Fall 1998jbm.johogo.com/pdf/volume/0602/JBM-0602.doc · Web viewPublished jointly by the Western...

ACCOUNTING / FINANCE

QUANTITATIVE METHODS/

MANAGEMENT INFORMATION SYSTEMS

SUMMER 1999

Summer 1999 Vol. 6, No. 2

Haroldene Wunder International Tax Reform and the Implication for

Multinational Financial Management

.0

Albert King A Crescendo Model of Career Motivation and Commitment Kimi King

Ed Bukszar Group Heterogeneity and Strategic Decisions: Do Informal “Consider the Opposite” Processes Improve Decision Efficacy?

Zvi Drezner A Computational Procedure for Setting Cutoff Scores for George Wesolowsky Multiple TestsWilli Wiesner

Mustafa Yilmaz Six-Sigma Quality in Small Businesses: A Genesis For GrowthSanjit Chaterjee

Published jointly by the Western Decision Sciences Institute and the School of Management, California State University, Dominguez Hills

1

MANAGEMENT

JOURNAL OF BUSINESS ANDMANAGEMENT

JOURNAL OF BUSINESS AND MANAGEMENT

JOURNAL OF BUSINESS AND MANAGEMENTTHE OFFICIAL PUBLICATION OF THE WESTERN

DECISION SCIENCES INSTITUTE (WDSI)

The Decision Sciences Institute is a professional society dedicated to the development and application of quantitative and behavioral methods to administrative problems. Most functional areas of business are represented among the membership. Through its journals, national and regional meetings, and other activities, the Decision Sciences Institute serves as a vehicle to advance and disseminate the theory, application, pedagogy, and curriculum development of the decision sciences.

Western Regional Officers 1999-2000President, Karen L. Fowler, University of Northern ColoradoPresident-Elect, Marc Massoud, Claremont McKenna CollegeVice President for Programs, Paul Mallete , Colorado State UniversityVice President for Programs-Elect, Eldon Y. Li, California Polytechnic State University, San Luis Obispo Vice President for Member Services, Krishna S. Dhir, Penn State HarrisburgSecretary/Treasurer, Richard L. Jenson, Utah State University

2

SUMMER 1999

JOURNAL OF BUSINESSAND MANAGEMENT

Vol. 6, No. 2 Summer 1999

EDITORS

EDITORIAL ASSISTANT

Franklin StrierBurhan F. YavasDeitra Temple

Editorial Offices:JOURNAL OF BUSINESS AND MANAGEMENTSchool of ManagementCalifornia State University, Dominguez Hills1000 East Victoria StreetCarson, California 90747Phone: (310) 243-3472, (310) 243-3501Fax: (310) 516-3664, (310) 217-6964

Published jointly by Western Decision Sciences Institute (WDSI) and the School of Management, California State University, Dominguez Hills. The purpose of the JOURNAL OF BUSINESS AND MANAGEMENT is to provide a forum for the dissemination of contributions in all fields of business, management and related public policy of relevance to academics and practitioners. Original research, reports and opinion pieces are welcome. The style should emphasize exposition and clarity, and avoid technical detail and jargon.

The views expressed in articles published are those of the authors and not necessarily those of the Editors, Executive Board, Editorial Board, WDSI or California State University, Dominguez Hills. All submissions will be reviewed initially by the editors and, if judged appropriate, will be sent to knowledgeable referees for review. The authors assume responsibility for the accuracy of facts published in the articles.

Copyright ©1999 WDSI and by the School of Management, California State University, Dominguez Hills. Subscriptions are $20/year. Manuscripts should be double-spaced and submitted in triplicate. Manuscripts and comments should be directed to the editors.

3



Executive BoardKaren L. Fowler, President, WDSIMarc Massoud, President -Elect, WDSI

Donald L. Bates, Dean, School of Management, CSUDH Franklin Strier, Editor Burhan F. Yavas, Editor

Editorial Board Dr. Joseph R. Biggs California Polytechnic State University, San Luis Obispo Dr. Henry Brehm University of Maryland Dr. Terry E. Dielman Texas Christian University Dr. Moshe Hagigi Boston University Dr. Ronald H. Heck University of Hawaii at Manoa Dr. Richard C. Hoffman Salisbury State University, Maryland Dr. Marc T. Jones University of Otago, Dunedin, New Zealand Dr. Erdener Kaynak Pennsylvania State University Dr. Thomas Kelly State University of New York, Binghamton Dr. George R. LaNoue University of Maryland Dr. George A. Marcoulides California State University, Fullerton Dr. John Preble University of Delaware Dr. Arie Reichel Ben-Gurion University of the Negev, Israel Dr. Elizabeth L. Rose University of Auckland, New Zealand Dr. Anne S. Tsui The Hong Kong University of Science and Technology, Hong Kong Dr. Michael Useem University of Pennsylvania

4

SUMMER 1999

Reviewer Acknowledgments

The editors of the Journal of Business and Management wish to express their appreciation to the following individuals who have reviewed manuscripts submitted for consideration in this issue of the Journal of Business and Management.

Vincent Apilato Laurence Barton Potkin BasseerJustine BellHamdi BiliciChing-Chang KuoChee ChowDorothy FisherSteven FerraroCharles FojtikJames HeathRonald HeckSwinder JandaSang-Hoon KimMiriam LaceyGeorge MarcoulidesIsaac MontoyaKurt MotamediRichard NehrbassGolnaz SadriMark SimkinElizabeth TrybusMichael Useem

5


6

SUMMER 1999JOURNAL OF BUSINESS

AND MANAGEMENT

TABLE OF CONTENTS

From the Editor’s Desk……………………………………………………….………….9

International Tax Reform and the Implications for Multinational Financial Management………….…………………………………………………………………11

Haroldene Wunder

Crescendo Model of Career Motivation and Commitment………………….………….37Albert KingKimi King

Group Heterogeneity and Strategic Decisions: Do Informal “Consider the Opposite” Processes Improve Decision Efficacy? …………………..…………………70

Ed Bukszar, Jr.

A Computational Procedure for Ssetting Cutoff Scores for Multiple Tests………….…86Zvi DreznerGeorge WesolowskyWilli Wiesner

Six-Sigma Quality in Small Businesses: A Genesis for Growth………………….…..100Mustafa YilmazSangit Chaterjee

7


8

SUMMER 1999

FROM THE EDITOR’S DESK

For many Multinational Corporations (MNCs), the corporate income tax rates of the countries in which they operate are major considerations. HAROLDENE WUNDER looks at the changing tax policies in fourteen countries and their impact on multinational financial management decisions.

Career motivation and commitment have attracted renewed interest. Utilizing data from a large aerospace firm, ALBERT KING and KIMI KING shed insight into motivation patterns formation and affective career commitment.

In team decision making, management faces certain key analyses: Does exposure to differing viewpoints improve decision making? Is the diversity of the team more important than the process? ED BUKSZAR examines these issues in a hindsight bias, debiasing experiment.

Employees often use cutoff scores to ascertain whether applicants meet minimum standards. ZVI DREZNER, GEORGE WESOLOWSKY and WILLI WEISER demonstrate a procedure for setting multiple cutoff scores that is arguably more cost-effective and accurate than the subjective methods currently used.

Six-Sigma Quality has become a popular tool in the management process of quality control. Although usually associated with large-scale production systems, MUSTAFA YILMAZ and SANGIT CHATTERJEE illustrate its application to small businesses that are actively striving for growth.

Franklin StrierBurhan F. Yavas

9

10

SUMMER 1999

INTERNATIONAL TAX REFORM AND THE

IMPLICATIONS FOR MULTINATIONAL FINANCIAL

MANAGEMENT

Haroldene F. Wunder *

Since 1986, a common trend in corporate income tax rates has emerged in the 14 countries studied in this research. On average, the highest corporate income tax rate in the 14 countries declined 11 percentage points, from 46.0 percent in 1985 to 35.0 percent in 1997. This convergence in rates should diminish the role of inter-country tax regimes in multinational corporations’ (MNCs’) repatriation decisions. However, significant inter-country differences still exist in tax holidays and in attitudes and policies toward foreign investment.

Multinational corporations (MNCs) possess a unique advantage: the ability to move funds and profits among various units through the use of internal financial transfer mechanisms, i.e., networks of financial linkages. These networks of financial linkages have value because of variations in national tax systems and in costs and barriers associated with international financial transfers.

Variations in national tax systems have historically played an important role in international financial management. For example, MNCs can engage in tax arbitrage and reduce their tax burden by shifting profits from units domiciled in high-tax countries to those in lower-tax countries. In the alternative, they may shift profits from business units in a taxpaying posture to those with tax losses.

Additionally, one of the most important uses of a judicious transfer pricing policy is to reduce taxes and tariffs, as well as to avoid exchange controls. Also, the decision to invoice an international transaction in one currency rather than another is affected by the relative tax rates in the foreign countries involved.

*Haroldene F. Wunder is affiliated with California State University, Sacramento

11


Finally, and most importantly, the design of a global remittance policy is affected by the variation in international tax rates. A global remittance policy requires the following interrelated decisions (Shapiro, 1996):

1. How much money (if any) to remit;2. When to remit;3. Where to transmit funds; and4. Which transfer method(s) to use.

Decision 3-Where to transmit funds-is largely determined by variations in international tax regimes. Although their relative importance has diminished, dividends are still the most important means of transferring funds from foreign affiliates to the parent company. Shapiro (1996, p. 439) estimates that dividends account for more than 50 percent of all remittances to U.S. firms.

Moffett and Davison’s (1996, p. 5) data are consistent with that estimate. Their study indicates that in 1993 dividends represented approximately 56 percent of total repatriated funds; whereas in 1989 they represented 74 percent of repatriated funds. While dividends’ role in repatriated funds has decreased, the roles of royalties and license fees, charges for services, and net interest received has become more prominent. Even though their relative importance has shifted, all types of repatriated funds should be sensitive to variations in international income tax rates. For example, repatriating earnings to the parent in the form of management fees is relatively more attractive than repatriation by dividend because some countries (e.g., Switzerland) permit a deduction for expenses paid to the home office (i.e., parent) (Scholes and Wolfson, 1992, p. 250).

Even though many countries, including the United States, permit MNCs tax credits for income taxes paid to foreign jurisdictions, the credit mechanism fails to make the MNC “whole” when the relevant foreign tax rate exceeds the U.S. tax rate. In this case, the MNC is in an “excess credit” situation because the credit against U.S. income taxes is limited to what the foreign income tax would have been if (the lower) U.S. rates applied.

The purpose of this research is to show that the relative role of variations in international income tax rates in certain aspects of international financial management is diminishing because of the convergence in international income tax rates that began in the mid-1980s. However, when combinations of tax regimes are considered [(e.g., income taxes, value-added taxes (VATs), capital gain taxes, tax holidays, and withholding-tax policies)], significant inter-country tax differences are found to still exist. Those differences continue to impact MNC’s financial and locational decisions.

The next section consists of a discussion of the following topics: repatriation models, relative tax rates, magnitude and trends in income tax rates in 14 developed countries from 1985 to the present, and other factors that affect MNCs’ financial decisions, i.e., general policies toward foreign investment, specific tax incentives offered to attract inbound investment, taxation of corporations’ distributions to foreign parents, scope of corporate income tax, presence of significant indirect taxes, and taxation of capital gains. The combined effect of corporate tax income tax, value-added tax, capital gains tax rates, and tax concessions is also discussed, along with implications for future research. The final section provides the conclusions of the research.

12

SUMMER 1999

DISCUSSIONRepatriation Models

The following discussion of repatriation models flows significantly from Moffett and Davison’s (1996) work. Only the more salient elements of each model are discussed here; Moffett and Davison (1996) discuss each model in greater detail.

The early literature regarding repatriation of foreign earnings assumes that MNCs are profit maximizers and focus on transfer pricing mechanisms to accomplish their objective. That body of literature views MNCs as exploiting tax differentials through transfer pricing and dividend repatriation (Lintner, 1956).

Kopits (1972) distills these philosophies into a theory of dividend repatriation behavior that focuses on differentials in tax rates, and empirically tests the theory with 1962 data on United States MNCs. As hypothesized, he finds an inverse relation between U.S. corporate income tax rates and dividend repatriation.

Adler (1979) develops a U.S. multinational taxation model based on the assumption that the MNC’s objective is to maximize consolidated worldwide after-tax cash flows, not profits. His model assumes tax effects not only from repatriations in the form of dividends, but also from royalty payments, intra-firm interest payments, and fees. He also addresses the consequences of U.S. MNCs being in an “excess credit” position when foreign corporate income tax rates exceed U.S. corporate income tax rates.

Mutti (1981) abandons the profit- and cash-flow maximization assumptions. Instead, he assumes that an MNC sets a target income from each foreign subsidiary and then allocates repatriated earnings among dividends, interest, and royalties in the most tax efficient way. His statistically significant nonparametric test results obtain with regard to both profit rates and total repatriations relative to earnings. His results suggest that other flows are also responsive to tax incentives (Mutti, 1981, p. 247).

In contrast, survey studies by Zenoff (1968) and Robbins and Stobaugh (1973) find that few MNCs manage by focusing on worldwide profit- or cash-flow maximization. Those studies conclude that most parent corporations allow their affiliates to retain enough cash to meet their fund requirements and require them to repatriate the rest.

