield, “Industrial Innovation in Japan and the United States, ” science, Sept. 30, 1988.


crucial. . . . You must propose these ideas at thebeginning of the cycle. . . . Halfway through is toolate . . . no matter how good the proposal, Thecompany is not going to interrupt the cycle, delayingthe whole project by a year and thus ending up witha noncompetitive product.20

It may seem that this constraint should apply toJapanese as well as to U.S. manufacturers. If it doesnot, or does less, one reason is that major Japaneseindustries contrive to keep the product cycle shorter.Thus, the point at which new ideas can be pluggedin comes around faster. As noted in chapter 5, U.S.and European auto manufacturers typically take 63months from design to introduction of a new model,while Japanese producers, on average, take 42months-and use fewer engineering hours to do it.Likewise, Japanese electronics companies gained acritical advantage in the early 1980s when they gotthe 64K Dynamic Random Access Memory (DRAM)chip to market faster than most American producers;their success in taking the lion’s share of this marketearly was one of the factors that drove all but threeU.S. companies out of DRAM production. A shorterdevelopment cycle can be a particular advantage ina fast-moving field. The company that gets a productincorporating the latest technology to market soon-est reaps the reward of the innovator-even if it wasnot the source of the new technology and has nomonopoly.

Two of the major factors that enable leadingJapanese companies to cut short the developmentcycle and get new products to market fast are inbrief: 1) the supplier group system, in whichsubcontractors take on some of the design anddevelopment burden; and 2) frequent, close commu-nication between the product designers and manu-facturing engineers and rotation of people fromdesign to production. This second feature may bethought of as a form of technology diffusion itself,one that takes place within the company.

The time constraints of the development cyclemean that people inside the firm must be instigatorsin collecting new technologies from the outsideworld. They are the only ones who know the cycle

well enough to bring new ideas in at the right time.Government can help make this easier, by removingimpediments to the transfer of technology fromgovernment-supported labs, and universities canstructure cooperative research programs to meshwith industry needs.21 But the main task of bringingthe results of research to industry lies with acompany’s own engineers. Encouraging their engi-neers to attend professional meetings, read theliterature, keep in touch with research in governmentand university labs, and learn about their competi-tors’ products are necessary steps for companies thatmean to keep up with the competition. Most bigJapanese companies do it. So do many U.S. firms.Still, many U.S. firms regard outside activities forengineers as indulgences that might advance theengineer’s own professional career but are of littledirect benefit to the company.

Staying abreast of technology advance meanskeeping up with developments abroad as well as athome. In the past, U.S. manufacturers were goodcollectors of technical information from other coun-tries and good imitators of new products andprocesses invented elsewhere. They had to be. Onlyafter World War II did the United States become sopre-eminent in scientific research, and U.S. technol-ogy pre-war was by no means superior to that ofother countries. Yet our dominance in manufactur-ing was established early in the 20th century, whenthe majority of scientific discoveries and a greatmany technological advances based upon them werestill being made in Europe.22

In the postwar period, American industry hascontinued to adopt and develop commercial technol-ogies of foreign origin (e.g., the jet engine, polyesterfibers, the CAT scanner), but in some cases adoptionby U.S. producers has been years behind thecompetition (e.g., radial tires and anti-skid brakingsystems for automobiles). A special problem isinattention to technologies from Japan. As theJapanese concentrate more and more on leadingtechnology advances, rather than following andimproving on what others have done, Japan’simportance as a source of innovation is rising fast.

%omory, op. cit., p, 14.21S- ~h, 7 for a disc~~si~ of how R&~ ~sulN f~m ~Cder~ l~~rat~e~ and ~vefnment.sup~~ university research might be more effedvdy

transferred to private industry.ZZA.S e~ly ss tie 1880.s, U.S. man~act~ng had ~ready begun its rise to dominance, in part because the continental scale Of the market ~lowed U.S.

manufacturers to benefit from economies of scale and learning curve effects earlier than the Europeans. By the 1920s, the United States produced twiceas much steel and electricity per capita as Europe’s leading industrial powers, Britain, France, and Germany.


Interest in technology transfer from Japan to theUnited States is growing. Several public and privateprograms encourage U.S. scientists and engineers tolearn Japanese, work in Japanese labs, and follow theJapanese technical literature. But the results of theseprograms are still modest. Most of the technologyflow still runs the other way.23

TECHNOLOGY DIFFUSION TOSMALL FIRMS

Technological sophistication in small Americanmanufacturing firms runs the gamut. Firms at thefrontier of new technologies often start small; theSilicon Valley computer company that started insomebody’s garage is legendary. On the other hand,the ranks of small manufacturing firms are also filledwith shops that make humbler items. Significantly,small companies are suppliers of thousands of partsand components for major manufactured productsthat are leading items in the U.S. market and worldtrade (e.g., cars, computers, farm and factory ma-chinery, medical instruments). The cost, quality, andprompt delivery of these supplies are key factors inthe Nation’s manufacturing performance. The levelof technology in small American manufacturingfirms-in product design, production equipment,organization of work, training and use of workers—is highly uneven. But technological backwardness iscommon enough to be a real drag on U.S. competi-tiveness.

For many small companies, the bedrock oftechnological competence is having up-to-date pro-duction equipment. It is not always easy for smallfirms in the United States to decide what equipmentbest fits their needs, or how to use it efficiently .24Added to that are difficulties in financing; gettingfunds for the purchase of new equipment is usuallyharder for small fins, even creditworthy ones, thanfor larger ones, and it costs more.

More important than simple possession of ad-vanced equipment is an educated grasp of how to useit. For example, staff members of several State

industrial extension services report that small com-panies fairly often buy computerized equipmentwithout fully understanding the training that work-ers-and managers as well—need in order to use theequipment; then, they often do not know where toturn to get the training.25 Note also the studiesmentioned above that compared German and Britishmetalworking plants and found productivity muchhigher in the German factories. The difference wasnot in the age or sophistication of machines, whichwere much alike in both places, but in training.

As noted, NC machine tools are about as commonin U.S. metalworking plants as in Japanese; and inboth countries, small to medium-size plants (50 to500 employees) have about the same proportion ofNC machines as larger ones—11 to 13 percent of allthe machine tools used in the shop (table 6-2). Butin using the machinery effectively-especially inapplying the soft technologies that involve organiza-tion of work and use of people—small Japanesefirms seem to outperform American firms, at least inthe flagship industries that have led Japan’s eco-nomic growth and export success. An example is inthe motor vehicle industry. Many U.S. suppliers ofparts and components have not been able to meet thestandards demanded by Japanese-owned auto com-panies operating in the United States. The small tomid-size U.S. companies that have establishedthemselves as suppliers to the Japanese transplantshave usually required months or years of training inJapanese methods (mostly soft technologies) beforethey could match the cost, quality, and deliverytimes of their Japanese competitors.26

Further evidence of the importance of things otherthan hardware to the performance of small manufac-turing firms comes from Tokyo’s Ota Ward, famousfor its thousands of innovative small factories (ofabout 9,000 plants, 95 percent have 30 or feweremployees). Only about one-third of the metalwork-ing firms responding to a 1988 survey had even oneNC machine.27 Evidently, most of Ota-ku’s verysmall firms still rely more on their traditional

23Fm discllssion of progr~s to encourage technology transfer to the United States from Japan, see ch. 7.zQFor a description of some of the problems small companies face in getting advice from consdting firms, w ch. 7.25]n tie P=t few ~em, agowingn~~rof States have established progr~s to extend t~~ic~ ~sis~nce and information to sm~lermanufacturing

fins. OTA examined five of these programs in visits and interviews in 1988, as discussed in ch. 7. Findings from this examination tdso appear in PhilipShapira, “Industrial Extension: kuming from Experience, ” contractor report to the Office of Technology Assessment, November 1988.

%k tie brief account below of the training of North American suppliers for NUMMI, the Toyota----OM joint venture.2TOf464 ~etal macfin~g fires ~esPnding t. tie Sumey, 150 (32.3 ~rcent) s~d they had at least one NC/CNC machine too1. This was UP from 18

percent in 1981, 22 percent in 1983, and 29 percent in 1986. Whittaker, op. cit.


strengths of flexibility, quick response to customers’needs, and worker skills than on advanced equip-ment.

The situation in Japan seems to be changing.Traditionally, Japan’s smallest fins-especiallythose in sectors with no direct connection to theleading growth-and-export industries--have beenbackward. Many of the tiny “street-corner facto-ries’ in Japan are still quite primitive, with no heat,no indoor toilets, and only the simplest equipment.However, purchase and sales data collected by theJapan Machine Tool Builders’ Association (JMTBA)suggest that small plants have recently kept up withtheir bigger brothers in purchases of computerizedequipment. According to the MITI survey of estab-lishments with more than 50 employees, 32 percentof machine tools bought in the 3 years 1985-87 wereNC. In the same 3 years, the JMTBA figures showthat 35 percent of all the machine tools solddomestically to all sizes of firms (including thosewith fewer than 50 employees) were NC.28

Anecdotal evidence also indicates that a widerange of up-to-date equipment can now be found inmany small family-run factories in Japan. Forexample, one investigator who interviewed morethan 100 small automotive subcontractors in Japanin 1986 reported that many were heavily equippedwith advanced technologies, including NC ma-chines, laser machines, robots, and computer-aideddesign. He described several scenes like this one:

In one second-tier subcontractor of Isuzu I saweight NC lathes, of which four were fed by robots.The rest were minded by two skilled workers, twosemi-skilled workers and a part-time worker. Thefirm was being run by an entrepreneur whose wifewas working as receptionist, secretary, financemanager and “Jack of all trades. ” These were theentire personnel of the firm!29

The success of small and medium-size Japanesemanufacturing firms in the soft technologies andtheir recent rapid advances in installing up-to-dateequipment owe a great deal to a web of supportinginstitutions, public and private. These include thetransmittal of new technologies by major manufac-turers to suppliers and a broad range of governmentprograms for all small and medium-size manufactur-ers. These forms of technology transfer are uncom-mon, incomplete, or missing in the United States.

Major Companies and Their Suppliers

One of the many strong points of close, collabora-tive, long-term relations between lead manufactur-ers and their parts and components suppliers is thatthey favor transfer of technical know-how from thelead company down the supplier chain to medium-size and smaller companies.30 In Japan, majorcompanies often lend engineers and technicians totheir first tier suppliers to help them learn how to usenew equipment or arrange work more efficiently. Itis also quite common for parent companies toadvance funds to their subcontractors for operatingcosts+ specially in cases where the subcontractor’ssales to tile parent company are expanding, but thesubcontractor has to pay his own suppliers before hefinishes work on the product, delivers it to the parentcompany, and receives payment.31

Sometimes parent companies help suppliers ob-tain financing for capital investment as well, but thispractice is less common than in the past.32 Japantoday has so much investment capital that banks areaggressively looking for business among small andmid-size firms, since larger ones are able to meetmost of their capital needs from retained earnings.However, small companies applying for a bank loanoften find it is still a help if they are stable suppliersto a large, famous company.

Z8jaPa Ma~hine Tml Builders’ Association, Machine TOOI Industry, Jupan 1988 (Tokyo, The Association, 1988). The domestic sales fiWes arederived from figures on production, less exports, plus imports, omitting the category “Other Machine Tools”; they are in numbers of machine tools,not value. The JMTBA figures show that NC/CNC machines accounted for 36 percent of all Japanese domestic machine tool sales in 1985, 39 percentin 1986, and 30 percent in 1987.

Zwostihiro Nishiguchi, ‘‘ Competing Systems of Automotive Components Supply: An Examination of the Japanese ‘Clustered Control’ Model andthe ‘Alps’ Structure,” paper prepared for the International Motor Vehicles program (Cambridge, MA: Massachusetts Institute of Technology, May1987), p. 22.

S%= Ch. 5 for f@er discussion of how major Japanese manufacturers transfer technology to their Suppliers.31yoShit& KwoMwa, Jap~ne~ Development Bank and John F, Kennedy School of Governrnen[, Harvard University, ~rsonnal cOItllIlticatlon,

Sept. 7, 1989.3zToyota spkeSmen, for exmple, told OTA in 1989 mat financial aid plays no part in their C1O% KhtlOIIS With SUPPllerS; tiey Concenwate entirely

on technical advice.


Technical assistance remains a prominent featurein the relation of major Japanese firms to theirsuppliers. For example, Toyota’s principles are toselect good companies to begin with, communicatewith them often from the very beginning of therelationship, and give technical assistance as oftenas needed to help the suppliers meet Toyota’sunbending requirements for low cost, high quality,and prompt delivery. The suppliers must take anactive part in raising their own standards. They knowtheir problems better than anyone else, and must beinvolved in the solutions.

That the Toyota system of technology transfer tosuppliers is no fluke, but is characteristic of Japanesemanufacturers, was shown in a 1984 survey ofmanufacturing subcontractors, done by the Smalland Medium Size Enterprise Agency (chusho kigyocho) of MITI. Some 45 percent of respondents saidthey received technical assistance from a parentcompany, 37 percent received information, 28 per-cent were loaned or leased equipment, 24 percent gottraining for their employees, and 14 percent receivedfinancial assistance.33 Moreover, 39 percent ofrespondents said they introduced new technology atthe urging of parent companies (77 percent said thereason was to raise their technological level).34

In their survey of computerized automation inU.S. manufacturing, Kelley and Brooks found thatclose links between supplier firms and their custom-ers, of a kind that would help or spur the suppliers toadopt computerized machinery, were not common inAmerica.35 But in the infrequent cases where suchlinks existed, they made a difference. Only 3 percentof suppliers got any financial help from customers inbuying new equipment; just 9 percent reported thattheir customers requested or required the use ofcomputerized machinery. However, 20 percent ofsupplier firms said that customer firms had loanedengineering or programming s t a f f . T h i s k i n d o f

exchange was linked with a higher probability ofhaving at least one computer-controlled machine inthe supplier firm, suggesting that the loan oftechnical people from a customer firm to a supplieris an important conduit in the transfer of up-to-datetechnology. 36

As noted earlier, the joint Toyota-GM venture,New United Motor Manufacturing, Inc. (NUMMI),is an outstanding U.S. example of technologytransfer from a lead manufacturer to suppliers. After4 years of interaction with NUMMI engineers, NorthAmerican suppliers of parts and components for theautos assembled in NUMMI’s Fremont, CA plantwere able to match Japanese suppliers in cost,quality, and delivery time.37 The NUMMI caseexemplifies technology transfer not only from autoassembler to supplier, but also from Japan to theUnited States.

In Japan, the vertical transfer of technologysometimes develops to such a point that supplierstake over major functions formerly performed by thelead manufacturer. For example, both Toyota andNissan have totally delegated assembly of some oftheir cars to companies that were formerly suppliersof major components. This strategy (itaku seisan, orconsignment manufacture) enables the lead manu-facturer to concentrate on high-volume productionof a relatively small number of platforms,38 whilespinning off to its deputies the production of carsthat are low or fluctuating in volume. In the Toyotagroup, for instance, Kanto Auto Works producesthree different platforms on one assembly line;namely, the high-volume Corolla, the luxury passen-ger car Mark II, and the low-volume sports car MR2.Thus, Toyota exploits the economies of high-volume mass production in its home factory, whilepreserving the flexibility to make a varied range ofproducts in the factories of its consigned assem-biers. 39

JqWhit~er, op. cit., p. 23, citing Chusho kigyo cho (Small and Medium Size Enterprise Agency) cd., Chusho kigyo haku.rho (Sm whi~ PaPr)(Tbkyo: Okurasho inSittSU kyoku, 1985).

341bid.

35Kd]ey and Brooks, Thz State of Computerized Automaton (1988) op. cit.

%e probability of a supplier’s adopting computer-controlled machinery with no technical support from customers was estimated at 0.49; withcustomer-provided technical support, the probability rose to 0.58—about 20 percent higher.

JTIn r,hi.scW, much of the t~hnoloa transfemed was soft. Suppliers learned to apply Toyota’s lean production system, with its emphasis m t-work,training, and getting it right the fmt time, rather than relying on a cushion of big inventories of parts and work-in-process, to compensate for late deliveriesand poor quality.

38A ‘*pl~f~’9 ~femto ~1 cm p~uc~on &e -e wheelb~; one platform may include several different models-cars with different sh=t metalSkitlS and interiors.

39Nishi@chi, Op. cit., pp. 10-12.

Chapter 6-Technology Transfer and Diffusion: Some International Comparisons 161

Companies farther down the chain of supplierssometimes employ a similar strategy of first transfer-ring technology to the level below them, and thenturning over major tasks to their feeder fins. Notinfrequently, talented employees of small third orfourth tier companies leave to form their owncompanies, but they still maintain close ties withtheir former bosses, working for them as sub-subcontractors. The ex-employers consider this hiv-ing off natural, and often help out the new firm withtechnical assistance, sometimes even financing.40 Intheir view, skilled, enterprising workers are likely tobe more productive when working for themselvesthan when working for somebody else, especially ina small family-run firm where advancement possi-bilities are limited.

Nishiguchi offers the example of a subcontractorwho specialized in prototype manufacture for theelectrical, motor vehicle, and precision instrumentindustries. His strong suit was meeting short dead-lines; for this he could command premium prices. Hefurnished his own factory with a facsimile machineand such up-to-date equipment as CAD/CAM sys-tems, laser milling machines, and CNC machines,and he cross-trained his workers on several kinds ofequipment. Beyond this, he set up an ‘‘educationalfactory” nearby, where he trained selected workers,lent them money to buy machines and, after a yearor two of training, provided financing for them to setup their own businesses, attached to the mother firm.In 1986, when Nishiguchi interviewed him, this manhad a network of 62 subcontractors-all equippedwith advanced machinery-30 of whom had beenincubated at his firm. When he received a rush orderon his facsimile machine, he could spread the workout among his own employees and his subcontrac-tors, and often deliver the order within hours.41 The

result of such ties between patron companies andsuppliers is superior flexibility, combined withadvanced technology.

Japanese Government Programs forSmall and Medium-Size Firms

In Japan’s combined public-private support sys-tem for small and medium-size manufacturing fins,the government role is pervasive.42 Spending and

loans by the national government for help to allsmall business (including non-manufacturing) amountedto about 4.4 trillion yen in 1989, or $31.2 billion at140 yen to the dollar. Of this, only $1.4 billionappeared in the regular general account budget,which is supported directly by taxes. The rest, $29.8billion, was in the Fiscal Investment and LoanProgram, a capital budget often called the secondbudget, which derives its revenues from governmenttrust funds and the country’s huge, government-subsidized postal savings program.43 Altogether,spending for small business programs amounted in1989 to nearly 5 percent of the total regular andcapital budgets of the national government.44 Thissum does not include spending by prefectures, cities,and city wards, which also contribute handsomely toprograms for small businesses, matching the na-tional government’s contribution in some cases.45

Modernization of small firms has long been aconcern of the Japanese Government; some loanprograms targeted to small businesses date backmore than 20 years. Reasons for the focus on smalland medium enterprises (SMEs) are social andpolitical as well as economic. SMEs play a very bigpart in the Japanese economy. In 1986, in themanufacturing sector alone, SMEs (300 or fewerregular employees, and capitalized at 100 million

%en-ichi Irnai, Ikujiro Nonaka, and Hirotaka Takeuchi, ‘‘Managing the New Product Development Process: How Japanese Companies Learn andUnlearn,” in Kim B. Clark, Rotxxt H. Hayes, and Christopher Imrenz, The Uneasy Alliance: Managing the Productivi&Technology Dilemma (Boston,MA: Harvard Business School Press, 1985), pp. 365-366; also, Mari Sake, “Neither Markets nor Hierarchies: A Comparative Study of InformalNetworks in the Printed Circuit Board Industry,’ Ixcturer, Industrial Relations Department, Imndon School of Economics and Political Science, mimeo,May 1988.

41 NishiW~, Op. Cit., pp. 2s-24.

4~he material in this section is drawn mostly from D.H. Whittaker, “New Technology Acquisition in Small Japanese Enterprises: GovernmentAssistance and Private Initiative,’ contract report to the Office of Technology Assessment, May 1989; and from OTA interviews in Japan in March 1989.Yoshitaka Kurosawa, on leave to Harvard University from the Japanese Development Bank, contributed additional information in a letter to Julie FoxGorte, OTA project Director, dated Sept. 7, 1989.

4~e m~ ~overment subsidy fa ~~~ ~vings is in the form of a ~ exemption for interest. A]so, dfing the many years that Japanese financialinstitutions were strictly regulated, the interest rate on postal savings was higher than for time deposits elsewhere.

~In @aI yea 1989, tie toud budget of the Japanese national government WSS92.7 trillion yen ($662 billion), including 60.4 tilllOIt yen in tie Wnerd

account, and 32.3 trillion yen in the Fiscal Investment and Loan Program. Japan Economic Institute, JE1 Report, May 12, 1989.A5For Cxmpje, in fisc~ year 1988, tie pref~t~es match~ fie nation~ government’s provision of 2 bi]lion yen ($154 million) fOr the @ipmCtlt

Modernization Loan System and the Equipment basing System for smaller enterprises.


yen or less) represented 99.5 percent of establish-ments, 74.4 percent of employees, and 56.5 percentof value added.46 At the same time, wages in thesesmall manufacturing firms are at least 25 percentlower than in the major companies, working condi-tions are frequently dismal, and technologies haveoften lagged behind the leaders. Besides thesereasons for government concern, there is the politi-cal fact that small business has been a steadfast,strong supporter of the ruling Liberal DemocraticParty. Every election brings new pledges of meas-ures to improve the climate for small business.

The Japanese national programs for SMEs includeboth financial and technical assistance, and the twoare intertwined. In the 1980s, special attention hasbeen given to programs aimed to help small businessadopt high-tech equipment such as computerizedmachinery and robots. Some key assistance pro-grams that encourage purchase of advanced equip-ment are open only to still smaller firms, with nomore than 20 to 100 employees.47

Among the multiple services the governmentoffers SMEs are a big program of direct loans foroperating funds or plant and equipment investmentand a still bigger program of government guaranteedloans. Other services include: a system to lease newequipment to SMEs on generous terms or sell it onthe installment plan; loans to groups or cooperativesof SMEs; management analysis for individual firms—a condition for government loan approval; publictesting and research centers, where SMEs can useexpensive equipment for a nominal fee and canconsult with engineers on technical problems. SMEsalso get tax breaks for investment in new equipment,especially high-tech equipment. For example, a1984 law allows SMEs the option of taking a specialfirst year depreciation of 30 percent for investmentsin electronic and ‘‘mechatronic’ technology, whichincludes NC machine tools, computers, and robots.

The national government, mainly through MITIand the Ministry of Finance, is the grand overseer ofthe SME programs and is the top provider of funds.The actual dealings with business people fall to theprefectural and local governments, and to quasi-public organizations such as chambers of commerce(in cities, or “societies of commerce and industry”in towns and rural places) and federations of smallbusiness associations.

In 1987, loans to SMEs via the three maingovernment financing institutions amounted to 3.8trillion yen, or $27 billion.48 Japanese loan guaranteeprograms for SMEs are still larger. The 52 nation-wide credit guarantee associations underwrote 7.8trillion yen ($56 billion) in loans to SMEs in 1987.By way of comparison, U.S. small businesses (up to500 employees) got $47.3 million in direct loansfrom the Small Business Administration in fiscalyear 1989, and loans were restricted to specialdisadvantaged groups. Federally guaranteed loansare available more generally to U.S. small busi-nesses; they amounted to $3.6 billion in 1989. Thesefigures are only illustrative; they do not include, foreither country, financial aid available from State (orprefectural) and local governments. And, to put thecomparison in perspective, small businesses play abigger part in Japan than in the United States. Evenconsidering the larger size of the U.S. economy,small and medium-size manufacturing firms aremore numerous and employ more people (10.7million v. 6.8 million) in Japan than in the UnitedStates. Finally, keep in mind that these figures forgovernment loans and loan guarantees are for allsmall businesses in both countries, not just formanufacturing firms. With all this, it is still notablethat the Japanese Government provides about 20times more financial aid to small business than theU.S. Government does.

Even so, government financing is not as importantto Japan’s SMEs as it was just a few years ago. (Box

46BY ~omp~Wn, in tie United Sutes in 19g6, sm~l businesses (enterprises with fewer than 500 employees) represented 85 Per~entof m~ufacturingestablishments, 35 percent of employment, and 21 percent of value added. An establishment is a single physical location where business is conducted.An enterprise is a business organization consisting of one or more establishments under the same ownership or control. The State of Srnutl Business:A Report of the President transmitted to the Congress, 1989 (Washington, DC: U.S. Government Printing Office, 1989), p. 21; table 13, p. 21; table A. 15,pp. 80-81; table A.20, pp. 92-93.

47Jap~es~ firms with fewer than 100 employees constitute 97 percent of establishments, 55 percent of employment, and 39 percent of value add~in private manufacturing in Japan; comparable figures for firms with fewer than 20 employees are 87 percent of establishments, 29 percent of employees,and 15 percent of value added.

48ThiS fi~ is net of repayments; it includes 1.80 ~l]lon yen from tie chu~o kigyo kinyu (sm~l Business Finance corporation), 1.85 trillion yenfrom the kokumin kinyu koko (Peoples’ Finance Corporation) and 128 billion yen from the shoko chukin (which is not always included in the groupof government financial institutions because it raises part of its funds from association members). The gross amount of loans made to SMES in 1987by these three institutions was 5.6 trillion yen—2.26 trillion, 2.89 trillion, and 493 billion yen respectively.

Chapter 6---Technology Transfer and Diffusion: Some International Comparisons ● 163

6-A offers a Yokohama factory owner’s account ofhis “graduation” from government financing andtechnology transfer programs over the years.) In thelate 1980s, with the quick recovery from the rise ofthe yen and the great prosperity that followed, Japanwas awash in capital. The august city banks, whichonce gave most of their attention and funds to largecompanies, were now scrambling to do businesswith SMEs. In March 1981, for example, 25 percentof city bank loans went to SMEs, but by August 1988the figure was 64 percent. Even though the govern-ment loans are usually pegged at lower rates-e. g.,4 percent instead of 5 percent to individual firms in1989, as low as 2.7 percent when provided throughcooperative associations, and zero for the govern-ment’s half share of certain equipment moderniza-tion loans-companies often prefer the greatersimplicity of dealing with a bank.

Government loans are still an essential source offinancing for small startup companies with no trackrecord, for firms changing direction, and as a safetynet in times of adversity. For example, many of the9,000-odd small manufacturers in the Ota ward ofTokyo were hard hit by the yen’s rise in 1986-87. Inthose 2 years, Ota-ku’s firms borrowed 1.5 billionyen ($11.5 million) in emergency loans to coveroperating costs. But the overall trend in the late1980s was for private loans to edge out governmentfinancing. Government loans dropped from 13percent of all outstanding loans to SMEs in 1980 to9 percent in 1988. These figures understate thegovernment role in financing of SMEs, however,because they omit the system of loan guarantees.And despite the decline of Japanese Governmentfinancing for SMEs, the volume remains huge inU.S. terms.

Besides its big, general program of direct loansavailable to all SMEs, the Japanese national govern-ment offers a whole menu of SME ‘‘measures,’funded at about 225 billion yen ($1.6 billion) in1987. Among these are two special programsdesigned specifically to encourage SMEs to acquiremodern technology. One of these, the EquipmentModernization Loan System, made 6,000 loans in1987, totaling 41 billion yen ($293 million) in 1987.The program is open only to firms with 100 or fewer

employees, as shown in table 6-4. It provides up tohalf the amount of the funds needed for themodernization project; notably, that half is interestfree. According to officials of MITI’s Small andMedium Enterprise Agency, no collateral is requiredfor these government loans because commercialbanks can provide loans requiring collateral.49

The Equipment Leasing System, through whichfirms can lease new equipment or buy it on theinstallment plan, is another key technology-promoting measure. Nothing better illustrates theJapanese policy of fusing financial assistance withpromotion of technological advance than this pro-gram. Founded in 1966 and open only to firms with20 or fewer employees, its direct purpose is to helpsmall, struggling companies invest in new equip-ment at affordable terms (easier terms than thoseoffered by private leasing companies, and easiereven than the Equipment Modernization Loan Sys-tem). The system has the added effect of providinga quite substantial, assured market for producers ofcapital equipment suitable for small shops, espe-cially machine tools. A high-tech equipment andmachinery leasing system, added in 1986, is open tofirms with as many as 80 employees, giving addedsupport to the market for such things as NC machinetools, robots, and computers. In 1987, about 4,500leases or installment purchases, amounting to 49billion yen ($350 million) were made under thisprogram. About one-third of the loans and leaseswent to SMEs producing machinery and other metalgoods, mostly for buying or leasing NC machines .50

In this connection, it should be noted that thegovernment is also a partner in quasi-private leasingcompanies that serve large as well as small compa-nies. For example, the Japan Electric ComputerCorporation (JECC), founded in 1961 to buy com-puters and lease them to users at subsidized rates, gothalf its capital from the Japan Development Bank, agovernment institution. The similar Japan RobotLeasing Company (JAROL) was founded in 1979,with 60 percent of its capital coming from the JapanDevelopment Bank. In addition, in 1980 the SmallBusiness Finance Corporation allocated funds spe-cifically for loans to small businesses buyingrobots. 51 The existence of these leasing and loan

@C)TA intemiew with Kaz~~o Bando and Kazumi Suda, Small and Medium Enterprise Agency (ChUShO kigy~ cho), MIT1, Ma. 16, 1989.sqn Tokyo, 37 ~rcent of~e loans m~e ~der the @uipment Modernization program in 1987 were for buying CNC mactines. (Tokyo Metfopolitan

Governrnent Labour Economics Office, untitled mimeo, 1989, cited in D.H. Whittaker, op. cit.)slEma Vogel, Comeback (New York, NY: Simon & Schuster, 1985), pp. 90, 122-123.


Box 6-A—A Small Plant in Yokohama

Showa Precision Tools Co., Ltd., of Yokohama, Japan makes plastic processing dies, blanking dies,progressive dies, and measuring and testing equipment.] The company’s name is well chosen. Everything about itsnewly built factory in Kanazawa Industrial Park speaks of precision, from the understated architecture of the frontoffice to the neatly pressed company uniforms worn by the company president and founder, Mr. Masanari Kida,and his chief engineer, Mr. Y. Yokoyama. Showa tools are esteemed for their quality and design. Because of thatreputation, the company is prospering. The first sentence in the company outline booklet says, “We are enjoyinga convenient life, thanks to the tools and machinery which have been developed. ”

Although Showa provides all its own capital now, Mr. Kida is well acquainted with Japanese Governmentprograms that offer financing for small and medium enterprises. Showa made frequent use of them from the timeit was founded 30 years ago until about 10 years ago. Even more recently, when Showa built a new factory inKanazawa Industrial Park, government financing filled a gap. Mr. Kida had the proceeds from the sale of his oldfactory and a substantial loan from a bank, but was still short of what he needed for new machinery. Financing fromthe government’s small and medium enterprise program made up the difference.

Although government financing is cheaper than a bank loan—the difference is a percentage point or so, orabout 4 percent instead of 5 percent-going through government programs is a hassle, Mr. Kida said. “If I go tothe bank, I can get the money today,” he explained. “If I borrow from a government program, it takes a month,and I have to fill out a lot of forms. This hassle is still worth it, he believes, for brand new businesses that haveno track record or an established relationship with a bank. Indeed, government financing was essential for Showain its earlier years.

One part of the government program is still useful to Showa———technical advice. When Mr. Kida last usedgovernment financing, advisors from the guidance center in the Yokohama city office gave him an analysis of hisfinancial arrangements. At his request, an advisor also evaluated some of his plans for new machine purchases. Hisrelationship with that advisor has lasted to the present day through the city’s yearly management service, whichprovides technical information and evaluation to small and medium-size firms. In return, the advisor uses theinformation he gets about the firm to enlarge his understanding of technology use and other conditions of smallbusinesses. The service also gives Mr. Kida general information on what his competitors are doing.

Firms like Showa can also get some training from the Yokohama city office. On request, the office will senda sensei (teacher, or master) to train the employees total quality control techniques. This training is fairly extensive.Between June and October 1988, the sensei came to Showa for eight 2-hour sessions to train 14 group leaders (theseare quality circle group leaders, not necessarily the formal authority figures). The sensei brings written materialsto every class, and then the group leaders are responsible for teaching the other people in the-group. The lessonswere:

●

●

●

●

●

●

●

●

What Are Small Group Activities?Why Are Group Activities Necessary?Small Group Activities and Total Quality ControlWhat Is Quality?How To Introduce Small Group ActivitiesLet’s Master Quality Control MethodsThe Way of LeadershipHow To Succeed in Small Group Activities

The lessons do not accomplish miracles. Although the classes may get the groupworkers are not always so enthusiastic. However, the group leaders do impart to others in

leaders ail fired up, otherthe group what they learn,

and eventually the lessons of Total Quality Control are learned by all. Mr. Kida did not think the services offeredby Yokohama prefecture were unique. He admitted that Yokohama and Kanagawa were more positive about suchactivities than other prefectures-but only a bit more.

Iinformar,im for this box comes from interviews conducted by OTA sttif in Yokoh~a, m~h 1989.

Chapter 6-Technology Transfer and Diffusion: Some International Comparisons ● 165

Despite its present independence from government financing, Showa is still part of a government-supportedcooperative association for small companies. Members can get up to 65 percent of their investment costs from thesmall and medium enterprise public corporation, at 2.7 percent interest. The maximum term of such loans is 15years, and the money is provided for additions to plant and equipment. The preferential financing is a strongincentive to join a cooperative association. There is also a down side to joining. Money borrowed as a group hasto be repaid as a group, so if one member fails or gets into trouble, all the other members are responsible for hisdebts and his recovery. Also, the land belongs to the group, and every inch of the precious stuff is used. So, if acompany wants to expand, it can do so only if someone else in the group goes under and their land becomesavailable, and even then approval of the group is needed. Others may want to expand, too.

In response to questions about the drive to innovate in small fins, Mr. Kida’s unhesitating answer wascompetition. ‘You must innovate or you get beaten,’ he said. Since 1986, Showa has bought 11 new NC machines.and now about 70 percent of all his machines are NC. He never leases the machines, on principle, because leasingcosts a bit more than buying. However, companies that can’t secure the capital up front need to be able to leasemachines. Like government financing programs, leasing is a nice option for fledgling companies.

Mr. Kida has an extra incentive to be right at the cutting edge of new technology. His business is an independentone, not in anyone’s supplier group or keiretsu. Companies in cooperative associations tend toward being moreindependent, according to Mr. Kida. Many firms would like to be on their own, but it is harder than beingsomebody’s supplier. ‘‘If you want to be independent, you have to study unceasingly,’ he says. He gets no technicaladvice from his customers, although engineers do come from customer companies to discuss their technicalrequirements. He has never gotten any financial assistance from a customer, either. But even in companies that arein a supplier group, the parent companies are giving less advice and less financing than they used to, perhaps becauseit isn’t necessary, and perhaps&cause of other changes in the environment of large companies—--moving offshore,for example.

Finally, Mr. Kida was asked why he didn’t just sell up. “You could be a millionaire, and live anywhere youwanted,” said the interviewer. “You could buy a ramen (noodle) shop, and stop the struggle. ” Mr. Kida seemedspeechless at the thought, so Mr. Yokoyama, the chief engineer, answered. He was horrified at the suggestion. “Wehave 100 employees here,” he said earnestly, “and they have families. That’s 400 people. We’re responsible forthose people. What would they do if the owner bought a ramen shop? Where would they go? No, we have to stayin business. Four hundred people depend on this business. ’

Table &4-Japanese Government Equipment Modernization Loan and Equipment Leasing Systemsfor Small and Medium Enterprises

.— -.Equipment leasing system

Equipment leasing (installment plan) Equipment Ieasing

Equipment modernization High-tech, information High-tech, informationloans system General equipment processing equipment processing equipment

Main recipients . . . . . . . . . Small and medium Small and mediumenterprises with 100 or enterprises with up toless employees 20 employees

Maximum amount ofloan or value ofleased equipment . . . . Half of funds required Equipment worth up to

up to 30 million yen 25 million yen

Interest or charge . . . . . . . Free 4.5% of the cost ofequipment as per annumcharge (an additional1 0°/0 guarantee money isrequired)

Period, ... , . . . . . . . . . . . . 5 years with l-year grace 4 years and 6 months (11

Small and mediumenterprises with up to80 employees

Equipment worth up to50 million yen

4.5% of the cost ofequipment as per annumcharge (an additional10% guarantee money isrequired)

6 years and 6 months(11period years and 6 months for anti-years and 6 months for anti-

pollution equipment) pollution equipment)

Small and mediumenterprises with up to80 employees

Equipment worth up to50 million yen

About 7°% as per annumcharge (including tax andinsurance premium)

Up to 7 years (84 months)

SOURCE: Mmstry of International Trade and Industry, Small and Medium Enterprise Agency (chusho klgyo cho) SMEA mimeograph, 198 (untitled)


programs assured equipment producers of a solidmarket, which probably encouraged them to gear upfor expanded output-even though, as it turned out,not all the programs were heavily used. For example,purchases of robots by Japanese firms turned out tobe so great that JAROL leased only 790 units in1982, when shipments were almost 10,000.52

Still other national government “special meas-ures” are designed to help bring SMEs up to speedtechnologically. According to MITI officials, theSME programs were originally formed with the viewthat small companies needed information more thanfinancing. To get public financing under someprograms, firms must have a management analysis,paid for by government funds and provided free bylocal governments, associations of commerce, orfederations of small business associations. Often theanalyses focus on finance, sales, and marketing, butadvice on technology and production methods isalso given. As illustrated by Mr. Kida’s experience(box 6-A), small businessmen may form a lastingrelationship with the person who does the originalanalysis for the loan, often coming back repeatedly forconsultation on technical or other business matters.

Public testing and research centers also play a bigpart in technology diffusion, Japan had 185 of thesecenters in 1985, with 7,000 employees and an annualbudget of 66 billion yen (about $470 million), halffrom the national budget and half from the prefec-tures. SMEs can come here and, for a small fee, useinspection equipment that is too costly or used tooseldom to make purchase worthwhile. They can alsofind consulting engineers for research and advice onspecial problems, and they can bring the consultantsto their own factories if necessary.

Local technology demonstration centers supple-ment the national testing and research centers. Theindustrial hall in Tokyo’s Ota ward is a goodexample. Advisors at the hall have regular consulta-tion hours for the Ota-ku’s thousands of tinybusinesses—Tuesday, Thursday, and Saturday, from10 to 4, on mechanical matters, and the alternateweekdays on electrical matters. The hall has about

500 consultation meetings a year in seven areas. Inorder of popularity, they are: machines, measuringdevices, materials, machining process, electricalproblems, controllers, and a miscellaneous categoryincluding legal problems. An example of an electri-cal problem: there are frequent, unpredictable dailyfluctuations in the voltage delivered by the city.Small businesses need to learn how to cope with thefluctuations and how to make machinery last in spiteof them. According to the managers of the hall, smallfirms could figure out many of these problemsthemselves, but they don’t have time.