Relative Tax Rates

Tanzi (1987) studies tax reform, beginning in 1986, in 14 different countries1 to discern if the 1986 U.S. tax reform was the impetus for tax reform initiatives elsewhere. He hypothesizes that the U.S. tax movement in 1986 may have “provided the officials of other countries with both a challenge and an opportunity to introduce changes in their own tax systems” (Tanzi, 1987, p. 339).

Instead, Tanzi concludes that tax reforms in both the United States and other countries in the mid-1980s were “political responses to accumulating evidence on the effects of tax systems on the economy” (1987, p. 339). Furthermore, he observes that “rarely has there been such a convergence of views on at least some of the aspects of the tax system that needed to be modified” (Tanzi, 1987, p. 339).

13


In other words, Tanzi found that, as of 1986, tax rates and tax regimes across countries were becoming more similar as a result of essentially contemporaneous tax reform initiatives. Grubert, Randolph and Rousslang (1996) agree that foreign tax reform occurred concurrently with, and not as responses to, U.S. tax reform in the mid-1980s. The initiatives occurred because of various countries’ assessments that existing tax

systems negatively affected their economies and were increasingly viewed as unfair (Tanzi, 1987, p. 339). Furthermore, the tax reform initiatives in the mid-1980s were similar across countries in that they were characterized by the lowering of tax rates and the broadening of the tax base (Slemrod, 1990; Whalley, 1990).

Tanzi’s comprehensive study appears only one year after the U.S.’s Tax Reform Act of 1986. He, therefore, lacks a lengthy vantage point. In contrast, this article benefits from a 12-year retrospective vantage point and examines the 1985-97 corporate tax rates of 14 countries hose orporate rates Tanzi studied. This time period provides sufficient comparative data to demonstrate that the cross-country tax reform momentum set in motion in the mid-1980s continues to the latter-1990s. In other words, the study period is of sufficient length to establish the existence of a pattern of converging corporate income tax rates. Value added tax (VAT) data are also provided because the post-1986 trend is to finance lower income tax rates with indirect taxes, e.g., VATs.

Magnitudes and Trends in InternationalCorporate Income Tax Rates

The data in Table 1 show that from 1985 to 1997 all 14 countries studied by Tanzi (1987) reduced their highest corporate income tax rate significantly. Until 1997, Italy was an outlier, raising its highest corporate rate from 46.37 percent in 1985 to 53.2 percent. In 1997, however, Italy followed suit and reduced the top corporate rate to 37.0 percent, a rate in line with the other countries studied.

The average corporate rate across the 14 countries dropped from 46.0 percent in 1985 to 35.0 percent in 1997. The most prominent rate decrease occurred in Sweden, which reduced its corporate rate from 52.0 percent in 1985 to 28.0 percent in 1997, a 24.0 percentage-point decrease! The smallest decrease in the top corporate rate occurred in Japan, where the rate decreased 5.8 percentage points, from 43.3 percent in 1985 to 37.5 percent in 1997.

Table 1 also demonstrates how similar the corporate income tax rates are in the 14 countries. Ten of the countries’ corporate rates cluster around low- to mid-30 percent. The lowest top rates are 28.0 percent (Finland and Sweden), while the highest rates are 45.0 percent (Germany and Ireland).

Clearly, the role of differential income tax rates in corporate repatriation decisions is diminished by the convergence of rates across countries that has occurred in the last ten years. However, Germany is unique in that, even though its corporate income tax rates have declined from 1985 to 1997, it taxes undistributed earnings more heavily than distributed earnings. The tax rate on distributed earnings was 36.0 percent in 1985 and declined to 30.0 percent in 1997. In contrast, undistributed earnings were subject to a 56.0 percent rate in 1985 and to a 45.0 percent rate in 1997. The discrepancy between

14

SUMMER 1999rates applied to distributed and undistributed earnings makes repatriation of earnings from German affiliates relatively more attractive.

Other Factors Affecting Multinational Decisions

Although repatriation decisions should be less sensitive to inter-country income tax rates because of the effect of post-1986 international tax reform, one should not conclude that differences in international tax regimes are no longer important in multinational decisions. MNCs are also affected by, at least, the following factors:

General policies toward foreign investment Specific tax incentives offered to attract inbound investment Taxation of corporations’ distributions to foreign parents Scope of corporate income tax Presence of significant indirect taxes (e.g., VATs) Taxation of capital gains

Each factor is addressed below as it pertains to the 14 countries studied.

General Policies Toward Foreign Investment

Table 2 describes the 14 countries’ general policies toward foreign investment.None of the 14 countries restrict the repatriation of capital or dividends. In fact, Canada, Ireland and Sweden actively encourage inbound foreign investment. The remaining countries strive for systems that are neutral toward foreign investment. Therefore, little inter-country variation exists in terms of general policies toward foreign investment since all are either neutral to or supportive of inbound investment.

Specific Tax Incentives Offered to Attract Foreign Investment

The tax hospitality with which new foreign investment is greeted varies considerably across countries. For example, Ireland’s tax holiday posture stands out in the data presented in Table 3.

The Republic of Ireland has put in place generous tax concessions to attract inbound foreign investment. Most notable is the 10 percent tax rate applied until December 31, 2010 to profits from new investments in manufacturing (consisting of 18 industries), computer services, engineering, research and development, and certain financial services. Scholes and Wolfson (1992, p.258-9) provide additional commentary on Irish tax concessions, particularly as they pertain to international financial service centers (IFSCs).

Table 3 reveals that Australia, France, and Japan also offer significant general tax incentives to foreign investment. Canada, Denmark, France, Japan, and the United States offer substantial regional tax incentives to inbound investment. In addition, various countries (e.g., Australia, Canada, Finland, Ireland, Japan, the Netherlands, and New Zealand) offer additional tax concessions to foreign investment in certain industries, particularly those that engage in research and development.

15


16

SUMMER 1999

Table 2Policies Toward Foreign Investment

COUNTRY POLICIES TOWARD FOREIGN INVESTMENT

Australia Government does not restrict the repatriation of capital or dividend payments on investments in Australia, thus contributing to Australia’s reputation as a secure location for investments.

Canada Canada imposes no foreign exchange controls and no other restrictions on the repatriation of profits or capital by foreign investors. The federal and provincial governments offer a wide range of tax and nontax incentives in the form of grants, loans, and guarantees.

Denmark The government neither favors nor discriminates against inbound foreign investment. Foreign-owned enterprises are treated on an equal basis with local Danish companies.

Finland There are no restrictions on the repatriation of capital or profits and labor costs. Except for certain tax incentives pertaining to development areas, the taxation policy emphasizes neutrality.

France Both profits and capital may be repatriated. Tax concessions are available to both domestic and foreign investors.

Germany There are no restrictions on the transfer of profits abroad or on repatriation of capital.

Ireland Foreign investment is actively encouraged through tax incentives, cash grants, interest subsidies, loan guarantees, and equity finance. Repatriation of profits and capital is not restricted.

Italy Capital and profits are freely remitted abroad. The industrial tax initiatives set up for Italian companies apply equally to foreign investors.

Japan Repatriation is not restricted. The basic taxation system neither favors nor discriminates against inbound foreign investment.

Netherlands There are no restrictions on repatriation. The tax system is neutral toward foreign investors.

New Zealand

Generally, the government’s policy is one of tax neutrality; most incentives have been removed. Repatriation is not restricted.

Sweden Free movement of capital is preserved. There is a generous and open attitude toward foreign investment.

United Kingdom

Repatriation is not restricted. Generally, the UK tax system is neutral towards foreign investment.

United States

The government attempts to remain neutral regarding tax treatment of foreign and domestic capital investment.

Source: Doing Business in . . . Series. New York: Price Waterhouse World Firm Services.

17


18

SUMMER 1999

19


20

SUMMER 1999

21


Following is a list of the types of tax incentives offered and the countries in which they are applicable:

General, Regional, and Industry-Specific Incentives: Japan

General and Regional Incentives: France

General and Industry-Specific Incentives: Australia, Ireland

Regional and Industry-Specific Incentives: Canada

Regional Incentives: Denmark, United StatesIndustry-Specific Incentives: Finland, Netherlands, New Zealand

The types of incentives provided are probably attributable to country-specific characteristics. For example, Canada’s organization as a federation of provinces likely accounts for its extensive regional tax incentives.

Taxation of Corporations’ Distributions to Foreign Parents

Eleven of the 14 countries subject corporations’ distributions to their foreign parent corporations to withholding tax. Table 4 describes each country’s withholding regime.

22

SUMMER 1999

Table 4Taxation Of Foreign Corporations’ Distributions

COUNTRY TAXATION OF FOREIGN CORPORATIONS

Australia A company incorporated outside Australia is not a resident of Australia unless it carries on business in Australia and either its central management and control is in Australia or its voting power is controlled by residents of Australia. Nonresident foreign corporations are taxed on Australian-source income but are exempt from tax on income derived from sources outside Australia.

Dividends paid by an Australian company to a foreign corporation are exempt from withholding if they are franked. Other dividends are subject to withholding tax (e.g., 15% rate if paid to a U.S. corporation).

Canada A corporation is generally considered nonresident for Canadian tax purposes if it is incorporated outside Canada and its central management and control reside outside Canada. Nonresident corporations are subject to Canadian income tax in respect of income derived from carrying on business in Canada and capital gains from the disposition of taxable Canadian property.

Dividends paid by Canadian subsidiaries to nonresident parent corporations are subject to nonresident withholding taxes (generally 10 or 15 percent to treaty countries).

Denmark In general, treaties limit the taxation of industrial and commercial activities in Denmark to the profits attributable to a permanent establishment in the country.

Dividends paid to a foreign corporation are subject to withholding tax at a 30%, or lower tax treaty, rate.

Finland A nonresident company having a permanent establishment in Finland is subject to Finnish taxes on the income earned through the Finnish permanent establishment.

Dividends paid to nonresident shareholders, excluding stock dividends, are subject to a withholding tax of 25 percent if there is no applicable tax treaty rate.

France Nonresident foreign corporations are subject to corporate income tax if they are engaged in a trade or business in France.

Dividends paid to foreign shareholders are subject to a 25 percent withholding tax or a reduced tax treaty rate (e.g., 5 percent if paid to a

23


COUNTRY TAXATION OF FOREIGN CORPORATIONS

U.S. parent).

GermanyGermany Cont.

A foreign corporation is considered a limited taxpayer in Germany and is taxable by reference to its German-source income only.

Dividends paid by German subsidiaries to parent corporations are subject to a withholding rate of 25 percent or reduced treaty rate (e.g., 5% if paid to a U.S. parent).

Ireland No withholding tax is levied on dividends paid to nonresidents.

Dividends issued to a parent resident in a treaty country are not subject to advance corporation tax (ACT) if the parent satisfies certain equity ownership requirements.

Italy In general, treaties limit the taxation of business activities to profits attributable to a permanent establishment in Italy.

Dividends paid to a foreign corporation are subject to various withholding tax rates. (The applicable rate for dividends paid to U.S. parents is 5 percent.)

Japan Dividends paid to a foreign corporation are generally subject to Japanese withholding at a 20% or lower treaty rate.

Netherlands Dividends are generally subject to a 25 percent withholding rate unless reduced (frequently substantially) under a Dutch tax treaty.

New Zealand

Dividends are subject to withholding tax on the same basis as that which applies to foreign shareholders generally, i.e. 15 percent for most treaty countries, 30 percent otherwise.

Sweden Dividends from subsidiaries to foreign parent companies are subject to a 30 percent, or lower treaty, withholding tax rate.

United Kingdom

Dividends: The United Kingdom does not impose a withholding tax on the distribution of profits.

United States

Generally, U.S.-source dividends received by a foreign corporation from its U.S. subsidiary are subject to a 30 percent, or lower treaty, withholding tax rate.


24

SUMMER 1999

Only Australia, Ireland, and the United Kingdom lack a withholding requirement. The withholding rates in the other 11 countries range from five to 30 percent, with lower rates applied to distributions to parents in countries with which they have double taxation treaties.

25


Scope of Corporate Income Tax

The data in Table 5 confirm the similarity in the scope of the 14 countries’ corporate income tax.

Table 5Entities Exempt From The Corporate Income Tax

COUNTRY EXEMPT ENTITIES

Australia 1. Municipal corporations and public authorities2. Mutual and nonprofit organizations3. Cooperative societies and companies4. Funds established for public charities or public education activities5. Registered trade unions6. Employers’ associations7. Nonprofit societies for encouragement of music, art, science, or

literature.Canada 1. Government-controlled corporations

2. Charitable organizations3. Nonprofit corporations4. Mutual insurance corporations

Denmark All entities except: public limited corporations, private limited companies, other entities with limited liability, and cooperative societies and associations.

Finland 1. State, municipal, and religious communities2. Nonprofit organizations for education, science, arts, public

defense, politics, and charity.France All entities except: corporations, LLCs, limited partnerships with

shares, silent partners of limited partnerships, and permanent establishments of foreign head offices.

Germany All entities except: all German corporations.Ireland 1. Charities if profits are used for the purposes of the charity.

2. Approved bodies formed to promote athletic or amateur games or sports.

3. Credit unions.Italy All entities except: corporations, LLCs, incorporated partnerships,

cooperatives, and mutual insurance companies with a legal or administrative headquarters or principal business purpose in Italy.

Japan Public welfare corporations

26

SUMMER 1999

COUNTRY EXEMPT ENTITIES

Netherlands 1. Nonprofit public and private welfare organizationsGeneral social security funds

New Zealand

1. .Charitable, educational, religious, or sporting associations2. Nonprofit organizations3. Associations formed for the mutual benefit of members

Sweden Foreign legal entities earning royalties from Sweden (de facto exemption)

United Kingdom

All entities except “companies” (which includes any body corporate, whether incorporated in the UK or elsewhere, or unincorporated association, but does not include a partnership or local authority).

United States

1. Federal and related agencies2. Religious, charitable, educational, scientific, literary, etc.

organizations3. Civic leagues and employee unions4. Labor, agricultural, and horticultural organizations5. Business leagues, chambers of commerce, real estate boards, etc.6. Social clubs7. Fraternal beneficiary societies8. Voluntary employees’ beneficiary associations9. Local teachers’ retirement fund associations10. Local benefit life insurance associations, mutual or cooperative

telephone companies, etc.11. Cemetery companies12. Credit unions13. Mutual insurance companies14. Farmers’ cooperatives15. Armed forces members’ posts or organizations16. Group legal service plans17. Religious and apostolic organizations18. Cooperative hospital service organizations19. Cooperative service organization of educational institutions20. Qualified state tuition programs


27


The majority of the countries exempt municipal and other public authorities, charities, and not-for-profit organizations. Therefore, the corporate income tax reaches similar economic segments of all countries studied.

Value-Added Taxes (VATs)

In Europe, the post-1986 trend has been to finance lower income tax rates by imposing or expanding indirect taxes, e.g., VATs. Table 6 contains data pertaining to standard VAT rates in the 14 countries.

Table 61997 Corporate Income And Value-Added Tax (Vat) Rates

COUNTRY PROFITS/INCOME TAX VALUE/ADDED TAXAustralia 36.00% NoneCanada 38.00% None1

Denmark 34.00% 25.00%Finland 28.00% 22.00%France 33.33% 20.60%Germany 30.00%/45.00% 15.00%Ireland 36.00% 21.00%Italy 37.00% 19.00%Japan 37.50% NoneNetherlands 36.00% 17.50%New Zealand 33.00% None2

Sweden 28.00% 25.00%United Kingdom 31.00% 17.50%United States 35.00% NoneAverage 34.52% 20.29%

Source: Corporate Taxes: A Worldwide Summary, 1997 Edition, New York: Price Waterhouse World Firm Services BV, Inc.

28

SUMMER 1999

Ten of the 14 countries have both a corporate income tax and a VAT. It is important to note that none of the ten countries replaced their corporate income tax with a VAT. Instead, VATs are used as a supplement to, and a means of financing, the lower corporate income tax rates.

Therefore, any conclusions drawn regarding the trend in decreasing corporate income tax rates must be tempered by the existence of VATs and the combined effect of the two tax regimes. The combined corporate income tax and value-added tax rates in the ten countries with both tax regimes range from 42.50 percent (Japan) to 60 percent (Germany).

Taxation of Capital Gains

Countries’ capital gains tax regimes also affect the inter-country cost of doing business. The overwhelming majority of the 14 countries studied tax corporate capital gains at ordinary rates. Table 7 provides a concise explanation of the taxation of capital gains in each country. The data in Table 7 are further evidence of the growing similarity in corporate taxation across international boundaries. Few inter-country differences exist.

Combination of Tax Regime Effects

The combination of income tax, VAT, and capital gains tax regimes has the potential to cause significant differences in the total tax cost of multinational activities. For example, the combination of no preferential treatment of capital gains and the existence of VATs make Denmark, Germany, Italy, the Netherlands, Sweden, and the United Kingdom relatively undesirable business locations from a tax-cost perspective.

In addition, these six countries are not among those offering generous tax concessions to encourage foreign investment. Germany, Italy, Sweden, and the United Kingdom provide no specific tax concessions to foreign investment. Denmark and the Netherlands provide only regional and industry-specific incentives, respectively.

Although the Republic of Ireland offers an array of tax concessions and tax holidays to new foreign investment, its combined income and value-added tax rate is a sizable 57 percent. Furthermore, Ireland’s standard corporate capital gains tax rate is 40 percent. Ireland’s tax hospitality stems from general and industry-specific tax incentives, not from low income, value-added, or capital gains tax rates.

29


Table 7Capital Gains Tax Regimes Corporations

COUNTRY TAXATION OF CAPITAL GAINS

Australia The capital gains tax applies to assets acquired on or after September 20, 1985. Capital gains are subject to tax at ordinary rates of tax. Capital losses are allowable as deductions only against capital gains and cannot be offset against other income .

Canada Three-quarters of a capital gain is taxable, included in the corporation’s income, and taxed at ordinary rates. Three-quarters capital losses are deductible, but only against capital gains. A three-year carryback and infinite carryforward provision exists.

Denmark Capital gains and/or losses are included in ordinary taxable income. A capital gain on the sale of shares held for more than three years is tax exempt. Special rules apply to investment subsidiaries in a tax haven and to sales of real estate.

Finland Capital gains and losses are generally included in taxable income. For example, sales proceeds are included in taxable income and the undepreciated tax basis of the asset sold is deducted in the year of sale.

France Short-term capital gains (on assets held less than two years) are generally taxable as ordinary income. Long-term capital gains on depreciable assets are taxed at the reduced rate of 19% on the excess of sales price over original cost; however, depreciation is recaptured as short-term gain. Long-term capital gain on nondepreciable assets held for two or more years and on shares representing at least 10% interest in a company held for more than two years are treated as long-term capital gains.

Germany Capital gains (and losses) are taxed as part of ordinary business income (or losses). Postponement of gain recognition on certain fixed assets is possible when the gain is offset against the cost of certain replacement items.

Ireland Since April 6, 1992, a single rate of 40% applies to all capital gains. Capital losses can only be used as offsets to capital gains in the same year or carried forward indefinitely.

Italy Capital gains are generally taxable in full in the tax period in which they are reported in the corporation’s accounting records. Tax on the capital gain from the sale of fixed assets may be included in taxable income ratably over a period not exceeding five years. (Prior to 1989 the

30

SUMMER 1999

COUNTRY TAXATION OF CAPITAL GAINS

installment period was not to exceed ten years.) For periods ending December 31, 1993, and later, installment recognition is permitted only for gains from dispositions of fixed assets held for more than three years.

Japan Generally, capital gains or losses are treated as ordinary income or losses. However, capital gains on the transfer of land and the right to land located in Japan made on or after January 1, 1992 are subject to additional or separate taxation methods depending on the holding period, as follows:

Holding period TaxationTwo years or less Separate taxation; standard corporation tax

rate plus 30%. Over two years but Regular corporation tax, plus additional tax of

20%not more than five yearsOver five years Regular corporation tax, plus additional tax of

10%

Netherlands Generally capital gains are taxed as ordinary income. However, special rules apply to capital gains from dispositions of certain stock investments and depreciable property. Capital losses are generally deductible.

New Zealand

There is no capital gains tax.

Sweden Capital gains and losses are generally taxed as ordinary income or loss. Capital losses on certain financial instruments (e.g., stock, stock options, and convertible debentures) are allowed only as an offset to capital gains on the same group of assets.

United Kingdom

Capital gains are taxed at the regular corporation tax rate of 33%. Capital losses are allowed only as offsets to capital gains; however, excess capital losses may be carried forward without limitation.

United States

After netting short- and long-term capital gains and losses, any net capital gain is included in gross income and taxed at the regular corporate tax rates. Capital losses are allowed only as offsets to capital gains. Excess capital losses may be carried back three years and carried forward five years to offset capital gains.

31


Source: Corporate Taxes - A Worldwide Summary, 1997 Edition, Price Waterhouse World Firm Services.

Even though Japan has both corporate income and value-added taxes, the combined rate is a relatively moderate 42.50 percent. Combined with its general, regional, andindustry-specific tax incentives, and its multi-level system for taxing capital gains from land sales (see Table 7 for details), Japan’s tax climate is relatively welcoming to foreign investment.

It is evident from the foregoing data and discussion that a trend in any particular tax factor (e.g., corporate tax rates) must be analyzed in conjunction with numerous other factors to determine the effect on MNCs’ financial decisions. Furthermore, not only general, but also regional, local, and industry, systems must be taken into account.

In addition, a country’s income tax rate is a measure of tax cost. The societal benefits accruing from the imposition of any particular tax are also relevant, although beyond the scope of this particular research.

Implications for Future Research

The international tax rate data provided herein suggest that future empirical research is in order to determine the response of individual MNCs to the international convergence of corporate income tax rates over the past ten years, the corresponding imposition (or expansion) of VATs, and the variations in capital gain tax regimes. Such research should focus on both repatriation and locational decisions.

Furthermore, the January 1, 1999, adoption of the Euro as the national currency of participating member states of the European Union (EU) signals a movement toward even greater economic uniformity among the 15 EU countries, which should result in a continuation of convergence in income tax rates. Future research should study repatriation and location decisions from both retrospective (i.e., previous tax reform) and prospective (i.e., anticipated tax, and other economic, reform).

32

SUMMER 1999

CONCLUSION

Since 1986, a trend in the corporate income tax has emerged in the 14 countries studied. On average, the highest corporate income tax rate in the 14 countries declined 11 percentage points, from 46.0 percent in 1985 to 35.0 percent in 1997. This convergence of corporate income tax rates should diminish the role of inter-country tax regimes in repatriation decisions involving dividends, royalties, license fees, charges for services and net interest received.

Other factors affecting MNCs’ financial decisions were identified and discussed. For the most part, the 14 countries were similar in terms of general policies toward foreign investment, taxation of corporations’ distributions to foreign parents, and the scope of the corporate income tax. Some differences exist in the tax holidays offered by the 14 countries, with the Republic of Ireland providing the most generous tax incentives for inbound investment.

In addition, European countries have responded to the decline in corporate income tax rates by imposing (or expanding) indirect taxes, e.g., VATs. Ten of the 14 countries have both corporate income and value-added tax regimes. Some inter-country variation also exists in the rates applied to capital gains. The combination of tax rates, tax holidays, and attitudes and policies toward converging international corporate income tax rates is apparent.

33


NOTES

1. Tanzi presents 1986 tax rates for: Australia, Canada, Denmark, Finland, France, Germany, Ireland, Italy, Japan, the Netherlands, Sweden, New Zealand, United Kingdom, and the United States.

34

SUMMER 1999

REFERENCES

Adler, M. (1979). “U.S. Taxation of U.S. Multinational Corporations: A Manual of Computation Techniques and Managerial Decision Rules.” In M. Salant and G. Szego (Eds). International Finance and Trade. Pp. 157-210. Cambridge, MA: Ballinger.

Grubert, H., W.C. Randolph, & D. J. Rousslang. (1996). “Country and Multinational Company Responses to the Tax Reform Act of 1986.” National Tax Journal 49 (September), 341-358.

Kopits, G.F. (1972). “Dividend Remittance Behavior Within the International Firm: A Cross-country Analysis.” The Review of Economics and Statistics 54 (August), 339-342.

Lintner, J. (1956). “Distribution of Incomes of Corporations Among Dividends, Retained Earnings and Taxes.” American Economic Review 46 (May), 97-113.

Moffett, M.H. & D.L. Davison. (1996). “Repatriating Foreign Earnings: Empirical Evidence for U.S.-Based Multinational Firms.” Paper presented at the Third Global Finance Conference, Honolulu, HI (April 1996).

Mutti, J. (1981). “Tax Incentives and the Repatriation Decision of U.S. Multinational Corporations.” National Tax Journal 34 (June), 241-248.

Robbins, S.M. & R.B. Stobaugh. (1973). Money in the Multinational Enterprise. New York: Basic Books.

Scholes, M.S. & M.A. Wolfson. (1992). Taxes and Business Strategy. Englewood Cliffs, N.J.: Prentice-Hall, Inc.

Shapiro, A.C. (1996). Multinational Financial Management. Englewood Cliffs, N.J.: Prentice-Hall, Inc.

Slemrod, J. (1990). “The Economic Impact of the Tax Reform Act of 1986.” In J. Slemrod (Ed). Do Taxes Matter? The Impact of the Tax Reform Act of 1986. Pp. 1-12. Cambridge, MA: MIT Press.

Tanzi, V. (1987). “The Response of Other Industrial Countries to the U.S. Tax Reform Act.” National Tax Journal 40 (September), 339-355.

Whalley, J. (1990). “Foreign Responses to U.S. Tax Reform.” In J. Slemrod (Ed). Do Taxes Matter? The Impact of the Tax Reform Act of 1986. Pp. 286-314. Cambridge, MA: MIT Press.

Zenoff, D.B. (1968). “Remitting Funds from Foreign Affiliates.” Financial Executive 10(1), 46-63.

35


36

SUMMER 1999

CRESCENDO MODEL OF CAREER MOTIVATION

AND COMMITMENT

Albert S. King*

Kimi King**

This study defines and measures a progression model of career motivation in a large aerospace engineering and manufacturing firm. The utility of the progressive phases is demonstrated by their capability for discriminating with 23 variables the propensity to career motivation and commitment and to distinguish regular variation on a continuum of 16 variables widely used to evaluate the quality of work organization. The analysis supports the notion of a graduated stepwise movement from identity to insight, to resilience in strengthening motivational attraction to career commitment, and demon-strates concurrent validity of progressive phases associated with perceived quality of work organization. Although the levels and phases are progressively prepotent and valenced in predicting employee attraction to career commitment, different patterns and paths through the phases for individuals are indicated. At a structural level, results shed insight into motivation patterns formation and affective career commitment.

otivation and commitment to career has attracted much rekindled attention for intriguing, compelling reasons (Schein, 1991; Tracy, 1993). Psychological forces

of self-identity, self-insight, and resilience in pursuing career goals represent core components for career motivation and commitment; and for building cooperation, cohesiveness, and consensus in organization. Personnel pull together for the common purpose in a strong career oriented organization where the unique attractiveness of its rich motivation bonds members tightly and commits them individually and organizationally to identify with its rules, rewards, and values (Katz and Kahn, 1978).