Besides these regular consultations at the hall,which are free, firms may ask advisors to visit theirplants for a fee of 10,000 to 20,000 yen a day (about$70 to $140). For knottier problems, firms may bereferred to the Technology Experimental Center inmetropolitan Tokyo, which has about 160 highlyqualified consultants—30 in technical fields—and200 technical advisors (this is one of the 185 nationalpublic testing and research centers). Another servicethe Ota industrial hall offers is use of specializedmeasuring and calibrating machines, at a fee ofabout $4 for half a day. In addition to all this, the hallputs on exhibitions three times a year showingmachines made in the wards to buyers in the area.Sometimes buyers from other countries are invitedas well. Occasionally, the prefecture exhibits Ota-made machinery at shows in other places.

According to surveys of small businesses, publicprograms rank low on the list behind parent compa-nies and machine and equipment makers as sourcesof technical information. This is no reflection on thepublic programs; services like those at Ota-ku’sindustrial hall are used by SMEs and seem to be wellregarded. 53 It is more an indication that the level oftechnology diffusion to SMEs in Japan, includingthe active role taken by parent companies, isextremely high. The role of parent companies maybe diminishing a bit, however, as the bonds betweenparent companies and subcontractors are weakeningsomewhat, The reasons are first, that major firms aredoing more subcontracting offshore; and second,that small supplier firms, more prosperous than ever

S2Kenne~ Fl~, “changing Pattern of Industrial Robol USe, “ in Richard M, Cyert and David C, Mowcty (eds,), The Impact of TechnologicalChunge on Employment and Economic Growth (Cambridge, W: Ballinger Publishing Co., 1988), p. 299. According to Vogel (op. cit.) virtually norobots were exported from Japan in the early 1980s because domestic demand was so great.

sqThe small b~iness owners interviewed by OTA staff in Japan spoke favorably about government technical assistance. tic, who stid he gener~lyprefers his own resources to government programs (though he had taken a large government loan to finance a new building for his factory), had noresemation in praising the Tokyo technology center. He goes (here about once a month for testing of materials and inspection services. The service ischeap-about 3,000 yen per visit—and the consulting engineer is very kindly and knowledgeable (’‘a good study person’ ‘),


before, are able to be more independent. In Ota-ku,officials say, about 1,000 of the 9,000 manufacturingfirms are now independent, with no strong ties tomajor firms. Many of these companies can makegood use of public technical assistance. The govern-ment is encouraging small independent firms toform cooperative associations, to work together onR&D and share technical, management, and market-ing information among themselves (see the discus-sion below on horizontal links between small firms).

Government programs offer specific help tostartups, in addition to loans. An example ofpublic-private partnership to encourage high-techstartups is the Kanagawa Science Park (near Yoko-hama). Building began in 1989. When completed, itwill provide common research facilities, includingprecision measuring and calibrating equipment, plusthe usual business incubator services such as ac-counting and payroll. It is intended to be a communi-cations center as well, the hub of an electronicinformation network that will extend to manybusinesses in the prefecture. Finally, there are plansto make the Science Park an international con-vention center--complete with hotels, banks, andrestaurants-designed especially to serve residentcompanies. The Science Park is set up as a stockcompany, with construction and initial subscriptionsfinanced by funds from the Yokohama Bank and theKanagawa prefecture. Other prefectures are plan-ning similar schemes, but Kanagawa is the first totake action.

To sum up, financing new technologies seems tobe no big problem for Japanese SMEs, and theabundance of government assistance is surely onereason. Where small U.S. firms may find theavailability of capital a real barrier to investing inmodern equipment (e.g., a CNC machine tool), theirJapanese competitors can turn their attention towhether the equipment precisely fits their needs,whether it is better to buy it or lease it, and whethergetting a 4 percent loan from the government ratherthan a 5 percent loan from the bank is worth thebother of waiting a month instead of a day. In

addition, technical assistance is very broadly availa-ble from many sources, often linked with some kindof financial assistance. Small manufacturers in theUnited States are not nearly so richly supplied withguidance in adopting and using new technologies.54

Horizontal Links Between Small Firms

Another way to promote the widespread adoptionof advanced technology, down to the level of tinyfamily-run firms, is through horizontal networks thatgive member firms help in developing and acquiringnew technologies, and advice on financing, manage-ment, and marketing as well. Such systems areprominent in the textile and metalworking industriesof both Japan and Italy. They can be found elsewheretoo, as in Denmark’s textile and furniture industries.These networks involve a considerable degree ofcooperation and information-sharing among com-petitive fins-practices that are quite foreign to U.S.business tradition. In some countries, the networks aresupported by a range of government programs that aremostly missing in the United States.

A well-known example of horizontal links amongsmall firms is in the northeast-central part of Italy,known as the Third Italy .55 Networks of small,technologically sophisticated textile and metalwork-ing firms began to develop in this region in the late1960s. By the early 1980s, these small enterpriseswere supporting a prosperous economy. In Emilia-Romagna, for instance, manufacturing wage rates in1980 were 125 percent of the Italian average. In1985, the region ranked second among Italy’s 21regions in per capita income, having risen from 17thin the 15 years since 1970.

The cooperative networks that were key factors inthe region’s economic success were founded withthe help of local governments, but later on werelargely financed and operated by the firms them-selves. Artisans’ trade associations, technicalschools and universities, and labor unions have alsosupported the networks’ programs. The networksprovide technical advice on new equipment, prod-ucts, and processes; financial help in acquiring new

54s= tie discussion in ch. 7.SsThe many writings on cooperative networks in the Third Italy include Giacomo Becattini, “ The Development of Light Industry in Tbscany: An

Interpretation, ’ Economic Notes, vol. 3, 1978; Sebastian Brusco, “The Emilian Model: Productive Decentralization and Social Integration,”Cambridge Journal of Economics, vol. 6, No. 2, 1982; Michael Piore and Charles Sabel, The Second Industrial Divide (New York, NY: Basic Books,1984); Edward Goodman, Julia Bamford, and Peter Saynor (eds.), Srntdl Firms and Industrial Districts in ltaly (Imndon and New York: Routledge,1989); Daniella Mazzonia and Mario Pianta, ‘‘An Innovation Strategy for Traditional Industries: Experience of the Italian Textile Districts of Prato andComo,’ mimeo, September 1986; Robert E. Friedman, “Flexible Manufacturing Networks, ’ and Richard C. Hatch, “Uxuming From Italy’s IndustrialRenaissance,” in Corporation for Enterprise Development, Entrepreneurial Economy, July-August 1987.


machinery and training in using it; business servicessuch as making up payrolls and sending out bills;and advice on markets and assistance in parcelingout work on large orders. Local governments,together with the artisans’ trade associations, havealso developed industrial parks where factory spaceis offered at reasonable, stable rents. The concentra-tion of small firms in the same area carries an addedbonus, making it easier for the firms to divide uplarge contracts or find subcontractors if they getjammed with too much work at one time.

A notable feature in the small firms that makeupthese manufacturing networks is their use of ad-vanced equipment. Part of the reason lies in thenature of the industries--cloth and clothing, shoes,furniture, metal parts for machinery or precisioninstruments. The investment needed for an efficientunit of production in such industries is not formida-bly high. A cluster of CNC machine tools orelectronic sewing machines or weaving machines isnot beyond the financial means of a family-runenterprise-especially when help in arranging fi-nancing is available to the small firm, as it is in thispart of Italy. Loan guarantee cooperatives (estab-lished by the trade associations) may arrange prefer-ential bank financing for buying the equipment;alternatively, members of artisans’ trade associationcan lease machinery. Not only is the equipmentaffordable but objective advice on what to buy andconsultation on using it is also available fromService Centers serving specific industries (organ-ized by trade associations together with local gov-ernments, labor unions and other business groups).

Government support of the networks is mostlyconfined to the regional and municipal levels. Thenational government has had little to do with it. Thedistinctively Italian Eurocommunist government ofEmilia-Romagna was the pioneer, but rightist re-gional governments, such as the Christian Demo-cratic one in the Veneto, have also lent their support.As noted, the major contributions from the regionalgovernment were made at the beginning, in the formof financial and planning support for starting upnetworks.

Whether these largely voluntary horizontal net-works are sturdy enough to last through changingeconomic conditions is an emerging question. Verti-cal as well as horizontal networks have always beena part of the scene in the Third Italy; many smallfirms are regular subcontractors for big enterprises(e.g., Benetton in apparel). However, the presence ofstrong horizontal networks has probably given smallfirms an extra measure of independence and bargain-ing power. Today there may be a trend towardgreater dominance by lead firms. A recent study ofthe textile districts of Prato (in Tuscany) and Treviso(in Veneto) and the food-producing machinerysector in Emilia-Romagna found increasing top-down control.56 The pattern is for small firms tocontinue decentralized production, but under thegrowing financial and strategic control (includingthe choice of technology and subcontractors) oflocally dominant firms or outsider corporations.

Japan also has regional centers that are outstand-ing examples of network manufacturing, especiallyin metalworking and textiles. Sakaki Township inrural central Nagano Prefecture is one such.57 Thismountainous little community, with a population of16,000, had 321 manufacturing enterprises in themid- 1980s, of which 257 had fewer than 10 employ-ees and only 4 had more than 300. Among them,these firms owned nearly 600 computer-controlledmachine tools.

Sakaki’s small metalworking firms began toflourish in the 1960s, at first on the basis of autosubcontracting. They have since become much morediversified, branching out into general machining,electronics, and plastics, thus escaping dependenceon the extremely demanding auto industry. Thefinancial underpinning for this growth was Japan’sextensive national program of government loans andloan guarantees to small business, administered bythe local association of commerce (shokokai). Theshokokai provides technical support along with itsfinancial aid, reviewing the plans of borrowers andoften proposing specific changes. It routinely ar-ranges classes in computer programming to supple-ment the basic introductory course given by themanufacturer of NC machine tools, and sometimes

3~Benne~ Htison, ‘ ‘Concentration WiMout Centralization: The Changing Morphology of the Small Firm hdustrial Districts of the Third It~y,’paper presented to the International Symposium on Imcal Employment, National Institute of Employment and Vocational Research, Tokyo, Sept. 12-14,1989.

sTFor adet~l~ de~~ri~ion of tie ~e@on~ met~working industry of Sakaki To~ship, ~ David Fri~~, T)w Mistiersrood Miracle: ZndU.$trblDevelopment and Political Change in Japan (Ithaca, NY and London: Cornell University Press, 1988), ch. 5,

Chapter 6-Technology Transfer and Diffusion: Some International Comparisons ● 169

brings in specialists to help individual companieswith particular problems. In Sakaki, factory opera-tors say that they know more about using theequipment than the large firms they supply .58

In the Japanese textile industry, big firms predom-inate upstream in fiber making and spinning.59 Butweaving and knitting is done mostly in smallfamily-run firms with no more than 20 looms(usually installed in a shed or annex to the weaver’shome), a few family-member workers, and two orthree employees. This system of family weaving isan outgrowth of the centuries-old custom of land-lords providing looms for tenant farmers to use in thewinter slack season. With land reform, the tenantsbecame owners. These tiny enterprises are well--suited to producing short runs to order—a good fitwith the Japanese textile industry strategy of com-peting on the basis of diversity, high quality, andresponsiveness to customers’ needs.

Most of these small firms are part of verticalnetworks; they are tied to one of the great spinningcompanies or to trading companies that supply themwith yarn, buy their cloth and, quite commonly, givethem free technical advice. A second importantsource of technological help is the regional industrycooperatives. These are voluntary associations, fundedmostly by members but aided by the many govern-ment programs for SMEs and cooperatives. Typicalactivities are to organize training programs in newtechniques and the use of new machinery, and tohelp firms apply to special industry banks that servesmall and medium-size firms for government guar-anteed loans. Some cooperatives are more active.For example, the Nishiwaki Weaver’s Cooperative,located in a rural area, owns and leases to membersabout 2,000 of 11,348 looms in use by the member-ship. Typically, the cooperative pays two-thirds ofthe purchase price and the weaver pays one-third,plus lease payments for the remainder. The coopera-tive may also guarantee loans for members who wantto buy looms outright.

The state-operated system of research institutesalso helps small firms keep abreast of new technolo-

gies. Japan has 46 textile research institutes in its 47prefectures. Besides collecting industry informationand providing a computer connection with theScientific Research Center in Tokyo, the institutesconduct experiments and research for small firms,charging a fee for service. The research is directedtoward practical problems (e.g., why a color mayfade), rather than broader, more basic topics thatwould interest a university research team.

The Japanese networks, much more than theItalian, have solid, consistent support from govern-ment programs, some available both to individualsmall firms and associations, and some targeted onlyto cooperative groups. The main program targeted togroups is the SME Upgrading Capital System,administered by the Japan Small Business Corpora-tion (JSBC).60 It lends money to the prefectureswhich, in turn, add funds of their own and makeloans to groups and cooperatives. Loans, for periodsof 7 to 16 years, are at low interest (2.7 percent) forgeneral activities and at zero interest for specialactivities. In 1987, government-supported upgrad-ing loans to groups and cooperatives amounted to395.3 billion yen ($2.8 billion). Another source oflow-cost financing for cooperatives is the shokochukin bank, which collects money from coopmembers, supplements it with government funds,and then makes loans to members. In addition, asmall government program (national and prefectu-ral) promotes joint R&D by small fins. It makesawards at the level of $2 million to $3 million yearlyto a couple of dozen cooperative associations.Cooperatives can also take advantage of the free orlow-cost public technology extension services.

The Japanese Government particularly encour-ages the formation of cooperatives in industries withmany very small, weak firms. Box 6-B describes theactivities of a cooperative of 18 plastic moldequipment manufacturers in and around Tokyo andYokohama, and some of the government programsthat support it.

5S1bj&, p, 192, WE 17s5~~ ~me~ for MS ~, ~ ~dition t. s~nd~(j works on be world textile industries, are Ronald kre, Flexible Rigidities: ~nduttial po@’ and

Structural Adjustment in the Japanese Economy, 1970-1980 (Stanford, CA: Stanford University Press, 1986); and The MIT Commission on IndustrialF%xiuctivity, “The U.S. Textile Industxy: Challenges and Opportunities, “ in The Working Papers of the MIT Commission on In&strial Productivity(Cambridge, MA: The MIT Press, 1989), vol. 2.

Whis program is in addition to the loan programs for individual SME firms,

170 ● Making Things Better: Competing in Manufacturing— —-— . . — — ..——.—

Box 6-B—A Plastic Mold Equipment Cooperative in JapanThe Keihin Plastic Kanagata, or plastic mold equipment cooperative, is an association of 18 small companies

in Ota-ku (a city ward in Tokyo), other places in Tokyo, and Kanagawa-ken (a prefecture near Yokohama). ] Thecooperative’s modest offices are located in a compact building in a pleasant but unpretentious Tokyo neighborhood.Within this rather humble exterior is a dynamo of activity.

The Japanese die and mold industry is characterized by a great diversity of products, custom manufacturing,heavy reliance on skilled workers, and a great preponderance of small and medium-size enterprises. Nine of tenplastics toolmakers are small firms with fewer than 19 employees. This kanagata is typical: the 18 membercompanies are all very small, and for them the 6 million yen ($43,000) price of admission is steep.2 The rewardsfor joining are large, however. Members can rely on the kanagata to collect orders from larger customer firms andapportion them to members so that all are kept busy, and customers can usually be accommodated even whenbusiness is booming. When business is slow, kanagata staff can pound the pavement in search of new orders. “Wetry to make sure all the members are working at full capacity, explained Mr. S. Sugano, director of the cooperative.

The kanagata also helps with purchasing, giving member firms both technical assistance in finding goodequipment and quantity discounts. The discounts are not inconsequential; on some machines they are as much as60 percent. (Discounts on quantity purchases are available not only to members of the coop but to a wider circleof 53 firms, in an organization the coop founded. ) Another benefit is in machine leasing. For example, 4 years ago,the kanagata bought 24 CNC machines and leased them out to members. Altogether, the machines cost 450 millionyen ($3.2 million). The kanagata used government loan programs to aid in buying them; one program providedtwo-thirds of the money at 2.7 percent interest over 10 years, and another provided the other one-third at 7.6 percentinterest, also over 10 years.

Even with quantity discounts and leasing on favorable terms, it doesn’t always pay for members to acquire theirown equipment, if it is used quite infrequently. For example, a few years ago, the kanagata bought a CAD system;a member of the coop staff who formerly worked for a plastic design company trains members to use it. Anotherlow-cost government loan, for 28 million yen ($200,000), helped the kanagata buy the equipment. Eventually, thecoop wants to be able to hook up the CAD system to computer- aided manufacturing in members’ plants. lt isexploiting government programs to establish computer networks to make possible the CAD-CAM connection.

In addition, the cooperative can provide both long and short term loans to member companies. Long-term loansare funneled through the kanagata from the shoko chukin bank, which collects money from coop members, addsgovernment funds, and makes loans on favorable terms to the members. A committee of the kanagata approves theloans. Typically, long-term loans are used for operating capital. Members can also borrow up to 6 million yen($43,000--the same as the membership fee) for 6 months at a rate of 1 percentage point above the commercial bankprime rate (about 5 percent in 1989). These short-term loans are used mainly for special purposes such as employeebonuses or debt service that firms are temporarily unable to cover (a common occurrence when firms were adjustingto the rapid rise of the yen in 1986-87). Also, members can buy insurance from the coop to cover possible lossesif one of their customers goes bankrupt.

Finally, the cooperative also provides many kinds of education and information sharing services. For example,members study CAD/CAM applications together, and in 1989 the kanagata had a study group examining theimplications to members of the new consumption tax.

The kanagata supports its staff and activities not only through membership fees but also by taking 1 percentof the order value of the customer orders it handles. Also, in selling equipment to members at the discount price,it adds a charge of 3 percent of the regular, undiscounted price and puts that into the coop’s operating fund (whichwas about 600 million yen, or $4.3 million, annually in 1989).

Throughout Japan, there are about 12,000 Kanagawa associations. In the kanagata prefecture alone are 1,300cooperative groups, with over 370,000 firms participating at some level. Probably most of the groups do not providesuch comprehensive services as the Keihin Plastic Kanagata, but they do typically offer financing assistance, if notpurchasing and order services.

IInfomauon for his INX comes from interviews conducted by OTA Staff h Tokyo, Much 1989,21n addition t. paying tie fee, companies must have the recommendation Of another membfX.

—.—. —— —

Chapter 7

Where We Stand:Public Policy and Technology

CONTENTS

INDUSTRIAL EXTENSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... ... ... ... .,+. +,. ,, +.4Federal Programs for Technology Diffusion to Small Manufacturers . . . . . . . . . . . . . .State Industrial Extension Programs . . . . . . . .

C0MMERCIALIZING TECHNOLOGY FROM FEDERAL LABORATORIES . . . . . .““” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

The Federal Laboratories: An Overview . . . . . . . . . . . . . . . . . . . . .DOE Labs

. . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .

DoD Labs● . . . . . . . . . . . . . . . . . ...............,,,.,,+,+

● . . . . . . . . . . . . . . . . . . . . . . . . . .Other Labs . . . . . . . . . . .

. . . . . . . . .............,.,,.,..+....+.,=,+,,. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Commercializing DOE’s Technology: Mechanisms +... . . . . . . .. . . . . . . . . . . . . . . .

General Applications . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .. ...........,...+.-,+,..,..ENGINEERING RESEARCH CENTERS . . . . . . . . . . . . . . . . s . . . . . . .

TAPPING INTO JAPANEsE TECHNOLOGY . . . . . . . . . . . . .. .......,,.,+.,++

People-to People Technology Transfer. . . . . . . . . . . . . . . . . . . . . . .

... ... ... ... .,. ..+. .4. .+. .. ., + $ ., ., +.$.,...Scanning Japanese Technical Literature . . . . . . . . . .Learning the Japanese Language . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

R&D CONSORTIA . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . .Collaborative R&D in U.S. High Technology: Electronics

. .........++,.,,,,

Cooperative R&D Ventures in Japan. ..........,....+...*,,*

. . . . . . . . . . . . . . . . . . . . . . . . . . .Making Successful Consortia . . . . . . . . . . . . . .

. .....,..**..+**. . . . . . . . . . . . . . . * . . . . .,, , + ., * .,,*,*..,

The Role of Government . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .INTELLEcTuAL PROPERTY . . . . . . . . . . . . . . . . . . . . . .How Much Can Increased Protection Help?

+ .*..*...*.....,..,......,,**. . . . . . . . . . . . . . . . . . .

Specific Problems . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . ....,.......*.,+*.*+,*ANTITRUST LAW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

The Changing Interpretation of the Law . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . ., ., ., . * ., + +*.*,,.*.

The Terms of the Debate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . +++.,,.+.Effect on Business Activity . . . . . . . . . . . . . . .

... ... ... ,***. m.. . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . .

Page173174177184185185186186187194195197198200200200202208210211211212217219220221225

BoxesBox

Page7-A. Five State Industrial Extension Programs . . . . . . . . . . . . . . . . . . . . . . . . .. . ● . . + . . . . ● . . 1797-B. DOE’s HTS Pilot Centers. ..,., +... +... . . . . . . .1887-C. The National Science Foundation Engineering Research Centers . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1967-D. Intellectual Property Rights Protecting Innovation . . . . . . . . . . . . . .2137-E. BTU Feds the Heat . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . .. . . . ., . ....*+.,. 230

TableTable

Page7-1. Expenditures on State Technology Programs, FY 1986 and FY 1988 . . . . . . . . . . . . 178

Chapter 7

Where We Stand: Public Policy and Technology

The science and technology policy of the U.S.Government has traditionally been concerned withbasic science, health, energy, agriculture, and de-fense. It has been described as big science deployedto meet big problems,l and as mission-orientedrather than diffusion-oriented.2 With few exceptions(the most important being agriculture and civilianaircraft), U.S. Government policy has not beendirected toward helping private enterprises makecommercial use of advances in technology. Onlyrecently, as it became painfully obvious that oneU.S. industry after another was losing technologicalleadership, have U.S. policy makers given seriousthought to a different approach. Some changes areoccurring, and of these, some are real departuresfrom the past. But they have been made in apiecemeal, ad hoc fashion. No comprehensive set ofgovernment policies has yet been adopted to pro-mote the use of technology for better performance inmanufacturing.

The Federal Government undertook a truly novelventure when it went halves with the semiconductorindustry in the Sematech R&D consortium, whichseeks to improve the manufacturing process for theindustry. Other government-supported R&D consor-tia have been considered (e.g., to promote R&D foradvanced television systems). Repeatedly, Congresshas enacted laws that urge the 700-odd Federallaboratories to make their research results moreaccessible to industry, and to undertake new R&Dprojects designed and operated in collaboration withindustry. In establishing Engineering Research Cen-ters in 18 universities, the National Science Founda-tion hopes to forge stronger links between academicengineering research and training and the world ofindustry. NSF is also encouraging U.S. scientistsand engineers to acquaint themselves with researchresults coming out of Japan, and to foster the flow oftechnology from Japan to this country. A growingnumber of States are establishing industrial exten-sion services to bring best practice technology tosmaller manufacturers, and the U.S. Government istaking some initiatives in the same arena.

These programs represent deliberate actions byFederal, State, and local governments in the UnitedStates to improve the use of technology by U.S.manufacturers. Other government actions, also in-tended to improve industrial performance, workmore indirectly. Among these are tax policies, suchas the present tax credit for increased R&D or thepast program of rapid depreciation for capitalinvestments in up-to-date plant and equipment.3

Laws protecting intellectual property (e.g., patentand copyright laws) are intended to reward innova-tion and thus to foster technological advance.Finally, Federal policies adopted for national goalsother than international competitiveness may stillaffect it indirectly. One of these is antitrust law andenforcement.

The following sections describe and analyzegovernment programs and policies as they existed in1990 from the standpoint of their effect on U.S.manufacturing technology. Chapter 2 of this report,analyzing policy issues and options, discusses pro-grams and approaches that Congress might wish toconsider for the future.

INDUSTRIAL EXTENSIONIn the United States, government technical and

financial assistance to small and medium-sizedbusiness is patchy and thin. Federal programs do notbegin to compare in size to the $31 billion per yearthat the Japanese national government pours into itscombined program of direct loans and technicalassistance to smaller businesses—not to mention theadded contributions from prefectures, cities, and citywards, plus the $56 billion in guaranteed loans forsmall firms underwritten by government institu-tions.4

The U.S. assistance programs are not only muchsmaller than the Japanese but also more hit-or-miss.Every city in Japan and most rural towns have theirindustrial halls, or federations of small business, orchambers of commerce, dispensing technical helpalong with plentiful funding for purchase or lease of

l~vin M- weln~rg, Reflections on Big Science (Oxford: Pergamon press, 1967).

zHenry Ergas, ‘Does Technology Policy Matter?’ Technology and Global /ndusmy: Compaw”es and Nutwns in the WorfdEconomy Bruce R. Guileand Harvey Brooks (cd.) (Washington, DC: National Academy Press, 1987).

s~scuwion of tax policies affecting R&D and capital investment is inch. 2.gFor a &scnption of Japanese national government programs to assist smaller businesses, see ch. 6.

- 1 7 3 -


the latest production equipment. Japan is blanketedwith government or quasi-public institutions at theservice of small and medium-size enterprises. In theUnited States, a small manufacturer in need oftechnical advice is lucky to find a State or localagency capable of providing it, much less a Federalprogram that fits his needs.

Small firms form a sizable minority in U.S.manufacturing. Some 358,000 small and medium-size firms (defined as those with fewer than 500employees) account for 98.8 percent of all manufac-turing enterprises, and 35 percent of the manufactur-ing work forces According to one estimate, thesesmall firms represented 21 percent of value added inmanufacturing in 1982.6 However, employment maybe a better gauge of the contribution of small firmsto manufacturing, since wages are the major compo-nent of value added and wages are lower in smallmanufacturing firms than in larger ones.

Many small outfits are suppliers of essentialmaterials and parts for large manufacturing firms,and they are especially important in metalworking—the fabrication and machining of metal parts. Over94 percent of the firms in five major metalworkingindustries are small plants with fewer than 100employees. 7 How well these firms do their jobsaffects the cost, quality, and marketability of majorproducts from kitchen appliances to automobiles tobulldozers, drilling rigs, and jet airliners. Small tomedium-size metalworking firms are also the heartof the industries making production machinery,from tools, dies, and jigs to block-long papermakingmachines. In other words, the technological upgrad-ing of small and medium-sized manufacturers hasnationwide economic implications.

Many of these firms need technological upgrad-ing. This does not mean that small factories need toinstall 21st-century computer-integrated manufac-turing systems. It does mean they need to acquireup-to-date equipment, train people to use it well, andorganize work efficiently. Getting best practicetechnology out to all corners of U.S. manufacturing

is not easy. Owners of small manufacturing firms areoften too busy doing a dozen jobs to find out forthemselves about technology improvements. Manydo not have their own manufacturing engineers,because the engineers cost too much, or are notneeded full time, or are unavailable in out-of-theway places where some manufacturing plants arelocated. Consulting engineering firms are usuallymore geared to serving large clients than small ones,and many small manufacturers don’t trust theirability to find a consultant who will tailor his adviceto what the manufacturer needs rather than what theconsultant has to sell. Vendors of production equip-ment can be good sources of technical advice, butoften they fall short of what is needed, especially inadapting software to fit particular firms’ require-ments and in training workers to use the equipment.According to one director of a State industrialextension service, you can’t just throw in a computerand read the manual-you have to train people.“We’ve had lots of companies with computers intheir closets. ” Finally, financing is the biggesthurdle for many small manufacturers. A small firmis less likely than a big one to have the contacts ortrack record needed to get loans or otherwise raisemoney for modernization, and financing is oftenmore expensive for small firms.

Federal Programs for Technology Diffusion toSmall Manufacturers

Recognizing the gaps in technology diffusion tosmall and medium-size manufacturers, Congress hasrecently created new programs of technical assis-tance to smaller firms. The Federal effort is still quitelimited, however, and there are no Federal loanprograms specifically aimed at promoting the adop-tion of new technologies by small manufacturers. Infiscal year 1989, financial aid administered by theSmall Business Administration amounted to $47.3million in direct loans (which are available only todisadvantaged people) $3.6 billion in loan guaran-tees, and a contribution of about $150 million to twoquasi-public financing agencies for small firms. This

57_’~ Stite ~~Sw/~ B~ine~~: A Report ~~r~ p~~~ide~ (Wmhington, DC: U.S. Gover~ent ~nt~g office, 1989), table /4.15, pp. 80-81, and tableA. 17, pp. 84-5. The Japanese sector is more heavily weighted toward smaller fins; esmbiishments witi fewer ~an 300 emPloY~s are 99.5 Wrcent ofall manufacturing establishments and employ 74 percent of the sectoral work force.

6Jw1 popkin & Co., “Small Business Gross Product Originating: 1958 -1982,” contract report to the Office of Advocacy, Small BusinessAdministration, cited in ibid., p. 31.

7~ 1986, here were 134,7~ entewn=s in tie five major z-digit met~wor~ng s~tors, Fabfica~ Met~ Products, Machinery except Ekct.(icd,

Electric and Electronic Equipment, Transportation Equipment, and Instruments and Related Products (SIC 34-38), and of these, 126,700 were smallenterprises with fewer than 1(X) employees. Ibid,, table A.18, pp. 86-87.

Chapter 7—Where We Stand: Public Policy and Technology ● 175

aid is given to all kinds of small and medium-sizefirms (most small businesses are in retail trade andother services) for all kinds of purposes which mayhave little to do with improving technology.

The biggest U.S. Government program promotingtechnology advances in small manufacturing is theSmall Business Innovation Research (SBIR) pro-gram, established by Congress in 1982.8 Under thisprogram, Federal agencies with R&D budgets ofmore than $100 million per year must set aside 1.25percent to help small and medium-size firms com-pete for Federal research contracts and support thesesmall firms in bringing their R&D results to the pointof commercialization. In 1987, 1,276 small compa-nies were awarded $350 million to do R&D work for11 Federal agencies. The first phase in the SBIRprogram is feasibility studies of promising ideas(2,189 awards in 1987, for a total of $109 million);the next is development of the ideas with the greatestpotential (768 awards, $241 million). SBIR does notfund the final stages of bringing a product to market,but the Small Business Administration does helpfirms that have gained a place in the R&D programfind private financing for commercialization.

SBIR has been given high marks for funnelingFederal R&D money to small fins, and for helpingyoung, innovative companies develop advancedtechnology products.9 Most of the projects are in theareas of defense, health, and energy, where FederalR&D is concentrated but where commercial possi-bilities are often limited. The program has beenespecially helpful, however, in at least one commer-cially oriented field-biotechnology .10 What SBIRdoes not do, and was not designed to do, is give bestpractice technical assistance to the great majority ofsmall manufacturing businesses, which are notinvolved in the development of products or proc-esses at the frontier of advancing technology.

The Small Business Administration runs a fewprograms that dispense business management and

marketing advice to the ordinary small company(which, as noted, is most often in services or retailtrade). One of these is the counseling and briefworkshops on business management offered byvolunteers, the Service Corps of Retired Executives(budgeted at $2.5 million). Another is the SmallBusiness Development Centers, mostly located onuniversity campuses, which provide counsel fromfaculty or students on particular problems, some ofwhich may be technical. There are 53 such centersnationwide, in all but four States; about half theirfunding comes from the government ($45 million infiscal year 1989) and the rest from the universities.Useful as these programs are, they are not focused onthe choice and use of technology in manufacturing.

Federal programs that concentrate on improvingmanufacturing fins’ use of technology come downto a very few. The oldest and largest is theManufacturing Technology (ManTech) program ofthe Department of Defense, funded at $175.5 millionin fiscal year 1990. ManTech was created toencourage the development and use of innovativemanufacturing technologies, and thus strengthen theU.S. defense industrial base. The program is directedto large companies as much as small ones, and isconcerned with production of military goods. Mostof the ManTech money goes to large defensecontractors, often for rather narrow projects promis-ing near-term savings. 11 However, some ManTechprojects have brought forth new manufacturingtechnologies of broad importance, civilian as well asmilitary. Numerically controlled machine tools weredeveloped in a ManTech project. More recentprojects with possible commercial applications in-clude work on near net shaping of metals andcomputer integrated manufacturing systems.

If the funding for ManTech programs (varying upand down from $130 million to $200 million in the1980s) seems a minuscule portion of the DefenseDepartment’s $40 billion R&D budget, it looms very

s~e sm~l Business Innovation Development Aet of 1982 established SBIR.gCompUoller Gener~ of tie Unitd States, Gener~ Accounting Office, Implementing the Small Business lnwvation Develome~ Ac+The First

2 Years, GAO/RCED-86-13 (Washington, DC: October 1985); A Profile of Selected Firms Awarded Small Business Innovation Research Funuk,GAO-RCED-86-113FS (Washington, DC: 1986); Effectiveness of Small Business innovation Research Program Procedures, GAO/RCED-87-63(Washington, DC: 1987); Small Business Innovation Research Participants Give Program High Marks, GAO-RCED-87-161BR (Washington, DC:1987).

1~.s. Congess, Offiu of T~hno@y As~ssment, New Development in Biotechnology: U.S. Investment in Bwtechnofogy, OTA-BA-360(Springfield, VA: National Teehnical Information Service, 1988). OTA found that “SBIR funds are one of the few sources of direct Federal supportfor applied researeh and development.

llMmufactW~g Studies Bo~d, Manufacturing Tec~/ogy: Corurstow of a Renewed Defense ]@tria/ Base (Washington, DC: NationalAcademy Press, 1987).


large compared to Federal spending for manufactur-ing technology on the commercial side-especiallydiffusion of technology to small manufacture.Technology diffusion programs include the 28-year-old Trade Adjustment Assistance, and the newlyminted Manufacturing Technology Centers (MTCs),created in the 1988 trade act and operated by theNational Institute of Standards and Technology(NIST, formerly the National Bureau of Stand-ards). 12

Trade Adjustment Assistance (TAA) for firms isopen only to companies that can show they were hurtby imports.13 It has usually been funded at about $15to $16 million per year but in recent years itsprospects were uncertain (the Reagan Admini-stration repeatedly proposed to abolish it) and itsfunding was cut. In fiscal year 1990 it received $9.9million in new and carryover funds. Nevertheless,until 1988 TAA was the major Federal programgiving one-on-one technical assistance to small andmedium-size manufacturers. The TAA program alsogives advice to its clients on such things asmarketing and advertising, inventory control, andfinancial management. Help is provided by 12 small,regional, non-profit centers that act, in effect, asindustrial extension agencies.

The new Manufacturing Technology Centers arecharged generally with transfer of advanced technol-ogy to industry, with special emphasis on U.S.-basedsmall and medium-sized manufacturers. The lawdirects the centers to make new manufacturingtechnology ‘‘usable” to these smaller firms; ac-tively provide them with technical and managementinformation about manufacturing; establish demon-stration centers for advanced production technolo-gies; and, for small firms with fewer than 100employees, make short-term loans of advancedmanufacturing equipment. So far, three federallyfunded Manufacturing Technology Centers (in TroyNY, Cleveland OH, and Columbia SC) have beenestablished in the United States and three more areplanned. The three existing centers got a total of $4.5million in Federal funds in 1989; matching fundsfrom local sources are required.

NIST expects the Manufacturing TechnologyCenters to serve primarily small firms with 200 orfewer employees, and to concentrate more onoff-the-shelf best practice technologies than onhigh-tech cutting edge systems fresh from the R&Dlab. NIST officials also say that the primary serviceoffered by the Centers will be modernization plans,customized to fit the needs of individual firms.However, the language of the law gives NISTlatitude to support Centers with varying approaches,and so far it has done so. The Troy MTC isconcentrating on transfer of high-technology sys-tems from labs to selected fins, though it alsocooperates with State agencies and communitycolleges in diffusing best practice to a broad range ofclient firms. Field agents of the Cleveland MTC areknocking on doors of thousands of small companiesin a concentrated industrial area and offering thosethat respond individual business and technical plans.The South Carolina MTC, which is closely linked tothe State’s technical college system, is installingcenters to demonstrate computerized metalworkingequipment.

NIST has its own small demonstration center inthe Shop of the 90s. This is a working machine shopthat fills job orders from government agencies butalso serves a technology extension purpose. It is anoffshoot of NIST’s highly automated, state-of-the-art Advanced Manufacturing Research Facility (AMRF),which was meant to serve in part as a learning centerfor manufacturers. However, many people fromsmall manufacturing firms found the AMRF entirelytoo advanced to have any practical application totheir businesses. The Shop of the 90s, using off-the-shelf technology, fits their needs and experiencebetter. Because it is a working shop, with 60employees and a business worth about $4 million ayear, the manager has credibility with small manu-facturers. State technology agents are brought in forpresentations, and the Shop is open for tours andphone inquiries.

One more small NIST program, also created in the1988 trade act, is intended to provide technical andfinancial assistance to State technology extension

lzNei~erprogr~ is st,rictly limi~ tO small and mtiium-size manufacturers, but in practice TM has mostly served small manufacturing fiis, andthe law creating the Manufacturing Technology Centers emphasizes dissemination of new technology to small and medium-size manufacturers. See theOmnibus Trade and Competitiveness Act of 1988 (Public Law 1O(H18), Subpart B, Sec. 5121(a).

lsTrade AdJWtment Assis~nce ~so includes a r~mployment ~d rewfiing program for workers losing ~eir jobs due to imports; ~S pm Of TAAis far bigger (recently funded at about $200 million per year) and better-known than TM for firms. For a description and evaluation of both programs,see U.S. Congress, Office of Technology Assessment, Trade Adjustment Assistance: New Ideas for an Old Progr@pecial Report, OTA-lTE-346(Springfield, VA: National Technical Information Service, 1987).


services. This program got no funding until fiscalyear 1990, when it received $1.3 million, but NISThad already begun some modest outreach to States.So far, it has mostly been a one-man show—a singleNIST official (sometimes accompanied by the man-ager of the Shop of the 90s) who travels to Statetechnology agencies explaining what resources NISThas to offer, referring them to other sources ofFederal help, and helping various State agenciesmake contact with each other.

Another federally funded technology demonstra-tion center has been in business since 1988. That isthe National Apparel Technology Center in Raleigh,NC, an outgrowth of the 10-year-old TC2 project.TC2 (Textile/Clothing Technology Corporation) beganas a combined government-industry effort to de-velop a flexible, automated sewing system able totake on a variety of complicated sewing jobs, suchas attaching the sleeve in a man’s suit jacket.Although it has fallen short of some of its ambitioustechnical goals, has produced some commerciallyusable automated sewing equipment. In addition,TC2 now supports the Raleigh center, which demon-strates a whole range of modern apparel-makingequipment to its member companies, large andsmall, and arranges seminars with apparel engineer-ing faculty of nearby North Carolina State Univer-sity. The Federal Government’s contribution to TC2

has been $3.5 million per year for the past few years.The Defense Logistics Agency also operates threedemonstration centers for apparel technology, eachfunded at up to $5 million per year, with three-quarters Federal funding. These centers are open tocivilian manufacturers as well as defense contrac-tors.

Altogether, these Federal technology extensionefforts are scattered and small. Up to now, theemphasis in Federal technology transfer programsfor small manufacturers has been much more onpushing out sophisticated new products and proc-esses (as in the SBIR program) than on helpingindividual firms adopt best practice technology.