M

*Albert S. King is affiliated with Northern Illinois University, DeKalb, Illinois.**Kimi King is affiliated with University of North Texas, Denton Texas.

37


This paper reports findings of a study designed to test a hypothesized linkage between generalized motivational domains of identity, insight, and resilience with affective career commitment. Based on the notion that an individual'’s behavior is determined primarily by past experiences and extending the analogy of individual development to that of organizational development, this paper discusses a series of developmental stages through which employees tend to pass in becoming “career” committed. First, an overview and perspective on career motivation and the advantages of commitment at the level of the organization is provided.

Motivation and Commitment: Benefits of Cooperation

The important source of commitment appears to be from within the intrinsically shared identities, insights, and values of the organization itself—where members are encouraged to turn inward and take initiatives from the imperatives of the organization’s own vision. Instead of gazing outward to copy what other organizations are doing, members of the strong career—oriented organization receive their impetus for action, not from admiring comparable organizations, but from emulating their own discoveries in fulfilling particular internal needs (Mintzberg, 1989).

As a resource, career motivation and commitment forms a centripetal force inward, protecting the organization from outside influence, drawing human resources toward countless acts of cooperation with each other. In this sense-by socially indoctrinating individuals into its norms and values--commitment may best be regarded as the central ethos of an organization, indeed the life-giving force or spirit that informs the formal framework of its fundamental function (Miller, 1978).

An important application appears to be that internalization of career values renders any particular organization more effective (Meyer, 1989). The concept continues to be of enduring interest to managers and organizations because of its connection to increased performance and motivation, lowered absence and turnover, along with heightened stability, satisfaction, and involvement (Porter, et al., 1974, 1976; Weiner and Vardi, 1990). Personnel get “juiced-up” to pursue the focus, structure, style, controls, or rewards or whatever else drives and determines the direction of the organization. The infusion of an organizational career ideology can alter even the most bureaucratic structure’s nature. It is the nature of the human commitment to customer responsiveness and sensitivity to employee and stakeholder needs that really counts (Peters and Waterman, 1982; Tracy, 1989).

38

SUMMER 1999

Another important implementation arises from using career ideology to resolve contradictory competing claims between people and departmental units, and to reconcile discrepant conflicts arising within individuals themselves (Reichers, 1986). At the organizational level, institutionalized career commitment helps forces and functions that diametrically dominate or oppose each other to pull together, work through differences, and facilitate adaptation and change. Strong career-conscious organizations solidify when threatened and when they have to because they are deeply seated in strong systems of beliefs (Kahn et al., 1964; Meyer, 1987). Such organizations readily reconcile conflicting interests and suspend debilitating rivalries since what matters is the organization itself, not any of its “special” parts. When career committed people believe in the organization over and above any of its specialized parts, the organization is powerfully empowered to adapt.

This dynamic is not construed to imply that the zero-sum rule is supported and the “more-less” hypothesis holds—that if an organization favors one particular function, others may fail; or if the organization is favored above all else, then individuals suffer. This may happen in a weak career ideology organization where functions and outcomes are managed merely as aggregations of different parts; and people are treated as means to an end, rather than an end in itself. But when the strong force of career ideology and commitment genuinely infuses an organization structure in a bone-deep belief in doing the right things for people authentically prevails, an organization takes on an institutional life and logical dynamic of its own and conflicts are reconciled.

That is what the concept of strong motivation leading to organizational and career commitment really conveys. Regardless of what function an individual or unit performs, each is treated as an embodiment of the total system and each is empowered to make decisions and take actions for the good of the whole. It is not just the salesforce that is responsible for revenue, nor the production members for controlling costs and efficiency. A “hands on, value driven” or all hands approach compels everyone to internalize many forces in carrying out his or her own duties. Reiterating a metaphorical epigram— “It is easy to change hats when all are emblazoned with the same insignia” (Mintzberg, 1989). Thus far, organization aspects of commitment have been subordinated to analysis of individual propensities (Hrebiniak and Alluto, 1972; Brown, 1969; Dubin, Champois and Porter, 1975; Kidron, 1978), but that emphasis is being shifted. Consider three studies of an emerging agenda. Thus, Buchanan (1974) and Jamal (1974) conceptualized level of integration in organization development in terms of positive association with

39


organizational career commitment. Moreover, O’Reilly and Chapman, and Allen and Meyer (1990) used an organization setting to illustrate how congruency of individual career and organization values leads to mutually rewarding relationships reducing the possibility of conflict. Finally, London (1983, 1990) specified an interactive model of career motivation components comprised of individual and situational characteristics coupled with decisions and behaviors that has important implications for designing new motivational strategies and exploring career commitment.

TOWARD A CRESCENDO MODEL OF CAREER MOTIVATION AND COMMITMENT

The model of career formation extends historical work (March and Simon, 1958; McGregor, 1967; Hall, 1970; Porter et al., 1974, 1976; Morris and Koch, 1979). Empiri-cal studies (Weiner, 1982; O’Reilly and Chatman, 1986; Meyer and Allen, 1987; Allen and Meyer, 1990; Mowday, Steers and Porter, 1979; London, 1990) separately tested for relationships of each of the three motivation domains with a wide range of individual and organizational outcomes. This study makes an additional use of the essential factors. It attributes differential motivation prepotency and the propensity to career commitment, uses norms from the employee population to distinguish high versus low propensity on each motivation domain, and then defines stages of organizational career motivation and commitment formation in terms of possible high versus. low combinations of scores on the three domains.

Stated simply, the model seeks to show an underlying pattern between identity, insight, and resilience in establishing the stronger construct of career ideology and commitment. In terms of prepotency, identity precedes insight which precedes resilience in the formation of career commitment. Stronger combinations of these three dimensions, in turn co-respond to successively higher states (or stages) of attraction to career ideology.

40

SUMMER 1999

PSYCHOSOCIAL STAGES OF CAREER MOTIVATION LEADING TO COMMITMENT

The current study adds to structural aspects on the psychology of commitment, setting forth the usefulness of a crescendo or stage model. In a preliminary study, these phases varied systematically with commonly used measures of the quality of work organization (Bateman, 1984; Allen, 1990). Results indicate that career ideology formations can be measured in large populations--easily and reliably. This compact analysis augments the anecdotal and judgmental focus of much of the available literature, and permits rigorous tests of it (Mowday, Porter and Steers, 1982; London, 1990).

The present study included over 1,800 respondents and involved employees of a large aerospace engineering firm focusing on a cross section of lower organizational levels and narrow range of jobs (logistics, production, maintenance, and service) at dispersed locations nationwide. In several ways, the study was designed to test the stability of the pattern of correspondence of career ideological forces with organizational variables. Attention was directed toward: (1) identifying characteristics of the workforce, (2) analyzing three motivation dimensions regarded as sequential linkages to career formation, (3) defining successive stages of career commitment, and (4) testing staged combinations of the dimensions with factors often used to evaluate the quality of work organization.

BACKGROUND OF STUDY

Employee respondents for the study came from several functional areas and include four job classifications in a large (20,000 employees) Fortune 500 aerospace firm with ten locations nationwide. The corporation provides eighty percent of its products and services to government, commercial, and industrial customers under dynamic market conditions that can be competitively challenging and rewarding. The four job classes included:

Production workers (equally skilled, semiskilled, and unskilled) constituting 50 percent of the total responding population.A cadre of logistics employees from one of the established position specializations (raw materials handling, supply and storage, transportation and shipment, distribution scheduling, and tools maintenance), constituting 25 percent.

41


Maintenance personnel (generally at equipment repairs, tools, supplies, and materials tasks) constituting 13 percent of the population.

Service employees in contact with suppliers, customers, contract agencies, production, maintenance, and operative personnel comprising 12 percent of the respondents.

With corporate human resources and plant management support, the organizational workforce of the design and assembly division of the firm provided voluntary responses to separate parallel survey instruments covering varying aspects of employee worklife. These aspects, identified as “covariants of motivation and commitment” included a set of sixteen variables to evaluate worksite quality. While these were intended to focus on the organization as a whole, analysis from the 23-item questionnaire at the end of this article was used to demonstrate that the sample was representative of a crescendo phase model.

The total population was 4,200, located at 20 sites nationwide, with plant manufactur-ing and assembly units having similar (not identical) production, scheduling, and cost containment roles. Although total N varied somewhat for different purposes due to item-wise adjustments made for missing data, N was never less than 1,800 for any analysis. The response rate averaged 50 percent, and comparisons of conventional demographics—age, sex, race, and so on—revealed that the responding workforce sample made up a reasonable analog of the total population. For example, women were only slightly (and nonsignificantly) under-represented in the responding sample containing 30.5% women and 69.5% men corresponding to total organizational percentages of 31.4 and 68.6, respectively. In all other demographics, the respondent to total population comparisons reflected similar closeness of fit.

42

SUMMER 1999

MOTIVATION DIMENSIONS OF CAREER COMMITMENT

To assess employees’ attachment to career goals and values and the formation of commitment within the overall organization context, the instrument included 23 items, which tap three motivation domains represented as subscales:

1. Identity, high scores on which distinguish individuals who tend to identify and define themselves in terms of their work. Individuals high in career identity are involved in their jobs and careers, seek upperward mobility, and do not wish to delay gratification in their development. Career identity reflects the direction of career goals--whether the individual wants to obtain a position of leadership, advance in the company, seek higher status, or, perhaps make more pay. This generalized value resulting from an internalized normative pressure is how central one’s career is to one’s identity. Individuals who are high on career identity are likely to find career satisfaction to be more important than satisfaction from other areas of life (primacy of work) (London and Mone, 1987).

2. Insight, high scores on which indicate respondents who are realistic about themselves and their careers as well as how well they relate these perceptions to their career goals. They see their own needs and capabilities and the organization career requirements as sufficiently related or similar that the possibility exists for a mutually rewarding employment relationship and providing path-goal clarity (Weiner, 1982; Allen and Meyer, 1990). In this case, because of path-goal clarity, self-objectivity, and goal flexibility, the individual finds it natural to deepen identity and sharpen understanding of and commitment to career. They look for feedback about how well they are doing, use this information to set specific career goals, and formulate plans to achieve their goals. Their sensory perception or career insight affects the degree to which they pursue their career goals.

3.Resilience, high scores on which are attitudinally related to the highest form of affective career commitment and come from individuals who are strongly attached to their career goals and values and to the organization and their career for its own sake, apart from purely instrumental work (Buchanan, 1974; Weiner, 1982). This is represented by an inherent willingness of individuals who have insightfully identified with their careers to make personal sacrifice, perform beyond normal expectations, work selflessly, and continue career contribution,

43


to endure difficult times, and not desire to leave the organization or their careers for self interest or personal noncareer gain. Resilience is the magnitude or extent to which the individual resists career barriers or disruptions affecting their work. People high in resilience have high self-efficacy, seeing themselves as competent people taking risks and responsibilities for their careers with low need for dependency and able to control what happens to them. Resilience, as a motivational domain, influences a person’s persistence in pursuing career goals (London and Bassman, 1989).

To grasp the concept of resilience within the context of career commitment more clearly, an understanding of its opposite or, career vulnerability is useful. This is the extent of psychological fragility (e.g., becoming disaffected and finding it difficult to function) when confronted with adversity (barriers to goals, uncertainties, poor relations with supervisors and co-workers). Being high on career resilience and commitment does not mean that the person is insensitive to adverse conditions, but rather that he or she will be able to cope more effectively with a negative work situation. Individuals will be more resilient the higher they are on self-efficacy, risk-taking, and competitiveness dimensions and the lower they are on the dependency dimensions. Those low on career resilience and commitment (high on career vulnerability) are likely to be motivated to avoid risks, be dependent on others, seek structure, and avoid situations in which outcomes depend on their behavior. Those high on career resilience and commitment are likely to do the reverse--take risks, be independent of others, create their own structure, and seek responsibilities for outcomes contingent on their behavior.

Because of the focal firm’s unique interests, data gathering was based on an exploratory study that generates its own items and the survey instrument was adapted from Weiner (1982) for identity, Allen and Meyer (1990) for career insight, and from Buchanan (1974) and London (1990) regarding career resilience.

44

SUMMER 1999

Survey respondents received simple instructions: “Write one number in the blank to the left of each statement indicating the extent you agree or disagree with each statement.”

Very Much Disagree 1 2 3 4 5 6 7 Very Much Agree

Low numbers describe statements with which you disagree.

High numbers describe statements with which you agree.

Factor analysis of the 23 items is shown in Table I, along with item numbers on the survey to reflect how subscale items are interspersed on the instrument. Table II shows item, subscale, and total alpha reliabilities. Information in these two tables suggest the usefulness of three factors--identity, insight, and career resilience--with acceptable reliabilities accounting for total intercorrelation in the data. The variable correlations of items with subscale scores (right column of Table II) suggest a major collective contribution of the three batches of items to the subscales.

Factor I (resilience) is loaded by four items from other subscales, but the other two factors (insight and identity) get loaded > .30 only by items classified apriori within their respective subscale classification. The communalities, H2, in Table I reflect the amount of each item’s variance included in the factor analysis.

As the factor analysis suggests, the three subscales have moderate correlations, thus:

Identity Insight Resilience

Identity 1.00 .31 .52

Insight 1.00 .29

Resilience 1.00

This pattern of correlations suggest that the three subscales make relatively independent contributions to defining the content of the psychology of attraction to affective career commitment.

45


Table IFactor Structure For Career Motivation

(N = 1,839; Items = 23)

Subscale Items Rotated Factor* I II III IV H2

Resilience

1. Not Difficult to Function .75 .60 .63

2. Persistence in Career .82 .51 .61

3. Positive Self-Image .84 .58 .61

9. Responsibility for Work Performance .61 .48 .71

14. Outcomes Depend on Behavior .59 .55 .46

15. Responsibility for Own Career Plan .47 .36 .25

21. No Need for Approval .66 .40 .46

Insight 5. Projects Affect Career .41 .47 .22

8. Unwilling to Modify Career Goals .52 .43 .27

10. Realistic Perceptions .43 .41 .23

13. Using Information .51 .38 .44

18. Awareness of Alternatives .41 .48 .31 .40

19. Seek Information .53 .41 .36

20. Accurate View of Strengths .51 .38 .45

22. Decisive in Career Decision Making .51 .37 .31

Identity

4. Career Satisfaction .47 .37 .24

6. Identifying and Defining .55 .40 .39

7. Involved in Work .36 .44 .31 .30

11. Seeking to be Acknowledged .45 .55 .35 .43

12. Need for Advancement .53 .37 .32 .36

16. Identify With Specialization .35 .31 .46

17. Unwilling to Wait .36 .32 .20

23. Sacrifice Activities and Responsibilities .34 .30 .23

Eigen Value 6.1 3.4 2.4 7.7

% Common Variance 26.5 14.8 10.4 33.5

% Cumulative Variance 26.5 41.3 51.7 33.5*Show only leading > .30

46

SUMMER 1999

Table IIReliabilities of Subscales and Total Scores

Career Motivation andCommitment Formation Subscale Items

SubscaleAlpha

Coefficient

ItemAlpha

Coefficient

Item/SubscaleCorrelationCorrected

A. Career Resilience .84 1. Not Difficult to Function 86 .70 2. Persistence in Career .83 .67 3. Positive Self-Image .83 .60 9. Responsibility for Work Performance .80 .77 14. Outcomes Depend on Behavior .80 .65 15. Responsibility for Own Career Plan .85 .61 21. No Need for Approval .82 .68

B. Career Insight .74 5. Projects Affect Career .73 .32 8. Unwilling to Modify Career Goals .68 .47 10. Realistic Perceptions .71 .35 13. Using Information .68 .48 18. Awareness of Alternatives .67 .41 19. Seeking Information .71 .43 20. Accurate View of Strengths .70 .44 22. Decisiveness in Career Decision Making .69 .37

C. Career Identity .71 4. Career Satisfaction .70 .33 6. Identifying and Defining .68 .46 7. Involved in Work .66 .44 11. Seeking to be Acknowledged .65 .48 12. Need for Advancement .72 .54 16. Identify with Specialization .70 .49 17. Unwilling to Wait .67 .29 23. Sacrifice Activities and Responsibilities .68 .28

D. TOTAL Career Ideology Score .82

47


Moreover, the model asserts a theoretical imperative that higher states of attraction to career motivation—which some observers, e.g., Mintzberg (1989), label “missionary” —derive from and transcend the dimensions of identity, insight, and resistence. A critical hypothesis is that the three dimensions or domains of motivation are indeed dimensions of a greater construct. To test this hypothesis, scores for respondents on all three dimensions were totaled and correlated with overall items. Results of factor analyzing all 23 items (forced) on only one scale “career commitment” on overall attraction to career ideology revealed an assuring pattern:

Career Commitment: Loyalty Value Congruence

Affective Commitment

Overall Ideology Attraction .34 .38 .43

To support the notion that this pattern is not measuring three different constructs, column four of Table I shows that all of the 23 items load fairly well on only one factor— “career commitment”--and so are really part of the same construct.

Differential Prepotency and Valence

The contributors to phases of motivation to career commitment can be described as ranging in prepotency from identity to insight to career resilience; and increasing in valence (need strength) from low to high involvement which correspond to an increasing career commitment. Increases in identity represent the most commonly prepotent and least valenced initiators of the process. Indeed, some degree of identity--as in “willing to put forth effort beyond what is normally expected” or “sacrificing activities and responsibilities . . . by working overtime” seems useful for effective performance in many occupations. Beyond a point, however, indigenous identity, as a generalized sense of motivational definition and obligation, augments the possibility of a mutually rewarding understanding of self-objectivity and path-goal clarity (insight) and feelings of intense inspiration and persistence (resilience). Of substantial interest is the proposition that early identity coupled with social perceptiveness of sufficiently related insights becomes magnified by processes of socialization, and so facilitate even stronger identification, which in combination with a future time orientation, lead to resilience and career commitment.

48

SUMMER 1999

Perhaps the classic case involves the skilled craftsperson assigned to a challenging position in a job class appropriate to previous trades training and requiring high participation. In episodic, regularly recurring contacts with similarly trained co-workers and supervisors, certain attitudes and beliefs appear increasingly common--an awareness of values that are embracing and enduring, involvement in a wide variety of tasks, distinct identity of members, significant purposes, familiarity with work outcomes, and so on (identity). With continued traditions and precedents, an increasing concern and satisfaction about shared goals and the ability to be a part of something highly important occurs. With repeated enactments of reinforcing meaningfulness, responsibility and knowledge of job results, social perceptiveness of path-goal clarity, self-objectivity, and realism of expectations build a momenta for future time orientation and value consensus (insight). As identity and insight mount, beyond a point, so can initiatives (self-esteem, adaptability, internal control) result in a cathexis surpassing an individual’s normal inner work standards and development orientation limits and generate the inducement surplus implied by the bone deep belief of persistence (resilience) in pursuing career goals.

It is not intended to indicate there is one pathway to career commitment, but rather to suggest why differences in valence (increased need strength toward commitment) occur. While ordering the phases in terms of increased valence, the proposed model does not require that all individuals go through exactly the same phases. The phases represent progressively valenced effects, with a variety of specific pathways (for individuals), leading to career ideology orientation and commitment (of organizations).

Percentile rankings derived from the three sets of subscale scores from the employee population of over 1,800 provided an empirically based estimate of high vs. low. The distribution of raw scores by percentiles yielded three cutting-points for high scores, or those greater than the median: 24 for identity, 32 for insight, and 28 for resilience. The use of the median percentile rankings was not taken simply to establish convenient cut-off points. Identification and development of these cut-offs were adopted from the examination for construct validity taken from Vaserhelyi (1977). Using these cut-offs, an 8-phase model of progressive career organization ideology was generated. High scores on career resilience are considered most highly valenced; high scores on career insight are more valenced than high scores on career identity. Basically, these cut-offs provide a ready interpretation for why employees can easily become indoctrinated and often become dedicated to a degree approaching complete commitment. Low to moderate identity will lead only to indifferent compliance, without strong need for recognition or advancement. Heightened identity, however, will result in increased sense of work involvement, identification with organization goals and path-goal congruence leading in turn to higher degrees of commitment implying surpluses beyond those associated with mere dependable role compliance. The focus on phases derives from the high vs. low

49


distinctions on three propensities to career commitment subscales. That is, eight combinations logically emerge from three subscales, each distinguished as two (high versus low) categories. Table III defines the phases in terms of this analysis and shows the distribution of assignments in the current study employee population.

COVARIANTS OF ORGANIZATION IDEOLOGY AND COMMITMENT PHASES

Judging from the literature on ideological organization, its covariants include a broad and enriched range of characteristics (Mintzberg, 1989). To illustrate, Schein (1991) associated formation of ideology and “strong culture” with a substantial catalog of effects which favorably compliment improvements in the quality of work life or commitment by those experiencing strong attraction to organization ideology: Higher productivity and performance, increased motivation and morale, reduced absenteeism and turnover, along with various self-reported indices of cooperation, including security, initiative, unambivalence, involvement, proaction, and affection. This central characterization suggests a convenient test of the progressive valence of the phases. Consequently, the quality of work life should co-vary regularly with the phases of valence (career motivation), if those phases represent conditions of increasingly ideological career commitment. To assess such concurrent validity, a three-stage effort was made to: (1) identify a set of variables commonly used to evaluate worksite quality, (2) extrapolate predictions concerning the expected relationships of these variables with phases of commitment, and (3) test these likely covariants to see if the phases of commitment “map” on them in expected ways.

Covariants were comprised of assorted measures, of which six should increase as ideological organization and career commitment progresses through the several phases:

Trust in supervision (Roberts and O'’Reilly, 1974).

Trust in employment practices (constructed for present study).

Job involvement (White and Ruh, 1973).

Participation in decisions regarding work (White and Ruh, 1973).50

SUMMER 1999

Willingness to disagree with supervisor (Patchen, 1965).Job tension, an additional scale which should decrease with advancing phases (Kahn, et al., 1965).

51

Table IIIHigh versus Low Career Motivation to Define Phases of Commitment

Ideology Subscales Phases of Commitment

Prepotency I II III IV V VI VII VIII

Identity Lo Hi Lo Hi Lo Hi Lo HiInsight Lo Lo Hi Hi Lo Lo Hi HiResilience Lo Lo Lo Lo Hi Hi Hi Hi

Low Valence HighAssignments(N = 1,839)

___390

___145

___231

___162

___145

___214

___147

___405

Interpretation

Prepotency ranges from Identity to Insight to Resilience in the formation of career commitment.

Valence ascends from Phase I (Low) to Phase VIII (High) for stages provisionally defined as:

CAREER OVERLAPPING MOTIVATION RANGES

Phase Meaning Identity Insight Resilience

I Awareness

II Concern

III Experimentation

IV Options/Alternatives

V Partial Acceptance

VI Momenta

VII Convergence/Congruence

VIII Affective Reinforcement

Job Diagnostic Survey (JDS), which measures satisfaction with ten facets of work (Hackman and Oldham, 1980) all of which should increase as ideological organization and career commitment increases. Facets of satisfaction include:

1. Meaningfulness of work 6. Growth satisfaction2. Responsibility for work 7. Job security3. Knowledge of results 8. Compensation4. General satisfaction 9. Co-workers5. Work motivation 10. Supervision

Table IV shows analysis of these results with an average alpha approximating .80 with reliability coefficients for only two of the 16 variables falling below .70 (responsibility for work and work motivation). The conceptual frameworks underlying the assorted scales are general and substantiated by considerably convincing justification (Hackman and Oldham, 1980). Consider, for example, job tension, which taps several important classes of alienation or dissatisfaction—e.g., those related to role conflict and ambiguity—that impact on career commitment phases, more or less directly.

Data from the workforce population permitted testing for the covariants of progressive phases of commitment and the results appearing in Tables IV-VI support the general usefulness of the crescendo phase model. On balance, about 18 percent of the variance in commitment phases x worksite descriptors is explained. For comparative purposes, simple correlations were run between worksite descriptors and four phase model scores—the three subscale scores and phase model total. On average, these 64 correlations explain 13 percent of the variance and this result reinforces the usefulness of the phase approach, which at once derives from the three subscales and yet transcends them.

Phases I and VIII have the lowest and highest scores on the subscales, but the 6 interior phases have total scores that do not vary directly with the phases. Tables IV and V showing paired comparisons of variables x phases support the overall summary that phase by phase, analyses of all possible comparisons indicate that: (1) over 90 percent of the differences (404/448) are in the expected directions, (2) over 55 percent of the expected differences (248/448) attain statistical significance, and (3) only 2 differences of 448 are in an unexpected direction and statistically significant. The paired comparisons utilized the least difference test, modified for unequal subpopulations. A conclusion flows easily--phases of commitment reflect regular and robust co-variation with the panel of descriptive reports about the image and quality of worksites.

Additionally, the data suggest most or all the phases discretely map significant differ-ences on target variables. Of course, the significant paired comparisons suggest this conclusion, but focusing on “distance” between phases highlights the point. Thus, Phases I vs. II, II vs. III, and so on, may be considered a distance of +1; Phases I and III, III and V, and so on, are +2; etc. Large proportions of expected and statistically significant differences shown in Table VI are evenly distributed among seven possible “distances” which support the utility of the full 8-phase model.

These data demonstrate that adjacent as well as distant phases tend to map discrete segments of the ranges of target variables. This holds most clearly for the five most distant pairs of phases, where over three-fourths of the paired differences are in the expected direction and attain statistical significance. Moreover, even distances of +2 and +1 generate 41 and 20 percent records in this regard. This suggests that even very close neighbors reflect substantial discriminatory power of the phases, given that a record of one in five statistically significant pairs conventionally signals noteworthy covariation.

In sum, data in Table VI significantly demonstrate the usefulness of all the phases. Phases I and VIII reflect the lowest and highest total scores and hence, +7 results can be interpreted as a total score effect. But phases of more proximate distance can have significantly different total scores and map on target variables in quite regular and robust ways.