State Industrial Extension Programs

Most of the action in industrial extension is in theStates, and even there it is limited, though increas-ing. Exactly how much it amounts to is uncertain,partly because surveys of State programs are incom-plete and quickly outdated, and partly because“industrial extension” is not very well defined inthe surveys. More than 40 States have programs to‘‘promote technology, “ but most of their effort andfunding goes for research and development inuniversities and for aid to high-technology startupventures-not for help to existing firms in adoptingbest practice technology. According to a survey ofState programs done for NIST in 1988-89, only 13programs in nine States had technology extensionprograms whose main purpose was direct consulta-tion with manufacturers on the use of technology .14However, this number is already out of date. At leastone new program, Nebraska’s, was established afterthe survey was completed.

One of the better recent surveys of State technol-ogy programs was done by the Minnesota Gover-nor’s Office of Science and Technology .15 It foundthat in 1988 States directly spent $550 million onvarious kinds of technology programs, but onlyabout 10 percent of that—some $57 million—wentfor technology transfer and technology/managerialassistance (table 7-l). Technology transfer, whichgot $46 million (8 percent) of the finds, was definedas facilitating “the transmission of new technolo-gies from the laboratory to the private sector. . . forthe creation of new businesses, the introduction ofnew product lines for established firms, or therevitalization of mature industries.” l6 Despite thislanguage, some activities that States call “technol-ogy transfer” might really be closer to industrialextension services. At a guess, the States arespending some $25 million to $40 million for suchservices.

As used here, industrial extension means a servicesomething like this: an accessible office staffed witha few engineers or people with experience inindustry invites telephone calls or visits frommanagers of small manufacturing firms seeking

ld~~d R. J~~~n, ~ting ~rWtor, T@-~olo~ Semices, National ~stitute of Standwds and Tec~ology, te~imony before the U.S. House ofRepresentatives, Committee on Small Business, Sept. 28, 1989.

15@vernor’s Office of Science and TW~OlOW, Stite Tec~/o~ Progr- in the United Smtes, 1988 (St. Paul, MN: ~nnesota Wpaltment OfEnergy and Economic Development, 1988).

IGIbid., p. 1.


Table 7-l—Expenditure on State Technology Programs, FY 1986 and FY 1988

Number of AverageExpenditures States with State

programs spendingFY 1986a FY 1988a FY 1988 FY 1988

Type of program $ Million Percent $ Million Percent $ Million

Technology/research centers . . . . . . 285.6 41.0 226.6 41.2 29 7.8Research grants . . . . . . . . . . . . . . . . 126.7 18.2 150.2 27.3 25 6.0Venture/seed capital . . . . . . . . . . . . . 159.6 22.9 37.4 6.8 18 2.1Research Parks/incubators . . . . . . . . 75.6 10.9 36.9 6.7 22 1.7Technology/managerial assistance . 10.5 1.5 11.0 2.0 30 0.4Technology transfer . . . . . . . . . . . . . . 8.4 1.2 45.7 8.3 26 1.8Other technology programs . . . . . . . . 30.1 4.3 42.4 7.7 41 1.0

700.0 b 100.0 b 550.0 100.0 44C 12.5Notes:a There are differences in accounting procedures between the 1986 and 1988 reports, For some states, the 1986 figures represented muiti-year appropriations.

The 1988 figures are all on an annual basis.b column sum does not add to total because of rounding.c Number of States with one or more technology Programs.

SOURCE: Calculated from: Governor’s Office of Science and Technology, State Technology Programs in the United States, (St. Paul, MN: MinnesotaDepartment of Energy and Economic Development, September 1986); Governor’s Office of Science and Technology, State Technology Programsin the United States, 1988, (St. Paul, MN: Minnesota Department of Trade and Economic Development, July 1988).

help. Promptly after the first interview, the officesends a technical specialist (either someone from itsown staff or an engineer from the State university)to make an onsite diagnosis. Then the extensionservice produces a customized client report, and itstechnical specialist or a consultant works one-on-one with the firm to put into effect the improvementsrecommended by the service and accepted by thefirm’s manager.

What small manufacturers need more than thenewest technologies fresh out of the laboratory isoff-the-shelf hardware and software and individualhelp in choosing and managing them. They needadvice on these choices from an independent sourcewith no financial stake in the selection. And theyneed to understand how much training is involved inadopting new equipment, and where to get it. Theseconclusions are drawn from the experience of peopleinvolved in technology extension, both the agentsproviding the services and the firms receiving them.In visits and interviews with five State industrialextension programs in 1988, OTA found that theprograms were serving genuine needs that were nototherwise being met, and that demand for theservices was high.17 At least two of the States—Georgia and Maryland-do not advertise the serv-ices they offer for fear of being swamped withrequests for assistance. (Box 7-A lists and brieflydescribes the programs OTA visited.)

Individual Problem Solving

Everyone interviewed took it as given that one-on-one contact between technical specialists and com-pany managers is the bedrock of industrial exten-sion. A good hard look at the company’s individualproblems is the starting point for all the programs.This often includes an intensive telephone interviewto begin with, followed by a site visit and adiagnostic report. Again and again, company man-agers remarked on the value of an objective,experienced outsider taking a fresh look at thecompany’s problems—something that managers ofsmall outfits are often too swamped to do. ‘‘I don’thave time to do research,” said Jerry Lipkin,Executive Vice-President of Moyco Industries, aPhiladelphia manufacturer of abrasives and dentalproducts. “I have to do sales, marketing, andpersonnel. ’

Sometimes, the diagnosis may find that a com-pany’s efforts to modernize are misdirected, or thatreal problems have escaped the manager’s attention.According to Travis Walton, director of Maryland’sTechnology Extension System (TES), some compa-nies think they need sophisticated computer equip-ment when they don’t. For example, “If you makethe same product year after year you don’t needCAD (computer-aided design)—you only need it ifyou customize. ’ One company, Travis added, cameto TES for aid in setting up a computer system to

l~~dings from ~e~ Vlslts and interviews are also reported in Philip Shapka, ‘‘Industrial Extension: hrning from Experience,’ contractor reportto the Office of Technology Assessment, November 1988.

Chapter 7—Where We Stand: Public Policy and Technology 179

Box 7-A—Five State Industrial Extension ProgramsIn 1988 OTA visited five industrial extension programs in four States, some with long experience and some

just a few years old. Through interviews with program managers, extension agents, and clients, OTA soughtinformation on the kinds of technical assistance small manufacturers need and how the programs are meeting theneeds. The five programs, with acronyms and year of origin, are:

Georgia Institute of Technology Industrial Extension Regional Offices (GTRI, 1960) is headquartered atGeorgia Tech in Atlanta and supports 12 regional offices, each with a field staff of two or three people givingindividual service to client fins, The regional offices also link clients with specialized services at Georgia Tech,including assistance on productivity, energy conservation, workplace safety, hazardous waste management, andtraining. Funded at $3.8 million in 1988, GTRI had 26 professional employees and served 960 firms. Days of fieldservice averaged 2 to 5 and the average cost per client was $4,000.

Maryland Technology Extension Service (TES, 1983), based at the University of Maryland, offersone-on-one client assistance at five regional offices. Field staff may refer problems to the university faculty. Witha full-time staff of seven people, and funding of about $400,000, TES served 250 to 300 clients in 1988, giving upto 5 days of service at an average cost per client of about $1,500.

Michigan Modernization Service (MMS, 1985) is a State-sponsored program, affiliated with Michigan’sIndustrial Technology Institute. Its services include intensive diagnosis and onsite visits from a field representative,experienced in industry and manufacturing technology, paired with a training specialist. Some 45 people staff theprogram, but most of the 25 professionals are part-time consultants. The 1988 budget was $2.8 million (expectedto rise to $3.9 million in 1989) and 140 clients were served (250 expected in 1989). Cost per client was about$20,000 for an average of 6 days of service.

Pennsylvania Technical Assistance Program (PENNTAP, 1965), a joint program of Penn State Universityand the Pennsylvania Department of Commerce, provides technical information from faculty specialists and someonsite visits, in response to client requests. Sometimes PENNTAP takes the initiative in acquainting firms with newtechnologies. Total budget in 1988 (including in-kind facilities and services donated by the University) was about$1.3 million and the staff was equal to 12 1/2 full-time slots. Some 850 firms and 450 local government bodiesreceived services; cost per industry client was $1,100 to $1,500. The length of service was not reported.

Pennsylvania Technology Management Group (TMG, 1984), a nonprofit corporation sponsored by theState, concentrates on bringing best practice technology to small manufacturers (defined as having fewer than 250employees, but in practice usually in the range of 20 to 40 employees). One of the small core staff (6 people)evaluates the client’s problems, and TMG then shares the cost of a consultant, if needed. With a budget of $350,000in 1988, TMG served about 40 clients, at an average cost of $8,800. The length of service averaged 8 days.

track inventory, but the real problem was that the company got the space it needed in only 25,000inventory was “totally chaotic” and far too big,tying up capital in unneeded items.

Another example comes from the Tnemec Co.,Inc. of Baltimore. This branch plant of a smallcompany ($1 3 million sales per year) makes indus-trial protective coatings for water towers, wastewa-ter plants, and the like. Tnemec wanted to expand tohandle a growing business, but the plant manager,Frank Lavin, recognized that he needed help inplanning the expansion. “I’m in a small businesswith a busy day-to-day routine, ’ he said. ‘‘I don’tknow how to build a new plant, ’ He called on TES.In a site visit, the TES engineer found that acomplicated, inefficient flow of materials had devel-oped over the years in the old plant, and suggesteda wholesale rearrangement. The result was that the

square feet, not 40,000 square feet as originallyplanned. ‘‘At $25 a square foot, we saved a lot ofmoney," Lavin said. He added that if he had askeda consulting firm for 40,000 square feet, they wouldhave built it without question. “Consulting engi-neers and architects build what you ask them to.

Trust

Lavin, like other company managers, praised the‘‘objectivity and the expertise of the State exten-sion service. Trust in the services’ impartiality-thefact they are not trying to sell the companiesanything or collect big fees—is a key element intheir success. This was the reason several plantowners and managers gave for turning to a Stateagency instead of a private consultant. Besides, theysaid, small firms have trouble getting competent


service from consulting engineers. One said bluntly:“They are a waste of time and expensive.”

Brooks Manufacturing is one company that struckout in trying to find the right private consultant. ThisPhiladelphia firm has a $6-million-a-year businessmaking electrical outlet strips, but it faces growingcompetition (especially from Taiwan) in its basicproduct line. Brooks is trying to build up its businessin more specialized, higher value-added items—electrical outlet strips for medical carts, for exampleand is developing a special power strip that iscompatible with sophisticated communications equip-ment. But the company is too small to support aresearch and development department to design itsnew products, and it failed to get what it needed fromthree different consulting engineers. “The engineer-ing service consultants usually send out the newguys to small fins, ” President Gary Brooks said.

Pennsylvania’s Technology Management Group(TMG) stepped in and helped Brooks find a capableengineering consultant, who developed new productdesigns and made blueprints for the company. TMGalso funded an evaluation of the company’s opera-tions to see whether it needed and could handle aMaterials Requirements Planning (MRP) system,which takes an order and breaks it down into theindividual components and material needed to fillthat order. On the basis of the evaluation, Brooksadopted the system. TMG also found a qualifiedconsultant to help the firm tailor the system to itsneeds.

At Moyco Industries in Philadelphia, Jerry Lipkinremarked that the intervention of TMG in finding aconsultant meant that the fees were predictable andthere was a cap on final costs. “We have beenburned by consultants in the past, and the program’sinvolvement helps reduce the risk of this happen-ing.” That TMG puts up a little money (maximumof $1 ,500) toward the consultant’s fee reassures thecompany that TMG too has a stake in the outcome,and that the consultant is qualified. For their part,consultants seem to welcome referrals from Stateextension services since this adds to their credibilityand opens doors to new business.

Extension services operating out of universityengineering departments can use members of theirown departments for consultations. For example,when American Bottlers Equipment Co. (Ambec) ofOwings Mills, MD, came to Maryland’s TechnologyExtension Service for help in computerizing its parts

list and linking the list with computerized drawings,the service used its university connection. TESworks out of five regional offices but is based in theEngineering Research Center of the University ofMaryland; it calls on engineering faculty membersin nearly half its cases. Travis Walton, director of theprogram, says TES has a “visiting nurse” approach—the engineers who staff the regional offices do whatthey know how to do and call for help when theproblem is beyond them.

Ambec is a small company specializing in themanufacture of stainless steel conveying and han-dling equipment for customers in the food andbeverage, pharmaceutical, electronics and otherindustries. It has sales of $10 million per year andabout 100 employees. Essentially a job shop, Ambecworks to customer specifications, using families ofparts which it assembles to meet a particularcustomer’s needs. Before consulting TES, Ambechad gone through a bad experience with a privateconsulting firm, which sold it a Material ResourcePlanning software system that was supposed to keeptrack of orders and parts, but never worked aspromised. Instead of trying that route again, thecompany called on the State extension service. TESlinked Ambec with a University of Marylandengineering professor and a student with goodcomputer skills. The student developed the programAmbec wanted and later went to work full time forthe company.

Confidence in an extension service’s competenceis as important to a company as trust in itsobjectivity. Connections with an institution that isalready well respected throughout the State help toestablish that confidence. In Maryland, for example,that institution is the University’s highly regardedengineering department. In Pennsylvania it is PennState University, in Georgia it is Georgia Tech, inMichigan it is the Industrial Technology Institute inAnn Arbor.

Sometimes, only experience will instill confi-dence. Terry Brady, president of Bradhart, Inc. ofHowell MI, consulted the Michigan ModernizationService--MMS) only as a last resort. Bradhart is asmall but top-of-the-line job shop, machining high-quality metal parts, especially bearings, to thespecifications of its customers in the aerospace,ordnance, and oil industries. To stay competitive inthe new global economy, the company decided tomodernize. Moving to larger quarters, it installed


several computer numerically controlled (CNC)machine tools and a computer system to integrateorders and office processing with production. Thisinvestment cost half a million dollars-a lot for acompany with sales of $3 million per year. Unfortu-nately, the company’s managers soon discoveredthat they had seriously underestimated the startupcosts for training workers to use the new tools.Further, the software for the computer system didnot run properly. The company had run out of credit.It was in a make-or-break position.

At this point, Brady called MMS, but withoutmuch hope of real help. He was surprised, frost atgetting a prompt businesslike response, and stillmore so at the quality of training and other assistanceMMS was able to provide. Finally, MMS gaveBrady a vital boost in confidence when its evaluationconfirmed that the company was right to investheavily in modern equipment, and was headed in theright direction. Brady remarked appreciatively onthe way MMS staff had served as a “soundingboard,” providing advisors who were not competi-tors but still had an understanding of business andtechnology. “I still don’t believe,” he said, “thatsomeone would want to help the little guy. ”

Training

Nothing could better illustrate the importance oftraining on new equipment than the Bradhart story.Because the managers did not appreciate how muchtraining would cost, the company almost went underin an otherwise sensible move to modernize. Fortu-nately, MMS was able to help Bradhart get Statetraining funds and find good training programs. (Asdiscussed below, MMS also helped the company geta bank loan to tide it over the crunch, before theinvestment in equipment and training began to payoff.)

With help from MMS, Bradhart set up trainingprograms for employees, both in-house and at a localcommunity college. Shopfloor employees receivedtraining on the CNC tools and in quality controltechniques; the office staff was trained in spread-sheet and database programs and job costing. Inaddition, the company sent four or five employees ata time through a local community college to learnbasic mathematics, quality control, and supervisoryskills. MMS also helped Bradhart untangle itssoftware.

Although the directors and staff in all fiveindustrial extension services stressed the importanceof training, MMS was the only one with a trainingelement routinely built into its services. It took timefor MMS to recognize the merits of marryingtraining with technology. In its early days (perhapsinfluenced by General Motors, which was thentrying to automate everything it could in autoassembly) the program concentrated on hardware,and training to use the hardware was not muchemphasized. Today, MMS takes care to emphasizethat it is not hawking technology per se but ishelping firms use technology, which means develop-ing management and training. On every site visit,MMS sends pairs of training and technology special-ists to make the diagnosis and write the report, whichincludes an assessment of training needs and optionsfor every client and actively helps clients design orprocure training. The Michigan program spendsroughly $1 dollar on training assistance for every $2dollars it spends on technology deployment.

Other industrial extension services, though lesssystematic about training than MMS, also knowwhere to refer clients for training advice andassistance. For example, Pennsylvania’s TMGlinked Brooks Manufacturing with a communitycollege to get training in quality control, statistics,teamwork, and basic math for its workers. However,some of the services have a harder time findingadequate training. Georgia Tech has a small indus-trial training unit able to provide limited training,mostly for frost line supervisors. But in some of itscases, training that extension agents recommend,and companies are eager to get, is not available.

For example, in a productivity audit of ImperialCup’s paper and plastic cup manufacturing plant inLa Fayette, GA, the Georgia Tech engineer includedseveral recommendations for improved training. Shefound the current training-2 days under a fret-linesupervisor-inadequate for working with the so-phisticated machinery in the company’s paper de-partment. Imperial tried to get a local vocational-technical school to train workers on the shop floor,but the school offered only classroom training. Thecompany also had trouble finding workers with theskills needed to maintain the machinery. The localvoc-ed school turned out electronics and autotechnicians, but not machinery repairers. In the past,the company sent small groups of workers to themachinery manufacturer in Wisconsin for training,but the manufacturer recently expressed reluctance


to continue it. At the time of OTA’s visit, no solutionto Imperial’s training needs was in sight.

Financial Aid

Industrial extension services do not provide fundsfor capital investment or operating expenses. Theyare in the business of giving technical, not financial,assistance. However, they can help small firmscoming to them with financial problems in twoways. First, their diagnosis may reveal that what thefirm’s manager thought was a need for funds is reallymore a problem of management that can be solved,say, with a better use of space, flow of materials, orcontrol of inventory.

Second, the State agency can be very useful indirecting firms to sources of funds, and supportingthem in dealings with banks. For example, theMichigan Modernization Service not only pointedthe Bradhart company toward State funds that couldhelp pay the big bills for their training needs. MMSalso helped the company get a bank loan, using Stateeconomic development funds as equity (the fundscame from Community Development Block Grants,contributed by the Federal Government to theStates). “This lessened the financial pressures,”said Terry Brady, Bradhart’s president. “We wouldhave gone down without the State’s help.”

Besides the block grants and other economicdevelopment funds, many States have special loanprograms for small businesses that extension serv-ices can tap. The extension services can also pluginto the Federal program of small business guaran-teed loans. It is safe to say, however, that finding themoney to modernize a factory is a serious hurdle formany small manufacturing firms in the UnitedStates. They do not have the many options of thesmall Japanese firm, which can afford to pass up alow-interest government loan in favor of a bank loanat a slightly higher rate, because the bank takes justone day to consummate the deal, while the govern-ment loan might take a whole month (see box 6-A,ch. 6).

Staff, Fees, Intensiveness, and Cost of Services

All the State agencies interviewed by OTAreported that they had found ways of getting goodstaff-even though most pay their engineers andother technically trained people below-market sala-ries. For their small core staffs, they look for peoplewith broad technical competence (rather than depthof knowledge in a narrow field) and an interest in

working with people as well as things. The Technol-ogy Management Group in Pennsylvania calls thekind of person they look for NYTE--not yourtypical engineer. TMG reports no trouble attractingand keeping staff, even though the pay (on average,$35,000 per year in 1988) is well below the medianfor engineers. The pay in the Georgia Tech extensionservice is higher (averaging in the mid-$40,000s),but the program’s directors say the satisfaction of thejob is at least as important as pay in attracting goodpeople. Most of the extension offices are in ruralareas where the agents get plenty of local recogni-tion, both for the job they do and as representativesof prestigious Georgia Tech.

The Michigan Modernization Service relies mostlyon part-time consultants for its field representatives,and has been through some periods of high turnover.The program directors say that although it is achallenge to get good people, it can be done. The payis pegged at the State rate for consultants—$250 perday, which compares with $800 to $1,000 per day forprivate engineering consultants. The field reps takethe work despite the uncompetitive rate, partlybecause it opens the door for more contracts later,partly because the State does much of the prelimi-nary work—and also partly because they enjoy it.Some of the field reps are retired industry engineers(often from the auto industry) and they are enthusias-tic about helping small firms learn how to solveproblems for themselves. The MMS has changed itsideas about what makes a good field representative.At first, they looked for people with specifictechnical qualifications. Now they look for breadthand the ability to establish trust, listen, and write agood analytic report.

With its large roster of part-time field representa-tives (25 in 1988), MMS does not often need outsideconsultants, but other programs (TMG in particular)use private consultants quite regularly. TES reliesheavily on its faculty connections (using them for 45percent of clients), and the Georgia Tech extensionoffices taps the resources of the Georgia TechResearch Institute in Atlanta for about 30 percent ofits clients. Thus, these programs are able to tackle ashifting variety of technical problems while keepingonly a small permanent staff for continuity and asense of mission.

None of the State programs charges a fee for itsinitial assessment. Only one, TMG, charges any feeat all; in this program, firms pay apart (usually about


two-thirds) of the fee for consultants. The fact thatthe firms pay nothing for the diagnostic assessmentmakes it easy for them to enter the program, evenwhen (like the Bradhart company) they don’t havevery high hopes for it. Many of the companymanagers interviewed by OTA said they would beglad-next time-to pay for services they got fromthe extension agencies. The problem with paying upfront is that they have no idea whether the agencywill deliver professional level services. In the case ofTMG, firms get their diagnosis before they are askedto share payment for a consultant. And about 60percent decide not to go ahead (though many of theseare able to make improvements on their own, basedon the diagnosis). Those that choose to go forwardknow what their cost will be, since TMG takesresponsibility for dealing with the consultant. TheMichigan program is considering a second phase ofservice that might charge user fees, but this wouldfollow the first, no-charge phase.

The cost per client of the five programs ran fromabout $1,000 to $20,000 in 1988 (box 7-A). Thereseemed to be a rough correspondence between thecost and the intensiveness of the services clientsreceive, although it is hard to say this definitivelybecause definitions of services differ, and so doallocations of cost. MMS and TMG, both of whichemphasize field visits and individual consultationsbased on a written diagnostic assessment, are at thehigh end. MMS reported an average of 6 days ofservice and a cost of $20,000 per client. TMG saidit gave an average of 8 days of service, at a cost of$8,800 per client—but the cost rose to $19,400 forthose clients (40 percent) who elected to use aconsultant.

Maryland’s TES and the Georgia Tech extensionservice both give up to 5 days service to their clients,though neither is rigid about “setting” the clockrunning.’ TES ‘ ‘usually’ makes field visits, thoughnot always. Georgia Tech may or may not; oneregional office reported having contact with 200companies in a year, helping 100 in depth, andmaking about 50 site visits. Another said that somefield officers are so familiar with a firm after dealingwith it over the years (Georgia Tech has been in theindustrial extension business since 1960) that a sitevisit isn’t necessary. Both tend to give their clientsoral, not written reports. And both rely for special-ized technical help on their university connections,not private consultants. TES pays its faculty advi-sors for their time only when asked, and then at their

university salary (not private consultant) rates. Thefaculty advisors may then use the money forprofessional purposes such as travel or researchsupport. Georgia Tech can call on its parent organi-zation for extra services to its clients-for example,a productivity audit from the State-funded GeorgiaProductivity Center program. The TES cost perclient is nominally $1,500, but most of the cost ofconsultation with engineering faculty at the Univer-sity of Maryland is not included in this figure.Georgia Tech reported a cost per client of $4,000.

The least expensive of these programs,PENNTAP, generally offers the least intensiveservices. Often, the problems that companies bringto it are narrowly technical and can be handled by atelephone call, a fax message, or group meetings.PENNTAP’s eight staff specialists (mostly engi-neers) do make site visits as well, however, and theytailor responses to clients’ individual problems.According to the program’s director, human contactis the key to technology dissemination. PENNTAPreported spending about $1,100 to $1,500 per clientfirm, with no estimate of the days of servicerendered.

Improvements in Services Offered

Most of the people OTA interviewed, includingthe staffs of the five extension services and theirclients, thought the programs were doing a good andmuch-needed job. If there is one change they allwant to make, it is to expand the programs and servemore firms, Two of the extension services, GeorgiaTech and Maryland’s TES, specifically stated thatthey don’t advertise for fear of attracting too muchbusiness. Georgia Tech asked the State legislaturefor funds to open five more regional offices.Michigan’s service was expanding in 1989, andPennsylvania established a new $10 million-a-yearprogram of Industrial Resource Centers, replacingthe much smaller TMG, which will serve as advisorto the new centers.

At one of the programs, MMS, the staff had givenserious thought to expanding services to individualcompanies, as well as extending service to morecompanies. MMS staff members believe that theaverage of 6 days of service they now give clients isabout right for a first bite. “Small and medium-sizefirms face a digestion issue,” said Alan Baum,director of research and analysis. “They can onlydeal with so much at a time. ” But the staff isseriously considering offering a second phase of


assistance of up to 20 days, with the firm paying forsome or all of the costs (the first phase, as notedabove, is free).

MMS has another idea in mind as well. That is tostrengthen horizontal links between small firms inthe same or closely connected businesses, freeingthem from too-great dependence on the larger firmsthat are their customers. Interestingly, managers ofJapanese Government programs for small manufac-turers are promoting more independence in much thesame way, through networks that provide coopera-tive product development and marketing services(see ch. 6). Michigan’s Industrial Technology Insti-tute, of which MMS is now a part, has made somepreliminary moves in this direction. Its PRIMEproject (Program of Research in ModernizationEconomics), started in 1985, is helping Michiganauto parts and components suppliers meet newdemands from the Big Three automakers—especially the demand for complete subassembliesrather than disparate parts. For example, PRIMEmight link a small foundry with a machine shop sothe two together could make a complete camshaftsubassembly.

Finally, some of the extension services-notablyGeorgia Tech-would like to do more with training.They believe that the training programs they cur-rently offer are too “off-the-shelf” and depend toomuch on the classroom. And they think that closerlinks between industrial extension and State voca-tional educational systems are a must.

It would be a mistake to consider the examplesdiscussed above as typical of industrial extensionservices in the United States. They are not. OTAchose these five programs to examine not becausethey are typical but because they are among the mostactive and the best. The purpose was to suggest whatcan be done with technical assistance to smallmanufacturers, not to suggest that it is being done

nationwide. The situation is patchy. Several Statesbesides the four mentioned here also have activeprograms, others are following the leaders andestablishing industrial extension services, and someare doing little if anything. An accurate count is notavailable, but it is likely that the State and Federalprograms combined are spending no more than $40to $50 million per year on industrial extension. Ifjust 24,000 small American manufacturing firmswere to receive industrial extension services eachyear (about 7 percent of small manufacturers roughlysimilar to the proportion that is served in Georgia bythe Georgia Tech extension service), the total costwould be $120 million to $480 million per year,depending on the level of service.18

COMMERCIALIZINGTECHNOLOGY FROM FEDERAL

LABORATORIESDuring the 1980s, the government has tried to

encourage the commercialization of technologyfrom the Federal labs by private industry, Congresshas passed several laws to promote it; scientificadvisers to the President and executive agencieshave strongly urged it; and President Reagan signedan Executive Order laying out guidelines to accom-plish it.19 The effects have been positive but modest.The Federal labs still have a long way to go beforerealizing their potential as a source of new ideas forindustry.

When the interaction works, lab-generated tech-nologies can have an impact. For example, althoughit focuses on nuclear weapons research for the U.S.Department of Energy (DOE), Sandia NationalLaboratories has also made contributions to civilianindustry. Sandia helped to develop important cleanroom technology and the hot-solder leveler, used inelectronics manufacturing. Each was worth over

lgs~ ch. 2 for more detail on these estimates.

l~e lawspmmfig t~hnolo~ tr~sferinclude the Stevenson-Wydler Technology Innovation Act of 1980, the Patent and Trademark AmendmentsAct of 1980, the Bayh-Dole Patent Amendments of 1984, the Federal Technolo~ Transfer Act of 1986, the Omnibus Trade and Competitiveness Actof 1988, and the National Competitiveness Technology Transfer kt of 1989. Also, during 1988-89, subcommittees of the House Committee on Science,Space, and Technology and of the Senate Committee on Energy and Natural Resources held hearings on technology transfer. Major reports to theexecutive branch include Report of the White House Science Council Fe&ra/ Laboratory Review Panel, Office of Science and Technology Policy,Executive office of the fiesiden~ 1983; Energy Research Advisory Board, Research and Technology Utilizatwn: A Report of the Energy ResearchAdvisory Board to the Unite dStates Department of Energy, DOQ&O067, 1988; and The Federal Technology Tran@er Act of 1986: The First 2 Years,Report to the President and Congress form the Secretary of Commerce, July 1989. President Reagan’s order establishing guidelines for tie Federal labson technology transfer was Executive Order 12591, Apr. 10, 1987.

Chapter 7—Where We Stand: Public Policy and Technology . 185

$100 million to industry by 1987 according toSandia’s estimates.20

Technology transfer is increasing, albeit slowly.Quantitative measures are elusive and fail to capturethe key ingredient of personal interaction. Nonethe-less, some trends are indicative. Active licenseagreements between DOE’s Oak Ridge NationalLaboratory and industry were up from 2 in 1985 to33 in June 1989.21 Industry increased its royaltypayments to DOE labs from $297,000 in FY 1987 to$908,000 in the first 9 months of FY 1989,22 and arelikely to rise further.

The labs were set up mostly to pursue missionsother than commercially promising R&D-notably,basic research and the development of science andtechnology related to weapons—so there are limitsto the potential for technology transfer. However,there are also barriers that are not integral to the labsthemselves. These can be overcome. Changes in thefunding, administration, and orientation of the labsare necessary, and should help the labs to increasetheir potential contribution to increase U.S. competitive-ness in manufacturing. The following sectionsexplore how the labs are responding to legislativeand executive mandates to improve technologytransfer. While progress has been made, more couldstill be done to make labs’ research available toindustry.

The Federal Laboratories: An Overview

The Federal Government spends approximately$21 billion on its labs, mostly through: the Depart-ment of Defense (DoD), $10.5 billion; DOE, $4billion; National Aeronautics & Space Admini-stration (NASA), $2.5 billion; and National Insti-tutes of Health (NIH), $1 billion. Various smalleragencies, such as the Agricultural Research Service,account for the remainder.23 Most of this money isspent on lab work for defense and basic research.Almost all of DoD’s money and about $2 billion ofDOE’s goes to defense-related R&D, largely weap-ons development. Most of the DOE labs’ remaining

resources (after defense-related spending) are spenton basic energy research.

Neither of these two predominant missions, de-fense and basic research, is directly connected to theneeds of the private sector. Not only is defense-related R&D designed to produce weapons systems(not usually transferable to civilian manufacturing),there are security-related barriers which tilt theinstitutional culture of defense-related researchersproducers away from technology transfer into thecivilian sector. Basic research faces different, butjust as significant, problems in forging links todevelopers and users of its technology. Basicresearchers are almost by definition interested in thepursuit of knowledge, not its application. This tendsto be true for both the institution and the individualresearcher.

Nevertheless, defense R&D and basic researchcan sometimes be made useful to commercialmanufacturing. Labs differ in their potential to helpthe private sector, and in their success in giving suchhelp; they come in different sizes and with differentstructures and orientations. It is therefore useful tobegin with a brief overview and some centraldistinctions.

DOE Labs

The DOE labs are key factors in any discussion ofthe Federal labs. Indeed, the nine multiprogramDOE labs are usually simply called the national labs,even though they account for only about a sixth oftotal government spending on Federal labs.

The DOE labs are funded primarily through threeprogram areas,24 which orient the work that theyfund in different directions: Defense Programs ($3billion) supports the DOE’s weapons work andnuclear materials production; Energy Research Pro-grams supports basic research in energy, mainlynuclear energy ($2 billion); and the Nuclear Energy,Fossil Energy, and Conservation and RenewableEnergy Programs (collectively referred to below as

20An& RePo~: Te~.lo~ Tra@er, Sad”a Natio~ ~or~orles, Fiscal yew 1987, SAND 87-0749, UC-13, April 1988 (Springfield, VA;National Teehnical Information Service, 1988), p. 7; Robert Stromberg, Technology Transfer and Policy Department, Sandia National Laboratories,personal communication, June 19, 1989.

zl~~d J*, Rogm A~inis~ator, Office of T~hnolo~ Applications, M~n Marietta Energy Systems, Inc., Oak Ridge National Laboratory,personal communication, June 20, 1989.

22Rees L. Dwyer, III, Executive Assistant to the Assistant Secretary, Management and Administration. ~Partment of Ener8Y~ Wrsoticommunication, Oct, 18, 1989.

zsNational Sci~tl~Fomdatim,~~&r~ Fu~forResearch a~~eve~p~e~t: Fiscal years 198P, 1988, util$l?$l, VO1. 3’7, ~taikd Statistical Tablcs,NSF 89-304 (Washington, DC: U.S. Government Riming Office, 1989), p. 29 (estimates for fiscal year 1989) (totals of figures shown for intramuralreseareh and researeh in all Federally Funded Research and Development Centers (FFRDC s)). These figures are by agency, not by lab; agenciessometimes spend money for research in other agencies’ labs. See ibid., pp. 4, 31.

zdsae o~r a~ncies also fund R&D in DOE labs.


Applied Energy programs) support various projectsbeyond basic research that are not related to weapons($1 billion). Only Applied Energy has commerciali-zation of technology as a specific part of itsinstitutional mission.

These three programs support the work done inthree sets of labs: the four big national labs primarilyconcerned with defense-related work (LawrenceLivermore, Los Alamos, Sandia, and Idaho Engi-neering); the five medium-sized national labs thatfocus primarily on basic research in energy (Ar-gonne, Brookhaven, Lawrence Berkeley, Oak Ridge,and Pacific Northwest); and 28 generally smallerlabs (e.g., the Princeton Plasma lab). Three of thesmaller labs —including the Solar Energy ResearchInstitute (SERI)—are run specifically by DOE’sApplied Energy programs. In general, the larger labsdo some work for each of the three programs.

DOE labs are unlike nearly all the rest of theFederal labs, in that all except two of the smallerones are operated by contractors (they are government-owned, contractor-operated, or GOCOs). Almost allother Federal labs are government-owned and govern-ment-operated (GOGOs). The contractors who oper-ate GOCOS vary: some are profit-making, othersnon-profit; some are industrial firms like MartinMarietta, other are universities like the University ofCalifornia. GOCOS face some specific problems oftheir own in the transfer of technology, as we shallsee later.

DoD LabS

There are some 68 DoD labs, and DoD spentabout $10.5 billion on lab R&D in 1989. These labsare run directly by the Departments of the Army,Navy, and Air Force.

Less is known publicly about the DoD than theDOE labs, partly for security reasons. However, theDoD labs have also been under increasing pressureto encourage commercialization of the technologythat they develop. The Stevenson-Wydler Tech-nology Innovation Act of 1980, the Patent &Trademark Amendments Act of 1980, and theFederal Technology Transfer Act of 1986 clearedlegal barriers blocking transfers from these labs and

promoted structural changes (like the delegation ofkey decisions) that would encourage transfer. SomeDoD labs are clearly making a major effort in thisfield and others have historically worked well withthe private sector. However, in October 1989 DoD’sOffice of the Inspector General published a reportthat was sharply critical of the extent to which theletter and spirit of the law had been implemented.25

Other Labs

NIH spends about $1 billion in Federal labs. AllNIH labs but one are GOGOS. NIH has a goodreputation for pushing its technology out toward theprivate sector and encouraging its scientists to doso.26

NASA spends about $2.5 billion in the Federallabs. All but one of NASA’s seven labs are GOGOS.NASA’s labs (and those of its predecessor, NACA)have been productive in collaborating with industry(see box 2-A). Some of NASA’s lab work is stilluseful to civilian aircraft manufacturers, but its mainfocus today is the national space program.

NIST (the National Institute of Standards andTechnology, formerly the National Bureau of Stand-ards) sees its work with industry as part of itsprimary mission. NIST spends about$110 million inits labs. The NIST labs have long worked closelywith industry in the areas of measurement, stand-ards, materials science, and computer systems, andNIST’s Center for Manufacturing Engineering (fundedat about $6 million) follows the tradition.

This section concentrates mostly on DOE’s ninenational labs. They are big, they work on a variety ofprojects that could be of commercial interest, andinformation about them is readily available. Forthese reasons, the report uses an analysis of DOE’snational labs to illustrate the problems and potentialof the Federal labs as a whole. As discussed below,the light cast by the DOE national labs helps toilluminate the positions of other agencies and labs.While some might argue that DoD labs cannot beexpected to follow the same path toward thecommercialization of technology, there is evidencethat the defense-oriented DOE labs provide somecommercially important technology.

~U.S. ~p~rnent of ~fense, C)ffice of the inspector General, Report on the Audit ofthe DoD Domestic Technolo~ Tran$er Pro8ram, No. w-(Arlington, VA: U.S. Department of Defense, Oct. 19, 1989).

W“he one area of special interest to manufacturing--biotechnology-is the subject of a separate OTA report, Biotechnology in a Global Economy,scheduled for release in late 1990.


Commercializing DOE’S Technology:Mechanisms

“Commercialization” here means making tech-nology developed in the Federal labs useful inindustry. In the past, that typically meant nothingmore than the publication of research results inconferences and journals, after which the resultswould make their way to industry and eventuallyfind application. Today, such delay is costly, as U.S.firms fall behind in applying the latest technology tomanufacturing. In these changed circumstances,faster commercialization takes on more importance,and several useful mechanisms to promote it haveemerged. Collaborative R&D is lab-industry team-ing to create new technology for industrial use.Spin-offs and startups transfer already existingtechnology to existing and new firms respectively.Various mechanisms (e.g., personnel exchanges)can prepare the ground for either form of commer-cialization.

Collaboration

Collaborative R&D-planned, performed, andsometimes funded jointly by the labs and industry—is a powerful means of commercializing technology.It is not entirely new for the Federal labs: NIH andNIST, for example, have done collaborative workwith industry for years. 27 However, it is not at allcommon in DOE’s national labs; only 57 collabora-tive projects were under way in all national labs in1987.28

Most of DOE’s collaborative R&D has beencarried out by its Applied Energy programs. Thesehave sometimes targeted particular industries forongoing R&D projects. This continuity allows thelabs and companies to get well acquainted with eachothers’ interests, abilities, and needs, and to smoothout ways of working together. For example, SERI,which has worked on solar energy applications formore than a decade, collaborated successfully withU.S. industry in an effort to catch up with Japan inthe commercialization of amorphous silicon tech-nology for solar cells. The SERI project lasted from

1984 to 1987 and had a 3-year budget of $19 million;four firms put up 30 percent of the funds. A second3-year program, lasting through 1990, is now underway, and half of its $40 million funding is industry-supplied. In both programs, the firms are givenpatent rights and certain proprietary rights to data,enabling them to get a jump on the competition.29

DOE’s HTS pilot centers, also run by AppliedEnergy, follow the targeting model. This experimentis discussed in box 7-B.