Table IVCovariation of Phases of Commitment and Target Variables Via One-Way Analysis of Variance

Progressive Phases of Commitment

WORKSITE DESCRIPTORS ILoLoLo

IIHiLoLo

IIILoHiLo

IVHiHiLo

VLoLoHi

VIHiLoHi

VIILoHiHi

VIIIHiHiHi

F-Ratio

F-Probability

ASSORTED SCALES Scores (390) (145) (231) (162) (145) (214) (147) (405) Participation .78 13.0 13.5 14.9 16.2 14.9 15.2 16.3 17.8 46.793 < .001 Job Involvement .85 24.7 25.4 29.8 30.9 29.7 31.8 32.9 34.7 72.227 < .001 Trust in Supervision .78 12.2 12.2 12.7 14.6 13.1 15.1 14.4 15.9 23.196 < .001 Trust in Employees .78 14.7 15.0 16.1 17.4 17.0 17.6 17.8 19.6 32.726 < .001 Willingness to Disagree with Supervision .79 14.9 15.0 15.3 16.4 14.6 13.7 15.6 14.6 2.996 < .01 Job Tension .86 23.4 22.9 22.0 21.0 19.5 18.2 17.0 16.1 49.240 < .001JOB DIAGNOSTIC SURVEY(JDS) SCALES Meaningfulness of Work .81 16.4 17.7 19.5 20.1 19.4 20.9 21.4 23.2 64.940 < .001 Responsibility for Results .67 29.0 30.3 31.9 33.6 32.1 33.4 34.6 36.4 51.462 < .001 Knowledge of Results .71 18.9 20.3 19.8 21.2 20.4 21.9 21.5 23.0 32.840 < .001 General Satisfaction .81 19.2 19.9 22.2 23.1 24.2 26.0 26.5 28.4 97.183 < .001 Internal Work Motivation .67 29.9 31.1 32.9 34.4 31.9 33.2 33.7 35.7 37.987 < .001 Growth Satisfaction .85 15.3 16.4 18.6 20.0 19.3 20.7 21.5 23.1 73.738 < .001 Satisfaction with Security .77 8.6 9.1 9.3 9.7 10.3 10.7 10.9 11.5 21.085 < .001 Satisfaction with Compensation

.87 8.2 8.6 8.4 8.7 8.7 10.2 10.1 10.5 18.987 < .001

Satisfaction with Co-Workers .76 13.9 15.5 16.0 17.6 15.9 17.9 18.0 18.8 76.314 < .001 Satisfaction with Supervision .80 12.2 12.5 13.5 14.7 14.1 15.9 15.7 17.1 38.385 < .001


56

Table VI

Four Measures of Covariation of Target Variables, by “Distance” Between Pairs of Phases, Percent, and Absolute Number of Observations (In Parentheses)

Total in Percent +7 +6 +5 +4 +3 +2 +1

Expected Direction 90.2 94%(15/16)

97%(31/32)

96%(46/48)

98%(63/64)

95%(76/80)

90%(86/96)

78%(87/112)

Expected Direction and Statistical Significance

55.1 94%(15/16)

94%(30/32)

92%(44/48)

78%(50/64)

58%(46/80)

41%(39/96)

20%(23/112)

Unexpected Direction 9.4 6% (1/16)

3% (1/32)

2% (1/48)

3% (2/64)

5% (4/80)

9% (9/96)

21%(23/112)

Unexpected Direction and Statistical Significance

0.5 0% 0% 2% (1/48)

0% 0% 0% 1% (1/112)

57

JOURNAL OF BUSINESS AND MANAGEMENTTable VII

Factors in Panel of 16 Target Variables, by Varimax Rotation Factors Suggesting Differences in Worksite Descriptors

I. II. III.High Energy, Positive Job Factors,

Peer-OrientedHigh Trust, upervisor-

Oriented, ension-AvoidingLow Energy, eer-Oriented,

Restricted Upward Feedback H2Assorted ScalesParticipation .49 .48 .52

Job Involvement .71 .63Trust in Supervision .79 .71Trust in Employees .40 .67 .62Willingness to Disagree with Supervisor -.36 .22Job Tension -.55 .43JDS ScalesMeaningfulness of Work .83 .72Responsibility for Results .73 .60Knowledge of Results .42 .33General Satisfaction .75 .37 .72Internal Work Motivation .66 .48Growth Satisfaction .44 .76Satisfaction with Security .73 .43 .27Satisfaction with Compensation .45 .39Satisfaction with Co-Workers .67 .35 .56Satisfaction with Supervision .82 .77Eigen Value (unrotated) 7.5 1.5 1.1Percent Common Variance 46.7 9.4

6.9

Percent Cumulative Variance 46.7 56.1 63.0

*Shows only loadings > .35

58

SUMMER 1999

Finally, the 16 target variables do not merely measure the same domain multiple ways. Varimax rotation factor analysis in Table VII reveals three domains which are provisionally labeled. One factor accounts for over 45 percent of the common variance to which the other two factors add some 16 percent. Cross loadings in the factors exist, but the convention of reporting only loadings > .35 highlights some differentiating tendencies. The first two factors suggest high energy and the second factor seems distinguished by its focus on supervisory versus peer level (peer orientation more common for factors I and II), as well as by lesser emphasis on job contributors to satisfaction. Moreover, the third factor seems characterized by low energy and peer defensiveness. This multidimensionality of the target variables reinforces the pattern of covariation discussed above, although not robustly.

CONCLUSION: PROGRAMMATIC RESEARCH DIRECTIONS

These results support four future initiatives. First, the data reveal a regular and robust covariation between the commitment phases and 16 common indicators of the quality of working sites. This clearly supports the usefulness of the phase approach, but the search for covariants should be extended to nonreactive and unobtrusive measures of individual and organizational behaviors.

Second, the incidence of career commitment phases among an organizational workforce requires attention for both human resource and cultural socialization reasons, the former if only because screening and selection surveys for workers’ training and aptitude potential due to values orientation (before undertaking expensive staffing programs) is de rigueur. Consider data from two organizations: The present focal study firm, (A), which is generally considered a moderately favorable place to work and a voluntary nurses association health services concern, (B), that is in most respects throughly modern, considered highly missionary, and human resources-oriented. Even the “missionary” organization faces a substantial challenge, to judge from the following distribution of employees by phases of career commitment:

PHASES OF ASSIGNMENTS BY PERCENTAGES

Organization I II III IV V VI VII VIII

A 21.2 7.9 12.6 8.8 7.9 11.6 8.0 22.0

B 10.9 10.6 7.7 4.7 13.9 9.1 7.3 35.8

59


Third, more information is needed about the locus of motivation and its connection to career formation and organization. The evidence, far from conclusive, suggests that most of the variance in commitment is accounted for by properties of precedent, tradition, and reinforcements of the immediate work unit and not by the overall organization or original career itself (Ouchi, 1980; Meyer, 1989).

Fourth, there may well be several types of commitment, even if features of the immediate work group and its style of control prove dominant. Progressively evoked seems to describe the dominant type of motivation to career ideology measured by the phases, given not only the strong associations with worksite descriptors, but also taking into account the substantial persistence of phase assignments observed in the study over a year’s interval. But, precipitous, or natural onset also no doubt exists and advanced and heightened bonding with career ideology might be induced by sudden critical identity enhancing events.

This core notion might develop in several ways. Although available data imply that Phase I-VIII can be considered progressively valenced, for example, this does not imply that a phase-by-phase entrance to, or exit from, advanced commitment always occurs. Powerful internalization might induce an acute “natural” progression of phases: IVIVIII, for example. One also can envision a basic pathway for “evoked” commitment due to gradually accumulating and indoctrinating involvement at work: IIIIVVIIVIII.

Finally, it is pure folly to assume that career forming motivations are based simply on expected economic gains; much deeper values are at stake:

When the informal processes of socialization tend to function naturally; perhaps rein-forced by more formal programs of indoctrination, then the ideology would seem to be strong. But when the organization is forced to rely almost exclusively on forms of calculated identification, then its ideology would appear to be weakening, if not absent to begin with (Mintzberg, 1989).

This conclusion that deeper values than money are at stake best illustrates how combinations of affecting factors of identity and shared insights among employee members echoes the same sentiment: The perceived instrumentality of participation, involvement, and commitment to career ideology truly promises that workers will understand an organization’s mission, philosophy, and policy and will be able to deduce or derive for themselves the proper objective for any conceivable situation.

For Human Resources and career development strategists, employees’ belief that intervention (career planning workshops, etc.) will yield positive rather than negative outcomes for them allows planners to predict, with incredibly high accuracy, the

- 60

SUMMER 1999

readiness and receptivity of organizations to new initiatives in organization change. In a sense, career commitment works like the basic postulates of an axiomatic system. They are the fundamental assumptions on which reasoning and sense making are logically derived, but in themselves are not logical. The real test of their worth is not their reasoning or logic, but the usefulness of the thinking and action that ensues. In strong career oriented cultures, everyone knows the importance of shared values and compelling commitment. It is the ideological drive for accomplishment pulling the organization together; providing continuity in what would otherwise be an autochthonous field of organization dynamics.

This line of research suggests, however, that strong cultures of career commitment are not to be found in many, or even most organizations. And, while earlier studies (Ouchi, 1978) have shown they are evident in most of the superior performers, all is not rosy in the world of culture either. As one observer has exclaimed, “A corporation doesn’t have a culture. A corporation is a culture. That’s why they’re so horribly difficult to change” (Kiechel, 1984). In established organizations, career ideologies are difficult to build and sustain, and can sometimes get in the way of organization effectiveness. Whereas strong career cultures promote change within themselves and become immutable, by forcing everyone to act within the same set of circumscribed beliefs, they themselves are not to be changed. Career commitment thus becomes an obstacle to change, and the very ideology of identity, insight, resilience, and positive affect that makes careers attractive and an organization so adaptive within its culture undermines its efforts to move to a new context.

The phase model approach to assessing processes of career formation promises to provide a useful organon or model for further understanding just how commitment serving to reconcile contradictory forces and promote change can paradoxically discourage, or even destroy it. The phase-wise analysis suggests that effective career ideologies are generated gradually and incrementally by patient leaders capable of establishing compelling career missions for their organizations, manage the paradoxes inherent in them, and are deeply committed to the people’s careers who perform them.

The phase model also suggests a call for career vision in managing organization mission. Successful corporate strategies are almost invariably guided by powerful corporate visions and realistic assessments of the company’s commitment and capability to attain them. Establishing this vision imposes great demands on leadership, but meeting the demands can produce a corporate renaissance. Paradoxically, the more people understand about career limits of commitment, the more they seem to expand the limits of careers. Such has certainly occurred in some organizations adopting the notion--“There is absolutely no limit to career quality”--as their credo for development. Such an expansion may provide the most effective way for a company to sustain itself and prosper in the future.

MANAGER APPLICATIONS: INTERVENTION STRATEGIES

61


Several active, organizational, and self-help strategies should be proposed to facilitate employee motivation to career development. If management wants to make careers more attractive to employees, it must consider making work conditions more satisfying. Management and human resources can enhance the progressive level of motivation by: (1) self-analysis and evaluation of commitment level; (2) greater participation, involvement, and commitment in selecting and achieving work goals; (3) originating and reinforcing positive work attitudes; (4) maintaining a “sense-making” morality of work and organizational culture; (5) providing realistic communications that create expectations that will be fulfilled; (6) designing jobs with variety, identity, significance, autonomy, and feedback which fully utilize and develop skills, knowledge, and abilities of employees; (7) structuring the social psychology of organization culture and work roles to satisfy employee personalities, interests, and preferences; (8) facili-tating and effecting personnel actions that tell employees management’s priority to operate through and with employees and that the organization is committed to treating employees with trust, open-mindedness, confidence, and respect; and (9) developing leadership, motivation, and communication encouraging Pygmalion-like practices of supervision (Hackman, 1980; Roberts, 1974).

Employees sometimes respond to low identity and lack of career insight with both upper managers and supervisors by neurotic behavior. They decide, in effect, that as long as they are going to be alienated and estranged in their disaffected work roles, they might as well withdraw and apathetically blame others for their indifference. Unfortunately, passive withdrawal does not treat motivation toward career adequately. Quite the opposite; one successful mode of intervention requires that employees selectively interact and spend more quality time with key supervisors and fellow employees to become aware of problems and cope with feelings (Rand, 1980).

Another effective strategy for heightening career attraction among employees is developing a social support system. Social support is defined as “the degree to which a person’s basic social needs are gratified through unconditional acceptance by a group and belongingness with others” (Kaplan, 1987). An employee social support system serves several specific functions. These include having co-workers available to: (1) listen actively to problems; (2) provide rewarding appreciation for skills and abilities; (3) serve as a basis for continuing socialization and education; (4) provide emotional support as well as a knowledgeable referent; and (5) facilitate testing social organization reality (Guzzo, 1985; Gorlin, 1984).

Whether in solo, assembly, or group/project operations, it is relatively easy for employees to become career isolated. This can lead to the individual employee feeling alone with his or her work problems. It cannot be over-emphasized that employees must learn to extend help to each other and to be more career supportive of their fellow co-workers (Pascale, 1988).

One advantage should be suggested to organizations seeking to increase commitment momentum through a career support mechanism. If a number of employees are experiencing strong satisfaction and fulfillment at the same time, they may reinforce their positive attitudes and behaviors. The net effect would be to increase the intensity of attraction for those involved (Schuler, 1982). Research findings extol the virtues of

- 62

SUMMER 1999

employees sharing and reinforcing their collective recognition and achievement by exchanging unremitting “tales of triumph and prophecies of promise!” (Jackson, 1987).