Until recently, DOE’s Defense Programs viewedcommercialization as a distraction from its missionof supplying the military’s needs, but it has come tobelieve that the military would benefit from strongercivilian industries.30 In 1989 it funded two lab-industry consortia for work on dual-use technologies(those having both military and civilian uses). Onegroup, working to improve the quality of specialtymetals such as nickel-based or titanium alloys, willuse Sandia’s specially instrumented research fur-naces to monitor and control the production process.During 1989-94, government will provide $2 mil-lion and the collaborating companies will contribute$4.75 million. The industry share will increasesteadily, rising to 100 percent after 5 years. Thesecond consortium, the Advanced ManufacturingTechnology Initiative, will work on next-generationmanufacturing technologies such as advanced con-troller software and artificial intelligence. DOE hasfunded this project at $500,000 for fiscal year 1990,a level that will be maintained for four more years.DOE funds for the two projects rose from $400,000in FY 1989 to $1.1 million in fiscal year 1990.

Several other lab-industry collaborations for dualuse technologies are under consideration. Theseinclude projects on plasma destruction of toxicsubstances, combustion synthesis of ceramics, andceramic metal composites. However, the two proj-ects noted above will entirely exhaust DefensePrograms’ funds for such collaborations for fiscalyear 1990.

‘z~,s, Gener~ ~cout~g Offjce, Tec~loU Tra~fer: ]Wiemen~tlon st~~ of the Federal Technology Tra@er Act of 1986, RCED-89-154(Gaithersburg, MD: 1989), pp. 29-31.

z~ner~ Research Advisory Board, op. cit., p. 21.zgIbid., pp. B5-B6.soMilltW depdence on Civilian tw~olo= is &ScuSsed in U. S. Congress, Office of Technology Assessment, ~o~ing the Edge: Maint~”~”n8 fhe

Defeme Technology Base, OTA-ISC-420 (Washington, DC: U.S. Governrnent Printing Office, April 1989).

21 -700 0- 90 - 7


Box 7-B—DOE’S HTS Pilot CentersAs part of its research program in high-temperature superconductivity (HTS), the Department of Energy (DOE)

started up HTS pilot centers at three of its national laboratories—Argonne, Oak Ridge, and Los Alamos—in October1988.1 These centers are planned as new ventures in lab-industry collaboration, a conscious experiment in rapidtechnology development and transfer.

Each center has government funding of $1,6 million for FY 1989 (total $4.8 million), and $2.0 million percenter (total $6.0 million) is planned for FY 1990. In their first year of operation, the pilot centers negotiated 20cooperative R&D agreements, with costs usually shared equally between the lab and industry. Industry was readyto join in many more projects than the centers could fund.

Several features of the pilot centers are designed to expedite technology transfer. First, the centers have atransfer-oriented mission and funds to accomplish that mission. The funds are spent only on projects requested byindustry. The labs and industry plan to collaborate over the whole R&D cycle, from basic research through productdevelopment, with lessons from development fed back into research, Each center has an industry advisory boardwhich DOE consults on the substance and procedure of lab-industry collaboration.

DOE has tried to speed up the negotiation process by offering a model collaboration contract, which carriesautomatic approval with changes requiring varying levels of clearance. At first, many firms found the modelcontract’s terms unacceptable, but DOE has been revising the terms to meet the fins’ objectives. DOE also agreedbeforehand to waive rights to inventions made in pilot center research, to a greater extent than for cooperative R&Dgenerally. Also, for work funded at least half by industry, DOE allows, on a case-by-case basis, the withholdingof technical data from publication for up to 2 years. This delay, not generally allowed in DOE cost-shared research,can give the firm a valuable head-start in the market.

The HTS pilot centers experiment will be evaluated after 2 years. DOE is committed to applying the lessonslearned to cooperative R&D in other programs,

1~~ AImos Natio~ Laboratory had rwmmxmied establishing these centers, when asked by DCIE to study how to involve industryin developing HTS technology. John T. Whetten, associate director, Ims Atamos Nationat Laboratory, testimony at hearings before the HouseCommittee on Science, Space, and Technology, Subcommittee on Energy Research and Development, July 27, 1988, Serial No, 100-122, pp.90-91.

In some cases of lab-industry collaboration, DOE Spin-Offs to Existing Firmshas put up all the money, with the private companyacting essentially as a contractor. This was the case Lab work done for purely research or defensein the collaboration between Cray Research Corp. purposes sometimes turns out to have valuableand Los Alamos National Laboratory. Cray pio- commercial applications. Firms that make a point ofneered the development of supercomputers. Its first staying in touch with the latest developments, in theand crucial customer was Los Alamos, whichneeded massive computing power to simulate the government labs and elsewhere, can find out early

operation of weapons and nuclear power plants. about such promising research results and can adapt

Although Cray did most of the R&D and Los them to commercial purposes ahead of the competi-

Alamos paid for it, the lab was more than a passive tion. A firm’s own engineers are in the best position

customer, spending several person-years studying to glean research results from outside labs, because

Cray’s machines and suggesting design changes to they know their own product development cycle,

better suit the lab’s needs. The lab’s purchases were and hence the best times for incorporating new ideas.crucial to Cray’s early survival. In 1976, when the However, monitoring the vast Federal labs system iscompany was on the verge of bankruptcy,31 Los difficult even for large firms and often impossibleAlamos bought the first machine sold by Cray. By for smaller firms with more limited staff, Without1989, Los Alamos had bought 14 Cray machines, for help from the labs, they are not likely to benefit froma total price (net of trade-ins) of about $200 million. spin-off. The labs can help in several ways.

Slcray had appll~ t. the swfities and EX~h~ge Commission in 1975 for penmsslon to go pubhc, but its application WaS rejected because SECbelieved that there was no market for the Cray machine and the company would not survive.

Chapter 7—Where We Stand: Public Policy and Technology . 189

Occasionally, DOE labs have encouraged spin-offby seeking out firms to apply the technology. Forexample, Los Alamos gave copies of its CommonFile System, software that lets different supercom-puters share the same data, to several other govern-ment and commercial labs between 1980 and 1988.To ease the burden of supporting the software andalso to reach a wider audience, Los Alamos found aprivate firm to develop the software into a commer-cial product, and in January 1989 concluded anexclusive licensing agreement providing for royal-ties and continued cooperation.32

Spin-off also takes place in less formal ways.Firms with technical questions often get modestamounts of free help from government labs. Forexample, Sandia receives 600 industry visitors permonth and believes that its “free, helpful consulta-tion” with industry is “probably the most produc-tive and yet hard-to-quantify source of technologytransfer by the laboratory.”33 For example, the labhas helped in designing high-pressure glass columnsfor liquid chromatography; assisted in testing thestrength of metals; and provided manufacturers withnew types of glass that it developed for sealing tometals. 34 Sandia staff even make house calls onoccasion. In one plant visit, the lab staff showed afirm how to use new equipment to duplicateSandia’s superconductor fabrication process. Thishelp, according to the firm, ‘leaped us months aheadof schedule. ’35 In turn, Sandia staff also learn howtheir technology works in the field.

Startups

A lab’s technology is sometimes commercializednot by an established firm but by a new firm startedfor that purpose. Startups often can get a newtechnology to market quickly and they may be more

committed to the technology than established firms,but they may lack internal funding, experience inmanufacturing, plant or equipment, and distributionchannels. From 1985 to 1987, 87 startups wereformed to commercialize technologies from DOElabs. 36

Researchers may leave a government lab to heador work in the startups. Some labs encourage this bygranting entrepreneurial leave, with the right toreturn to their old jobs within a stated time.37 Theselabs see the movement of researchers into startupfirms as a good way to commercialize technologyquickly. However, some people are concerned thatlab research teams could be depleted and also thatlabs might improperly favor their own researchersover established firms for commercializing thetechnology.

Some labs have gone farther in encouragingstartups. The Tennessee Innovation Center (TIC)was formed in 1985 with $3.5 million from MartinMarietta Energy Systems, the operator of Oak RidgeNational Laboratory .38 TIC provides numerous serv-ices to entrepreneurs, including office and lab spaceand help in forming business plans and incorpora-tion. TIC typically contributes capital of $30,000 to$100,000 in return for a minority interest in the firm.Its stock in its most successful investment was worthabout $7 million by June 1989.39

Another approach is offered by the non-profitARCH Development Corp., formed in 1986 as anaffiliate of Argonne National Laboratory and theUniversity of Chicago. ARCH is given patent rightsto virtually all inventions at Argonne and theUniversity of Chicago.40 It identifies those worthpatenting, bears the expense of obtaining patents,and tries to license the inventions or, where it makes

SZC ‘Gener~ Atomics to Mtiet Los Alarnos Computer”softwme, ’ Los Alamos National Laboratory Public Affairs Office, Jan. 26, 1989; RaymondElliott, Computing and Communications Division, Los Alamos National Laboratory, personal communication, July 3, 1989.

33R. Geer, “Ta~olog Transfer Is a ~ocess of Q~et Matchm&ing, ” ~b N~s, vol. 41, No. 12, June 16, 1989, p. 1; Annual Report: TechnologyTransfer, Sati”a National Laboratories, Fiscal Year 1987, op. cit., p. 9.

sdAn~ Report: Techno~~ Transfer, San&a National L.uboratories, Fiscal Year 1987, OP. cit., PP. 16, 23-25, 28-29.35~.er from [author is ~onfidenti~] t. Dr. Dan Doughty, SuFWIWr, ~organic Materi~s, chemis~ Division 1846, Sandia National Laboratories,

June 8, 1989.sGEnergy Re~~h Advisory Board, op. cit., p. 42.3Tu.s. Department of Energy, Technolo~ Tra@ers~ ry, July 1988, p. 6; see also David Kramer, “Two Los Alarnos Scientists Forma Spin-off

To Develop New Cell-Probing ‘Tweezers,’” McGraw-Hill’s Technology Tran.@er Report, February 1989, p. 3.3EThc funding in turn came from the management fee paid by DOEsg~n~d J~ed, ~ogm A&n~istrator, Office of T~hnolo~ Applications, M~~ M~ietta Energy Systems, OA Ridge National LtdX)ratory,

personal communication, June 20, 1989.%incetheUniversity of Chicago, which operates Argonne, is a nonprofit organization, DOE waives its patent rights on request, with some exceptions.

The waiver process is discussed later in this section.


good business sense, forms a startup firm itself tocommercialize the invention. The startup’s initialcapital comes partly from a $9 million venturecapital fund managed by ARCH, but ARCH usuallywaits to get additional capital from an unrelatedparty, as an objective check on the proposedcompany’s worth. ARCH is seeking to replicate theenvironment at MIT and Stanford, which has donewell in supporting startup fins. MIT, with itsresearch budget of only $700 million and only sevenprofessional staff working on patents and licensing,produces about the same number of licensingagreements and new firm startups as all of DOE’slabs combined, with their government budget ofmore than $5 billion.41 The success of MIT andStanford owes much to the infrastructure of entre-preneurs, venture capitalists, business planners,lawyers, and bankers, which ARCH is seeking toreplicate.

Other Forms of Technology Transfer

A common and relatively simple way of makinglab technology available for commercial purposes isto let firms use the labs’ specialized facilities. Thisis not a new idea. Before World War II the NationalAdvisory Committee on Aeronautics made its wind-tunnels and other test facilities available to commer-cial aircraft companies, and NASA continued to doso after the war. Today, DOE’s national labs allowprivate firms to use an array of expensive special-purpose facilities. In 1987, about 185 scientificfacilities in the national labs were used by 1,623industry and university participants.42 As of March1989, Brookhaven National Laboratory’s two syn-chrotrons, set up as advanced X-ray sources, werebeing used by more than 80 American universities,23 U.S. fins, 14 other government labs, and 22foreign institutions.43 The Combustion ResearchFacility at Sandia National Laboratories offersspecialized lasers and computers for studying howfuels burn. Its users include General Motors, Ford,Chrysler, Exxon, Mobil, Conoco, Unocal, Combus-tion Engineering, AT&T, and GE.44

The Federal labs are also putting new emphasis ontechnology transfer in their formal communications—published papers, conferences, and so on. Several ofDOE’s national labs, for example, publish semi-technical brochures to acquaint industry with tech-nologies which may be of interest. Meetings andworkshops focused on technology transfer are in-creasingly common.

The Federal Laboratory Consortium (FLC), com-posed of representatives from Federal laboratories,also promotes communication with industry .45 TheFLC guides firms into the Federal lab system,showing them where to go for help on a particularproblem--often within a day or so of the initialinquiry. In conjunction with the Industrial ResearchInstitute, the FLC held lab-industry conferences toidentify possible areas of collaboration in manu-facturing technology (in 1988) and in hazardouswaste management (in 1989). The FLC also fundsprojects to demonstrate technology commercializa-tion. For example, the University of Utah has adatabase on specific interests of high-technologyfirms, using it to market the University’s owninventions. The FLC paid the university to adapt thisdatabase for experimental use by three Federal labs.Finally, the FLC, the Department of Commerce, andDOE all maintain computerized general-purposedatabases on technologies of possible interest toindustry. Some of the labs also maintain specializeddatabases, such as one on superconductivity at OakRidge National Laboratory.

Many of the mechanisms described above restimplicitly or explicitly on personal contact betweenlab employees and private industry, and indeed theexchange of personnel between labs and industryoffers another mechanism for technology transfer.Lab researchers can take sabbaticals or visitingpositions to spend time (perhaps a year or two) in anestablished company, and vice versa—with benefitsboth of immediately transferring information in bothdirections and developing personal contacts for thefuture. Such formal exchanges have been rare in the

41 JohII T. Reston, Director, MJT Technology Licenstig Office, ‘ ‘Creating New Companies and Business Units Within Existing Companies viaUniversity License Agreements, ” presented to the European Venture Capital Association 1987, modified April 1989; Senator Pete V. Domenici,testimony at hearings before the Senate Committee on Energy and Natural Resources, Subcommittee on Energy Research and Development, May 11,1988, SeriaJ No. 100-602 (Part 2), pp. 34.

qz~er~ ReWarch Advisory Board,, Op. cit., pp. 21, 61.dsDavid timer, “For Mre: Lab Facilities,” McGraw-Hill’s Tech Transfer Report, Mwh 1989, P.1~~bid.; An@ Report: Tec~~gy Transfer, Sandia Natwnal Laboratories, Fiscal Yeu 1987, op. cit., p. 8.45~@~ly e~~blish~ by tie ~fe~~p~ment in 1971, the ~C evolv~ ~ an info~~ coordinating ~oup ~ti] it was given an offki~ mandate

by the Federal Technology Transfer Aet of 1986 (see 15 U.S.C. 3710(e)).


national labs. In 1987 just 19 industry researcherscame to them, and 4 lab scientists went to compa-nies, in an exchange program underwritten by DOE.However, about 400 more industry scientists andengineers worked less formally at the national labsat some time in 1987, using funds from industry andDOE R&D programs. Lab researchers can also serveas consultants to industry-a practice that increasedin the 1980s (from 266 consulting projects in 1981to 697 in 1987).46

Barriers to Technology Transfer

A number of factors limit the Federal labs’transfer of technology to industry. There are prob-lems related to the labs’ historical mission, thebureaucracies that run the labs and supervise them,and the nature of technology transfer itself (espe-cially in the area of exclusive rights). Industry itselfis not blameless: for example, both U.S. universitiesand foreign corporations send more visitors to thelabs than does U.S. industry .47

Mission--The lion’s share of DOE labs’ fundingcomes through the Defense and Energy ResearchPrograms. For these programs, commercializationtends to be low priority. In contrast, DOE’s AppliedEnergy projects are usually planned with commer-cial application as an integral part of their mission,and it is on the whole accomplished effectively.However, Applied Energy has a small and decliningshare of DOE lab funding.

Funding-Technology transfer does not comecheap. Identifying technologies with commercialpossibilities, finding firms that might be interested,and exchanging information with those firms taketime and effort, but are necessary parts of aggressivetechnology transfer. Negotiating terms with firms

interested in licenses—and fighting through red tapeback at the lab or agency—takes still more effort,indeed probably requires some full-time technologytransfer staff. Encouraging startup firms can also beexpensive. Patenting is also expensive, especiallyoutside the United States. And if the labs go in forcollaborative R&D projects with industry, the labs’share must be funded--often a a level greater thancould be justified by the labs’ defense or basicresearch missions.

On the whole, DOE’s technology transfer efforthas been underfunded. Collaborative R&D hasrarely been funded outside the Applied Energyprograms and technology transfer offices have beenthinly staffed. DOE is not alone in this. DoD, forexample, has required its labs to fund technologytransfer activities out of overhead.48

Lab directors and agencies can hardly be expectedto embrace technology transfer enthusiastically ifthey have no money to pay for it, or have to rob Peterto pay Paul. Low spending is also a signal. Skimpyfunding leads companies to question the labs’commitment. 49 Dependability is important too. De-lays in expected funding have caused industry toview the labs as unreliable collaborators.5o I naddition, firms may hesitate to pledge themselves tomulti-year projects when the government will com-mit funds only year by year.

Incentives—Incentives for collaboration in thelabs are sometimes weak or even negative. Timespent answering a fro’s questions is usually timespent away from research; and help to industry doesnot always count in a researcher’s performanceevaluation, even though the law specifically directs

46Enera RexMch Adviso~ B~~d, op. cit., pp. 21.22. @Iy 45 indus~ re~~chers visited the DoD labs in 1986, while 291 visited the much SmidlerNIST (then called NBS) labs; U.S. General Accounting Office, Technology Transfer: U.S. and Foreign Participation in R&D at Federal Laboratories,RCED-88-203BR (Gaithersburg, MD: U.S. GeneraJ Accounting Office, 1988), p. 20.; Rees L. Dwyer, III, Executive Assistant to the Assistant Secretary,Management and Administration, Department of Energy, personal communication, Jan. 4, 1990.

d?David fi~er, “TriVelpl~e: Visits Give Rise to Tech Transfer,” McGraw Hill’s Tech Transfer Report, March 198~, p. 5.48u.s.~p~ment of ~fe~, Office of tie ~swtor Gener~, Report on be Audit of the DOD Domestic Technology Transfer ~ogr~, Report No.

90-006, Oct. 19, 1989, pp. 8-9.dgJo~ Whetten, act~g dir~t~, LOS Almos National Laboratory, testimony at hearings before the House COmmitti on Science> SPace~ and

Technology, Subcommittee on Energy Research and Development, July 27, 1988, Serial No. 100-122, p. 90.s~llli~ Black, Jr,, Senior Vice President, Biomagnetic Technologies Inc., testimony at hearings before the House COmmittw on scie~e~ SP~el

and Technology, Subcomrni tt= on Energy Research and Development, June 23, 1988, Serial No. 100-118, pp. 79-80; William Gallagher, manager,Exploratory Cryogenics, Thomas J. Watson Research Center, research division, International Business Machines Corp., testimony at hearings beforethe House Committee on Science, Space, and Technology, Subcommittee on Energy Research and Development, June 23, 1988, Serial No. 100-118,p. 156; Harold Hubbard, Director, Solar Energy Research Institute, testimony at hearings before the House Committee on Science, Space, andTechnology, Subcommittee on Energy Research and Development, July 27, 1988, Serial No. 100-122, p. 97.


that it should.51 Collaborations with industry maybeunattractive if the work is proprietary and theresearcher cannot publish his results. In addition,time that researchers spend on sabbatical in industryis often not counted as pensionable.

Recently, researchers and their labs have beenpermitted to keep portions of patent royalties paidfor their inventions. While the amount of money isoften modest, it does offer recognition for work thatis useful to industry .52 Some agencies and labsprovide added incentives. At least one lab (OakRidge National Laboratory) sets aside an extra 4percent of royalties to reward lab researchers otherthan those named as inventors on licensed patentsfor extraordinary contributions to technology trans-fer.53

Slow Negotiations--Speedy negotiations for li-censing of technology, and also for collaborativeR&D (which typically includes licensing provi-sions), are important to fins. They have to fitinnovations into their product development sched-ules and hold on to earmarked funding (their own orinvestors’). Delays can cause deals to collapse as thefirm’s strategic situation changes, or the peopleinvolved move on. Startups are especially vulnera-ble.

Negotiations with labs can often take manymonths. Some delay may be hard to avoid but someis caused by bureaucratic slowness and government

reluctance to grant exclusive rights. Both are largelyavoidable. Reviews by agency headquarters thatconvert two-way negotiations between a lab and afirm into three-way negotiations have often been theculprit.54 For GOGO labs, agency review of collabo-rative R&D agreements was in principle short-circuited by the Federal Technology Transfer Act of1986 55 and an Executive Order in 1987,56 whichrespectively permitted and required agency heads todelegate to lab directors the authority to negotiatecollaborative R&D agreements, subject to agencyveto within 30 days. However, many agencies havebeen slow to implement this delegation.57 Moreover,these provisions did not apply to DOE’s GOCO labs.After complaints by labs and industry about DOEred tape, Congress in November 1989 amended thelaw to permit similar delegation of authority toGOCO labs.58 DOE will probably make such adelegation. 59

Exclusive Rights—Many delays revolve aroundthe companies’ desire for exclusive rights, to helprecover the cost of expensive R&D efforts. Exclu-sive rights may also carry certain social costs,including higher prices and reduced use of thetechnology by others.60 These costs and benefitsmust be balanced case by case.6l This sort ofdecision might be made by the labs themselves,subject to agency guidelines and audits. However, inmany cases the labs’ hands are tied.

slThe F~er~ Technology Transfer Act of 1986 directs lab directors tO “ensure that efforts to transfer technology are considered positively in . . .evaluation of. . . job performance.” 15 U.S.C. 3710(a).

s2The F~er~ T~~ology Tr~sfer Act of 1986 ~lo~s re~~chers in GOGOs to collect 15 percent of the roy~ties from their patf311tS, Up tO $]~,oooper year. Many agencies, including DoD, voluntarily give inventors a greater share. The lab gets much of the rest. Many of DOE’s GOCO labs also givethe inventors a share of patent royalties. U.S. Congress, General Accounting Office, Technology Tram~er: Implementation Status of the FederalTechno@yActof 1986, op. cit., pp. 37-38; Energy Research Advisoxy Board, op. cit., p. 44; U.S. Department of Energy, Technology Transfer Summary,July 1988, p. 5.

Ssclyde HwfiM, ~sident, M~in M~ena Energy Systems, Inc., testimony at hearings before the House COmmitw on science, Space> andTechnology, Subcommittee on Energy Research and Development, Mar. 25, 1988, Serial No. 100-136, p. 45.

54JoWh ~len, &=tor, Offiw of F~er~ T~~oloa Management, U.S. ~p~ment of Commerce, ~rson~ communication, MM. 9 ~d 21, 1989.55Se 15 u,s.c< qTIOao ~s au~ori~ applles o~y to projats in which the lab contributes only personnel, se~ices, facilities, quipment or other

in-kind resources; the lab cannot pay money to its industrial partners.sGEx~utive Order 12591, Apr. 10, 1987.s~.so Congess, ~ner~ ~comung office, Techno~gy Trawfer: Implementation Status of the Federal Technolou Tra~fer Act of 1986, oP. cit.,

pp. 23-30; U.S. Department of Defense, Office of the Inspector General, op. cit., p. 10.58Nation~ Comwtitiven=s ‘r’~hnolow Transfer Act of 1989, fib~ic Law 101-189, SCC. 3133 (amending 15 U.S.C. 3710a).59D0E h~ ~Eviou~ly ~upW~~ ~ bill which ~o~d have made s~h delegation mandatory for the nation~ labs. (The bill Wm not enacted.) btter

from John Herrington, secretary, U.S. Department of Energy, to Senator Pete Domenici, Sept. 28, 1988, supporting S. 1480, as reported in Senate ReportNo. 100-544, Sept. 23, 1988. See Sec. 205.

~hese costs and benefits apply to intellectual property protection (patents, copyrights, trade secrets) in general, not just in lab-industry agreements;see the section below entitled Intellectual Property,

61ExcluS1ven@~ c~o~n~ liml~ to ap~c~~ applic~on of We t~hnology. For ex~plc, ~ engine manufacturermight be gh%tt the fSXChlSkright to use a patented alloy in engines, but be given only a nonexclusive right, or no right at all, to use the alloy in other products.


. Patents. In order for DOE labs to give firmspatent rights, DOE must generally first waivethose rights. In the past few years, labs haveexperienced long waits in obtaining waivers. Itappears that it typically took 6 to 12 monthsfrom the lab’s application until DOE approval.In early 1989 DOE’s patent counsel worked toeliminate any backlog of applications over 6months old, but the waiting times have againgrown longer pending resolution of policyissues. Some labs have complained about thepaperwork DOE requires for waivers. DOE’sview has been that it is required by statute toconsider certain factors in granting waivers.62

In the Bayh-Dole Patent Amendments Act of1984, Congress tried to cut this red tape forDOE’s labs with non-profit operators. The Actprovided that, with certain exceptions, theselabs need not apply for waivers but can simplyclaim the right to government-funded inven-tions.63 Congress specifically exempted inven-tions that are classified for security reasons (forwhich DOE rarely if ever grants waiversanyway), and also unclassified inventions atdefense-oriented labs that relate to weapons ornaval nuclear propulsion. Congress also per-mitted DOE to exempt other inventions under“exceptional circumstances.”64 After DOEimplemented this provision in its operatingcontracts with these laboratories,65 DOE haddisagreements with the Commerce Departmentand the University of California over the properscope for DOE’s ‘exceptional circumstances”exemption.

DOE has supported extending the Bayh-Dole approach to national labs with for-profit

●

●

�

operators, 66 and also to unclassified weaponsinventions unless they are designated as sensi-tive technical information-all subject to guide-lines and safeguards such as restricting the usethe operator may make of royalties.67 However,this legislation was not enacted.

Proprietary Rights. The right to keep dataproprietary may be as important as patent rightsto firms. Until recently, the Freedom of Informa-tion Act (FOIA) was a major obstacle, at leastcalling into serious question an agency’s abilityto keep secret the results of collaborative R&D.This discouraged firms from participating.68

However, Congress recently largely removedthis obstacle, exempting the results of collabo-rative R&D from release under FOIA for 5years-usually enough time to get a head startin the market.69 DOE’s organic statute (theprovisions that set up DOE and its predecessoragencies) provides that DOE should not hinderthe dissemination of technical data.70 Thecourts have not ruled on how this might applyto results of collaborative research. DOE be-lieves that the Act might apply, but only to dataactually in the custody of DOE or the lab.

Copyright of Software. Firms that developgovernment software into a commercial formor who collaborate with the government tocreate software are also likely to insist onexclusive rights. Often secrecy is not practical,as software can be duplicated once it exists.Copyright could provide effective protection.However, it is generally not possible in collabo-rations with or licenses from a GOGO, becausematerial developed in whole or part by govern-

6~e law ~~nc. ~E t. follow the ~o~~ of ~romot~g Commercialization, fostefing competition, making tie benefits of R&D widely availablein the shortest possible time, and encouraging fiis’ participation in DOE research, and to consider such factors as the firm’s investment, ability tocontribute to research or commercialization, and the need to grant rights as an incentive to participation. See 42 U.S.C. 5908.

6335 UOS,C. 202(a). Al~ou@ AT&T T~~ologies, tie AT&T subsidi~ that ~s Sandia, takes no management fee, it is considered a fOr-prOfh fi~for this purpose. .

~35 U.S.C. 202(a); see also 35 U.S.C. 200.65s~ Energy Rese~ch Advisory Board, Op. cit., p. 49.

@These include Martin Marietta Energy Systems, Inc., which operates Oak Ridge; AT&T Technologies, which operates Sandia; and EG&G Idaho,Inc., Westinghouse Idaho Nuclear Co., Inc., and Rockwell-INEL, which operate Idaho National Engineering Laboratory.

67~~er from Jo~ Hern~gt~n, s=re~, us, ~p~ment of Ener~, to senator Pete ~menici, Sept. 28, 1988, suppofiing S. 1480, as reported inSenate Report No. 100-544, Sept. 23, 1988. See Sees. 207,209.

MU,S, Congess, Gener~ ~comting office, Tech~~gy Transfer: lrr@ementation Status of the Federal Technology Tra~fer Act of 1986~ OPO CitOIp. 49; US. Congress, General Accounting Office, Technotbgy Transfer: Constraints Perceived by Federal Laboratory and Agency ~ciah, op. cit.,pp. 15-17.

69Nat10~ Competitiveness Tw~o~on Transfer Act of 1989, ~blic Law 101-89, &x. 3 133(a)(7) amending 15 U.S,C. 3710a.T~e law stites, for exmple, that ~mgements for conducting research shall not “contain any provisions or conditions Which Prevent the

dissemination of scientific or technical information except to the extent such dissemination is prohibited by law. ” 42 U.S.C. 2051.


ment employees is not copyrightable.71 Hence,officials at several labs and agencies favorchanging the law.72

This problem does not arise in GOCOcollaborations, since GOCO lab staff are notgovernment employees. However, DOE ini-tially permitted firms to copyright softwarecreated partly by a lab only if the firm agreed todeposit the source code for public inspection—which firms were sometimes unwilling to do. In1989, DOE changed its policy to permit fins,on a case-by-case approval, to make publiconly an abstract of the software.73

Additional Concerns—Even if labs and parentagencies make it a part of their mission to putgovernment research at the service of industry, andif they get funds for the purpose, other concerns stillcan stop efforts to promote commercialization un-less a strong voice within the agency favors suchefforts. Moreover, balancing other concerns, such asU.S. national security, against the benefits of com-mercialization is likely to require intra- and inter-agency coordination and indeed Presidential leader-ship.

One concern is fairness. In offering licenses totechnology and opportunities for collaborative work,labs and parent agencies try to avoid favoringparticular fins. The practical matter of avoidinglawsuits or complaints to Congress is involved, aswell as the ethical issues of fairness. But attempts tobe fair can slow commercialization.74

Also, lab-industry collaboration has the potentialfor conflicts of interest. For example, the collaborat-ing lab researcher may also have done privateconsulting for the firm, may have once worked forthe firm, or may seek royalty payments for himselfor the lab from the firm. Guarding against conflictsof interest takes careful planning and judgments.Agencies without a strong commitment to technol-

ogy transfer might prefer to avoid the wholeproblem.

Some labs, such as NIH, place a high value on freeexchange of ideas within the lab and with peopleoutside. This poses problems for collaborationsinvolving proprietary research with industry .75

National security needs may also clog the freeflow of information out of the labs. In response tothis problem, DOE’s Defense Programs office as-signed responsibilities for information security andtechnology transfer to the same staff, thus helping toensure that the two concerns are fairly balanced.Sandia did the same.76

Finally, there is the tricky double problem ofdefining a U.S. firm and determining Federal labpolicy toward non-U.S. firms.

General Applications

The story of the DOE labs has implications for allthe Federal labs, despite the differences amongthem. One is simply that technology transfer can bedone. There are some success stories from DOE,most of them rather unpublicized. Technologies didemerge from the labs and were exploited by U.S.fins, often with help from the labs. On the otherhand, the story also suggests that even the DOE labs,which have faced considerable congressional scru-tiny on this issue in recent years, have a long way togo in improving their performance.

The HTS pilot projects illustrate both sides of thestory. DOE took significant steps forward in settingup the projects and committing to apply the lessonsthat may emerge from them to other lab programs.However, the process is likely to be a slow. Theexperiment lasts 2 years, evaluation will take time,the development of DOE-wide policy will takelonger, and implementation of that policy will takelonger still. This is in the nature of the beast, andDOE should not be faulted for working methodi-

7117 LJ.s.c. 105, 101.72u.so Congess, Gener~ ~comting office, TechM~gy Transfer: Implementation Statu of the Federal Technolou Tra~fer Act of 1986* oP. cit. >

pp. 4849; U.S. Congress, General Accounting Office, Technology Transfer: Constraints Perceived by Agency ~icials, op. cit., pp. 11-12.73~cording t. DoE’s ~oul, it is Wsslble hat a cow wo~d nevertheless compel disclosure of tie entire source code under the NE’s OrgaIliC

statute. Richard Constant, Assistant General Counsel for Patents, U.S. Department of Energy, personal communication, Feb. 23, 1989. However, DOE’sposition is that publication of the abstract satisfies the agency’s dissemination requirement.

WU.S. ConHess, Gener~ Accounting Office, Technology Transfer: Implementation Statw of the Fedemf Techhv Tra~fer Act of 19~6~RCED-89-154 (Gaithersburg, MD: May 30, 1989), pp. 49-50. In general, collaborative R&D agreements contracts are not subject to the stringent fairnessrequirements of government procurement contracts.

75u.s. Congess, Gener~ Accounting C)ffice, Technology Tran@er: Constraints Perceived by A8enV o~ic~~~ oP. cit.> P. 17”76Ad Report: Tec~~~ Tramfer, Sa&’a Natio~l ~boratorles, Flsc~ Year 1987, op. cit., ~. 9.


cally. However, the process could easily take 4 to 5years-the equivalent of two generations of prod-ucts in some high-technology sectors.

Attitudinal barriers need to be further dismantledas technology transfer becomes an organic elementof most programs, rather than remaining the prov-ince of isolated specialists or even special programs(like Applied Energy at DOE). Commercializationneeds to be supported with funding, which includesfunds for appropriate people. For example, therecent Inspector General’s report on the DoD labshighlights the paucity of patent lawyers, leading toa backlog of applications and hence of technologytransfers that cannot be made until applications havebeen filed. Legal obstructions need to be addressed,such as the problems surrounding software copy-rights and DOE’s ability to maintain proprietaryinformation. Also, authority must be delegated tolevels low enough to get the job done. DOE’s labshave suffered long delays at agency headquartersand, according to the Inspector General’s report, theDoD has not delegated sufficient authority or givenpolicy guidance to a low enough level in the DoDhierarchy.

There are many reasons-financial, legal, practi-cal, and philosophical-why the gears still grindslowly in bringing new technologies out of Federallabs and into manufacturing companies. It is mucheasier for both labs and parent agencies to go ondoing things the traditional way than to tackle newproblems in government-industry interaction— suchas justifying extra funding for technology transfer,wrestling with conflict of interest issues, or negotiat-ing collaborative research. It is also evident that realdifficulties stand in the way of making the necessarychanges. The labs’ success in transferring technol-ogy will depend very much on funding for thispurpose and on the will and attitude of senior labmanagers and top officials of parent agencies, alongwith continued leadership from Congress and thePresident.

ENGINEERING RESEARCHCENTERS

The idea of creating university-based, multidisci-plinary engineering research centers (ERCs) cameout of discussions in 1983 between the NationalScience Foundation (NSF), the National ResearchCouncil, and the President’s Office of Science andTechnology Policy.

77 It was hoped that these center%

would help the performance of U.S. industry bystrengthening some of the weak links in Americanengineering: the link between engineering educationand the real world of manufacturing, the linkbetween university engineering research and indus-try engineering problems, and the links between theengineering disciplines.

NSF began to setup the program in 1984, and by1988 had funded 18 ERCs at an average of about $2million per center annually .78 (Box 7-C lists theERCs and their areas of research.) The NSF fundscover about half the costs. Industry contributes aboutone-third, and the rest comes from university, State,and local finds. Each center gets NSF funding for aninitial 5-year period, with a review after the thirdyear. If the evaluation is positive, the ERC gets 5more years of funding, starting with year four.Another review after the sixth year leads (if it ispositive) to a final 5 years’ funding from NSF—atotal of 11 years, after which the ERC has to competefor new funds with proposed centers or else findsome other source of money.

For an innovative n-year program that wasdeliberately planned with a long time horizon, it istoo early to draw definitive conclusions about theprogram’s success in meeting its goals. A fewobservations based on experience so far are inorder. 79

Overall, the centers have attracted impressivelevels of financial support and participation fromindustry—a crucial element in their success. Indi-vidual centers get from 9 to 61 percent of theirfunding from private companies, and about 420companies are taking part. However, most of the

TTMuCh of tie material in this section is drawn from ~hp Shapira, “The National Science Foundation’s Engineering Research Centers: Changingthe Culture of U.S. Engineering?” contract report to the Office of Technolo~ Assessment, March 1989. Additional material is drawn from DavidSheridan, “The Engineering Research Centers,” contract report to the Office of Technology Assessment, June 1989.

T81n January IW th.rtx more ERCS were established.T~ew ob~ryations we b- on inte~iews with NSF officials, and site visits to ERCS at four universities, including interviews with fac~ty

members, students, and industry participants in the ERCS. The universities were Carnegie-Mellon, the University of Illinois at Urbana, the Universityof Maryland, and Purdue. For more details of the visits and interviews, see Philip Shapira, op. cit.


Box 7-C—The National Science Foundation Engineering Research Centers

In June 1984, the National Science Foundation (NSF) invited proposals for the creation of EngineeringResearch Centers (ERCs). The Foundation received 142 proposals from over 100 universities. Six centers wereselected in 1985:

. Columbia University, telecommunications Massachusetts Institute of Technology, biotechnology process engineering● Purdue University, intelligent manufacturing systems● University of California-Santa Barbara, robotics systems in microelectronics* University of Delaware, composites manufacturing● University of Maryland/Harvard University, systems research

Another round of 102 proposals was evaluated in 1986; NSF awarded five additional ERCs:● Brigham Young University, University of Utah, advanced combustion. Carnegie-Mellon University, engineering designQ Lehigh University, large structural systems (for construction). Ohio State University, net shape manufacturing. University of Illinois, compound microelectronics

In 1987 three more centers were designated:. Duke University, emerging cardiovascular technologies. University of California-Los Angeles, control of hazardous wastes* University of Colorado/Colorado State University, optoelectronic computing systems

in the fourth round, 1988, four more centers were awarded:. North Carolina State University, advanced electronics materials processing● Texas A&M University, University of Texas-Austin, offshore technology for recovery of oil and other

resources* University of Minnesota, interracial engineering

Madison, plasma-aided manufacturing● University of Wisconsin—

In 1988, the third year review of the first generation of ERCs resulted in decisions to phase out twocenters-Delaware and Santa Barbara--over the following 2 years. The other four original centers were continuedfor another 5 years.

And in January 1990, three more centers were added:* University of Montana, interracial microbial proves engineering. Mississippi State University, geometrically complex field problems● Carnegie-Mellon University, data storage systems center

companies are large (over 500 employees). The operation for more than a couple of years can all citeprogram does not reach many small or medium-sizefirms. 80

Industry participation ranges from short-term helpwith specific problems, to recruitment of well-trained engineering graduates, to collaborations inlong-term strategic research (e.g., several firms areparticipating in the optoelectronics program with thetwo Colorado universities, as a way of getting intofuture generations of semiconductor manufactureand application). The centers that have been in

specific examples of technology transfer to industry.For instance, an advanced engineering design sys-tem developed at the Carnegie-Mellon Center is nowbeing used by General Motors. Most of the technol-ogy transfers so far. though, have tended to be highlyspecific technologies. For example, a performanceanalysis workstation developed at the University ofMaryland center has been commercialized by AT&T-SUN. NSF hopes that the centers will develop‘‘whole new technology systems rather than piecesof systems. ’

‘ao~ ~xception is MT’s biot~hnology program, In this field, many of’ the leading ~ornp~lcs we Srnal!


While comments by industry representatives onthe ERCs were nearly all favorable, observations byuniversity faculty members on industry’s involve-ment were more mixed. Many faculty membersemphasized that contacts with industry had a posi-tive influence on their own research, and that theprogram had established new relationships or en-riched existing ones. On the other hand, somefaculty members criticized industry’s short-termoutlook and unstable participation. Some (not all)companies seemed interested only in getting imme-diate answers to particular problems and avoidedrisky or long-term research. More generally, ERCfaculty were concerned about the constant turnoverof industry representatives, which obliges them tokeep training new industry people. Said one: ‘Thereis a constant educational process. ’

The evidence so far shows the ERCs are makinggood progress in educating engineers in new ways.They are giving students opportunities to work withindustry while they are in training; exposing them toan array of engineering disciplines and methods;giving them access to sophisticated research facili-ties; and fostering an interest in manufacturing. ERCgraduates seem to have little trouble finding jobs,and in several cases corporate sponsors have activelyrecruited students before they graduated. Some ofthe students fear, however, that they will not beproperly recognized by industry since their educa-tion has broken the mold of traditional disciplinaryboundaries.

The number of students affected by the programis still small. In most of the universities with ERCs,only about 1 percent of engineering undergraduatesare taking part in the ERC program (MIT, withnearly 14 percent undergraduate participation is anotable exception); between 2 and 14 percent ofengineering graduate students in universities withERCs are participating. And only 18 of the 280-plusU.S. colleges and universities offering engineeringeducation have ERCs.

The ERCs have far less funding from NSF thanoriginally planned, and this has caused problems forsome of the centers. Individual ERCs are getting$300,000 to $1 million less per year than expected.Some have been able to makeup the difference fromindustry contributions, but others have had to reducethe scope of research and cut funds for equipmentand students. One ERC director said that theshortfall in funding had curtailed efforts to build

relationships with smaller businesses, and forcedhim to spend more time in fund-raising and less inresearch. It is possible that the industry share of ERCfunding will continue to rise. However, companiestend to emphasize short-term projects, and theirsupport over the long term is uncertain. In a surveyby the General Accounting Office of companiessponsoring ERCs, 85 percent of respondents saidthey would continue support for the following year,but only 41 percent were willing to commit support4 years in the future. Thus, it is likely that withgreater industry funding would come less stabilityand more pressure for short-term results.

The ERC program is mostly at the research end ofthe R&D spectrum in industry. Whether it will leadto successful commercialization of new products ormanufacturing processes is unknown. On this point,there is some skepticism within the program itself.As one ERC program manager with NSF said: “Ithink the ERCs will make clear the next generationof technology systems in their particular areas ofresearch, but who in the United States will becapable of manufacturing those new technologies?’A faculty member at the University of Illinois ERCsaid: “It will be Sony, Toshiba, and other Japanesecompanies that will commercialize it. ’

Possibly the ERCs’ biggest impact on industrywill be the caliber of the engineering students turnedout. ‘‘When they move into industry,” said one NSFofficial, “those engineering students will be wellprepared to take on the engineering problems ofindustry in a real world industrial context. ” Headded: “I look for them to move into managementeventually where they will make their greatestcontribution. About half the managers in Japan havea technical background, but the proportion in theU.S. is much lower. I’m hopeful the ERCs will playan important role in correcting this imbalance. ’

TAPPING INTO JAPANESETECHNOLOGY

Until recently, U.S. industry gave rather scantattention to research results and new technologiesdeveloped in Japan, for several reasons. First, manypeople in U.S. industry were hard to convince thatJapanese technology had much to offer. This skepti-cism is now rare. Second, much of the Japanesesuperiority stems from excellence throughout themanufacturing process, and this involves things thatare hard to copy. It is no easy matter to imitate a


whole interrelated system of organizing work andmanaging people. However, many U.S. managersare trying to adopt various aspects of Japanesemanufacturing practice, and some are making head-way.

Today, interest in Japanese technology goesbeyond the factory into the laboratory. Japaneseengineers and scientists are adding strength inresearch to their proven abilities to adopt foreigntechnologies and improve on them. Thus, keepingup with research results from Japanese labs is takingon new importance.

People-to People Technology Transfer

The Japanese have long been adept at keeping upwith foreign scientific and technological research bysending people to study in other countries. For years,a great many Japanese scientists and engineers haveundertaken graduate studies in American universi-ties, attended scientific meetings in the UnitedStates, visited U.S. national laboratories, and wonfellowships in U.S. Government laboratories. Butthe flow has mostly been one way. For example, in1988 there were over 6,700 Japanese scientists andengineers working in U.S. Government and univer-sity facilities. The number of Americans working inJapanese labs was probably 800 at most.81

Several factors account for the meager presence oftechnically trained Americans in Japan. First, U.S.engineers have not been particularly eager to workin Japan. Not many speak Japanese and until quiterecently, few were interested in learning it. For thoseengineers and scientists who do want temporaryassignments in Japan, high living costs and thedifficulty of finding jobs for spouses are otherimportant obstacles. Moreover, very few U.S. com-panies or institutions have wanted to send technicalpeople to Japan for extended stays, nor do theyespecially reward scientists and engineers who have

experience in Japan. For example, MIT graduateengineers who take MIT-sponsored internships inJapanese Government, industry, or corporate labsusually find on their return that they are hired onmuch the same terms as engineers with no Japaneseexperience or Japanese language.82 However, thepersonal relationships the interns form in their yearor two in Japan may prove of great importance overthe years in learning about the latest Japaneseadvances in technology. One company manager saidthat these young people may well turn out to be theindustry leaders 25 years later.

The nature of Japanese institutions also detersU.S. researchers from doing work there. Much R&Din Japan—including some of the best—takes placein private industry, and since a good deal of thiswork is proprietary, acceptance of outsiders incorporate labs can be difficult. In government anduniversity labs, the quality of basic research has beenuneven, very good in some fields but less so inothers. Furthermore, foreign researchers’ access togovernment labs was rather limited until recently. Inthe United States, university and government labshave the reputation for consistently high-qualitywork. Positions in the United States interest foreignresearchers, and foreigners are generally welcome.Japanese scientists win many of these positions onmerit, often drawing stipends from the U.S. Govern-ment.83

Since 1962, the United States and Japan have hadbilateral exchange programs in the field of scienceand technology. The U.S.-Japan Cooperative Sci-ence Program, established by executive agreementthat year, has supported hundreds of joint seminarsand short-term cooperative research projects eversince. In the late 1980s emphasis in these bilateralexchanges shined to longer term projects and moreresearch by American scientists and engineers inJapan. A new agreement signed in 1988 reflectedthis changed emphasis.84

81Nation~ science Fo~ation, $Jt&tic~ Rese~ch s~ices.swn~r its Japan Scienw and Technology program, the Massachusetts Institute of Technology has sponsored 1-or 2-year internships in Japan since

1983. Returning interns reported to an OTA-MIT workshop in 1988 tiat, while employers took a positive view of the interns’ Japanese experience, theywere not always interested in making immediate use of that experience, or able to do SO. Representatives of American companies that support the MITprogram confirmed the point; the interns are treated like other newly hired engineers and are expected to fit into existing patterns of work assignmentand rewards. (U.S. Congress, Office of Technology Assessment, Techwlogy Transfer to the United States: The MIT-Japan Science and TechnologyProgram, background paper, April 1989).

83For exmple, 327 Jap~e~ did ~mh at the Nation~ Institutes of Health in 1986-87, compared to 72 West Germans and 68 French. StiWnds forfive out of six Japanese werepaid by the NIH, at a cost of $6,8 million; fewer than half of the Germans and two-thirds of the French got NIH stipends.See Marjorie Sun, “Strains in U.S.-Japan Exchanges,” Science, July 31, 1987.

~The A~ment BetW~n the Uniti States of America and Japan on Cooperation in Research and Development in Science and TNhnoIogY, firstsigned in 1980 and revised in 1988.


One goal of the U.S. negotiators in the newagreement was “equitable contributions and compa-rable access to each Government’s research anddevelopment systems.”85 In 1988, the JapaneseGovernment established two award programs tobring as many as 100 young (under 35) post-doctoralor master’ s-degree American scientists and engi-neers to Japan each year for research lasting 6 to 24months. Placements are in university and govern-ment labs, some of which rank as world leaders (e.g.,the Institute for High Energy Physics at Tsukuba).The awards pay for airfare to Japan, travel withinJapan, a stipend, housing and family allowances,medical insurance, and Japanese language instruc-tion. Each award is worth about $50,000 per year;100 awards would amount to about $5 million peryear.

In addition to founding these two programs, theJapanese Government also made a one-time gift of$4.8 million in 1988 to enable U.S. investigators todo research in Japan.86 The National Science Foun-dation administers the fund, using it mostly forlong-term visits for U.S. researchers (of any age, notlimited to post-docs) in all kinds of Japaneselabs—university, government, or corporate-withwhom NSF concludes agreements. For example,NSF has an arrangement with the Japanese Ministryof Industry and International Trade (MITI) to offerU.S. applicants up to 30 research spots per year in the16 laboratories directed by MITI’s Agency ofIndustrial Science and Technology.

NSF also provides awards covering tuition, fees,and a stipend for researchers undertaking intensivestudy of the Japanese language. The program isprimarily for graduate or post-doctoral scientists andengineers, but is also open to senior researchers,including people in industry; it can accommodateabout 50 people per year. In addition, NSF supportsprograms at four universities to improve the teach-ing of Japanese, and about 50 more individualstudents get tuition and stipend awards in connectionwith these programs. Altogether, NSF set aside$800,000 in fiscal year 1988 for its JapaneseInitiative programs, and $725,000 in 1989; spendingin 1990 is expected to stay at the 1989 level. Most of

the NSF funds are spent for bilateral seminars,short-term visits, and the Japanese language pro-grams.

In late 1989, NSF spokesmen said that theJapanese language programs were oversubscribedand ‘‘competitive, ” and that qualified people arebeing turned down. Participation in the new pro-grams for long-term visits and research in Japan wasspottier. NSF estimated that of the 100 placesavailable from April 1989 to March 1990 in the twoJapanese Government programs, about 60 to 65would be filled. NSF’s own program supportinglong-term visits to Japan has had 18 participantssince May 1988, but some seemingly attractive spotshave had few takers. For instance, only one of the 30slots offered in the MITI labs was occupied in 1989.None of a possible three posts in the Fifth Genera-tion project was filled (one was the previous year).Only one researcher so far has been posted to aJapanese corporate lab.

The reasons mentioned above—the high cost ofliving in Japan and ignorance of the Japaneselanguage—are still important deterrents to manypotential candidates. The age limitation may beanother; American researchers find it easier to takea year abroad when they are already established inacademic or research positions than when they arejust starting out. But a major factor may beunfamiliarity. These programs are barely more than1 year old. As their reputations grow, they could fillup, as have some of private programs that sponsorplacement of U.S. engineers and scientists in Japan.One of these is the Japan Science and TechnologyProgram of the Massachusetts Institute of Technol-ogy, which sends MIT graduate engineers andscientists to corporate, government, or universitylabs in Japan for 1- or 2-year internships. In its first6 years, 1983-89, the MIT program had 53 partici-pants (an average of fewer than 10 per year). In1989-90, it sent 47 interns to Japan.

Even assuming fairly rapid growth, all theseprograms together, public and private, will send onlya few hundred researchers to Japan per year. Addingin those who go on their own, the numbers are stillsmall compared with the thousands of Japanese

85~~er from~e HO~O~~bl~Gw~~~p. ShUltZ, swretw of Smte of the Ufited States of America, to His Excellency, Soustie Uno, Minister for ForeignAffairs of Japan, June 20, 1988; letter from Mr. Uno to Mr. Shultz, June 20, 1988. See alSO the Omnibus Trade and Competitiveness Act of 1988, whichdirected that federally supported international science and technology agreements should ensure “equitable and reciprocal” access to technologicalresearch, to the maximum extent practicable (Public Law 1O(M18, Part II, Sex. 5171, ‘‘Symmetrical Access to Technological Research”).

We gift was arranged by then Prime Minister Takeshita.


scientists and engineers who study and work in theUnited States. Moreover, relatively few Americansin other fields related to industry and technology—economics, business administration, current busi-ness experience-spend time in Japan acquaintingthemselves with Japanese management and businesspractice. A few university programs (e.g. Stanford’s)encourage exchanges of this kind by offeringintensive training in the Japanese language.

Scanning Japanese Technical Literature

U.S. acquaintance with written research resultsfrom Japan does not begin to match Japaneseknowledge of U.S. research. One reason is the idea,still current in some companies, that anythingimportant will be published in English.87 A moreimportant reason is the scarcity of technicallytrained Americans able to read Japanese. Companiesthat want to keep up with Japanese research oftencannot find someone to do it.88 Job-seekers whooffer this skill may be highly valued. For example,one American specialist with experience in scanningJapanese journals, translating titles and abstracts,and using on-line Japanese databases was hired by ahigh-technology company that told her to name herown price. The experience of this informationspecialist contrasts with that of the MIT engineersreturning from Japanese internships, whose experi-ence in Japan and knowledge of Japanese wereusually not much used or specially rewarded in theirfirst jobs back home. Companies may set a highervalue on knowledge of Japanese in a full-timeinformation specialist than in a freshly mintedengineer, whose main value to the company istechnical competence.

Government and private efforts to provide serv-ices that scan and translate Japanese technicalliterature have been only modestly successful so far.In the Japanese Technical Literature Act of 1986,Congress directed the U.S. Department of Com-merce to set up an office to provide such services.The office established to do the job is small, staffedby two people and funded at less than half a milliondollars per year, reprogrammed from other depart-

ment funds. Initially, the office arranged for transla-tions, but the service was so expensive ($60 perpage) that there was little demand for it. Services stillprovided by the office include a directory oftranslation and monitoring services, a listing ofimportant Japanese documents available in English,and a yearly report on important Japanese advancesin science and technology.

A more direct and focused effort to learn aboutJapanese accomplishments in high-technologyfields is JTECH, managed by the National ScienceFoundation in collaboration with other Federalagencies and funded at $600,000 in fiscal year 1990.JTECH sends teams of leading scientists and engi-neers to Japan to evaluate R&D in areas such ascomputer-assisted design and manufacturing ofsemiconductors, complex composite materials, andsupercomputing. Workshops at NSF discuss theteams’ preliminary findings, and the panel reportsare distributed by the National Technical Informa-tion Service. In 1989, JTECH published reports onthe much-discussed topics of superconductivityapplications and high-definition television.

Learning the Japanese Language

For the long run, broader knowledge of Japaneseamong Americans is the best assurance that scien-tists, engineers, and business managers will be ableto keep up with technological advances in Japan.And the best way to learn Japanese is to start early.Japanese school children get 10 years of instructionin English, from the elementary grades through highschool. (Though the instruction is weak in conversa-tional skills, most Japanese professionals learn toread some English.) It is the rare American highschool that offers Japanese courses, and instructionin the elementary grades is practically nonexistent.

R&D CONSORTIATraditionally, consortia have played a much

greater role in technology development in othercountries, such as Japan and Korea, than in theUnited States. Antitrust law and the prevailing free

sTone yougen@nWr, a formermA Japan fellow who now works for Hewlett-Packard, told the OTA-MIT workshop that he reads Japanese twhnicdarticles on his own, but few of his colleagues see the need. The company does not use his Japanese beyond asking him to translate occasional messages.U.S. Congress, Office of Technology Assessment, Technology Transfer to the United States from Japan, op. cit., p. 11.

ssIt mi@t be thought hat ~me Of tie JapaneW scientists and engineers who study in the United States would stay and work for U.S. ftIIIKS (SS Komnand Taiwanese researchers have done in large numbers), thus providing a source of technically trained people able to read Japanese. However, mostJapanese have been little inclined to stay in America and work for American companies, and some Korean and Taiwanese are returning to their homecountries even after many years of working for U.S. firms.


market ethos combined to make cooperative re-search appear inefficient or even illegal.

When American technology led the world, meansof improving not just the technology but the processfor creating it had little place in the public policyagenda. Yet as America’s competitive position hasdeteriorated, and a state of crisis has emerged,especially in certain high-technology sectors likesemiconductors, some now argue that R&D consor-tia are a critical element in the return to internationalcompetitiveness. The argument contains the follow-ing points.

First, as manufacturing processes become morecomplex and the technology more sophisticated, thecost of R&D rises. In particular, the sheer size of theinvestment necessary to advance to new generationsin microelectronics implies risks unacceptable to allbut a few large firms. For example, developingX-ray lithography technology runs into hundreds ofmillions of dollars. Such investments are beyond thereach of smaller firms, and even IBM is balking atthat on its own. Consortia can allow the maximumleveraging of resources, by giving a company accessto substantial R&D returns for a relatively smalloutlay. Companies can also ensure that they are in aposition to appropriate the results of the research inthat field-reducing a different risk, that they will befrozen out of a key development.

Second, U.S. industry is known for short-termthinking-which is a particular handicap in develop-ing new technology. Consortia can reorient theperspective of participants toward longer term in-vestment.

Third, there are externalities. Single firms may notbe able to capture benefits from research that wouldnonetheless benefit the community as a whole. If anumber of firms join together to do the research, therisks are spread and diluted.

Fourth, research consortia often have an impor-tant training function, even when they do not reachthe technological goals they originally aimed for.

Fifth, consortia may improve the diffusion of newtechnologies by increasing the speed or the breadthof diffusion or both, a very important attribute. Thismay be especially true for consortia designed to helpcompanies to catch up in areas of technical weak-ness.

Sixth, the creation of significant alliances andeven a consensus among participants in the face offoreign competition can be useful. In textiles, forexample, the Textile and Clothing TechnologyCorporation (TC2) is credited with developinginter-and intra-industry linkages that have strength-ened the domestic industry, even though the originaltechnological goal of the project was not achieved.

All these benefits are important. If they were theonly side of the story, strong backing for R&Dconsortia would bean obviously appropriate goal forpublic policy. But three main sets of drawbacks havebeen put forward. Some have stressed an anti-competitive and hence antitrust element of coopera-tive R&D; this argument becomes more telling forconsortia that are further downstream toward manu-facturing. Alternatively, some argue that R&Dconsortia have minimal effects-they simply don’twork and are not a useful means of furtheringcompetitiveness. Finally, there are questions aboutthe relationship of the government to R&D consor-tia.

The problem of antitrust is discussed in the lastsection of this chapter and in chapter 2. However,since R&D consortia are under discussion here, theantitrust argument is not very relevant; few peoplesee antitrust problems in nonproduction cooper-ation.

The second criticism is more cogent. Not allconsortia are successful, but some are. The problemis to identify the circumstances that make forsuccess, rather than offering simplistic generaliza-tions. Some of the key questions are:

●

●

Goals. Are consortia designed to attain somegoals more successful than those aimed atothers? For example, is basic research a moreappropriate goal than research closer to com-mercial application? Does a consortium dobetter trying to produce new technology orshould it simply focus on catching up withtechnology that exists elsewhere?Players. Who needs to be involved? Must thebiggest firms in an industry be part of theconsortium? Should all participants be roughlythe same strength or size? Should the industry’stechnology leader participate? Do consortiawith vertical participation fare better than thoseinvolving only firms from a single stage of theproduction process?


●

●

●

●

Financing. Are there optimum forms of financ-ing? Should the government help? How much?Technology transfer strategies. R&D consortiahave two primary purposes-the creation ofnew technology, and the diffusion of technol-ogy. Can successful diffusion strategies bedefined?Personnel. Firms are typically reluctant to sendtheir best people to consortia. Does this matter,given that in some consortia most scientists arehired directly rather than being seconded fromparticipants? How does this affect technologydiffusion to participating companies?Structure. Does the structure of the consortium—timeframe, forms of participation, location ofresearch labs, accrual of patent rights to insid-ers and outsiders, etc.-affect its success?

The third set of criticisms concerns the role of thegovernment. In particular, the use of governmentmoney for R&D inevitably means that the govern-ment will have a say in which technologies tosupport. Critics argue that the U.S. Government inparticular lacks the institutional capacity to makesuch choices.

This section examines some of the more impor-tant cases involving R&D consortia, focusing on theUnited States and Japan. It then offers some possibleguidelines for cultivating successful consortia.

Collaborative R&D in U.S. High Technology:Electronics

The electronics industry accounts for the majorityof joint R&D activity in the United States as well asin Europe and Japan. In the United States, most jointactivity has occurred in the last 10 years. Early jointR&D efforts in this country were centered inuniversities, mainly because of antitrust concerns.Over time, and with relaxation of antitrust prohibi-tions, more joint efforts have been undertaken byprivate companies and those tend to be targetedfurther downstream.

This section looks at three different types of jointR&D in electronics: basic research (industry-university collaboration), long-term strategic re-search (MCC), and manufacturing R&D (Sema-

tech). The following section looks at collaborativeR&D more generally in Japan.

Basic Research: Industry-University Consortia

This form of research collaboration has grownrapidly in the microelectronics industry during the1980s. 89 Usually, the projects focus on basic re-search and on training students in subjects that fit theindustry’s needs. Member firms are granted accessto all research findings. They are also encouraged tosend technical people to the university to do researchfor extended periods. They often use their universityaccess for recruitment; this may be the mostimportant aspect of cooperation for the firm. Univer-sities benefit because the extra research fundinghelps them to attract and keep faculty and graduatestudents and to upgrade their laboratories andequipment. Also, it encourages interdisciplinaryteaching and research-something that is hard toaccomplish with the university’s own resources.

Some programs are designed to promote regionaldevelopment. An example is the North CarolinaMicroelectronics Center (NCMC), which draws onuniversity faculty from the Research Triangle toconduct R&D in a center constructed and operatedin part with State funds. Member firms worktogether in vertically integrated teams: NCMC'sinitial sponsors included a semiconductor manufac-turer (General Electric), a telecommunications equip-ment maker (Northern Telecom), a semiconductormanufacturing equipment firm (GCA), and a sup-plier of manufacturing process gases (AIRCO).Although NCMC’s success at economic develop-ment has been questioned, it appears to have beeneffective in achieving technical goals.90

In microelectronics, the Semiconductor ResearchCorp. (SRC) is a key case.91 It plays the role ‘ f

broker for the semiconductor industry’s basic re-search activities. An early goal was to stem theproliferation of expensive and duplicative universityfacilities for R&D on integrated circuits, and in thisSRC had some success. Through the SRC’s techni-cal advisory boards, member firms have also ap-proached some consensus on the main technologiesto push for rapid advance. In addition to shaping theresearch agenda in microelectronics, this team-

8!?Much of the ~aten~ ~this ~tion is b- on David C. Mowe~, ‘ ‘Collaborative Research: ~ Assessment of Its potenti~ Role k the ~veh)pmentof High Temperature Superconductivity, “ contract report prepared for OTA, January 1988.

wan Dirnancescu and James Botkin, The New Alliance: America’s R&D Consortia (Cambridge, MA: Ballinger, 1986), pp. 9-10; 75.glIn 1989, SRC had 28 mernbm companies and a budget of about $30 million, $20.4 million of that from indusn.


building exercise helped lay the groundwork forSematech.

SRC has had less success in transferring results ofthe research it funds to member firms. The reasonsare not altogether clear, but a likely one is the typicaldifficulty companies find in making immediate useof basic research. Another is the separation betweenR&D and manufacturing in many member compa-nies, and a third is the lack of a reward system withincompanies for adopting ideas developed outside.

Long-Range Strategic Research: MCC

The Microelectronics and Computer TechnologyCorp. (MCC) was founded in 1982 by leaders of thecomputer industry, galvanized by the threat ofJapan’s Fifth Generation computer project.92 Theidea was to share resources and risks, and toundertake mid to longer term R&D where individualcompanies might not venture. More than at itsfounding, MCC today conducts numerous special-ized projects tailored to the needs of its members,and it is putting more effort into meeting the needsof smaller companies and into technology transfer.However, it has so far kept the ability to do somecore, longer range R&D projects.

MCC funding is almost entirely private. It was thefirst U.S. industry consortium in a non-regulatedindustry, and is a large one, with a staff of 430 andan annual budget of around $65 million. It currentlyhas 20 member firms (shareholders), drawn largelyfrom the computer, semiconductor, and aerospaceindustries .93 MCC’s five research programs areapplication-driven; they are in advanced computingtechnology, computer-aided design, packaging andinterconnect, software technology, and high-temperature superconductivity. They operate on 6-to10-year horizons, with an increasing emphasis onspinning off interim products.

MCC originally expected to draw staff from itsshareholders but the firms were reluctant for com-petitive reasons to assign their best people. AdmiralBobby Ray Inman, the first CEO of MCC, initiallyrejected 95 percent of the researchers sent by the

member companies, instead hiring highly respectedoutside scientists who were attracted by the largeR&D budgets, high wages, and the central missionof long-term R&D.94 These direct hires now com-prise 85 percent of MCC’s staff.

The structure of MCC is also an accommodationto competitive rivalries. Each of the five mainresearch programs is operated independently, andthere have been strict rules (recently somewhatloosened) about information exchanges among sci-entists across programs. Shareholders can pick fromamong the programs, joining as few as one. Thiscafeteria structure allows member firms to work inareas where they are weak and keep their strengthsto themselves.

Both MCC’s structure and the large percentage ofdirectly-hired staff have impeded the transfer oftechnology, particularly within the consortium. Inmannoted that although these factors made managingMCC much more difficult, they also helped MCC toattract and maintain a sufficient number of share-holders.

The Six-Year Mark—As MCC ended its sixthyear, evaluations were mixed. Membership washolding steady, and MCC managers believed thatexisting shareholders represent a generally solidcore of supporters. But shareholders continue towithhold their best people and their best ideas fromMCC: virtually every good research idea pursued byMCC has come from within the consortium. More-over, member firms are demanding a more immedi-ate bang for their buck, and some have said they arelooking to lower their contribution to MCC. (Mostof the shareholders pay at least $1.5 million per year;some 20 associate members pay annual dues of$25,000 for limited access to MCC research.)

For members, a basic problem is the dearth ofclearly usable research results. Only three commer-cial products have resulted from MCC technology .95However, some firms also use MCC technology lessdirectly, as Honeywell did to develop an internalproduct designed to place components on a multilay-

X2The ~dW~ gl~~, ~M and AT&T, did not join, possibly fOr antinst re~ns.g3Me~r accomt for one.h~f to t~o-~irds of ~1 firms in Mose industries, ad most have R&D budgets of $100” million or more. Merton J. P~k,

“Joint R&D: The Case of Microelectronics and Computer Technology Corporation, “ Research Policy, 15, 1986, pp. 224-225.gdInterview M* ban, NOV. 1, 1989.9SNCR cow. rwently in~~uc~ ~sl= Advisor, an Cxwfi system for ~tegrat~ circuit designers b- on MCC’S work in artificial intelligence.

The consortium has also licensed its laser bonding technology, a technique for comwting the leads of semiconductor chips to the circuit board. Mostrecently, the Digital Equipment Corp. (DEC) amounced plans to use MCC’S tape-automated bonding technology in one of its VAX computer systems.


ered printed circuit board. Boeing has set up fourlabs in Seattle to develop technologies that it takesfrom MCC.96 Other benefits are apparent but hard tomeasure. For example, access to MCC has allowedmember firms to delay capital investments and thenmake the right ones when the time comes. And evennegative results help shareholders to avoid blindalleys.

Nevertheless, MCC members and executivesalike feel that the consortium should be spinning offimmediately usable technology even as it pursueslong-term projects. This pressure for results has beenintensified by a change in the corporate level ofinteraction with the consortium; MCC executivesrefer to this as the ‘‘kings, dukes, and baronsprogression. In place of CEOs with long-rangevisions (the founding members, or kings), responsi-bility for interacting with MCC has migrated down-ward to the managers of profit-and-loss centers(dukes and barons) in many member firms. Thesemanagers have much more immediate needs, andgenerally press for nearer-term payoffs.

A few shareholders--Digital Equipment Corp.,Control Data Corp., and Boeing, for example-havemade major technology transfer efforts. DEC spendshalf again its investment in MCC seeking ways touse the consortium’s results.97 But others largelyignore MCC. Scientists at MCC describe someshareholders as “black holes’ because of thedifficulty of locating-and then maintaining contactwith-the appropriate recipient for a particulartechnology. “Too many [shareholders] are waitingaround for a virtual product design to emerge beforethey examine what’s happening and why they mightuse it, ” Inman observed after he left MCC.98 Inaddition to diverting MCC resources away fromlong-term research, this demand for neatly packagedresults creates tensions, according to one programmanager, because ‘the weak sisters want us to bringthe technology damn near to market,” while thestrong ones don’t.

Some technology transfer problems arise in strongfirms as well as weak ones. For example, MCC’sCAD program serves a group of semiconductormanufacturers who have become increasingly de-pendent on the emerging software vendor industry.When MCC gave CAD members research algo-rithms instead of completed software tools, it was“like feeding grass to tigers,” according to MCC’schief scientist, John Pinkston.99 The CAD programdirector stepped down and the program was substan-tially reorganized. A similar problem occurred withMCC’s much-praised laser-bonder. Most MCC share-holders could not use the technology in ‘raw’ form.It was eventually licensed to a non-member firmwith the sophisticated capacity to make use of it.

Mid-Course Corrections-MCC has changed itsstructure to combine shorter with longer termprojects. l00 The CAD program and two others wereeach reorganized into a core unit working towardlong-range goals, plus several satellite projects,designed to produce ongoing results for sharehold-ers. In these programs, shareholders must buy intothe core project and at least one satellite. However,MCC’s Advanced Computing Technology (ACT)program (by far the largest) recently eliminated thecore structure altogether. A shareholder can nowselect from 12 medium-term projects—includingneural networks, optical computing, and artificialintelligence—at an annual price of $125,000-$700,000 apiece plus a one-time fee of $250,000 foraccess to the program. Although some of ACT’s$1.5 million contributors are sure to trim theirinvestment, MCC hopes that new participants willmore than offset that loss.

MCC has begun to seek government money incases where shareholders fail to exploit its researchor where the government funds complementaryresearch. For instance, the shareholders did not pickup the parallel processing work of the advancedcomputing program, so MCC instead attracted a $6million DARPA contract. Toward the end of 1989,MCC estimated that government contracts would

WMamgemeti Review, February 1989* P“ 26-~~ong o~er ~ngs, DEC rwufies hat evew MCC project it supports have an individual sponsor within the company. BY including tie funding

for external R&D in the budget for internal research projects, DEC encourages managers to pay close attention to the work of consortiums. ScientificAmerican, May 1989, p. 100. DEC also works hard to put researchers returning from a tour with MCC into positions where they can help the companythe most.

g~r~ Guterl, “MCC: The Dilemma of Joint Research,” Business Month, M~ch 1987, p. 50.

991ntetview with Pinkston, MCC, May 12, 1989.l~m ~ discmsion of MCC’s rargafization, ~ J. Robefi Lineback, ‘ ‘Mcc, After Five yews Of R&D, Refocuses To Em Its K&p,’ E/ectrom”cs,

December 1988.


grow from 2 or 3 percent of its budget to 10 to 15percent in 1990. MCC now also does proprietarywork for individual member firms. The packagingand interconnect program (MCC’s most successful)runs such projects for five of its seven shareholders.These projects exploit other ongoing research andcurrently total less than 20 percent of the program’seffort.

MCC is putting more resources into transferringits results. In 1988, for example, there were some 80technology transfers to shareholders compared to ahandful in 1985. Fully a quarter of MCC’s budgetnow goes into technology transfer activities. Other,more qualitative changes include relaxing the barri-ers between programs, formal voting on programresearch to increase shareholder commitment toMCC’s work, and attempting to increase the share-holder portion of MCC’s staff to 35 percent. On themembers’ side, most do keep some people on thepremises in Austin-and not just to see what othershareholders are up to, as in the early days, but to doreal work. Although staff seconded from sharehold-ers are still a small minority, those who are assignedthere could be used to transfer technology back tothe company.

MCC’s Future—The shift in MCC toward moreclient-centered and more immediate results is in parta response to the needs of weaker members. Onefunction of MCC is to help C companies become Bcompanies, or help A and B companies strengthenweak areas. Thus the trend toward shorter term,more specialized R&D has positive aspects. At thesame time, the trend could upset the balance betweenMCC’s original goal, to take on long-term andrelatively risky research, and the need to generateproducts that are more immediately or more nar-rowly useful to members of the consortium. It is thisbalance that distinguishes MCC from institutionsthat are devoted mostly to serving individual cus-tomers with proprietary R&D.

Manufacturing R&D: Sematech

SRC and MCC notwithstanding, microelectronicsindustry observers were skeptical about the 1987announcement of a proposed manufacturing re-search consortium to be funded equally by industry

and government. Twenty years of intra-industrycompetition would not be easily set aside. Thenewfound cooperation was partly based on fear, asJapanese inroads into the market for dynamicrandom access memory (DRAM) chips—the work-horse of the chip business—threatened U.S. firms’very existence.

Despite the similarities between MCC and Se-matech, including a handful of common members,there are major differences. The Federal Govern-ment, consciously excluded from MCC, is a fullpartner in the 5-year chip consortium: DARPA iscontributing $100 million per year, roughly half ofSematech’s budget.lO1 More important, Sematech’sfocus is narrower and more applied; it goal is todevelop 0.35 micron manufacturing technology by1993. Sematech’s membership is relatively homoge-neous: 14 semiconductor manufacturers, both mer-chant and captive, which together represent 80percent of U.S. chip production capacity. Sematech’s15th member is Semi/Sematech, an organization ofU.S. equipment and materials producers.

Sematech’s members include IBM and AT&T,the industry giants that have stayed away fromMCC. Sanford Kane, then IBM’s vice president forindustry operations, explained his company’s ration-ale for supporting Sematech:

The survival of the U.S. semiconductor industrywas critical to us for several reasons. Number one,we were one of the largest purchasers of chips in theworld. We liked to source locally, and we didn’twant to be in a position where we had no choice butto be dependent on our competitor. Second, IBM wasthe largest manufacturer of chips in the world. Weproduced in-house those chips that gave us atechnological edge. In order to stay state-of-the-m%we needed to have sophisticated equipment to makethe semiconductors. If the U.S. chip makers go, sowould the U.S. equipment companies. We knew itwould be difficult to establish close relationshipswith the Japanese, especially since most of theirfirms are associated with chip companies. We wouldbe forced to share information and it would bedoubtful whether we could get access to state-of-the-art equipment as quickly as our Japanese counter-parts. l02

IOl~D’S SUm~ is sCh~Ul~to end in 1993, Ro~~ Noycc anticipates ~at, if Sematwh is successful, kdus~ will continue to fund the effOll ~lhOUlgovernment support, albeit on a smaller scale. If industry is unwilling to fully fund Sematech after 1993, he maintains, it should be ended. “I’m a fmnbeliever in sunset provisions,’ says Noyee. Interview with Noyce, May 11, 1989.

lm’’semat~h,” Harv~d Business School Cm #N9-389-057, 1988, p. 10.


So important is Sematech’s success to IBM andAT&T that the two firms shared their respective 4megabit (M) DRAM and advanced 64 kilobit (K)SRAM processing technologies to the consortium,along with the engineering support necessary to getthem into operation. These contributions have al-lowed Sematech to establish baseline manufacturingwith 0.7 and 0.8 micron technology less than a yearafter moving to its Austin facility. Although Se-matech’s fabrication facility (fab) will turn out onlya few hundred wafers a day, just a fraction of acommercial fab’s output, the consortium considersthat sufficient to achieve rapid process learning.Sematech’s strategic plan calls for high-yield, pilotapplication of 0.35 micron processing technology anestimated 6 to 18 months ahead of leading foreignchipmakers. 103

Three strategic objectives are central to achievingthis goal: 1) improving suppliers’ technologies, 2)improving chip makers’ manufacturing skills andtechniques, and 3) strengthening the manufacturingtechnology base for semiconductor production.

Objectives-First of all, Sematech must strengthenU.S. materials and equipment suppliers. The indus-try includes hundreds of small supplier fins, mostwith sales of less than $10 million per year. Thesefirms have traditionally had an arm’s-length, andoften adversarial, relationship with semiconductorproducers, who preferred to keep their chip designsand manufacturing processes secret. A recent reportdescribes the industry’s situation:

Compared with captive equipment makers inintegrated Japanese and European electronics firms,U.S. equipment makers lack the advantages ofpredictable internal markets, access to broad scien-tific expertise, and deep pockets for high-cost R&D.They also lack the opportunity for joint developmentand internal site testing of new equipment, and thebenefit of systematic high-quality feed-back onproduct performance.l04

Sematech represents an attempt to overcomesome of these structural handicaps. Through com-petitive R&D contracts to selected suppliers, it willtry to promote long-term alliances between chip-makers and suppliers. Although chip producers are

the source of two-thirds of the innovations insemiconductor equipment, they have traditionallykept these innovations secret, so as to preserve theircompetitive advantage in process technology.l05

The consortium structure encourages chipmakers toreveal their secrets to equipment producers.

By awarding contracts to multi-company teams,the consortium is also trying to promote cooperationand consolidation among suppliers. For example, ateam composed of three rivals in high purity gastechnology was recently awarded a contract todevelop gas pipelines, filters, and other technologyfor Sematech.

The R&D funds awarded to selected supplierswill also be important, although probably less sothan the knowledge it generates and improvedrelationships with chipmakers. Sam Harrell, presi-dent of Semi/Sematech, expects over half of Semat-ech’s budget to filter down to equipment andmaterials firms in the first few years of the consor-tium. lO6

Finally, Sematech will provide a high-quality betatest site, where suppliers can test run their newequipment and processes under realistic manufactur-ing conditions. Currently, a supplier firm must testits equipment on an actual production line; since itcan take weeks to debug a new piece of equipment,chipmakers are understandably reluctant to act asguinea pig. This test facility will also serve to certifythe quality and composition of chemicals and otherinputs.

Sematech’s second objective is to improve manu-facturing skills and techniques among chipmakersthemselves and, in so doing, to change the veryculture of semiconductor manufacturing in theUnited States. Sematech’s director of strategicplanning, A. S. Oberai of IBM, describes the problemthis way:

In Japan, engineers spend 70 to 80 percent of theirtime on “continuous improvement programs. ” Theprocess operator is king-the first line of attack. It ishe who keeps the equipment in order and decideswhen to call in the engineers. In the United States,engineers spend 70 to 80 percent of their time oncrisis management as opposed to crisis avoidance.

1~’lle consortiums interim goal is to apply 0.5 micron manufacturing technology by 1990, roughly even with leading fomi~ chipmakers.1~’’sem~h: ~o~esws and prospwts,” Re~rt of the Advisory Council on Federal Participation in Sematech, 1989.IOS~c Von Hip@, The SoWce~ of Innovation (New York, NY: Oxford University Press 1988).1~’’sema~h,” Harvard Business School, op. cit., p. 12.


The system encourages that by rewarding doers—problem solvers—rather than problem avoiders. [Incontrast to Japan, in the United States] processoperators do no maintenance or planning-they justpush wafers.107

Sematech should also provide an arena for thecooperative development of standard equipmentinterfaces-a key problem up to now—and mem-bers will be able to use their substantial marketpower to get those standards accepted industry wide.The ultimate goal is a computer-integrated manufactur-ing system, which will provide diagnostic informa-tion about chip manufacturing to a computer in astandardized format.