Finally, it appears that the intensity of a “reengineering” campaign (and such campaigns are escalating in popularity) does not sway employee predispositions to identity or insight one way or the other. Since employees’ motivations toward their careers, in general, are quite stable and well formed, their orientations tend not to be influenced much by campaign slogans and “quick-fix” tactics (Getman, Goldberg, and Herman, 1986). Managers may simply fail to recognize the signs and symptoms of serious identity and insight problems early enough to make the appropriate form of intervention needed (Kotlowitz, 1987; Janus, 1982). The following symptoms are considered significant: (1) disaffection; (2) powerlessness; (3) meaninglessness; (4) normlessness; (5) social isolation; (6) value isolation; (7) self and/or work group estrangement; (8) disciplinary and grievance difficulties; (9) blaming others; (10) subgoal formation; (11) lack of awareness of real common problems; (12) acting contrary to data, information, and company policies and actions; (13) displaying extreme displeasure with trifling circumstances; and (14) behaving differently outside of work organization. Organizational diagnosis and action interventions should not be confused with unfair practices of interference. Although employers may not directly ask employees what they think and feel about organization career vs. other central life interests (e.g., family or union), they can ask how satisfied they are with work and other conditions; and how much “say so” they have in confronting important problems (Kochan, 1979; Patchen, 1965).

AFFECTIVE CAREER COMMITMENT: DECISION AND NOT FEELING

On the long, difficult, and complex train of box car jobs in life, affective feeling of emotion is only the caboose, the engine is the decision of commitment. Consider the wedding vow “ . . . will be your husband/wife as long as I love you” versus “will be your husband/wife until death us do part.” The former affect is feeble, fickle, fragile, and unstable. The latter cognition is stable, enduring, anchored in concrete conviction of decision. Thus, commitment is a choice decision, not a feeling.

Now imagine a parent saying, “I will be your mother/father as long as I love you.” Less ephemeral and more inspiring is, “I am your parent always and the only parent you will ever have.” In both institutions of marriage and family, feeling is subordinate to the

dominance of decision and affective commitment is the unfailing action. So it is that commitment as a value involved in affective career commitment, like that of family and marriage, is a choice and not a feeling. Career commitment is a verb!

SUMMARY CONCEPT

Patterns of attraction toward career commitment are perhaps more common than generally accepted at the organizational level. Fortunately, the effects are very discernible and real. This approach has emphasized adapting a survey inventory for the

63


employee work role to facilitate recognition of early signs of the attraction to affective career motivation profile as a special subset of the formation of strong career ideology and commitment. Utilizing this knowledge, it is hoped that individual managers in human resource programs will find it easier to develop strategies to examine and monitor this evolutionary phased and progressively patterned linkage impacting the quality of work organization.

The method presented is a tool that will permit Human Resource managers and professionals to make first order assessments of quality of work life to discover which personnel and/or jobs are most strongly bonded to career organization and which are the most likely candidates for intervention and revitalization. It is a tool that managers will find familiar in general design, for it is squarely based on individual self-analysis and intuitive processes, that ultimately are among the most critical to human resource effectiveness. It is intended that articulation and self-discovery of the subtle human ways in which employees have unconsciously become socialized into the career culture will make a mystifying situation not magic, but manageable and improvable.

SURVEY QUESTIONNAIRE

Write one number in the blank to the left of each statement indicating the extent you agree or disagree with each statement.

VERY MUCH VERY MUCH DISAGREE 1 2 3 4 5 6 7 AGREE

Low numbers describe statements High numbers describe statementswith which you disagree. with which you agree.

_____ 1. Confronted with unfavorable conditions, I do not become upset or find it difficult to function.

_____ 2. Persistence in pursuing career goals is more critical than resisting barriers or disruptions affecting my work.

_____ 3. In order to see myself in a positive self-image, I need to be a competent person able to control what happens to me.

_____ 4. Career satisfaction is more important to me than satisfaction from other areas of life.

_____ 5. In order to work toward future goals, I work harder on projects that will affect my career than on routine tasks.

_____ 6. Identifying and defining myself in terms of my work, I am willing to give up something of value for my career.

- 64

SUMMER 1999

_____ 7. Highly involved in my work, I put forth effort beyond what is normally expected in order to further my career.

_____ 8. Comparing my current career with opportunities for new and different career experiences, I am unwilling to modify or alter my present career goals.

_____ 9. Under uncertain, unstructured situations, gaining responsibility for work performance is more important to me than secure employment.

_____ 10. I am realistic about myself and the organization and can relate these percep-tions to career goals.

_____ 11. Seeking to be appreciatively acknowledged, I volunteer for important assign-ments and report results to higher managers.

_____ 12. I have a need for advancement and seek important rewards by striving for achievement.

_____ 13. Using information about how well I am doing, I formulate plans to achieve specific career goals.

_____ 14. Without fear of living up to mine or others’ expectations, I do not avoid situations in which outcomes depend on my behavior.

_____ 15. I take responsibility for my own career plan by taking on jobs and assign-ments for which rewards are based on competition.

_____ 16. I identify more strongly with my area of specialization than as an employee of the organization.

_____ 17. Having met time and experience requirements for rewards, I am unwilling to wait for promotion, pay increases, and other opportunities.

_____ 18. My awareness of career alternatives has helped to clarify my career goals and means for achieving them.

_____ 19. Being sensitive to factors affecting career progress, I seek information and contacts to alter my behavior to the situation.

_____ 20. Having an accurate view of my strengths, weaknesses, and motives helps me to have realistic expectations for career outcomes.

_____ 21. I do not try to impress supervisors or co-workers by being influenced by them and have no need for supervisor or peer approval.

_____ 22. Being thorough and decisive in career decision making, I have not regretted my decisions after they were made.

65


_____ 23. I sacrifice activities and responsibilities, including time with family, by working overtime for my career.

REFERENCES

Allen, N. J., & Meyer, J. P. (1990). “The Measurement and Antecedents of Affective, Continuance, and Normative Commitment.” Journal of Occupational Psychology, 63, (2), 1-18.

Bateman, T. S., & Strasser, S. (1984). “A Longitudinal Analysis of the Antecedents of Organizational Commitment..” Academy of Management, 27(1), 95-112.

Brown, M. E. (1979). “Identification and Some Conditions of Organizational Involvement.” Administrative Science Quarterly, 14, (4), 76-89.

Buchanan, B. (1974). “Building Organizational Commitment: The Socialization of Managers in Work Organizations.” Administrative Science Quarterly, 19, (8), 68-73.

Dubin, R., Champour, & Porter (1975). “Central Life interests and Organizational Commitment of Blue Collar and Clerical Workers.” Administrative Science Quarterly, 20, (12), 32-43.

Getman, J. G., Goldberg, S. B., & Herman, S. B. (1986). Union Representation Elections: Law and Reality. New York: Russelsage Foundation.

Gorlin, H. (1984). Innovations in Managing Human Resources. New York: Confer-ence Board.

Guzzo, R. A. (Summer 1985). “The Effects of Psychologically Based Intervention Programs on Worker Productivity.” Personnel Psychology, 18, 275-293.

Hackman, J. R., & Oldham, G. R. (1980). Work Redesign. Reading, MA: Addison-Wesley.

Hall, D. T., Schneider, B., & Nygren, H. T. (1970). “Personal Factors in Organizational Identification.” Administrative Science Quarterly, 15, 48-59.

Hall, D. T. (1996). Career Development in Organizations. San Francisco: Jossey-Bass.Hrebniak, G., & Alutto, J. A. (1972). “Personal and Role-Related Factors in the

Development of Organizational Commitment.” Administrative Science Quarterly, 17, 26-37.

Jackson, E. (1987). “Stress Management and Personal Satisfaction.” Clinics of North America, 21, 559-576.

Janus, C. J. (November 22, 1982). “A Push to Unify the New Unions.” Business Week, 93- 94.

Jamal, M. (1974). “Task Specialization and Organizational Commitment: An Empirical Examination among Blue Collar Workers.” Academy of Management Proceedings,

34th Annual Meeting, Seattle, Washington, 226-239.Kahn, R., Wolfe, D., Quinn, R., Snoeck, J., & Rosenthal, R. (1964). Organizational

Stress: Studies in Role Conflict and Ambiguity. New York: Wiley.Kahn, R. L., Wolfe, D. M., Quinn, R. P., Snoeck, J. D., & Rosenthal, R. A. (1965).

Organizational Stress. New York: Wiley Publishers.Kaplan, B. H. (1987). “Social Support and Health.” Medical Care, 15, 47-58.

- 66

SUMMER 1999

Katz, D., & Kahn, R. L. (1978). Social Psychology of Organization. 2nd Edition. New York: John Wiley and Sons, 359-397.

Kidron, A. (1978). “Work Values and Organizational Commitment.” Academy of Management Journal, 21(2), 26-34.

Kochan, T. A. (March 1979). “How American Workers View Labor Unions.” Monthly Labor Review, 103, 4, 23-31.

Kotlowitz, A. (1987. August 28). “Labor’s Turn.” The Wall Street Journal, 1, 14.London, M. (1983). “Toward a Theory of Career Motivation.” Academy of Management Review, 9, (1), 620-630.

London, M. (1987). “Employee Development in a Downsizing Environment.” Journal of Business and Psychology, 2, 60-73.

London, M. (1990). “Career development.” Developing Human Resources, in K. N. Wexley and J. R. Hinrichs, eds., Washington, D.C.: BNA, 350-385.

London, M., & Bassman, E. (1989). “Retraining Midcareer Workers for the Future Workplace.” Training and Development in Organizations, I. L. Goldstein and Associates, eds., San Francisco: Jossey-Bass, 333-375.

London, M., & Mone, E. M. (1987). Career Growth and Survival in the Workplace. San Francisco: Jossey-Bass.

London, M., & Bray, D. W. (1984). “Measuring and Developing Young Manager’s Career Motivation.” Journal of Management, 3, 3-25.

March, J. G., & Simon, H. A. (1958). Organizations., New York: Wiley McGregor, D. (1967). The Professional Manager. New York: McGraw Hill.

Meyer, J., et al. (1989). “Organizational Commitment and Job Performance: It’s the Nature of the Commitment That Counts.” Journal of Applied Psychology, 74(1), 152-156.

Meyer, J. P., & Allen, N. J. (1987). “Organizational Commitment: Toward a Three-Component Model.” Research Bulletin 660, The University of Western Ontario, Department of Psychology, 68-80.

Miller, J. G. (1978). Living Systems. New York: McGraw Hill.Mintzberg, H. (1989). Mintzberg on Management: Inside Our Strange World of

Organizations. New York: Free Press, 120-125.Mintzberg, H. (1983). Power In and Around Organizations. Prentice Hall, 330-349.Morris, J. H., & Koch, J. L. (1979). “Impact of Role Perceptions on Organizational

Commitment, Job Involvement and Psychosomatic Illness among Three Vocational Groupings.” Journal of Vocational Behavior, 14, 88-101.

Mowday, R. T., Steers, R. M., & Porter, L. (1979). “The Measurement of Organizational Commitment.” Journal of Vocational Behavior, 8, (6), 224-247.Mowday, R., Porter, L., & Steers, R. (1982). The Psychology of Commitment, Absentee-

ism and Turnover. New York: Academy Press, 185-230.O’Reilly, C., III, & Chatman, J. (1986). “Organizational Commitment and Psychological

Attachment: The Effects of Compliance, Identification, and Internalization on Prosocial Behavior.” Journal of Applied Psychology, 71(3), 492-499.

Ouchi, W. G., & Jaeger, A. M. (1978). “Type Z Organizations: Stability in the Midst of Mobility.” Academy of Management Review, 3, 305-314.

67


Ouchi, W. (1980). “Markets, Bureaucracies, and Clan.” Administrative Science Quarterly, 25, 129-141.

Pascale, R. (May 28, 1988). “Fitting Employees into the Corporate Culture.” Fortune, 28- 40.

Patchen, M. (1965). “Some Questionnaire Measures of Employee Motivation and Morale.” [Monograph No. 41]. Ann Arbor, Michigan, Survey Research Center, 141-179.

Porter, L. W., Steers, R. M., Mowday, R. T. & Boulin, P. V. (1974). “Organizational Commitment, Job Satisfaction, and Turnover Among Psychiatric Technicians.” Journal of Applied Psychology, 5, 603-609.

Porter, L. W., Crampton, W. J., & Smith, F. W. (1976). “Organizational Commitment and Managerial Turnover: Longitudinal Study.” Organizational Behavior and Human Performance, 15, 87-98.

Rand, J. F. (June 1980). “Preventive Maintenance Techniques for Staying Union Free.” Personnel Journal, 498.

Reichers, A. (1986). “Conflicts and Organizational Commitment.” Journal of Applied Psychology, 71(3), 508-514.

Roberts, K. H., & O’Reilly, C. A. (1974). “Failures in Upward Communication in Organizations: Three Possible Culprits.” Academy of Management Journal, 17, 104-115.

Schein, E. H. (1991). Organizational Culture and Leadership. San Francisco: Jossey-Bass.

Tracy, L. (1984). “A Dynamic Living-Systems Model of Work Motivation.” Systems Research, 1, 191-203.

Tracy, L. (1989). The Living Organization: Systems of Behavior. New York: Praeger.Vaserhelyi, M. A. (1977). “Man-Machine Planning Systems: Style Examination of

Interactive Decision Making.” Journal of Accounting Research, 42, 310-316.Weiner, Y. (1982). “Commitment in Organizations: A Normative View.” Academy of

Management Review, 7(3), 418-428.Weiner, Y., & Gechman, A. G. (1977). “Commitment: A Behavioral Approach to Job

Involvement.” Journal of Vocational Behavior, 10, 47-52.Weiner, Y., & Vardi, Y. (1980). “Relationships Between Job, Organization, and Career

Commitments and Work Outcomes--An Integrative Approach.” Organizational Behavior and Human Performance, 26, 81-96.