Sematech’s third objective is to strengthen themanufacturing technology base. Through expertworkshops, the consortium tries to identify the mostpromising paths to its various technical goals. Inaddition, 10 percent of Sematech’s budget is goingto 11 university centers of excellence,108 whoseactivities are being directed by the SRC. The centersare conducting research in a limited number of areasthat will be critical to Sematech’s post-1990 activi-ties: for example, contamination/defect assessmentand control, and advanced plasma etch processingtechnology.

Preliminary Assessment—Sematech got off to arocky start. Member firms clashed over the kindsand volumes of chips to produce.l09 There wereproblems in recruiting a CEO, and DARPA rejectedthe consortium’s first operating plan. However, aftera year-and-a-half of operation, Sematech has madesignificant progress. It has built a world-class cleanroom and fab in less than half the normal time and

at lower cost. In March 1989, several days ahead ofschedule, the consortium produced its first chips,using AT&T’s SRAM technology.

Member company assignees to Sematech aregenerally high caliber, and the consortium has,unlike MCC, achieved its goal of balance betweenassignees and direct hires. 110 There is other evidenceof members’ commitment. National Semiconductorhas built a pilot production line to apply the toolsthat Sematech is developing.lll And some supplierfirms appear to have made preliminary plans tolocate R&D and production facilities in Austin.112

According to Sematech officials, members’ com-mitment went ‘from casual to urgent” following theFederal Government’s decision to participate.113

Whether or not DARPA’s financial contribution iscritical, 114 Federal participation is certainly impor-tant symbolically: it gave Sematech credibility andencouraged industry members to believe that gov-ernment officials will take seriously their concernsabout unfair Japanese trade practices.

The participation of IBM and AT&T is at least asimportant. Their contribution of leading-edge tech-nology represents an enormous benefit to merchantfirms-one that has probably outweighed any coststo them of Sematech membership. However, the realtest of commitment will come only later, whenmerchant firms will receive much less relative totheir contributions, financial and otherwise.

Even now, not everyone would agree that Se-matech is a conditional success. One concern is withSematech’s decision to limit actual production ofchips in its fab, in keeping with strong oppositionfrom IBM and Texas Instrument to high-volume

lm~temiew with A. S.Oberai, May 10, 1989.10SAS of January 1990.109Sme ~~ysts ~We that, in Choosing [email protected] C h i p s , Sematech is attacking the ~ong problem. mey see the indus~’s fUtUR h

application-specific integrated circuits (ASICS), which are custom-made in small quantities.11OAccording t. LW Novak, a Texas ~s~ents assign= and sema[~h’s dir~tor Of l~hnology ~ansfer, an msignmenl to sematech iS .CXXn a.s

“career enhancing. Slow growth of the industry has kept many employees from moving up the career ladder in their home companies. Sematechprovides an alternative ascent route<

lllN~ York Times, July 2, 1989.1 lz’’sema~h: %o~ess and Prospects,” Op. cit., p. ESA$.113~id,

114claUde B~le]d, Americ-nEnt@se ~tltute, inte~fiony ~fore the Joint ~onomic committ~, J~e g, 1989, argued that industry was pEpWdto fund Sematech on its own. However, Robert Noyce argues that semiconductor companies can ill afford the $100 million per year they conrnbute, sincetheir profits are among the lowest in manufacturing.


production. 115 Some critics believe that a high-volume operation is essential for testing yield andreliability, and that frictions between design andmanufacturing teams can otherwise be swept underthe rug.

High-volume production would require attentionto every step in the process chain instead of thecurrent selective emphasis, principally on lithogra-phy. That would cost considerably more money thanSematech has available. Some believe that theconsortium’s budget is too small to ensure success.Financial constraints do force Sematech to place itsbets on a limited number of technologies aimed atachieving 0.35 micron circuitry. If those bets provewrong, as one member company liaison said,“Sematech will have bought the farm.”

Even if Sematech’s technology wagers pay off, asmost experts expect, the consortium still faces majorproblems. Technology transfer is one: Will membercompanies actually adopt the manufacturing toolsthat Sematech develops, let alone the new ‘culture’of semiconductor manufacturing? Similarly, will thenew closer relations between chipmakers and theirsuppliers be sufficient or lasting? Finally, will all ofSematech’s work translate into significantly greaterU.S. market share?

Cooperative R&D Ventures in Japan

In a nation where corporations are famous for theferocity of their competition, the continued use ofcooperative research ventures could not have comeabout by accident. Several factors have combined toproduce the level of joint research activity seen inJapan.ll6

First, the nation has a tradition of governmentefforts to promote cooperation between competingfirms that dates back to the beginning of industrializa-tion in Japan. This history has conditioned firms to

accepting joint R&D.117 Second, the stability offirms within industries fosters the development of acertain level of trust. Wakasugi describes member-ship in a research consortium as ‘effectively perpet-

ual."118 Third, firms want to participate becausethey are afraid of letting rivals gain a competitiveedge. 119 Companies do not invariably join R&Dconsortia when invited; 120 however, reluctance toflout powerful, respected government agencies suchas MITI, combined with fear of missing the bus,usually win out.

On the government side, Japanese ministries areconstantly engaged in turf battles. One way for themto gain size and prestige is to become the promoterof more and more cooperative R&D ventures. TheScience and Technology Agency established theJapan Research and Development Corp. in 1961, thesame year MITI started the Engineering ResearchAssociation program (ERA). When MITI announceda 10-year biotechnology research consortium in1981, three other agencies responded with their owncooperative biotechnology projects. Governmentagencies provide substantial financial inducementsfor joint R&D, including loans whose repayment iscontingent on the venture’s success, rapid deprecia-tion of equipment, R&D tax credits, and outrightgrants.

Finally, Japan’s legal climate is extremely favora-ble to cooperative ventures. One of the clearestexpressions of Japan’s attitude towards antitrust isembodied in the regulation that, even if the JapanFTC feels they have a legitimate case, they cannotact if it would “cause a loss of internationalcompetitiveness” for that firm.121

Private cooperative research ventures are com-mon in Japan, but they usually do not involvegovernment participation. Although fully one-thirdof industrial R&D is collaborative, 90 percent of that

1ls~g~e Plaming ~age5 of Semat=h, may indus~ offici~s argued that a high-volume production operation wfi essential foresting yield and

reliability, but Texas Instruments did not want competition in the DWM market. IBM also opposed high-volume production for a variety of reasons.Sematech members eventually agreed on a facility capable of high-volume production but with an actual output of only 200 wafers a day. New YorkTimes, Mar. 5, 1987.

llsThissectiondraws p~fily on the following sources: Mark Eaton, ‘‘MIT1 and the Entrepreneurial State: The Future of Japanese Industrial Policy,’unpublished monograph, 1987; George R. Heaton, Jr,, ‘‘The Truth About Japan’s Cooperative R&D,’ fssues in Science and Technology, fatl 1988; JonahD. I.Awy and Richard J. Samuels, “Institutions and Innovation: Research Collaboration as Technology Strategy in Japan,” MIT Japan Science andTechnology Rogram, WP 89-02, April 1989; Daniel I. Okirnoto, “Regime Characteristics of Japanese industrial Policy,” Japan’s High TechnologyIndmtries, Hugh Patrick (cd.) (Seattle, WA: University of Washington Ress, 1986); Richard J. Samuels, ‘ ‘Research Collaboration in Japan,” MIT JapanScience and Technology Program, WP 87-02, 1987; and Ryuhei Wakasugi, “A Consideration of Innovative Organization: Joint R&D of JapaneseFirms,” Shinshu University, Faculty of Economics, Staff Paper Series 88-05, March 1988.

llT~~ and Samuels, op. Cit. p.%’l18~@i Wbugi, 1$)87, cited in by and Samuels, Op. Ck, PI 38.

119Kom YmmUa, ‘Joint Rese~chand ~tl~t: Japanese vs. Amefican Swategies, ’ J~an’sHigh Techrwlo~Industries :Lessons andLim”tatwnsof Industrial Policy, Hugh Patrick (cd.) (Seattle, WA: University of Washington Ress, 1986), p. 187.

l~smue~s (op. cit., p- 39) Offem sever~ exmp]e5 of leading fms who shunn~ jo~t re~~ch when they believd they were f~ ahead of their fivdS.

lzlyamua in Patrick, 1986, p. 196.


is simply two-firm contracts between users andsuppliers (i.e., firms that would not compete any-way). Only one-fifth of all joint research----or about6 percent of total industry R&D-occurs betweenrival firms. These are the cases in which governmentparticipation is most common; because of competi-tive pressures, such alliances tend to succeed onlywith government sponsorship.

The typical vehicle for a joint project involvingrival firms is the Engineering Research Association(ERA). The kenkyu kumiai ho (Cooperative Re-search Act), passed in 1961, gave ERAs the samelegal standing as industry associations. Early ERAswere designed to help small and medium-sized firmscatch up technologically, and so aimed to import anddistribute technology, rather than develop it fromscratch. In the early 1970s, ERAs entered a newphase. Although technological catch-up was still thegoal, large firms began to play a bigger role and thefocus shifted to more advanced research and producttechnologies. Critical to this change was a new MITIpolicy encouraged use of ERAs for “large-scaleprojects” involving research too extensive for anysingle firm to undertake. The projects (31 to date)were initially designed to meet specific goals,including creation of a prototype in some cases. Therisk was primarily financial, since the technologiesthemselves had almost all been proven in the UnitedStates or Europe. A celebrated example of this kindwas MITI’s VLSI project (1976-79), which helpedJapan’s electronics firms master the manufacture oflarge-scale digital integrated circuits.

Around 1980, cooperative R&D in Japan entereda third phase, as the Japanese Government began toshift from catch-up to state-of-the-art projects. TheFifth Generation Computer Project, a successor toVLSI, is a 10-year national project focused onartificial intelligence and other leading edge tech-nologies designed to make computers far moreaccessible to untrained users. Large-scale projectshave become increasingly risky: for example, opto-electronic elements had not been proven when theOptical Measurement and Control System Projectbegan in 1979. Similarly, the Next GenerationIndustries Program has turned increasingly to uncer-tain technologies such as bioelectronic integratedcircuits.

As cooperative R&D in Japan entered its thirdphase, new vehicles were set up to promote coopera-tive research, and existing institutions are evolvingto meet the new challenges. MITI and the Ministryof Posts and Telecommunications established theJapan Key Technology Center (KTC) in 1985 tofund promising proposals from research consortia.Much like a venture capitalist, KTC buys equityshares in the new firm-up to 70 percent of totalcapitalization. The research group, not the govern-ment, retains all patent rights.

Like ERAs, KTC consortia will have initial termsof 7 to 10 years and budgets for term of around $100million. Unlike most ERAs, KTC projects mustbreak new ground in basic or applied research, andthe work is more likely to be conducted in jointfacilities. 122 So far, KTC has provided more than$250 million in capital to 61 research projects. Eatonsees the program as a watershed: “It is difficult tooverstate the significance of the KTC . . . It signalsa new willingness by the Japanese, led by the state,to risk resources for basic industrial research. ’’123

For all its contributions, joint R&D has not beenthe primary means of technical advance for Japaneseindustry. It has always complemented rather thandominated the research that companies were simul-taneously doing in their own labs. Nevertheless,cooperative R&D ventures have proved technologi-cally significant, especially in electronics. DuringJapan’s period of catch-up, they provided an effi-cient way to rapidly raise the overall technologicalbase of Japanese industry. As research consortiashift their activities to more exploratory research,success will be less predictable and more elusive,since uncertainty is the price of doing things that arereally new. However, consortia do have the virtue ofspreading the risks in uncertain ventures.

R&D consortia have also helped to speed thediffusion of technology between Japanese compa-nies. Getting firms to share technology can bedifficult. Though intellectual property laws areweaker in Japan than in the United States, they doprovide some protection. The problem of technologysharing can be avoided if all the major playerparticipated in its development in the first place.That way key capabilities are less likely to becomeproprietary, and the overall level of technological

l~Eaton, op. cit.l~Ibid.


competence rises faster. No major firms are leftbehind in the technology race, and more firms meanmore competition. (For further discussion of compe-tition and technology diffusion, see the sectionIntellectual Property.)

As for the Japanese government’s part in R&Dconsortia, the idea that Japan’s technology advanceis driven by a massive, government-directed program--a view fairly widely held in the West at one time-isuntrue and largely discredited. It is a mistake,however, to underestimate the government’s influ-ence. Although government’s annual spending perproject is typically rather modest (the $300 millionMITI spent over 4 years for VLSI was unusuallyhigh), the government commitment is steady andlong-term, and this counts for a great deal.

Moreover, the fact that the government enters intorelatively few R&D consortia should not be taken asa sign that the government’s role is insignificant.The projects in which government participates arecarefully chosen. Usually they are the upshot ofcontinuing discussions between government agen-cies and business councils. They are consistent withthe “vision,” also developed by government andindustry, of what kinds of technologies and indus-tries are essential to the Japanese nation. Projects inthe 1970s and 1980s were chosen for their contribu-tion to the Japan’s becoming a knowledge-intensivesociety.

Thus, the government’s choices are both strategicand symbolic. They also give a signal. Private banksand financial institutions follow MITI’s lead. Andfunding of joint R&D is only one of a whole raft oftools at the government’s disposal for supportingstrategic technologies. Besides the special loans,grants and tax breaks companies can get as induce-ments for joining R&D consortia, they may also getsimilar benefits from government programs in thecommercial development that follows.

Making Successful Consortia

The innumerable factors (ranging down to thepersonalities of key participants) that affect theoutcome of R&D consortia prevent the developmentof a recipe. Nonetheless, it is possible to offer someguidelines.

Because companies in the same industry areprimarily competitors, minimizing conflict betweenconsortium participants is critical. It is always

difficult to get participants-often with long histo-ries of competitive relations-to work together onanything, although successful cooperative researchdemands that they do. Consortia which fail to reduceconflicts to workable levels either collapse duringthe planning stages or find their effectivenesssharply reduced.

Conflicts can be avoided in more than one way.First, the evidence from Japan suggests that coopera-tion is more easily established when a technology isalready known. Catch-up consortia have the advan-tage of avoiding certain conflicts by definition; forexample, the participants need have little fear thatany monopoly-creating technological breakthroughis at stake. Catch-up consortia also benefit from theirclear goals, which make them inherently more likelyto succeed than new technology consortia. Thiscomparison by itself is misleading, however. Catch-up consortia should be compared with other catch-upmechanisms, not with new technology consortia.Likewise, new technology consortia should becompared with other mechanisms of technologicalinnovation.

It may also be easier to avoid conflicts when acooperative project is aimed at goals far from thecompetitive arena. For this reason, some claim thatR&D consortia should be aimed at basic industrialresearch rather than applied research. Yet if partici-pants can agree on well-defined areas of precompeti-tive research, they can overcome the potential forconflict. Indeed, international competitive pressurescan be so strong in some cases that participantsbecome exceptionally interested in making majorcooperative R&D efforts that go right through pilotproduction (e.g., Sematech) perhaps even into com-mercial manufacturing (e.g., Airbus). Moreover, theresults of applied research may be more useful toconsortium members than yet another increment ofbasic research-in which the United States isalready strong.

Conflict is not the only impediment to success.Ultimately, success comes only when the productsof a consortium are adapted and integrated into themainstream of participating fins’ operations. Thereare several ways of encouraging diffusion of theresults from cooperative R&D. Most important, asubstantial financing commitment from participantsseems to be necessary. Firms pay attention whenenough of their own money is at stake. Of course,defining “enough” is possible only case by case;


Sematech defined it as 1 percent of each fro’srevenues. Another funding strategy is to make surethat the resources given to the consortium are takenfrom the budget of the department in the firm that isresponsible for using the consortium’s results.

Another issue is personnel. Firms do not oftensend their very best people to R&D consortia. But ifthey send fourth-best personnel, the consortium willhave no credibility within member fins. Firms mustrecognize the problem and send at least theirsecond-best people, if the consortium’s results are toget an attentive hearing. Also, a powerful patron forthe consortium within the member firm helps toensure that the results are exploited; without such apatron, the consortium can easily fall victim to the‘‘ilot-invented here’ syndrome.

To establish closer links between consortia and atleast some participants, the EC (almost always) andthe Japanese (sometimes) have located their consor-tia research within the labs of participating firms. Incontrast, MCC and Sematech have their own labs.Another diffusion strategy is the promotion ofparallel research. Japanese firms taking part inconsortia often have entire research labs devoted toshadowing the consortiums results.

On a slightly different point, one key reason whyJapanese companies take part in consortia is to keepan eye on what their competitors, domestic andforeign, are up to. Technology diffusion is anacknowledged weakness of U.S. manufacturing (seech. 6). If R&D consortia succeed not only intransferring their own results effectively to mem-bers, but also in raising members’ awareness morebroadly of technology advances in their field butoutside their members’ own area of emphasis, thenthe consortia will have served a useful purpose.

The Role of Government

Government should try to ensure that consortia itsupports are designed appropriately, taking to heartthe lessons described above. But this is not sufficientgrounds on which the government can make a choiceof projects. For example, it is necessary but not at allsufficient that the project have enthusiastic partici-pation (including a hefty financial commitment)from industry. The government—with its inevitablylimited funding-must also have some notion of theextent to which the projects it supports are importantfor the economy as a whole.

Some areas of technological advance offer long-term benefit to society but do not attract sufficientprivate investment because they are high-risk activi-ties with uncertain pay-offs, and are very expensive.In the past, industries and technologies have beensupported by the government on the grounds thatthey were vital for the nation’s security. But todayother grounds appear more pressing. Certain indus-tries have vitally important spill-over effects: knowledge-intensive industries have ramifications far beyondtheir own (fast-expanding) boundaries; the wholecomplex of industries that the Europeans call‘‘telematics’ directly affects the competitiveness ofmany other industrial sectors. And there are other“key” technologies. It is not an accident that bothJapan and the Europeans have targeted the same keyindustries and technologies.

The notion that some consortia are more worthyof government support than others implies that somegovernment agency must make that determination(see ch. 2). It is fair to argue that any large-scale shifttoward cooperative R&D with government supportwill, in the end, imply the kind of choices that acivilian technology agency would be designed tomake. Only such an agency could have the necessaryexpertise.

This brings up a final important point. While thegovernment must seek to ensure that its support goesto projects that are both valuable and likely tosucceed, government support is itself a factor in theequation. In Japan and Europe, government supportprovides both cash and credibility to a project, andmay make the difference between success andfailure. That, after all, is why these governments’support for R&D projects can be justified: ifgovernment money did not make the differencebetween success and failure, there would be no needfor that spending in the first place.

INTELLECTUAL PROPERTYIntellectual property rights, once largely ignored

in our Nation’s trade policy, are now an importantissue. Foreign firms skilled at imitation are said to bestealing our inventions, and weak protection ofintellectual property is held to blame. The UnitedStates is negotiating for stronger rights worldwide(including better enforcement), both bilaterally andin multilateral negotiations of the General Agree-ment on Tariffs and Trade (GATT) and the WorldIntellectual Property Organization (WIPO). There is


also concern to maintain strong protection in thiscountry-especially against infringing imports.

Attention has been focused on the whole range ofintellectual property rights, including (among otherthings) patents for inventions; copyright for books,records, and computer software; and laws concern-ing trademarks used to identify a fro’s goods. l24

This section, however, addresses only those intellec-tual property rights that directly protect technologi-cal innovation, including patents and software copy-rights. Generally, the term “intellectual property”in this section denotes only intellectual property ofthis sort. (Box 7-D describes those intellectualproperty rights and explains how their effectivenesscan vary.)

Inventors do not always need legal protection tokeep imitators out of the market. Sometimes theycan keep the new technology secret long enough toget a good lead in the market.125 For example, it ishard for an outsider to determine the exact composi-tion of many chemicals, and still harder to determinethe process used to make them. In other cases,secrecy is not feasible. Many products are rathereasily examined and duplicated—machinery, forexample. Even computer chips, with their micro-scopic maze of interconnecting circuits, are surpris-ingly easy to copy. Information can also leak outthrough employees switching firms, and throughsuppliers and customers. In many industries, de-tailed information about new products and processesoften leaks out to competitors within a year afterthey are developed.126 When secrecy breaks down,inventors may turn to the legal protection ofintellectual property rights.

The U.S. interest in intellectual property rights toprotect technological innovation is not surprising.While other countries have surpassed us in imple-menting some new technologies, the United States isstill a world leader—perhaps the leader—in discov-ering new technologies. The inventor of somethingnew has an obvious interest in keeping to himself theright to make and sell the thing-or in extractingroyalties if he wishes to sell that right to others.

The case that stronger protection would yieldgreat dividends is not clear, however. Protection ofinventions is only one factor—by no means the mostimportant-in competitiveness. Furthermore, thenet effect of stronger protection on the advance oftechnology is uncertain. While it might encourageR&D by rewarding inventors, it can limit thediffusion of the new technology. Finally, it is noteasy to persuade less developed countries, whereintellectual property protection is relatively weak, toraise their level of protection. While intellectualproperty protection does matter, even the bestforeseeable changes in protection here and abroadwill probably have at best a modest positive effect onU.S. manufacturing. The most promising avenues ofchange are: 1) streamlining enforcement of patentrights in the United States and Japan, and 2)harmonizing patent procedures among differentcountries.

How Much Can Increased Protection Help?

This section considers whether: 1) stronger pro-tection for technological innovation can preventlarge trade losses to imitators, 2) stronger protectionwould make the U.S. economy (and other econo-mies) more efficient by encouraging research anddevelopment, and 3) the United States can convinceother countries to increase protection.

Preventing Losses

It is not known how much the U.S. trade deficit isincreased by gaps in intellectual property protection.The U.S. International Trade Commission (ITC)estimated that for 1986 these gaps caused U.S. firmsto lose revenues of $43 to $61 billion.127 This figureis extrapolated from survey responses by selectedfirms with estimated losses totaling $23.8 billion.The sum includes at least $1.8 billion in sales lostdirectly in the United States to infringing imports,$6.1 billion in sales lost directly abroad because of

l~For a more gener~ ~scussion of in~ll~tu~ property rights, see U.S. Congress, Office of Technology Assessment, fnteffecrud Properfy Righfs inan Age of Electronics andlr@ormation, OTA-CIT-302 (Springfield, VA: National Technical Information Service, 1986).

l~Tr~e ~re~ 1aw Cm pro~t against competitors’ use of information gained by unauthorized Mcas.126~wln Mansfield, ‘ ‘How Rapi~y ~s Technolo~ ~~ out?” The Jour~iof/~lr~/Eco~mics, VO1. 34, No. 2, December 1985, pp. 219-221.

Mansfield reported on a survey of 100 fms in 13 industries, chosen at random from those industries’ high R&D spenders. Ibid., p. 217, fn. 2.12’7UOS. ~ternation~ Trade Comi5sion, Foreign Protection of lntellect~lprope~ Righfi ad the Effect on fJ.S. I&t~ and Trade (Washington,

DC: USITC Publication 2065, February 1988), pp. H-2 through H-3.


Box 7-D-Intellectual Property Rights Protecting InnovationIntellectual property rights follow national boundaries. They are granted by national (or state or provincial)

governments, and apply only within the national territory. A firm that desires rights in other countries can seek suchrights from those countries, If a firm’s rights are being violated in a foreign country, it must go to that country’scourts for monetary compensation or an order stopping future violations.

The most important intellectual property right for protecting technological innovation is the patent. Patents aregranted by a national government for inventions of new, useful, and non-obvious products or processes (includingnew, useful, and non-obvious improvements of existing products and processes). A patent grants the right for a fixedperiod (in the United States, 17 years) to stop others from doing the following things within the national territory:making covered products; using or selling covered products, wherever they were made; using covered processes;and (in some cases) using or selling products made abroad by covered processes.

Other rights include copyright, semiconductor mask work protection, and trade secrets+ Copyright is the rightfor a fixed period (in the United States, the term varies depending on circumstances but is at least 50 years) to controlthe copying and other dissemination of textual, artistic, or other expressions; it is significant for protectingtechnology because in the United States and many other countries, copyright protects against the copying ofcomputer programs and data. Mask work protection, a new form of protection created by the United States in 1984and since adopted by other countries, protects against copying the layouts of semiconductor chips, which involveelaborate interconnections and are very expensive to design. Trade secrets law concerns the stealing of a firm’ssecrets, including technical know-how. Depending on the law, a firm which has tried to keep its knowledge secretmay be able to stop other firms which gained unauthorized access to this knowledge from using it.

The scope and effectiveness of protection varies from country to country. For example, Argentina, India,Brazil, and Mexico do not grant patents for pharmaceutical products, and Brazil does not grant patents for processesfor manufacturing pharmaceuticals. Also, the duration of patent protection varies. The United States generallygrants protection for 17 years from the date of the patent grant, while in India patents concerning foods,pharmaceuticals, veterinary products, pesticides and agrochemicals last only 5 years from the patent grantor 7 yearsfrom the application, whichever is shorter. Some countries require that, when it is in the national interest, patentowners must grant a license to a local producer. If the parties cannot agree on a royalty rate, an administrative orjudicial authority will set one.

Thus, the laws vary in scope; but the effectiveness of protection might depend still more on procedures forenforcing legally defined rights. For example, applying for and enforcing patents can be costly, especially in foreigncounties where companies have to hire foreign patent lawyers and pay to have documents translated. Applicationand enforcement proceedings often take several years, during which time others are in most cases free to imitatethe invention. Moreover, the rules of procedure and evidence in some countries might make it difficult to prove thata violation has occurred.

inadequate protection, and $3.1 billion in lost estimates of their losses. It is hard for a firm to stateroyalties. 128 It also includes other items, such asreduced profit margins, business never attemptedabroad, loss of manufacturing economies of scale,and the effect of a weakened sales force on otherproduct lines. The ITC’s report did not quantifythese other items separately.129

As the ITC acknowledges, projecting from the193 companies that estimated lost revenues to thewhole U.S. economy is problematic. Even for these193 companies, the ITC depended on firms’ own

confidently how much it has lost even in direct sales,and even more so for other items such as hypotheti-cal losses from sales never attempted. The estimatescould also be too high if the firms incorrectlybelieved that certain goods were infringing. Inaddition, the ITC study concerned infringement ofall intellectual property. If losses were restricted totechnology-based intellectual property, the subjectunder investigation here, the ITC’s estimate wouldprobably be at least 10 percent lower.130

l~Ibid,, pp. 4.5, A-c. These Mm fi~res arc compiled from firms tha[ estima[ed these items separately. Other responding firms may have inCUrrdthese losses but did not estimate thcm separately.

1291 bid., p. 4-4.130Ente~~ent, f~ and ~verages, pUbllShlng and ~rlntLng, a[ld textl]es and apparel comprise $2..Q~ billion oftie total losses of$23.8 billion. Ibid.,

p. 4-3. These particular categories probably mvolvc almost exclusively trademarks and hterary or arustic copyright.


More fundamentally, the existing forms of protec-tion cannot always protect against imitation. Evenwhen a firm has a patent, others can often inventaround it—i.e., develop alternative technologies toget the same result.131 The patent does give theoriginal inventor a headstart while others catch up.During this time, the inventor might be able toimprove the manufacturing process, expand produc-tion to exploit economies of scale, and developmarketing channels ahead of his competitors. Ifcompetitors are better at manufacturing or market-ing, however, they can overtake the inventor. Thishas happened many times. A classic example is theCAT scanner, whose inventor was eventually re-warded with a Nobel prize. The invention wascarried out for the British firm Electrical MusicalIndustries (EMI) Ltd. in the late 1960s. AlthoughEMI produced and sold the CAT scanner success-fully at first, the company lost its lead in the U.S.market half a dozen years after introducing it there;its biggest, most successful competitor was GeneralElectric, which quickly developed a somewhatimproved scanner and provided hospitals withsuperior training and servicing. A couple of yearslater, EMI dropped out of the CAT scanner busi-ness. 132

Promoting Economic Growth

The theoretical basis for intellectual propertyprotection is to promote economic growth byencouraging research and development—a factor ofspecial importance to the United States, since U.S.manufacturing competitiveness depends heavily ontechnological superiority. Without strong protec-

tion, the argument goes, many innovations thatwould benefit society as a whole might never bemade because the innovator could not make enoughprofit to recover the costs of research and develop-ment and compensate for the risks of failure.133 Itshould be borne in mind that not all R&D pans out.Profits on the successes must be great enough tocover the costs of the inevitable failures as well. Inaddition, the world today is so full of competentmanufacturers that imitation is occurring faster thanever, making it increasingly harder for innovators torecover enough profits. If they are to get sufficientreward for their inventions, they need strong protec-tion of the law.

This argument for intellectual property protectionhas some force. However, there is evidence thatmany inventions would still be made even if legalprotection were unavailable.l34 In an opinion survey,100 firms in the United States were asked whatpercentage of their patentable inventions commer-cially introduced in 1981-83 would have beenintroduced even if patent protection were not availa-ble. The answers, by industry, were: 35 percent inpharmaceuticals, 70 percent in chemicals, 80 to 90percent in petroleum, machinery, and fabricatedmetal products, and over 90 percent in primarymetals, electrical equipment, instruments, officeequipment, motor vehicles, rubber, and textiles.135

In another opinion survey, 130 firms rated patents asfairly ineffective in many industries, includingelectronics; secrecy and lead time were generallyconsidered more important.136 Another study alsofound that patents provide relatively little protection

131~ven~g ~Omd ~atentS iS relatively emy in elatronics, and relatively difficult in ph~aceutic~s. Rich~d Levh, Alvin Klevorick, RichardNelson, and Sidney Winter, “Appropriating the Returns From Industrial Research and Development,” Brookings Papers on Economic Activity, No.3, 1987, p. 811; Edwin Mansfield, Mark Schwartz, and Samuel Wagner, “Imitation Costs and Patents: An Empirical Study,” Economic Journal, vol.91, 1981, p. 913. In general, broader interpretation of what a patent covers can make the patent harder to invent around.

lszDavid J. T-e, 1‘Captting Value From Technological Innovation: Integration, Strategic Partnering, and Licensing Decisions, ’ Technology atiG/obaf Industry: Companies and Nafions in the Wor/dEconomy, Bruce R. Guile & Harvey Brooks, (eds.) (Washington, DC: National Academy Press,1987), pp. 65-66,85-86.

lssM~t Ofien tie major exwme is not ~ m&ing an invention or discovering a new res~t, but rather in Subquent development work culminatingin commercial production.

l~~e li~ratm on this pint is review~ in Wesley Cohen and Richard hvin, ‘‘Empirical Studies of Innovation and Market Structure, ’ Handbooko~lndusm”al Organization, Richard Schmalensee and Robert Willig, (cd.) (New York, NY: North Holland, 1989), vol. 2, pp. 1091-1093.

135~~n Masfield, ‘cp~ents and ~ovatlon: An Empficd Study,” Ma~gementScie~e, vol. 32, 1986, pp. 174-175. The firIIIS were dSO askdwhat percentage of their patentable inventions made during 1981-83 would have been made even if patent protection were unavailable; the responseswere similar, The fms were chosen at random from 12 U.S. industries (excluding very small firms). Of the 100 fins, 96 responded,

136Richmd ~vin, Alvin Klevorick, Richmd Nelson, and Sidney Winter, op. cit., pp. ‘7~, 792, ‘794, 796-797,811. The flrms d.%) rated process piMeIItS

less effective than product patents. The data suggests several reasons: processes are easier to keep secret and fms prefer to keep processes secret ratherthan disclose them in a publi~ed patent; processes are patentable less often than products; competitors can find alternative processes rnon easily manahemativeproducts; competitors’ uses of patented processes are harder to detect and prove than competitors’ manufacture and sale of patented products.Ibid., pp. 794 (table 1), 797 (table 2), 803 (table 5).


in electronics, compared to drugs and chemicals.137

Yet the electronics industry spends about 8 to 9percent of sales on R&D, compared to an average ofabout 3 percent for all R&D-performing manufactur-ing industries.138

In addition, the social benefit of encouragingR&D must be balanced against certain social costs.One cost is reduced diffusion of new technology—including technology that would have been devel-oped even without legal protection. By exercisingintellectual property rights, an innovator can preventothers from using the new technology. Thus, peoplewanting to buy products dependent on the newtechnology might find the products expensive orunavailable because competition has been stifled.Perhaps even more important, other researchers willbe reluctant to build on a protected invention (e.g.,by improving it or by applying it in a new way), sincethe original inventor may try to stop other firms fromusing any derivative technology. (And if the inven-tor is willing to license the invention, the royaltiesmight be more than most firms would pay.) Forexample, the semiconductor industry would havebeen slower to take off if AT&T had been able tokeep others from using the key technology coveredby its early transistor patents.139

While all patents to some extent risk inhibitingtechnology diffusion, some recent patents with abroad sweep have aroused particular concern. Theseinclude patents on software to display multiplewindows on a computer screen and to compare thetexts of different versions of a document,l40 and apatent on a new mathematical technique to solveproblems such as routing of telephones and schedul-ing of airlines.141 The courts might interpret suchpatents narrowly or find them invalid altogether—but first some firm must have sufficient stake in theissue to mount a court challenge. In some fields, thedanger of inhibiting technology diffusion might berelatively low. For example, patents on existingpharmaceuticals generally do not stop rival firmsfrom developing and marketing new pharmaceuti-

cals based on different active ingredients (althoughpatents would stop rival firms from developing, forexample, new dosage forms of existing pharmaceu-ticals).

Patents can also help diffusion of technology,because they are published and must explain how topractice the patented technology. Upon reading apatent, an expert in the field might think of follow-upwork or might think of a way to achieve the sameresult in a different way, outside the patent’spurview. Many Japanese firms routinely track patentapplications (which in Japan are published 18months after filing) to learn about new technology.In the United States, applications are publishedwhen a patent is granted—typically about 2 yearsafter the application is filed. However, this pub-lished information often could be learned instead byexamining the products.

There are also costs of running the legal system.Patent applications can be expensive for inventors tofile and for the government to evaluate. Lawsuitsbetween patent owners and alleged imitators areexpensive for the parties and take up valuable courttime. Moreover, in many lawsuits the court finds thatno patent rights have been violated, so the time andexpense was for naught---a, worse, the suit mighteven have been brought deliberately to harasslegitimate competitors and scare their customers.Another cost is that, to the extent the law is unclearor its enforcement unpredictable, business planningis hindered for both inventors and possible competi-tors.

Since intellectual property protection entails so-cial costs as well as benefits, providing ever strongerintellectual property protection does not necessarilypromote economic growth. The best results comefrom protection strong enough to encourage innova-tion but not so strong as to greatly inhibit technologydiffusion—with some attention to making enforce-ment predictable and inexpensive. Unfortunately,determining what level of protection would work

lgT~win Mansfield, Mark Schwartz, and Samuel Wagner, op. Cit., p. 913.ls8Nation~ Science Fomdatlon, Nafiow/Patterm ~fR&f) Resources: ]989, Find Report NSF 8$)-308 (Washington, DC: U.S. Government finting

Office 1989), p. 65. This source gives ‘‘Company R&D funds as a percent of net sales in R&D-performing manufacturing companies by industry’ for1986. The percentage for electronic components (SIC 367) is 8.5; for drugs and medicines (SIC 283), 8.8; for other chemicals (SIC 284-285, 287-291),2.6; and for all industries combined, 3.2.

IWU.SO Congess, office of Twhno@y As~ssrnent, /nternutionu/ Competition in Services, OTA-lTE-328 (Sprinxleld, VA: Nation~ TCX~C~Information Service, July 1987), p. 216. As part of a 1956 consent decree settling an antitrust suit against it, AT&T agreed to license its transistor patentsto other firms.

I@Lawrence Fisher, ‘‘Software Industry in Uproar Over Recent Rush of Patents,’ The New York Times, May 12, 1989, p. D1.lqlGina Kiolata, “Ma~ematicia~ Are Troubled by Claims on Their Recipes, ” The New York Times, Mar. 12, 1989, p. E26.


best is largely a matter of guesswork. Empiricalstudies of how the level of protection affects the totalamount of innovative activity (both invention anddiffusion) are few and inconclusive.142 Other effectsmight also be weighed in the balance. For example,strong patent and trade secret protection can make itharder for employees in an established firm to leaveto found their own firm in the same field oftechnology.

143 Whether this effect is desirable de-pends on the characteristics of particular industries.

Another effect is that strong protection at homecan be a bargaining chip abroad. In the 1960s, bothIBM and Texas Instruments gained permission toproduce and sell in Japan only by agreeing to grantJapanese firms licenses under their key U.S. patentsfor computers and semiconductors respectively. Ineffect, IBM and Texas Instruments gained access tothe Japanese market in exchange forgiving Japanesefirms access to the U.S. market.l44

Finally, in considering the optimum level ofintellectual property protection, it should be borne inmind that intellectual property protection is not theonly way to encourage innovation. For example, thegovernment can fund research and development,give tax breaks or preferential financing to industryinvesting in R&D and in modern equipment, andcollaborate in industrial R&D projects. On thewhole, compared with some other countries, theUnited States has chosen to rely less on these othermeans of encouraging innovation and more onprotection of intellectual property. This choiceprobably arises from the fundamental, widely heldview in this country that government and civilianindustry should be separate. Intellectual propertyprotection is seen as proper: it simply lets firmsmake profits in the marketplace based on theirinventions. The alternative that the United States hasmost strongly embraced—support for basic researchand for defense R&D-involves little interactionbetween government and civilian industry.

If the United States were to put more emphasis onvarious other means of encouraging innovation,

intellectual property protection might become lessimportant.

Convincing Other Countries

Assuming that stronger worldwide protection ofintellectual property could improve U.S. competi-tiveness, is there a realistic prospect that othercountries can be persuaded to change their laws?Generally, the less economically developed a coun-try is, the less it desires intellectual propertyprotection.

Less developed countries are not much moved bythe argument that imitation is a form of stealingbecause it takes the benefits of R&D without sharingthe costs. These countries are apt to reply that theyare already much poorer than we are, and that strongintellectual property protection would just aggravatethe difference. Stronger protection would benefitinnovators in rich countries while driving priceshigher for consumers and stopping capable localimitators in poorer countries.

The argument that protection is good for eco-nomic development in the long run, because itencourages local innovators, also falls on stonyground. Imitation of existing technologies is at leastas well proven as a springboard for economicdevelopment as local innovation. Korea and Taiwanare modern examples. The United States, as coloniesand as a nation, based much of its own earliereconomic growth largely on imitation of Europeantechnology.

Finally, developing countries are urged to con-sider that they may not always be able to imitate.Sometimes they will need to buy technology fromforeign firms, and these firms might refuse to licenseor sell technologies in countries that lack adequateintellectual property protection. This possibilitydoes not seem to scare developing countries much.For example, many complaints have been lodgedabout unlicensed imitation or “piracy” of inven-tions in Korea, yet Korean firms have found willingsellers of technology among U.S. innovators, espe-cially in the semiconductor industry.

Idzwesley Cohen and Richd kin, “Empirical Studies of Innovation and Market StIUChKG” Hana%ook of Industrial Organizatwn, RichardSchmalensee and Robert Willig, (cd.) (New York, NY: North-Holland, 1989), vol. 2, pp. 1089-90, 1094-95.

143when a ~uent fm sues a spinoff for ~leged patent or trade s~ret vlolations, tie cost of fighting tie suit can intimidate the spinoff, regardless ofthe merits of the case. The parent’s motivation is often more to prevent hiring away of employees than to protect intellectual property--as shown by theterms of settlements. Professor John Barton, Stanford Law School, personal communication, June 12, 1989 and Feb. 1, 1990,

144u.s. Conmss, Office of T~~o]on Asx55ment, /~erMtwM/ co~etitiveness in E/ec~om”cs, OTA-ISC-ZOO (Springfield, VA: NationalTechnical Information Sewice, November 1983), pp. 193-194.


If the United States cannot persuade other coun-tries that strong intellectual property protection is intheir own interest, it can resort to carrots and sticks.The carrot is often exemption from tariffs for certainimports from that country under the GeneralizedSystem of Preferences.145 The stick can be denial orwithdrawal of such benefits, or flexible retaliation(often in the form of punitive tariffs on certaingoods) under the recently strengthened Section 301of the Trade Act of 1974. This approach hasachieved some success. For example, partly becauseof sustained U.S. pressure, Singapore strengthenedits copyright protection generally and also applied itto software.l46

If other countries under our urging do passtough-sounding laws, that does not guarantee theirenforcement. Enforcement requires sophisticatedgovernmental apparatus. A country that has troublefeeding its population cannot be expected to spendlarge amounts of money to ensure speed and fairnessin processing patent applications and trying patentsuits. Moreover, a country that has been pressuredinto granting intellectual property rights might beparticularly inclined to give enforcement efforts alow priority.

Specific Problems

Efforts to strengthen intellectual property protec-tion are likely to be most effective when aimed atprotection in the United States and in other devel-oped countries. Both have large markets for theproducts of U.S. technology; and developed coun-tries have more interest in granting strong protec-tion. The weak spots in these countries largelyconcern procedures for administering and enforcingthe law, rather than the law’s substance.

U.S. Patent System

Courts in the United States have ruled morefavorably toward patent holders in the 1980s than inthe 1960s. This applies to rulings on patent validity(in particular, whether an invention was sufficiently

non-obvious to merit a patent); scope of coverage(how broad a range of possible imitation is prohib-ited by the patent grant); permissible conduct by thepatent holder (e.g., whether the patent holder mayimpose various marketing restrictions on its licen-sees); and compensation to be awarded for infringe-ment. The changed legal climate in part reflects ashift in viewpoint. Traditionally, judges enforcedpatents narrowly on the ground that patents createdundesirable monopoly rights. Increasingly, how-ever, judges have viewed patent rights as simply alegitimate incentive and reward for innovation. Inaddition, in 1982 patent-related appeals were consoli-dated in one court, the U.S. Court of Appeals for theFederal circuit. That court’s rulings have strength-ened and clarified patent law. In fact, some believethat this court has tilted the law too far in patentowners’ favor.

The effectiveness of U.S. patent law is limited bydelay in enforcement. It takes over 2 1/2 years, onaverage, to bring a patent case through trial for aruling. 147 Only then (with fairly rare exceptions) canthe patent owner get a court order to stop theimitation. 148 A firm whose patent is being infringedmight not make it to the end of the trial, and even ifit does the court might not fully compensate for theharm.

Delay is particularly troublesome in suits againstimported goods. Often it is hard to trace theseimports back to their source, to find out whom to sue.Moreover, the U.S. courts have no way of enforcinga ruling against a foreign manufacturer who hasneither assets nor employees in the United States. Ifthe manufacturer is beyond the court’s power, thepatent owner must sue domestic distributors instead.(The same is true for other intellectual propertyrights, including copyright, mask work, and tradesecret.) The patent owner might hesitate to sue, forthe foreign manufacturer’s distributor might also behis own, for the same product or others. Even if thecourt rules for the patent owner and orders thedistributor to stop selling the infringing goods, the

Idswhile gant~g such preferences might ~ conson~t with o~ over~l foreign policy obj~tives, ~e~ preferences do m~e foreign ~mpelillOnstronger in the U.S. market, thus to some extent offsetting the gain in our total competitive position due to stronger intellectual property protection abroad.

146R. ~chWi G@bawmdT1mothy Richads, /nte//e~r~/ProPer~ Rights: G~b~Co~e~~, G/oba/Co@ict? (Bolllder, CO: WestviewPress, 1988),pp. 313,329.

~41An&RepO~Of tie Director of t~ A~in~trati~e ofice of the United States Courts ]988, p. 221. This report shows a rfldian tilIIC Of 31 mOnl.hSfrom filing to disposition of patent cases after trial; the arithmetic mean, or average, would probably be somewhat greater, since 90 percent of the casestake at least 11 months and 10 percent of the cases take at least 62 months.

1481t is ~ Pficiple ~ssible to get s~h an order ~fore ~i~, c~l~ a prelimin~ inj~ction. However, in practi~ it is very had for a patent Ownerto get such an order unless he has already won a prior lawsuit based on that patent.


foreign manufacturer need only switch to anotherdomestic distributor. The patent owner will thenhave to start all over again, first identifying and thensuing the new distributor.

The owner of a patent or other intellectualproperty right can avoid these multiple lawsuits byfiling a complaint with the U.S. International TradeCommission (an independent Federal agency) underSection 337 of the Tariff Act of 1930, as amended.149

If the Commission finds that imports are violatingthe complaining party’s intellectual property rights,it can issue an exclusion order, enforced by theCustoms Service, barring importation of thosegoods. Moreover, the Commission is required torender a decision within a year (18 months in aminority of cases deemed “more complicated”), aconsiderably shorter time than the average a courtsuit takes.

The GATT has ruled that Section 337 violatesGATT treaty provisions by providing special, harsherenforcement of the patent laws where foreign goodsare concerned.150 Section 337 is still in effect.However, other GAIT members could now retaliateagainst future use of Section 337 against their fins.The GATT decision found many aspects of Section337 inconsistent with the GATT treaty, and it will bedifficult to amend Section 337 to bring it into linewith the GATT decision while still keeping the coreadvantages of: 1) a quick decision, and 2) anexclusion order.

Japan’s Patent SystemU.S. firms have often been frustrated by the

ineffectiveness of Japan’s patent system in protect-ing their inventions.l5l Part of the problem hasstemmed from U.S. fins’ lack of familiarity withhow the system works. Difficulties have also oc-curred due to the language barrier. But part of theproblem stems from the nature of the system,

especially from delays and a public policy thatfavors granting of licenses to those who improve ona basic patent.

While precise figures are not available, it seemsthat an application for a Japanese patent, if opposedvigorously by another firm, will generally be tied upfor at least 6 years before an inventor can proceedwith a lawsuit. (Patent applications in the UnitedStates take an average of less than 2 years.) After apatent is issued in Japan, a firm accused in court ofpatent infringement could probably cause the law-suit to take at least 3 to 4 years.152 (U.S. patent trialsare not much quicker. They average somewhat over2 1/2 years, and a determined defendant often can adddelay.) This puts an inventor in a poor bargainingposition: grant a license, or wait at least 9 to 10 yearsto get anything from a lawsuit.153

In addition, other firms might seek patents forvarious improvements. This practice is very com-mon in Japan. After an application is first published(generally 18 months after the application), firmsfrequently file many applications for improvements.Under Japanese law, a firm that receives a patent foran improvement can apply to the Patent Office for acompulsory license under the basic patent, if theowner of the basic patent refuses to agree on licenseterms. The Patent Office has discretion to grant ordeny the request. l54 While this law has neveractually been used, its presence can weaken thebargaining position of a patent owner.155

The Japanese system, which encourages licens-ing, might work well for Japan’s economy bypromoting diffusion of technology. However, itoften does not serve U.S. firms well. To succeed atall in Japan, a U.S. firm might have to be the solesupplier of the item in question. As discussedelsewhere in this report, Japanese firms have astrong tradition and bias in favor of buying from

1491(3 USC. 1337. This ~em~y WM ~~en~en~ by the ~lbus Trade and competitiveness At of 1988, Public Law 1OO-418, SW. 1342.ISOA GA~di~pute ~e~olution panel ~~ in November 1988, upon the complfit of the EC, that Swtion SST violates the ‘ ‘national Weatnlent” ClaUSe

of the GATT treaty, which requires that a member country treat imports from another member countxy in a reamer ‘no less favorable than that accordedto like products of national origin in respect of all laws, regulations and requirements affecting their internal sale, offering for sale, purchase,transportation, distribution, or use. ’ GATT Article H. The United States subsequently accepted the panel’s ruling, making it an officiaI GATT deeision.

151~e Semte Comittm on Comer&, Science, and Transpo~ation, Subcommittee on Foreign Commerce and Tourism, has held hearings on theJapanese patent system on June 24, 1988 (Serial No. 100-59) and Feb. 28, 1989 (Serial No. 101-19).

152Y01c~ro Yamawhi, patent attorney regstered in Japan, Bevefidge, DeGran~ & Weilacher, personal communication, Jan. 30, 19$X).lss~erdelays we ~ssible in ~th the Japanese and us- systems. The systems are hard to compare ~rudy. In general, both systems involve Sifik

types of delays (though sometimes in a different order), but the delays are longer in Japan. Ibid.154Jap~ese Patent Law, IAW No. 121 of Apr. 13, 1959, as amended, SeCS. 7292.

lssyolc~ro Yamaguchi, op. cit.


other Japanese firms. Japan’s complex distributionsystem reinforces the preference for buying Japa-nese. An exclusive patented technology might be theonly way for an American firm to get into theJapanese market if it is forced to license Japanesefirms, they might capture the whole market.

The United States has been negotiating with Japanto fix these and other problems. Already, Japan hasincreased the Patent Office staff to reduce delay,although not nearly as much as the United Statesbelieves is necessary. Japan has also lengthenedvarious deadlines for non-Japanese parties, to allowsufficient time for communication and translation.Because Japan is now producing many inventions, itmight be receptive to granting stronger power topatentees to exclude competition. The keidanren, aninfluential Japanese association of businesses, hasalready urged that patent systems worldwide shouldprovide “effective patent enforcement,” including“preliminary and final injunctions [i.e., court ordersagainst infringement] as well as monetary awardsadequate to compensate patentees fully and serve asan effective deterrent.’’156

Critics of Japan’s patent system might keep inmind that the U.S. system has its own drawbacks.Resolution of patent cases that go to trial is slow (onaverage over 2 1/2 years). In addition, patent litiga-tion can be expensive, and our patent law is quitecomplex. Thus, foreigners may feel that our systemputs them at a disadvantage. While patents are issuedrelatively quickly in the United States, they are fairlyoften ruled invalid by the courts. Moreover, 20 yearsago the courts enforced patent rights more narrowlythan they do now. It should therefore not be toosurprising that Japan and other countries, followingin our economic footsteps, do not grant as strongprotection as we might wish.

Patent Office Procedures Worldwide

Procedures for issuing patents differ from onecountry to the next. This raises the cost of filingapplications in more than one country. The UnitedStates has been negotiating in WIPO and with Japanand the countries of the EC on terms of possibleharmonization. The negotiations have already bornefruit. At U.S. urging, Japan in 1988 started allowinginventors to put multiple claims (in effect, multiple

variants of the same invention) in one consolidatedapplication, as has long been the practice in theUnited States and Europe.

Most of the changes, however, will probablycome only as part of a comprehensive settlement. Aspart of any package deal, it is likely the United Stateswill have to change from a first-to-invent system (inwhich the first person to make an invention isentitled to a patent) to a first-to-file system (in whichthe first inventor to file an application is entitled toa patent). Only the United States and the Philippinesfollow the first-to-invent system.

A fret-to-file system has some advantages. Sincepatents disclose the invention, early patenting couldincrease technology diffusion. A fret-to-file systemalso avoids extensive legal fights over who was thefirst inventor. However, switching to a first-to-filesystem in this country could disadvantage smallinventors (either individuals or startup or smallfirms), who are probably more important in theUnited States than elsewhere. Having no patentdepartment, small inventors usually have a hardertime filing applications and therefore prefer to delayfiling, to see if the invention warrants filing and tohave more time to prepare the application. Under afirst-to-invent system, these inventors can delayfiling applications without fear of being preempted.Under a first-to-file system, a small inventor mightlose out to a later inventor in a large firm that gets itsapplication filed first. This hardship on smallinventors could be lessened by making initialapplications easier and cheaper to file, as they are inmany fret-to-file countries.

ANTITRUST LAWFederal antitrust law prohibits a wide range of

business conduct that restrains trade or monopolizesa market. Its core provisions, Sections 1 and 2 of theSherman Act of 1890, as amended, prohibit bothbusiness “combinations . . . in restraint of trade”and the “monopoliz[ation], or attempt to monopo-lize," trade.157 The Sherman Act and other antitruststatutes are worded in general terms, and could bytheir literal language prohibit a great deal ofinnocent business activity. The courts therefore havetaken the statutes as an invitation to fashion a body

l~e ~~ll=tu~ ~~y mi~~ (USA), Keidanren (Japan), and UNICE (Europe), Baic Framework of cm provisions on l~ellect~Property: Statement of Views of the European, Japanese and United States Bwiness Communities, June 1988, p. 33.

15715 Uos,c. 1.2.

2 1 - 7 0 0 0 - 9 0 - 8


of precedent to explain more clearly what conduct isprohibited.

U.S. antitrust law has long been an effectiveshield against the power of monopoly and has keptmany fields open to enterprising, innovative new-comers. Today, as foreign competition looms largein the U.S. economy, some have questioned whethertraditional interpretation and enforcement of anti-trust law may need some changes. Antitrust lawcould potentially prohibit firms from merging,forming joint ventures, and cooperating in variousother ways—such as setting industry standards andconducting joint R&D projects. This section as-sesses the extent to which antitrust law mightprohibit or discourage such behavior even when itwould enhance the competitiveness of U.S. manu-facturers.

While direct evidence is scanty, it appears thatantitrust law does discourage some competitiveness-enhancing conduct, and would somewhat impedegovernment and private efforts to increase suchconduct. In most cases, the law does not actuallyprohibit the behavior in question; but because thelaw is unclear and involves stiff penalties, business-men are often afraid to do anything that even lookslike it might be an antitrust violation.

The Changing Interpretation of the Law●

The types of conduct prohibited by antitrust law,and the philosophical justification for the prohibi-tions, have changed somewhat through the years. Inthe late 1800s and early 1900s, antitrust law wasaimed in part at keeping businesses small. Smallbusinesses were seen as more humane and moreresponsive to local needs. The Supreme Court, forexample, criticized the transformation of “an inde-pendent businessman, the head of his establishment,small though it might be, into a mere servant or agentof a corporation, ’’158 and noted the “widespreadimpression that corporate power had been and wouldbe used to oppress individuals and injure the publicgenerally.’ ’159

In recent years, antitrust law has been aimed notso much at preventing bigness as such but rather atensuring fairly free competition. Under neoclassicaleconomic theory, a free market-in which manyfirms compete and no one firm is large enough toaffect the market for its product—is most efficientfor society. When one or a small number of firmscomes to dominate the market; those firms tend toreduce output and raise prices compared with freemarket levels. This market power is called oligopoly(if only a few firms are competing) or monopoly (ifonly one company sells the product).

Since the 1960s, the primary purpose of antitrustlaw has been to promote competition by minimizingthe creation and exercise of market power.l60 Todayboth the courts and the Federal enforcement agen-cies acknowledge that some kinds of cooperationcan often be justified by compensating benefits tosociety. l6l For example, suppose several competingfirms with large combined market share in metalalloys create a joint venture to develop and sell aparticular new alloy. Despite the joint venture’smarket power, society might be better off with thejoint venture than without it because on their own thefirms might have taken much longer to develop theproduct.

In evaluating a firm’s conduct, enforcementagencies and the courts most often use a balancingtest, or “rule of reason. ” Conduct that threatenssubstantially increased exercise of market power ispermitted if the societal benefits outweigh thesocietal costs (or, as sometimes phrased, if thepro-competitive effects outweigh the anti-competitive effects). The rule of reason is not alwaysused. For example, certain egregious conduct, suchas agreements among sellers to fix prices or dividemarkets, is deemed to be a per se (Latin for “byitself’ violation. In such cases no balancing test isperformed, on the ground that the conduct rarely ifever can have any social benefit.

The wide adoption of the rule of reason by thecourts and the enforcement agencies has madeantitrust law more accommodating than it once was.

lS8Fr~~ck Row, “me ~l~e of ~ti~t ~d tie ~l~ions of Models: The Faustian Pact of Law and ~OnOmiCS,” Georget~n @ JOIW@vol. 72, June 1984, p. 1517, quoting United States v. Trans-Missouri Freight Association, 166 U.S. 290, 319, 323 (1987).

l~~id., p. 1517 footnote 32, quoting United States v. Standard Oil Co., 221 U.S. 1. so (191 ~).16Dp~llip ma @ ~uis K@oW,A~~frwfAW@~~: frobfe~, Text, c~es, Ah ~. (Boston, w: Little Brown & L’o., 1988), pMS. 111, 130, pp.

13-14, 44-45; Frederick Rowe, op. cit., pp. 1524-1535.161Reportof the Amrican B@Asso~~tion section ofAntip~t~ T~kForce on the Antitr~tDiv&n of the U.S. Department ofJwtice, hdy 1989,

pp. 8-16.

Chapter 7—Where We Stand: Public Policy and Technology .221

Congress also modified the antitrust statutes in the1980s because of concern for U.S. manufacturingcompetitiveness. The Export Trading Company Actof 1982162 provided for advance antitrust approvalfor firms working together through export tradingcompanies. In the National Cooperative ResearchAct of 1984,163 Congress mandated a rule of reasonapproach, and in some cases lessened penalties, forjoint R&D. But whether these recent changes areenough, or whether further modification of antitrustlaw is called for to enhance U.S. competitiveness, isstill an issue.

The Terms of the Debate

There are arguments for maintaining the statusquo. In the past decade very few antitrust lawsuitshave been brought, let alone won, that challengeactivity which arguably should be encouraged oncompetitiveness grounds (e.g., joint R&D). 164 More-over, it is hard to find examples of firms’ giving upany such activity because of antitrust concerns.

In addition, antitrust enforcement has substan-tially lessened over the last two decades, especiallyin the early years of the Reagan administration. Forexample, civil cases filed by the Justice Department,around 30 per year in the late 1970s, dropped to thelow teens by 1983. Cases against conduct other thanprice-fixing and bid-rigging165 have become rare,and the Justice Department’s guidelines, testimonyand other public pronouncements often express astrong concern to ensure that antitrust law notprohibit desirable business activity. Private suitsalleging antitrust conduct have also decreased inrecent years—from about 1,110 cases filed in the 12months ending June 30, 1984 to about 660 in the 12months ending June 30, 1988. Further weakening ofantitrust enforcement could send the wrong signal tobusiness, and invite anti-competitive behavior.l66

There are also arguments for further modification.One argument challenges the neoclassical premisethat free markets always benefit society. Althoughthe premise might be true generally, it does not applyin all cases. Specifically, perfect competition maynot be conducive to innovation in today’s businessand technological conditions.

Firms will perform less innovation than would bebest for society if they cannot capture substantialbenefits of their innovations. In a perfectly freemarket, other firms imitate the innovator, take awaysome of the business, and drive the price down to alevel that typically does not let the innovating firmrecoup its investment. If the innovator can getmarket power, at least for a while, he can recovermore of the value to society of his innovation. In thelong run, a society in which innovators can expect togain some market power, at least temporarily, isprobably better off than one in which perfectcompetition always prevails.167

The patent system provides one way of gainingsuch market power. In fact, patent systems areusually justified on the ground that they encourageinvention. A patent owner has the legal right to stopothers from using the patented technology for a termof years (in the United States, generally 17 years),and this right often yields some market power, atleast until others find a way around the patent.

Sometimes patents are not a very effective way ofgetting enough market power to repay an innovator(see the section Intellectual Property in this chap-ter). Corning out first with a new or improvedproduct is an alternative way of achieving marketpower, at least for a time. But it is often the casetoday that neither patents nor the advantages ofbeing first to market are enough to encourageadequate innovation by single companies. Increasingly,the kind of innovation nations need to enhance

l~pub]ic LSW 97-290, 15 U.S.C. 4001 et seq.

l~~blic ~W 98-4.62, 15 U.S.C. 4301 A305.1640TA is ~wwe of n. ~~h ~-., exapt hat ~me C=s have ~n fil~ (but not won) against grOUpS of f~s ~tting industry stidards.16SBid ~%ing involves tie exe~i~ of m~et ~wer by buyers, which has un&sirable eff~~ similar KI tie exerci~ of market power by @kXS.ltiRePort of t~ Amr&an B~ &50c~twn section of Antitr~t ~ T~k Force on the Antitrust Division of the U.S. Department of Jusn”ce, Op. Cit.,

pp. 4-5, 16-18, A7 (source for chart on page A7 is U.S. Department of Justice, Antitrust Division Workload Statistics FY 1978 to F’Y 1987); Annua/Report cfthe Director of the Ati”nistrative Q?ice of the United States Courts: Z988 (Washington, DC: U.S. Government printing Office, n.d.)t PP. 181t185; see also Patrick Marshall, “Do Antitrust Laws Limit U.S. Competitiveness?” Congressional Quarterly’s Editorial Research Reports, vol. 2, No.1, July 7, 1989, pp. 368-70, The Reagan Administration did file a high number of criminal cases, but largely against small local businesses such asconstruction. Report of the American Bar Associatwn Section of Antitrust Luw Thsk Force on the Antitrust Division of the U.S. Department of Justice,op. cit., p. 17; Patrick Marshall, op. cit., p. 368.

167ThommJordeand DavidTme, “~ov~ion, CwPr~ionand~ti~t: B~ancing ComWtition andc~ration,’ ‘High TechnofogyL.uwJournaf,vol. 4, No. 1, spring 1989, pp. 8-13,


competitiveness requires resources far beyond themeans of small companies.168 Even for very largefirms, the expense and risks of R&D may be toogreat for the firms to go it alone. For example, bothhigh definition television and optoelectronics re-quire expensive technology development in severalareas at once.

Capital costs for manufacturing can also be veryhigh. It costs at least $250 million to build a plant toproduce the present generation of DRAM semicon-ductors, and for the next generation capital costs willbe much higher, perhaps $500 million per plant.Such costs may be too great for most U.S. firms tobear on their own. The manufacture of semiconduc-tors and other high technology products also re-quires technical expertise in many fields-oftenbeyond the capability of a single firm.

There are other reasons as well for cooperationamong fins. Small manufacturing firms sometimesbenefit from pooling their resources and biddingtogether on large jobs. Voluntary establishment ofindustry standards requires cooperation betweenfirms. Sometimes mergers are necessary to matchforeign fins’ economies of scale.

All of these cooperative activities might run afoulof our antitrust laws. The law does not invariablyprohibit activities like these. In fact, if firms canadvance technology or otherwise improve theircompetitiveness only by teaming up, then courtsmay well judge that the benefits to competitionoutweigh the harm. Under the rule of reason, therewould then be no antitrust violation. The trouble isthat firms cannot be sure of this ruling in advance.Because elements of the law are vague, and thepenalties for antitrust violation can be severe, firmsoften shy away from cooperative deals out of acombination of fear and ignorance.

Several factors can make it hard to predict theoutcome of antitrust cases involving cooperation.Under the rule of reason, the beneficial and harmfuleffects of the deal must be compared. The harmfuleffects are determined largely by how much marketpower the deal creates, but in practice, market poweris often very difficult to determine. For example, a

proposed joint venture might be expected to sell 80percent of laptop personal computers, but only 10percent of all personal computers, in the U.S.market. The venture’s vulnerability to antitrust willdepend heavily on the extent to which laptops andother personal computers are substitutable. Even ifthe products do not readily substitute, some manu-facturers of non-laptop personal computers might bewaiting in the wings, ready to produce laptops if thejoint venture tried to raise prices. In that case thejoint venture would have little market power. Ingeneral, the substitutability of products and theability and willingness of firms in neighboring fieldsto enter a market could be points of contention incourt. Also in contention could be the deal’s claimedbeneficial effects. Will the firms substantially ad-vance technology to develop a new product, or is thedeal really just a front for pooling market shares?And is it really true that the individual firms couldnot profitably develop the product in question ontheir own?

How these points are resolved at trial againdepends on several factors. The judge or jury mayunderstand the need for firms to pool their resourcesbut they may not. Facts about the market are hard todetermine and are often the subject of conflictingexpert testimony. Even after the facts are resolved,the weighing of positive and negative effects is nota precise calculation. It inherently involves theexercise of judgment.

In addition, while the rule of reason is widelyused, alternative legal tests might in some cases cutshort the full consideration of the activity’s benefits.Antitrust doctrine contains the per se test (an activityis condemned without any consideration of itsbenefits), the “quick look” test (an activity’sbenefits are considered only if on a quick look itappears reasonably likely that such benefits exist);and the “least restrictive alternative” test (anactivity is condemned if the court believes itsbenefits could have been achieved by anotherarrangement with less restrictive effect on competi-tion). These doctrines might, for example, be appliedto some joint production cases-especially if the

l~~ordingto~e National Science Foundation, 200companies accounted for90percent of all industrial R&D spending in the United Staks in 1986.The average R&D spending among this group was $273 million. If the average R&D intensity of the firms was 10 percent (a very high figure) the 1986average net sales for companies in this group were $2.7 billion. William L. Stewart, “Effects of Corporate Restructurings on R&D Support,” testimonybefore the House Committee on Science, Space, and Technology, Subcommittee on Science, Research, and Technology, July 13, 1989.

l@Jorde and Teece, op. ci~, pp. 4M2, 47A8.


court is skeptical that joint production can yieldbenefits for society.l69

These are some of the complex issues of fact andlaw that make the outcome of a particular casedifficult to predict. They can also make a trial quiteexpensive. On the average, antitrust cases takelonger than other cases filed in Federal district court.For example, of the cases that go to trial, antitrustcases take a median time of 35 months, comparedwith a median time of 19 months for all cases.170

Firms therefore have reason to be cautious aboutactivities that might be considered antitrust viola-tions. The severe enforcement regime in antitrustlaw, which includes multiple enforcers and stiffpenalties, reinforces caution.

Both the Justice Department and the FederalTrade Commission enforce Federal antitrust laws.These agencies can file civil suits to stop theoffending conduct. Also, if the Justice Departmentestablishes that an antitrust violation has causedeconomic harm to the government, the defendantmust pay the government actual damages.171 From1984 to 1987 the Justice Department filed a total of46 civil suits. Of the 50 civil suits terminated in thatperiod, the government won 39, lost 2, and negoti-ated an agreement in the remaining 9.172 For mergersand some joint ventures above certain dollar thresh-olds, firms must give the government advancenotice. 173 If the government announces its intentionto challenge the deal in court, the firms involved willusually either modify the deal to satisfy the govern-ment’s concerns or abandon the deal altogether.

The Justice Department can also file criminalsuits with potentially large fines for the corporationand culpable officers and employees, and imprison-ment for culpable officers and employees,174 al-though criminal suits have been reserved for egre-gious attempts to fix prices, divide markets, or rigbids. From 1984 to 1987 the Justice Departmentfiled a total of 219 criminal cases. Of the 237 casesterminated in that period, 210 resulted in convic-tions.175

While the U.S. Government’s civil enforcementin recent years has not been aggressive, firms areoften reluctant to gamble that government policywill remain the same. Firms can seek approval inadvance for a particular course of action from theJustice Department or the FTC,176 but these approv-als often take several months and can involveconsiderable legal expense. Firms usually save thisprocess for substantial projects (justifying high legalfees) that can wait several months. Also, governmentapproval does not insulate firms against privatesuits, though in practice it lessens the likelihood thatprivate suits will be filed or will succeed.

Private parties can file antitrust suits if they claimto be threatened or to have suffered some economicharm caused by an alleged violation. Private suits arefar more numerous than government suits. In the 12months ending June 30, 1988, private parties filedabout 660 private antitrust suits in Federal court,compared with 20 civil and 70 criminal cases filedby the government.177 Private suits are on average

~~~An~ Report of the Direc@r of the A&ninistrative @ice of the United States Courts: 1988, OP. cit., pp. **O, **1.ITIsee 15 U.S.C. 4, 15a, 25. The F~r~ Tr~e Commission proceeds under 15 U.S.C. 21 (mergers) or under the Federal Trade Commission ~t, 15

U.s.c. 45.172u.s. Dep~mentofJu~tlce, ~ti~t Division WorNoad Statistics 1978 to ~ 1987, reprint~ in Reportof the American BarAssociation Section

of Antitrust L.uw Twk Force on the Antitrust Division of the U.S. Department of Justice, op. cit., p. A7 (total civil cases: filed, won, lost, dismissed).17315 U.S.C. 18a.17415 u-s-c. 1.175u.s. ~p~mentof Justice, ~ti~t Division WorkJoad Smtistics ~ 1978 to FY 1987, repl-int~ in Reportof the~ericanBarAssoc;ation Section

of Antitrust Luw Tmk Force on the Antitrust Diviswn of the U.S. Department of Justice, op. cit., p. A8 (total criminal cases: filed, terminated, won).In 1987,42 individuals were fined a total of $1,636,(XKI; 15 individuals were sentenced to serve time in jail and 33 more were given probation; 1,994jail days were served; and 66 corporations were freed a total of $16,265,000. Ibid., reprinted in Report of the American Bar Association Section ofAntitrtat Law Tmk Force on the Antitrust Division of the U.S. Department of Jwice, op. cit., pp. Al I-A13.

176Nei~erJmtice ~p~mentapprov~snor FTC st~f level approv~s ~e bindingon the government, butno firm h~ everbeen SUed for conduct withinthe scope of such an approval. Janice Rubin, Library of Congress, Congressional Research Service, American Law Division, “The Impact of U.S.Antitrust Law on Joint Activity by Corporations: Some Background,” May 1, 1989, p. 7 (Department of Justice); Carl Hevener, Justice DepartmentLiaison, Federal Trade Commission, personal communication, Dec. 1 and 5, 1989 (Federal Trade Commission). During 1984-87, the Justice Departmentreceived about 25 requests for approval per year and granted roughly 90 percent of them. U.S. Department of Justice, Antitrust Division WorkloadStatistics FY 1978-1987, reprinted in Report of the American Bar Association Sectwn of Antitrust Law T&k Force on the Antitrust Division of the U.S.Department of Justice, op. cit., p. A5 (business reviews),

~77An& Report of the Direc~r of the Administrative Q@ce of the United States Courts: 1988, OP. Cit., pp. Igl, *W.


less successful than government suits, although thestatistics are somewhat unclear.178

Private parties that bring suit are usually eithercustomers or competitors. If the suit is successful,the offending conduct must stop and the offendermust pay to the complaining party: 1) trebledamages, i.e., three times the amount of economicharm the complaining party can show he suffered,179

and 2) the reasonable cost of the complaining party’sattorneys for the successful claims. These are severepenalties. In most other areas of U.S. law, acomplaining party, if successful, is entitled only tosingle damages, i.e., the actual amount of harm heshows he has suffered, and is usually not entitled toreimbursement of the expense of hiring attorneys.

The prospect of treble damage and attorney feeawards can encourage lawsuits by competitors andcustomers. Even if the defendant believes he couldprobably win in court, the prospect of paying theselarge awards—as well as the time and money neededto fight the case-might scare him into payingsomething to settle the case, and someone contem-plating filing a suit knows this.180

Antitrust laws may be enforced by State govern-ments as well. Under Federal antitrust law, Stategovernments may file civil suits on behalf of theircitizens--e.g., on behalf of a large class of consum-ers who allegedly paid excessive prices because ofan antitrust violation. The penalties are the same asif the citizens themselves had filed suit.181 Stategovernments also can enforce the State’s ownantitrust laws, if the State has any.

In sum, even if antitrust law does not usuallycondemn activities outright that could improvemanufacturing competitiveness, it can often discour-age such activities. For large firms, able to get expertlegal advice, antitrust risk is often considered as onefactor among many. Cooperative projects have

ordinary business risks as well. Will the technologywork? Will the market be there? Antitrust addsanother risk, and makes the project that much lessdesirable. Similarly, the need to pay for a legalanalysis of the antitrust risk and for ongoing legalsupervision adds to a project’s cost at various stages.For small fins, often unable or unwilling to pay forlegal advice, antitrust fear is more likely to act as anabsolute bar. If a small firm suspects that a projectinvolves antitrust risk, it might drop the ideaimmediately-even if there is no real risk. Theambiguity and complexity of antitrust law aretherefore particularly troublesome to small firms.

While the chilling effect of antitrust is plausible,it is hard to tell how important it is compared withother factors that discourage cooperation. It isdifficult to find examples in which antitrust actuallykilled a cooperative project, and the examples givenhere are not overwhelming. More telling examplesmay exist. Firms might hesitate to offer thembecause word of their actual or contemplated activi-ties could provoke suits or give away strategicinformation to competitors. More important, busi-ness decisions typically depend on many factors, andit is hard even for those involved to say whether fearof antitrust changed a decision. A businessmansensitized to antitrust concerns might even avoid orquickly abandon ideas for cooperative projects.Those ideas will never be counted or noticed asactivity discouraged by antitrust.

Some activities-e. g., R&D consortia, joint man-ufacturing, and resource pooling by small firms—have only recently received serious attention ingovernment and industry as ways to enhance com-petitiveness. Even if antitrust in the past has notvisibly discouraged very much activity, it might doso more in the future as more of these projects areproposed.

ITSOf the 891 priv~e anti~st suits termi~ted in the 12 months ending June 30, 1988, 145 are Ikted as “settled,” 434 ss “other dismissed,” @ ss‘‘other non-judgment, ’ and 248 as going to judgment by the court. Of those 248,67 are listed as “for plaintiff, ” 143 ‘‘for defendant, ” and 38 “other.”David Germy, Statistical Analysis and Reports Division, Administrative Office of the U.S. Courts, personal communication, Dec. 7, 1989. It appearsthat at least 577 cases (“otherdismissed” plus ‘‘judgment for defendant’ went entirely for defendant, and that in 212 cases (’ ‘settled’ plus ‘judgmentfor phdifr plaintiff recovered something.

l?g~y @le d~~es M-C @d in stits concerning R&l) projects registered under, and within the scope of, the National Cooperative Resemh Atof 1984, discussed below.

l~smtisucs fm the 12 months end~g Jwe 30, 1988 show 145 private anti~st suits “se~led” out of 891 ~rmina~, or 16 percent. David Gentry,Statistical Analysis and Reports Division, Administrative Office of the U.S. Courts, personal communication, Dec. 7, 1989. However, some cases wheresome settlement was paid might be reported under a different heading.

18115 U.SOC. 1%.


Effect on Business Activity

Joint Research and Development

Traditionally, joint R&D has posed relativelylittle antitrust risk. Such projects offer obviouspotential benefits in spreading risks and improvingfirms’ efficiency and competitiveness. Moreover,member firms are generally free to manufacture andmarket products on their own (although there mightbe agreements restricting use of the resulting tech-nology). Experts in the field know of no antitrustcase brought against genuine joint R&D, and theJustice Department acknowledges that “[a]s a gen-eral matter, joint R&D activities can have substan-tial procompetitive effects. ’ ’182

The National Cooperative Research Act of 1984lessened the legal risks of joint R&D.183 First, theAct provides that such activity will always be judgedby the rule of reason, balancing its beneficial andharmful effects. l84 While joint R&D would ingeneral have been judged by the rule of reasonanyway, this provision did remove some uncertaintyand sent a signal from Congress that cooperativeR&D can yield important benefits. This probablyincreased judges’ sensitivity to the benefits inbalancing the pro- and anti-competitive effects ofjoint R&D.

The Act also provides that, for joint R&D projectsregistered promptly for publication in the FederalRegister, only actual damages (rather than trebledamages) may be awarded in a private lawsuit.(Attorney fees can still be awarded.) 185 As ofJanuary 1990, 160 separate projects had filed 323registration statements including amendments.186

Some of these projects probably would not havegone forward without the 1984 Act.187 Registrationgreatly reduces the financial exposure in undertak-ing joint R&D and by the same token greatly reducesthe incentive for parties to file private suits. The Act

further discouraged indiscriminate private suits byproviding that one who files a private suit breed ona claim that is “frivolous, unreasonable, withoutfoundation, or in bad faith” must pay attorney feesof the accused party.

One consortium registered under the Act is theNational Center for Manufacturing Sciences (NCMS),which started operations in 1987 and by 1990included over 100 manufacturing firms.188 Themembership encompassed large firms such as Gen-eral Motors and AT&T; smaller firms such asKinefac Corp., a 70-employee metalworking firm inWorcester, Massachusetts; and even some firmswith fewer than ten employees. (Many of the smallerfirms joined at the urging of their larger customers.)

It is not clear whether NCMS would have beenformed without the 1984 Act. Even with the Act,antitrust has been a major concern for present andprospective members. In its startup period, throughearly 1989, NCMS spent about $200,000 for anti-trust advice from a law firm. In the organizations’first months, the director of NCMS spent most of histime on antitrust issues. NCMS’s early meetingswere devoted largely to antitrust concerns, and intoearly 1989 NCMS was still receiving about twoqueries a week from members. Antitrust concerns ●

have gradually lessened as 1)NCMS became famil-iar with the issues and could more easily addressmembers’ concerns, 2) members noted that no firmhad been sued for R&D registered under the 1984Act, and 3) competition became more intense andthe benefits of joint R&D became more apparent, sothat members were willing to accept some antitrustrisk.

Before joining NCMS, most members were not inthe habit of sharing technical discussions or R&D-partly from unfamiliarity, partly from antitrust fear.NCMS has discovered that its members’ pressingR&D concerns overlap considerably, so that coopera-

182u.s. ~~m~t of Justice, Antitrust Enforcement Guidelines for Internutwrud operations, NOV. 10, 1988, P. 56.

l~public LAW 98462, 15 U.S.C. 4301-$305. There are some limitations, discussed below, on what constitutes joint R&D un&r the ~t.l~The ~t requires that the tie of r-n also be used in judging joint R&D under State antitrust law.l~~e kt simi]~ly limits damage awards in cases based on State antitrust law.186u,S. ~@~ntofJu~ti~repre~nt~ive, ~wn~cou~cation, Feb. 1,1990, The first 125 or so proj~ts~descri~ itl “National (kiOpertUiVe

Research Act of 1984 Consortia,” New Technology Week, Special Supplement, June 12,1989.l~T~GovernentRo/e in~oi~Ve~Wes, he~ngbeforethe Houx committ~ on Science, Space, ~dT~hnology, Sept. 19,1989, serial No, 101-58,

testimony of Mauro DiDomenico, director, technical liaison office, Bellcore, p. 101; and testimony of Peter Mills, chief administrative officer, Sematech,p. 122.

188Thematen~ on NCMS comes fmm ~ Miller, Dir~tor, NCMS, Wrson~ communication, Apr. 10 ad 27, May 3, Aug. 5, Sept. 6, 1989, ad J~,29, 1990; and Patrick Ziarnik, counsel, NCMS, personal communication, Jan. 26, 1990.


tive R&D can yield substantial savings. One exam-ple is R&Don laser beam splitting-i. e., how to useone laser beam for several processes at once. NCMSalso facilitates informal technical exchanges amongmembers—for example, while discussing whetherto fund proposed projects.

NCMS has sometimes made matches betweencompanies with complementary abilities. It did so,for example, in the development of ductile iron as aninexpensive substitute for structural steel in thenon-moving parts of machine tools. A consultingprofessor had recommended the alternative of duc-tile iron, but NCMS could not find any U.S.company that could both design the specializedmolds needed for pouring the iron and do thepouring. NCMS then brought together firms with theCAD ability to design the molds and a company thatcould pour the ductile iron once it had the molds. Theeffort cost NCMS about $50,000. It has saved oneNCMS member about $250,000.

Although the 1984 Act lessened fears of antitrustsuits arising from joint R&D, these fears have notcompletely vanished—as shown, for example, bythe continuing concern of NCMS’ members. Also,the Commerce Department has from time to time

- provided a “safe house” in which competitorsafraid of prosecution for merely discussing a possi-ble R&D collaboration could frost come together tohold discussions under government supervision.189

The concerns are understandable. Even firms thatregister their projects may still be sued by thegovernment or by private parties for single damagesand attorney fees. Firms that prefer to maintainsecrecy and do not register are subject to privatetreble damages. Even under the rule of reason thefirms are not necessarily home free. For example, theJustice Department’s 1980 guidelines for jointresearch ventures state that “[a] joint venturebetween directly competing companies in a highlyconcentrated industry . . . will be subject to veryclose antitrust scrutiny."190

In cases where cooperating firms need to ex-change cost and marketing information in order toguide the project in a commercially useful direction,

the 1984 Act covers such exchanges only when“reasonably required to conduct the research anddevelopment.” If, despite the firms’ belief thatcertain communications were necessary, a courtshould hold otherwise, then those communicationswould not be protected by the Act’s provisions.

Also, the Act covers manufacturing only for“experimental and demonstration purposes. ’’191 Some-times sizable runs are needed to demonstrate aprocess, and the firms involved would take asignificant loss if they could not sell the itemsproduced. Yet a court might rule that such sale is notcovered by the Act.

More fundamentally, the Act does not covercommercial manufacturing. Yet there may be casesin which the scale of the enterprise needed tocapitalize on R&D is beyond the means of singlef ins .

Joint Manufacturing

Joint manufacturing can sometimes make firmsmore competitive, for example, by allowing econo-mies of scale in production. This could be especiallyimportant in fields such as semiconductors where aproduction facility costs hundreds of millions ofdollars. Firms might sometimes need to sharetechnical as well as financial resources. Manufactur-ing, like R&D, often requires expertise in manyfields. Again, semiconductor fabrication offers anexample. Future examples might be devices basedon optoelectronics or high temperature supercon-ductivity. Sharing both financial and technicalresources can mean the difference between beingfirst to market or being an also ran.

Joint manufacturing can have another sort ofbenefit when it follows joint R&D performed in acentral organization. Technology transfer from thatcentral organization back to the member companiescan be difficult. It might be easier for the centralorganization to proceed with manufacturing. Forexample, some say that it is unfortunate thatSematech will not perform commercial manufacture(see the section in this chapter on R&D consortia).Innovation ideally consists of repeated feedback

189L~fi~ F~~~~, f)i~~r, ~ufi~ TwhrIo]ogy p~er~ip PK)$JHUII, U.S. ~p~mertt of co~e~e, Prsond comm~ication? 3! 19$9’IWOSO ~pwment of J@Ce, Gt&fe~~r Research joint ventures (1980), Illustrative Examples, CztW B—red Re=wch and ~vdOPtnent JO~t

Venture in Concentrated Industry, reprinted in Trade Regufatwn Reports par. 13,120 (Commerce Clearing House 1988), While these guidelines mightbe somewhat outdated in view of the more liberal tone of the U.S. Department of Justice, Am”trust Enforcement Guideiinesforlnternatwrud Operations(1988), p. 56, the 1980 guidelines were relied on by NCMS’ antitrust counsel in 1989 in giving cautious advice regarding a proposed project.

19115 U.S.C. 4301 (a)(6)(C).


between R&D, design, manufacturing, and market-ing. Many rounds of feedback between a joint R&Dventure and its members might be much moredifficult than many rounds of feedback would bewithin a venture that did both R&D and manufactur-ing.

In general, joint manufacturing carries moreantitrust risk than joint R&D, because it can directlyreduce competition. When firms manufacture jointly,purchasers may have fewer choices of products andsuppliers. However, in some risky high technologyventures, joint manufacturing might be the only wayto encourage new entrants. High-definition televi-sion (HDTV) offers an example of antitrust concernsin manufacturing joint ventures. Starting in early1989, the American Electronics Association (AEA)sought to promote R&D and manufacturing consor-tia for HDTV. During the first several months, AEAhad trouble even getting firms to talk with eachother, largely because of antitrust fears. Antitrustthen became less of a concern—partly because thefirms with AEA’s help were able to think through theantitrust issues, and partly because the many con-gressional hearings on possible changes in antitrustlaw made it seem likely that Congress would amendthe law, or at least that enforcement agencies and thecourts would interpret the law with more apprecia-tion of the benefits of joint production.192

Some analysts question whether U.S. firms reallyneed large joint manufacturing ventures, arguingthat Japanese firms have done very well withoutthem. Even if Japanese firms do not do much jointmanufacturing (and this point is subject to dispute),the two countries are not the same. Japanese firmsare much better able to finance large manufacturingprojects on their own (see ch. 3). They alsosometimes have a wider range of in-house techno-logical expertise. 193 In addition, Japanese hightechnology firms are sometimes protected againstforeign competition-either by the government orby customers who buy Japanese products preferen-tially.

Antitrust concerns can also discourage smallmanufacturing firms from cooperating in marketingand in performing jobs that are too big or require toomany specialized capabilities for the firms to handleon their own. Japanese manufacturing firms cooper-ate a great deal in this manner, with the blessing andactive encouragement of their government; someEuropean firms have done so as well (see ch.6). SomeU.S. industry groups, encouraged by the CommerceDepartment, are seeking to increase such coopera-tion in this country. 194 Antitrust concerns seem tohave impeded these efforts somewhat.195

An example comes from the Flint River Project,Inc., a subsidiary of Efficient Enterprises, Inc., inTroy, Michigan. In 1988, Flint River began trying toform a network of small manufacturers of spare partsfor automobiles, heavy equipment, and defense.Flint River proposed to market the fins’ productsdomestically and abroad by finding jobs to be done,selecting a suitable team of firms for each job, andperforming any needed technical coordination, in-cluding design and project management. The net-work still had not formed as this report was written.While antitrust was not the only problem, it was asignificant one. The firms were afraid that participat-ing in such a network could be deemed an antitrustviolation--e.g., a conspiracy to fix prices.196

Another example: in the early 1980s, a fewmembers of the Milwaukee chapter of the NationalTooling and Machining Association discussed a bidsolicitation from the U.S. Department of Defense forabout 40 million dollars’ worth of special-purposecarts to transport bombs. The members believed thatby combining their production capacities and theirvarious specialized abilities (such as welding, preci-sion manufacturing, design, and possession of alarge crane) they could do the job as well as andmuch more cheaply than traditional large defensecontractors. However, early in the discussions some-

l~pat Hill Hubbwd, Viu president, EIA, personal communication, Sept. 8, 1989.193GmgoVTm.y,6‘SncW~ ~mge and Competitiveness: The U.S. Semiconductor ~dusv, ‘‘ TechnologicalForeca.stingandSocialChange. vol.

38, 1990 (forthcoming); Richard Elkus, chairman, Promernx Corp., personal communication. Dec. 1 and 7, 1989 (elec~onics ind~m).l~~~we ~Ues, offiW of T~hnology policy, U.S. Department of Commerce, personal commun.icadon. May 3. Sept. 7* 1989.195Ro~Fn~m, “FlexibleNctworksand ~ti~st,’ T~En~eprene~i~Eco~myRevi~, VO]. 7, No, $), May 1989(published by t.hecorporation

for Enterprise Development, Washington, D.C.).l%Mich~l H~ler, president md chief executive officer, Efficient Enterprises, Inc., Troy, ~, PrSOnd communication, Apr. 26* MaY 3? ‘d SePt. 14>

1989, and testimony at hearings before the House Committee on Small Business, Subcommittee on Regulation and Business Opportunities, Sept. 13,1988, Serial No. 100-74, pp. 125-28.


one mentioned antitrust, and as a result the idea wasquickly dropped.l97

In neither of these two examples was the fearbased on a legal analysis of the particular circum-stances. Rather, the firms’ managers said in effect atthe outset, ‘This might have antitrust problems, andI can’t afford a lawyer to find out. Unless I somehowget assurances that there will be no problem, I won’tproceed.”

Standards-Setting

Voluntary industry standards are necessary forindustrial efficiency. Without them, for example,light bulbs would not fit into sockets and regionaltelephone networks could not exchange information.However, it is possible for a standards-settingassociation to be dominated by a clique of firms thatuse the process of establishing standards to shutother firms out. For example, a clique might developstandards in secret, so that their competitors wouldnot be able to conform their products to the standardpromptly. In addition, a clique might pick onestandard not because it is the best, but because it isdifficult for competitors to meet. Such practiceswould probably be deemed antitrust violations, aswell they should be.198

However, even if firms perform standards-settingwith no anti-competitive intent, other firms maynevertheless file an antitrust suit claiming that thestandard somehow unfairly discriminated againstthem. A court might take these claims seriously. TheSupreme Court recently stressed that standards-setting can easily be abused to harm competition:

[T]he members of [standards-setting] associationsoften have economic incentives to restrain competi-tion and . . . the product standards set by suchassociations have a serious potential for anticompetitiveharm. . . . Agreement on a product standard is, afterall, implicitly an agreement not to manufacture,distribute, or purchase certain types of products.199

To forestall accusations, many U.S. associationshave adopted elaborate procedural rules for setting

standards, including open meetings at which allfirms are free to express their opinion. According toBell Communications Research (Bellcore), theseopen meetings are cumbersome and can slow downadoption and implementation of standards. In fast-moving fields such as telecommunications, delaymight mean missed opportunities and reduced com-petitiveness. Firms could speed up progress by alsomeeting informally in smaller groups to iron outdifficult technical problems; however, they arereluctant to do so for fear of an antitrust lawsuit.200

This problem arose in connection with communi-cation standards for ways in which telephonenetworks in different regions or countries canexchange various information--e. g., route callsaround congested lines, determine whether a calledparty’s line is busy, and verify credit card numbers.These standards are handled by the T1.S1.3 WorkingGroup (formerly the T1.X1.l Working Group) of theT1 Standards Committee, which is accredited by theAmerican National Standards Association. Accord-ing to Gary Schlanger, that working group’s chair-man for 1984 to 1987, U.S. and foreign approachesto exchanging such information started to diverge in1986. The working group’s efforts to resolve thosedifferences and harmonize the U.S. and internationalpractice were deadlocked for 2 years. Mr. Schlangerbelieves that with smaller, informal meetings har-monization could have been achieved. Instead, in1988, the T1 Standards Committee adopted a U.S.standard inconsistent with the international standardused by the rest of the world. Now U.S. equipmentmanufacturers and phone companies must cope withtranslating between the U.S. standard and the worldstandard. 201

Joining Forces Against Foreign FirmsIn some cases, a fragmented U.S. industry faces

competition from a much more powerful foreignindustry. By combining forces, the U.S. firms mightachieve similar advantages and hold their own. Butsome mergers or joint ventures between U.S. firms,each of which hold substantial shares of the same

197CUI ~qulst, Resident, c~lson Tixd & Manufacturing Co., Cedarburg WI, Personat comm~ication, APr. 28, 1989,

198se for exmple R~”antBur~rs, Inc., v. People’s Gas, Light and Coke CO., 364 U.S. 656 (1961).~99Allied T&e & Conduit Corp. v. Indian Head, [rLc., 108 S. ct. 1931, 1937 (1988).200Joe ~e~, Ge~r~ Atto~ey, Bellcore, ~rWn~ commulcatlon, May 3 and Sept. 13, 1989; Ma~o DiDomenico, director, technical litiSOn OffiCX,

Bellcore, testimony at hearings before the House Committee on Science, Space, and Technology, Subcommittee on Science, Research and Technology,Sept. 19, 1989, Serial No. 101-58, pp. 106-109.

201GW sc~~nger, DiviSiOn M~ager, c~er Intercomwtions, s@n&@ and Nllm~fing Plm Mamgernent hp~ment, Bellcore, ~rsondcommunication, Sept. 15, 1989.


market, are either blocked or never attemptedbecause of antitrust law. Under the Justice Depart-ment’s Merger Guidelines, the firms involved mustprovide a high level of proof of any claimed benefits.Also, the Department will consider whether similarbenefits could be achieved by other means.202 Onemerger blocked by the Department was the attemptof BTU International, a manufacturer of furnacesused to produce semiconductor chips, to purchaseThermco, a subsidiary of the Allegheny Corp. (box7-E). While the Justice Department in this case mayhave simply followed judicial precedent and theMerger Guidelines, these rules may be out of tunewith the realities of failing U.S. competitiveness.

Overall, the uncertain cost of keeping antitrustlaw as it is today must be measured against theuncertain cost of proposed changes. Several possible

modifications would leave intact the basic doctrineof antitrust law and the basic enforcement machin-ery, but would adjust around the edges. For example,points of law could be clarified; safe harbors andadvance approvals could be provided; and trebledamages could be reduced in some cases to single.203

It is not clear that changes such as these would sparksignificant anticompetitive activity, particularly inlight of today’s economic conditions. On the whole,it is harder to maintain market power today than itwas earlier in the nation’s history. In some fields,products and processes have shorter life cycles sotoday’s monopolist might find his position erodedtomorrow by competitors’ new technology. Mostsignificantly. foreign firms are more likely than everbefore to compete against U.S. firms that try to raiseprices above competitive levels.

~u.s. Upwmmt of Justice, “Merger Guidelines,” June 14, 1984, sw 3.5, PP. 35-36.~~e~ and otier proposals are discussed h ch 2.


IThis SCtim is M on ~foma~on fr~ the following personal communications: Paul van der Wansem, CEO, BTU kterntiiond,Ncwember2, 1989; Paul CYDonnell (the lawyer representing BTU to the Department of Justice), Ropes and Gray, January 18, 1990; Bob Cole,President, Varian-TELLtd., Deeember, 1989; Anthony Mtdler, CFO, Sillicon Valley Group, December5, 1989; Ken Phillips, Public Relations,Motorola, Deeember, 1989; Bob England, Vice-President, Semiconductor Group, Texas Instruments, January 26, 1990; Representatives of theDepartment of Justice, December 13, 1989, and January 31, 1990.

2~rding ~ m~ket ~e data ga~e~ by Dataquest and filed by Thermco with the Justice Department, BTU had 46 percent adTherrneo 47 pereent+ BTU, defining product categories differently, filed data showing much lower market shares.

3WIe no il~ fi~s arc available, tie estimates by TEL’s U.S. distributor and TEL’s U’S, competitor, B’I’U, agree.4MUller ~int~ out th~ ThermEO h~ incre~ti its market share since its purchase by SVG, so WG is not re~lY ~aPPY Wiw @

Department of Justice’s ruling. Nevertheless, he concedes that the overall strength of U.S. manufacturers in this field, and their ability to mistJapanese entry into the American market, would have been greater if BTU and Therrnco had combined foxes.


old loyalty to TEL equipment did not represent a sufficiently large share of the market to prevent BTU from gaininga monopoly position.

Immediately after BTU’s takeover effort failed, the Silicon Valley Group approached Thermco and wasallowed to buy the firm without ado. SVG makes other types of semiconductor production equipment and hadpreviously attempted and failed to enter the diffusion furnace business. Although BTU’s market share in the UnitedStates is currently larger than TEL’s, SVG sees TEL and not BTU as its principal competitor. TEL’s 1988 worldsides of diffusion furnaces were bigger than the combined world sales of BTU and SVG, making it the world giant.

Though it is impossible to know for certain what the results of a BTU/Thermco buyout would have been, thefollowing seem likely: BTU and Thermco could have combined their current technological strengths, improvingproduction of the most advanced systems; BTU and Thermco combined could have significantly increased theirR&D by eliminating duplication; and the increased size of a united BTU/Thermco would have allowed thecompanies to realize savings in production, marketing, and administration. Whether such advantages could haveslowed or stopped TEL’s penetration of the U.S. market cannot be known for sure. Further, the beneficial effectsof the merger might not have been as great as originally hoped.5 But both BTU’s van der Wansem and SVG’s Mullerthink that combination would have been stronger than the current U.S. configuration of BTU on its own and SVGwith Thermco, In this case, the Department of Justice’s refusal to permit the merger seems to have hampered thecompetitiveness of the U.S. manufacturers.

5~~C. ~d ~o~ ~mt t. ~ ~CqU~~+& ~=e B~ ~ ~n an ~ch fiv~ ad ~a~ layoffs were bound (O result as tk tWOoperations were mmhed. SVG’S Mtdler suggests that enough key personnel might have quit that BTU would have acquired only the hoilow shellof ThermCo.

Index

Index

accelerated depreciation (see also investment tax credit)accelerated cost recovery system 22, 44-46

administrative guidance and signaling 99, 100, 102advanced computing technology 204Advanced Manufacturing Research Facility (AMRF) 176Advanced Manufacturing Technology Initiative (DOE)

187Advanced Technology Program (NIST) 33,54, 73,77Advanced Television (ATV) 80-89agriculture

Cooperative Extension Service 55, 71technology policy 71, 72

Agricultural Research Service 185Airbus 72, 210American Bottlers Equipment Co. (AMBEC) 180American Electronics Association (AEA) 227American Telegraph and Telephone Co. (AT&T) 205,

206, 207, 225analog and digital electronics (see electronics)antitrust law and enforcement 17, 30-31, 66-69, 219-231

barriers to cooperation 225-228enforcement 68-69, 223-224manufacturing 226-228mergers and joint ventures 67, 228-231research and development 225-226small firms 224, 227-228standards setting 228treatment of foreign firms 67, 229

apparel (see textile and apparel industries)Applied Energy Programs (DOE) 185-186, 187ARCH Development Corp. 189-190arm’s length relation with suppliers 13-14, 19, 151automobiles 6-7, 129-135, 158

Europe 131, 157Japan 129-135productivity 7New Unified Motor Manufacturing, Inc. (NUMMI)

133-134, 160product development cycle, innovation 6, 131-133suppliers: dependence on, relation to, and cooperation

with 18-19, 25, 129-135suppliers to Japanese transplants, GAO study 134vertical integration 129-130

banksJapan 98, 99-100, 101-104relationships with clients (see capital suppliers,

relationships with firms)basic educational skills 50, 117-118basic research 71Bayh-Dole Patent Amendments Act of 1984193Bell Communications Research (Bellcore) 228Bradhart, Inc. 180-181

Brookhaven National Laboratory 186, 190Brooks Manufacturing 180BTU International 230budget deficit 9, 11,22,39,42-43, 106-107buffered production system 6

capital costs 9-11, 22-23, 42-46, 93-95, 97, 100-101budget deficit 9-11, 42-43in Japan 93-94, 96-105interest rate 9, 22, 42, 94savings 22, 43-44tax structure 22-23, 44-46, 96

capital formation (see capital costs, investment)capital gains tax 23, 47-48capital investment (see capital costs, investment)capital suppliers, relationships with clients 9-11, 46, 93

banks 9-10,42,96, 101-103hostile takeovers 10-11,23, 106-107shareholders 10, 103-105

CAT scanner 214catch-up consortia 89, 209cathode ray tube (CRT) displays 84, 97chip on glass 86,civilian aircraft industry

technology policy 71, 72, 74-75, 79civilian technology agency 33-35, 73-78civilian technology policies 32-33, 71-73, 74-75CNC machine tools (see machine tools, numerically

controlled)collaborative R&D (see cooperative R&D)Combustion Research Facility (CRF) 190commercialization

technology from Federal laboratories 29-30, 61-64,184-195

Community Development Block Grants 182computers (see also electronics, cathode ray tube

displays) 80-88computer-aided design and manufacturing (CAD-CAM)

159, 161, 204Congressional U.S.-Japanese Fellowship Program 65consortia (see cooperative R&D)consumer electronics (see also high-definition television)

importance 83-85, 89, 147consumers 80, 99, 100consumption 43-44

consumption tax 43mortgage interest deduction 43-44

cooperation (see also cooperative R&D)antitrust problems 17, 30-31between competitors 210, 226-228between suppliers and customers 13-16, 131-133

cooperative networks or associations (see small andmedium-sized firms)

-235–


cooperative R&Dantitrust law 225-226consortia 16-17, 77-78, 202-211Federal laboratories and industry 19-20,29-30,62-64foreign firms’ participation 78-79government-industry collaboration 78-79in Japan 88, 208-210in Europe and the European Community 17, 78-79industry-university consortia 64-65, 202-203intellectual property rights 192-194supplier-customer cooperation 131-133

cooperatives of small firms 29, 61, 167-170copyright (see also intellectual property)

software 64, 193-194corporate debt

effect on capital expenditures 47, 110-111effect on R&D 47, 109-111

corporate financedebt 101-103equity 103-105

Cray Research Corp. 147-148, 183, 188critical technologies (see strategic R&D)cross shareholding (see stable shareholding)

data compression 82, 86-87Defense Advanced Research Projects Agency (DARPA)

34, 75-78, 89, 106, 205-207Defense Logistics Agency (DLA) 177Defense Programs (DOE) 185, 187defense-related technology (see also national security)

32, 39,40, 71, 153-155, 187Department of Defense (DoD) 185-186

commercializing technology from 186dependence on foreign firms 40, 71support for technology development 32, 39-40, 71,

153-155Department of Energy (DOE) and DOE laboratories 62,

64, 184-195Department of Industry and Technology 73design for manufacturability 125-126developing countries

intellectual property 20, 69, 216-217digital and analog electronics (see electronics)Digital Equipment Corp. (DEC) 204digital filters 87digital modulation techniques 82, 87-88digital signal processor chips 86-87

Economic Recovery Tax Act of 198145-46education (see also worker training) 12-13, 24, 50-53,

115-117, 118-120engineers and scientists 52, 120-126, 198-200primary and secondary 13, 24,50, 52, 115-117student loans 51test scores 116-118

vocational 118-120electronics (see also computers, semiconductor industry,

semiconductor equipment industry)analog and digital compared 81-82overlap or synergism between different fields 80,83-88

endaka 23, 104Energy Research programs (DOE) 185-186Engineering Research Centers (ERCs) 64-65, 173, 180,

195-197engineers and scientists 120-126

career paths 125education and training 24-25, 52-53exchange programs with Japan 65foreign language skills 66, 198-200interaction between design and manufacturing 52-53job rotation 125minority engineers and scientists 52, 120, 122-123numbers 121-123on the shop floor 25, 125-126simultaneous engineering 131work force demographics 24-25, 52, 120, 122-123

equipment leasinggovernment-supported equipment leasing programs

28, 57-59Japan 155machine tools, including numerically controlled (NC)

58, 155, 181equity financing (see corporate finance)equity holders (see shareholders)Europe and European Community (EC) 9, 21,41,78-79

automobiles 131, 157industry and trade policy, 79R&D consortia 17, 78-79Single Market, 19929strategic technology policy 78-79

exclusive rights (see also intellectual property) 192-194Export Trading Company Act of 1982221extended definition television (EDTV) 83, 84external or foreign technologies (see also not-invented-

here) 19, 30, 156-158, 197-200

family incomes 39Fanuc 156Federal budget deficit 9, 11, 22, 106-107Federal laboratories 19-20,28-30, 184-195

cooperative research and development (collaboration)29-30

earmarking funds for technology transfer 63; 191exclusive rights (patents, copyright, proprietary data)

64, 192-194red tape 29, 64, 192technology commercialization 29-30,61-64, 184-195

Federal Laboratory Consortium (FLC) 29,63, 190Federal technology extension (see also technology

extension) 54-56, 174-177Federal Technology Transfer Act of 1986186

Index ● 237

Federal Trade Commission (FTC) 223fiber optic communications 88fifth-generation computer project (Japan) 203, 209financial assistance (see small and medium-sized firms)financial environment 9-11,22-23,41-42,93, 96

stability of 11, 23, 96, 106financial markets 9-11,96-112

in Japan 96-105Fiscal Investment and Loan Program (Japan) 161flat panel liquid crystal displays (see liquid crystal

displays)flexible manufacturing networks (see small and medium-

sized firms)Flint River Project, Inc. 227foreign suppliers

components and supplies 15-16, 134-135, 148production equipment 138-142

Freedom of Information Act 193

General Agreement on Tariffs and Trade (GATT)211,218

Generalized System of Preferences (GSP) 217Georgia Tech Industrial Extension Service (GTRI) 55,56,

179-184Germany, Federal Republic of

financial environment 23, 42NC machine tools 151, 155, 181vocational and technical education 118-119, 120, 155

government-industry collaboration (see cooperativeR&D)

government-owned, contractor-operated laboratories(GOCO) 186, 192, 194

government-owned, government-operated laboratories(GOGO) 186, 192, 193-194

high-definition television (HDTV) 80-89, 147, 227high-speed counters 83high-temperature superconductivity (HTS) 77

pilot centers in DOE laboratories 188, 194home mortgage interest deduction 43-44hostile takeovers (see mergers and acquisitions)household savings 22, 43human resources (see education, engineers and scientists,

worker skills, worker training)human resources policy 23-24, 50-53

imitation 8, 69, 212-214, 216Imperial Cup Co. 181-182improved definition television (IDTV) 83, 84, 147Individual Retirement Accounts (IRAs) (see savings)industrial extension (see technology extension)industrial policy (see industry and trade policy)Industrial Technology Institute 184industry and trade policy 21, 79-80Inman, Bobby Ray 203

innovation (see also product development cycle)effect of patents 214-216, 221

institutional investors 23, 47, 106-107intellectual property 20, 31-32, 211-219

copyrights for software 193Federal laboratories 64, 192-194first-to-file and first-to-invent patent systems 219harmonization of patent systems 32, 70-71, 219incentive for innovation 214-216, 221Japan’s patent system 70,218-219patent rights and enforcement 212-213, 214, 217protection of patent rights abroad 70, 212-213relevance to GATT 70, 218

interest rates 42, 97-99, 101, 106International Business Machines (IBM) 142, 146, 205-

206,216International Trade Commission (ITC) 212-213investment (see also capital costs, long-term investment,

time horizons) 9-13,22-23,41-50,96-98effect of takeovers 23, 46-49in equipment and facilities 6, 9-11, 41-46in R&D 9-11, 44, 201long term 9-11,41-50,93-112tax measures to promote 22-23,44-46, 96, 154

investment tax credit (see also accelerated depreciation)22,44-46, 154

Italycooperative networks 167-168

Japan (see also other entries beginning with “Japan” or‘‘Japanese’

automobiles 6-8, 129-135, 158-159banks 98, 99-100, 101-104bilateral exchange programs with U.S. 65-66, 197-200capital costs 10-11, 42, 93-94, 96-105component and parts suppliers 15-16, 129-135, 148consignment manufacture 160consumers 80cooperatives 169-170corporate finance (see main entry)Engineering Research Association Program 208, 209Equipment Leasing System 57, 155, 163Equipment Modernization Loan System 155, 163fifth-generation computer project 199, 203,209financial environment 9-11, 23,42,46, 94-96financial markets 96-105Fiscal Investment and Loan Program 161Industrial Bank of Japan (IBJ) 99, 102industrial policy, industry and trade policy 79-80,

100-102Kanagata 170Kanagawa Science Park 167lean production system 6long-term investment 93, 96-105main-bank system 96, 101-103metalworking industry 168


Ministry of International Trade and Industry (MITI) 88,100, 163, 199, 208-210

Ministry of Posts and Telecommunications (MPT) 88,209

NHK 85,88Next Generation Industries Program 209NSF-Japan programs 65oil shock 104Optical Measurement and Control System Project 209Ota-ward (Tokyo) 163, 158-159, 166patent system 218-219patient capital 93postal savings system 98premier competitor 4-5product development cycle 6, 131-133productivity growth 3R&D consortia 88,208-210rising yen 23, 104savings rate 43small and medium enterprises (SMEs) 27, 155, 158-

159, 161-167Small and Medium Size Enterprise Agency 160, 163Small Business Finance Corp. 163small business programs 27, 152, 154, 161-167SME Upgrading Capital System 169stable shareholding 10, 46, 96, 104-105supplier-customer relations 13-15, 129-135, 157, 159,

168, 169tax incentives for equipment investment, SMEs 155,

162Technology Experimental Center 166technology extension services 53, 54, 56, 151-156,

173-174Temporary Interest Rate Adjustment Law 99Tokyo Stock Exchange (TSE) 103-104U.S.-Japan Cooperative Science Program 198vertical integration 11, 142-145VLSI Project 139-140, 209, 210, 216X-ray lithography 15, 78, 142

Japan Development Bank 100, 101, 102Japan Electric Computer Corporation (JECC) 163Japan Fair Trade Commission 208Japan Key Technology Center (KTC) 209Japan Research and Development Corp. 208Japan Robot Leasing Co. (JAROL) 163Japan Science and Technology Program (MIT) 198, 199Japanese Small Business Corporation 169Japanese technical literature 30, 65, 200Japanese language studies 66, 200joint manufacturing 226-228joint R&D (see cooperative R&D)JTECH 200junk bonds 46-47, 107-108Justice Department 221, 223, 226Justice Department’s Merger Guidelines 229

Koreastrategic technology policy 79

lean production system 6less developed countries (see developing countries)leveraged buyouts (see mergers and acquisitions)liquid crystal displays 85-86long-term investment (see a/so investment) 9-11, 33,

41-50, 93-112effect of capital cost on 9, 42-46, 94effect of high debt on 47, 108, 111in Japan 10-11, 46, 93, 96-105

long-term perspective (see time horizons)Los Alamos National Laboratory 188, 189

machine tools, numerically controlled (NC) 58, 60,152-155, 158-159, 160, 161, 181

military role in development 153-154, 155-156management 5, 12, 20, 50, 52manufacturing engineering and technology 39, 52-60

definition and importance 5incremental improvements 125-126, 151, 156improvement by producing consumer electronics 83,

84, 89, 147Japanese system 6, 145

manufacturing jobs 3, 5manufacturing process (see manufacturing engineering

and technology)Manufacturing Sciences Directorate 53, 65Manufacturing Technology Centers (NIST) 26, 53-56,

176Manufacturing Technology (ManTech) program 175-176Martin Marietta 189Maryland Technical Extension Service (TES) 178-180,

182-183Massachusetts Institute of Technology (MIT) 190mergers and acquisitions (M&A), including hostile

takeovers 23, 46-69, 98, 101, 106, 107-112, 138causes of increased activity in 1980s 46-47, 107-108effects in general 23, 47, 108-112effects on research and development 23, 47, 109-111junk bond financing 46-47, 108tax measures to control 23, 47-49

metalworking firms 174Michigan Modernization Service (MMS) 179-184microelectronics (see electronics)Microelectronics and Computer Technology Corp. (MCC)

203-205military requirements and technology policy 39-40,

153-155Ministry of Finance (MOF) (Japan) 99-100, 105Ministry of International Trade and Industry (MITI)

(Japan) 72,77,86,88, 100visions 77, 199, 208-210

index ● 239

Ministry of Posts and Telecommunications (MPT) (Japan)88,209

Minnesota Governor’s Office of Science and Technology177

mortgage interest deduction (see home mortgage interestdeduction)

Motorola 50, 142, 146Moyco Industries 178, 180mutual shareholding (see stable shareholding)

National Advisory Committee on Aeronautics (NACA)71,74-75, 190

National Aeronautics and Space Administration (NASA)71, 74-75,76, 185-186

National Bureau of Standards (see National Institute ofStandards and Technology)

National Center for Manufacturing Sciences (NCMS)225-226

National Cooperative Research Act of 1984 (NCRA)225-226

National Institute of Standards and Technology (NIST)(formerly National Bureau of Standards) 26,73,77,176-177, 186

National Institutes of Health (NIH) 76, 185, 186national laboratories (see Federal laboratories)National Research Council 195national savings initiative 22, 43National Science Foundation (NSF) 76, 77

Engineering Research Centers 64-65, 173, 180, 195-197

JTECH 200Manufacturing Sciences Directorate 53,65NSF-Japan programs 65-66, 197-200

national securityand technology transfer from Federal laboratories, 194dependence on foreign firms 40, 71changing definition of 32, 40, 7

National Tooling and Machining Association 227NC machine tools (see machine tools, numerically

controlled)NHK (Japan) 85,88North Carolina Microelectronics Center 202not-invented-here syndrome 19, 151, 156

Oak Ridge National Laboratory 185, 188, 189, 190Office of Management and Budget (OMB) 106Office of Science and Technology Policy (OSTP) 195Omnibus Trade and Competitiveness Act of 198826,

53-55

parallel processing 87Patent and Trademark Amendments Act of 1980186patents (see intellectual property)patient capital 10,93

Pennsylvania Technical Assistance Program(PENNTAP) 179, 183

Pennsylvania Technology Management Group (TMC)179-183

pension funds 106-107Perkin-Elmer company 141picking winners 32-34,71-73pixels 82, 86Prairie State program 51product development cycle 156-157

cooperation between suppliers and customers 131-133Japan 19, 156-157length of cycle 19, 157simultaneous engineering 131

productivity and productivity growth 3-4automobiles 7

Program of Research in Modernization Economics (PRIME)184

proprietary data rights 193

R&D tax credit (see research and development tax credit)reporting requirements 105, 107-108research and development consortia 16-17, 77-78, 200-

211research and development spending, international

comparisons 44research and development tax credit 22-23, 44-46retained earnings 101, 159robots 152, 159

Sandia National Laboratories 184-185, 189, 190savings 22, 43-44, 97-98, 106-107

business savings and retained earnings 101consumption tax 43household savings 22,43in Japan 98-99incentives to raise 43Individual Retirement Accounts 43national savings initiative 22, 43taxation or deferred taxation of interest income 43,

science education (see education)Section 301 of the Trade Act of 1974217Section 337 of the Tariff Act of 193031, 218Securities and Exchange Law 104securities transfer excise tax 48Sematech 15,40,71,78, 142, 146,205-208,210-211semiconductor equipment industry 15, 77, 78, 137-142

GCA 138-141Nikon 138-141Perkin-Elmer 141Sematech 142, 146SEMI/Sematech 205wafer steppers 138-141X-ray lithography 142,201

semiconductor industry 11, 15-17, 142-148


capital costs 11, 141, 145, 146DRAMs 145, 146, 157, 205,206, 222Japanese industry and trade policy 145links with consumer electronics 146-147links with equipment suppliers 138-142manufacturing technology 86, 145, 205-208NMB Technologies, Inc. 145Sematech (see main entry)Semiconductor Research Corp. (SRC) 202-203SRAMs 206, 207vertical integration, large size 142-145

Service Corps of Retired Executives 175shareholders 101, 103, 104, 105

stable shareholding (Japan) 10, 46, 96, 104-105Sherman Antitrust Act of 1890219Shop of the 90’s 176-177short-term focus (see time horizons)Showa Precision Tools, Co., Ltd. 164small and medium-sized firms (see also other entries

beginning “small”)cooperative networks or associations of 29, 61, 167-

170financial aid 18,27-29,57, 158, 161-163, 174-175financial aid tied to technical assessment 28, 59-60,

166, 182importance for large customers 25, 130-132Japanese support for 18,27, 152, 161-167, 168-170tax incentives for capital investment 60, 155technology extension services for 25-27, 53-56, 158,

166-167, 173-184Small Business Administration 60, 163, 174, 175Small Business Development Centers 175small business guaranteed loans 57, 60, 174Small Business Innovation Research Program (SBIR) 57,

175Small Business Investment Corp. (SBIC) 57software

copyright 64, 193-194Federal lab-industry cooperation 64, 189, 193-194patent protection 215

Solar Energy Research Institute (SERI) 186, 187Sony Corp. 84,86Spin-Off 188-189stable shareholding (see shareholders)standards setting 228Stanford University 190Startup firms 189-190State industrial extension services (see technology

extension)statistical process control 51Stevenson-Wydler Technology Innovation Act of 1980

186strategic technology policy 32-35, 71-80

picking winners 32-34,71-73relation to industry and trade policy 79-80

supercomputers 87, 147-148, 188

Cray Research, Inc. 147-148, 188superconductivity (see high-temperature

superconductivity)supplier-customer relations (see automobiles,

cooperation)Synfuels Corp. 72

T1 Standards Committee 228Taiwan

strategic technology policy 79takeovers (see mergers and acquisitions)targeted industries 97-98, 187tax credit (see investment tax credit, research and

development tax credit)technical assistance (see small and medium-sized firms,

technology extension)technology diffusion (see also small and medium-sized

firms, technology extension) 25-32,53-60, 151-160, 215

stable supplier-customer relationship 14, 131-133,159-161

technology extension 18, 25-27, 53-56, 166-167, 173-184, 179-184

Federal assistance to States 55-56, 176-177Federal programs 26, 53-56, 174-177Japan 53, 54, 56, 151-156, 173-174Manufacturing Technology Centers (NIST) 26,53-56,

176National Institute for Standards and Technology (NIST)

26, 53-56, 176-177State programs 53-56, 177-184tied to financial aid 59-60, 182training 181-182

technology transfer 159-161, 203-205, 210-211from Federal laboratories 29-30,61-64, 184-195from Japan to the United States 65-66, 157-158,

197-200telematics 211television (see high-definition television)Tennessee Innovation Center (TIC) 189Texas Instruments 207, 216textile and apparel industries 136-137, 169, 177, 201

Dan River 137Greenwood Mills 137National Apparel Technology Center 177Quick Response 136-137textile machinery industry 15Textile/Clothing Technology Center (TC2) 177

Thermco 230-231Thomson Consumer Electronics 87, 88time horizons 9-11, 23, 33, 93

R&D consortia can lengthen 16,201,205Tnemec Co. 179Tokyo Electron (TEL) 230-231Toyota 130, 160Trade Adjustment Assistance 28, 176

Index ● 241

trade and industry policy (see industry and trade policy)trade deficit 3, 39

importance of manufacturing 3trade policy (see industry and trade policy)training (see worker training)transaction tax (see securities transfer excise tax)TV (see high-definition television)

wages 39, 133work force (see engineers and scientists)worker skills (see also education, worker training)

115-120changing technology and work environment 12,23-24,

117foremen 120

worker training 24, 51, 155, 158, 181-182U.S. International Trade Commission (USITC) 212-213U.S.-Japan Cooperative Science Program 198

X-ray lithography 15, 142, 201venture capital 189-190VLSI project (Japan) 139-140,209, 210, 216vocational training (see education, worker skills, worker

training) Zenith 84, 88, 147

u. s. Government PRIbTI’ING OFFICE : 1990 0- 21-700

Date post:	02-Dec-2014
Category:	Documents
Upload:	gaurav-shakya
View:	59 times
Download:	4 times