White, J. K., & Ruh, R. H. (1973). “Effects of Personal Values on the Relationships Between Participation and Job Attitudes.” Administrative Science Quarterly, 18, 506-514.

- 68

SUMMER 1999

69


GROUP HETEROGENEITY AND STRATEGIC DECISIONS:

DO INFORMAL “CONSIDER THE OPPOSITE” PROCESSES

IMPROVE DECISION EFFICACY?

Ed Bukszar Jr. *

The effect of discussion in heterogeneous groups on the individual group members is measured in a hindsight bias, debiasing experiment. Results indicate that exposure to differing viewpoints did not improve decision making, but lead to polarization of viewpoints instead. Exposure alone may not stimulate the level of information processing necessary to bring about improved decision making. It is suggested that the nature of a decision process may be more critical to improvement in strategic decision making than heterogeneity of the decision team. Implications for strategic management are discussed.

ne fairly robust finding from research on executive team demographics is that performance, particularly in relatively dynamic environments, is enhanced by

executive team heterogeneity. Executive team heterogeneity is correlated with organizational adaptiveness, a critical factor in dynamic environments (Bantel & Jackson, 1989; Lant, Milliken & Batra, 1992; Murray, 1989). Heterogeneous teams are thought to gather information from a variety of sources representing diverse interpretations and perspectives leading to greater creativity and innovation, which may enhance a firm’s adaptability (Wiersma & Bantel, 1992).

O

Diversity also presents problems for organizations, such as constraining efforts to take decisive action (Goodstein, Gautam & Boeker, 1994). In contrast, similarity of schemata and cognitive structure in homogenous teams can be expected to enhance

*Ed Bukszar, Jr., is affiliated with Simon Fraser University, Vancouver, Canada.

- 70

SUMMER 1999

cohesion (Michel & Hambrick, 1992), facilitating both consensus decision making (Priem, 1990) and effective decentralization (Bourgeois & Brodwin, 1985).

On balance, positive effects of heterogeneity are thought to predominate as these groups quickly develop norms which facilitate interaction (Hambrick, 1994). Over time, culturally diverse groups experience process and performance improvements (Watson, Kumar & Michaelsen, 1993). Researchers attribute benefits from group heterogeneity to naturally occurring ‘devil’s advocacy’ or ‘dialectic inquiry’ thought to take place when executives with broadly different backgrounds work together through Hegelian processes to solve strategic problems.

The formal process of dialectic inquiry has three steps (Churchman, 1971). First, a plan (thesis) is presented and its underlying assumptions identified. Then a counter-plan (antithesis) is developed which is generally credible, but rests on different assumptions than the initial plan. Finally, a structured debate is conducted and a final plan (synthesis) is developed.

Devil’s advocacy is less structured than dialectic inquiry. The devil’s advocate, an appointed dissenter, develops a critique of the prevailing plan but offers no counterplan (Herbert & Estes, 1977). The adversary pressures decision makers to be more thorough in evaluation by forcing consideration of potential failure.

Schwenk (1991; Valacich and Schwenk, 1995) found that formal use of either dialectic inquiry or devil’s advocacy improves decision making. (In stable, well-understood environments both lead to reduced performance due to added time for decision making and increased uncertainty raised by the processes.) However, evidence on dialectic inquiry and devil’s advocacy applies to the formal use of these decision tools. The findings of research utilizing executive team demographics presumes the mere presence of executive team heterogeneity is sufficient to stimulate similar, albeit informal processes which capture the essence of formal decision aids. The finding that executive team heterogeneity leads to enhanced organizational adaptiveness would appear to bear this out. It is useful to know whether the mere presence of heterogeneity is sufficient to improve decision making. This paper tests these limits.

Previous Tests of Informal Dialectic-Type Processes

Using subjects who supported or opposed capital punishment, Lord, Lepper and Ross (1979) reported cause for pessimism as to whether simple exposure to opposite perspectives improves decision making. Subjects were presented with credible studies supporting each position. Reading the two studies further polarized beliefs on the death penalty. In general, polarization is thought to occur when confirming evidence is given greater weight than disconfirming evidence (Einhorn & Hogarth, 1978). Supporting evidence bolsters one’s initial position to a greater degree than disconfirming evidence casts doubt upon it.

71


Schoemaker (1993: 202) compared the forecasts of subjects required to write positive and negative scenarios for “strategically important issues from work or home” with subjects who merely read the same scenarios. He elicited forecasts from both groups two weeks prior to and two weeks after exposure to the scenarios, enabling a within-subject comparison. He reports no within-subject or between-subject differences in point estimates of forecasts.

Wider subjective confidence ranges were evident for both groups. Schoemaker interprets this as evidence of greater uncertainty, supporting the use of multiple scenarios. However, multiple scenarios are meant to improve decision making, not merely increase uncertainty. Improvements in forecasts would appear to be lacking in his study, given that there was no change in the subjects’ point estimates. As no benchmarks for appropriate forecasts were available, the lack of change in point estimates cannot be conclusively interpreted.

These studies suggest that advantages attributed to exposure to opposing perspectives may be somewhat elusive. However, the use of simple business scenarios or non-business scenarios may have hindered the generalizability to strategy as they lacked the contextual richness usually accompanying strategic decisions. The studies also used only written scenarios. It is reasonable to assume discussion would be an integral, and perhaps primary part of informal processes stimulated by group heterogeneity.

Hindsight Bias

Studies show we recall outcomes of most situations as more predictable than they seemed in prospect (Fischhoff, 1975, 1977, 1982a; Fischhoff & Beyth, 1975). When given outcome information, our remembered or reconstructed probability estimates are biased. We have a sense of having “known all along” that the outcome would occur. Fischhoff labeled this “hindsight bias.”

Hindsight effects are shown to be both wide-ranging and robust. (For reviews, see Christensen-Szalanski and Willham, 1991; and Hawkins and Hastie, 1990).) Attempts at reducing hindsight bias have had little success (Fischhoff, 1982b; Davies, 1987). The most noteworthy method is described by Slovic and Fischhoff (1977) who found that having subjects generate scenarios to explain possible outcomes that did not occur reduced, but did not eliminate, hindsight bias. Generation of alternate scenarios increases the salience of other possible outcomes and thus their subjective probabilities, with a corresponding reduction in the perceived inevitability of the known outcome. We refer to this approach as “considering the opposite.”

While considering the opposite reduces hindsight bias, it is counterintuitive to construct scenarios explaining outcomes which did not occur. A less formal way to achieve similar results is needed. One possibility is to combine individuals with differing perspectives on outcomes. In this way, rationales may be made more salient by discussion in a heterogeneous group, and hindsight bias may be reduced. Bukszar and Connolly (1988) suggested this as an area for investigation after demonstrating that discussion in homogeneous groups failed to have an effect on the bias.

- 72

SUMMER 1999

The parallel between hindsight debiasing and assumptions underlying the benefits of heterogeneous executive teams is striking. In both cases formal dialectic-type processes were shown to improve decision making. Informal processes, such as those stimulated by discussion in heterogeneous groups, may produce similar results. An investigation into the hindsight debiasing potential of heterogeneous groups would shed light on the ability of heterogeneous groups to improve decision making in general, absent implementation of formal “consider the opposite” processes.

This study tests the following hypothesis:

Discussion groups, consisting of subjects heterogeneous with regards to outcome information, will reduce hindsight bias and improve decision making efficacy.

METHODS

Forty-five MBA students in two sections of a Strategic Management seminar in Vancouver, Canada, participated in the experiment. Nineteen subjects had little managerial experience and non-business undergraduate degrees. Subjects hailed from ten countries other than Canada and seven of the ten Canadian provinces.

The remaining 26 subjects were enrolled in a separate Executive MBA program. Executives were from middle and upper management, and from large and small companies in industries that included forest products, mining, retailing, banking, government, regulated utilities, health care and software development. Educational backgrounds were diverse. Of the 45 subjects, 21 were female.

Subjects were more representative of groups within business organizations than is typically the case in student samples. They were familiar with each other’s abilities and interacted comfortably. The MBA programs from which the subjects were drawn operate on a cohort model where all coursework is taken together. The executive cohort met on alternating weekends, taking meals together and staying in the same hotel for the previous 20 months. Subjects from the non-executive cohort studied together over the previous eight months.

Subjects completed an in-class, written case analysis worth 10% of their final marks. The six-page case, an updated and slightly modified version of “Hygeia International,” (Newman & Logan, 1980) detailed a pharmaceutical company’s opportunity to enter chick-hatching and egg production in Nigeria. Although Hygeia (a pseudonym) had the necessary expertise and technology and sold pharmaceutical products in Nigeria for agricultural purposes, the decision required evaluation of several factors.

73


The case outlined the decision process, the information available to decision makers, and Hygeia’s past experience in chick-hatching (minimal) and in Nigeria (significant). The case provided positive and negative aspects of the project, and previous studies indicate that subjects were evenly split on whether to undertake the project (Bukszar & Connolly, 1988; Connolly & Bukszar, 1990).

Two versions of the case were distributed randomly to the subjects. The versions were identical until the final statement: “Following careful consideration, management at Hygeia decided to enter this market in 1990. Results for the first year of full operations (1992) indicate a __ R.O.I.” One version had a 4% R.O.I., the second a 36% R.O.I.

Subjects were asked to “evaluate the Nigerian project in terms of suitability for Hygeia, and provide what you believe to be the most likely explanation for its outcome.” Subjects, working alone, had 90 minutes to analyze the case and write a report.

Upon completion, subjects were separated randomly into two groups. The first (the control group) completed a questionnaire prior to discussing the case. The second group (the “discussion” group) immediately engaged in group discussion and then completed the questionnaire (see appendix).i Control subjects were moved to another room prior to distribution of the case, with their names displayed on an overhead to eliminate confusion, and given an explanation that separation would provide more work space.

The provision of differing outcomes (4% ROI versus 36% ROI) was intended to create two sets of distinct and variant perspectives on the project. Wide differences in judgment about the project, corresponding with the outcome information, have been produced in previous studies using this manipulation. The combination of these two sets of perspectives in discussion groups provides the manipulation of heterogeneity in this study.

- 74

SUMMER 1999

The study is a standard, 2 x 2 design with “Outcome” and “Discussion” the two factors, and “4% ROI / 36% ROI” and “Control / Discussion” the two treatments, as shown in Figure 1.

Figure 1Study Design

DISCUSSION

Control Discussion

OUTCOME

4% 4% / Ctrl 4% / Dis

36% 36% / Ctrl 36% / Dis

Comparisons between the discussion and control groups constitute the critical tests of the hypothesis. Convergence of opinion for the high and low-outcome subjects was expected to occur in the discussion group compared with the control group as a result of exposure to opposing perspectives. Thus, in contrast to the Schoemaker (1993) study, this manipulation provides benchmark expectations for estimates made after exposure to opposing perspectives.

The facilitator began the discussion by asking whether this was a good project for Hygeia to have entered. Subjects were told to ignore the outcome and evaluate the project based on merit alone.ii Vigorous discussion lasted 45 minutes, with high quality arguments equally for and against the project.

75


RESULTS

Data were analyzed using MANOVA, ANOVA, MANCOVA and ANCOVA.

Hindsight Effects

The test for hindsight effects involves a comparison of high-outcome subjects (36%) and low-outcome subjects (4%). MANOVA results reveal a significant main effect for ‘Outcome’ (F = 8.745, p < .001). ANOVA tests were conducted to delineate the MANOVA results. Summarized ANOVA results are reported in Table 1.

TABLE 1Mean Responses and ANOVA Results

LO LO HO HODependent Variables 4%/Ctrl 4%/Dis 36%/Ctrl 36%/Disa. Probability of success *** .22 .29 .65 .63b. Predicted first year ROI *** 6.3 6.8 21.0 22.0c. Adequacy of information 2.5 2.6 2.6 2.9d. Prediction confidence * 1.9 2.4 2.9 2.6e. Perceived riskiness * 3.6 3.5 2.8 2.6f. Readiness to invest * 3.5 3.5 2.6 2.5g. Willingness to enter * 0.5 0.55 0.73 1.0h. Confidence in decision 4.3 4.0 3.8 3.8i. Second yr. ROI *** 8.3 11.0 29.0 33.0j. Project continuance + 0.64 0.55 0.45 1.0k. Decision process rating ** 2.3 2.5 3.1 3.3l. Credit/Blame 3.2 3.1 3.2 3.4m. Cause of error 2.8 3.3 4.1 3.6n. Timing *** 1.9 2.0 2.6 3.0o. Personnel evaluation ** 3.3 3.6 2.6 2.2p. Autonomy shift ** 3.3 3.6 2.9 2.4q. Decision review 2.6 3.7 3.6 2.5r. Information search 2.5 2.1 2.8 3.3

|___________| | ______ | Low-Outcome vs. High-Outcome

Main effects for Outcome (LO 4% / HO 36%):* p<.05** p<.01***p<.001

Interaction effect for Outcome and Discussion:+ p<.05

Dependent variables are derived from the corresponding questions in the Appendix.

- 76

SUMMER 1999

Subjects given, but asked to ignore, a high-outcome found the project lower in risk (question e) than their low-outcome counterparts (mean of 2.7 for high-outcome subjects versus 3.6 for low-outcome subjects on a scale from 1 to 5 with very low risk equal to 1 and very high risk equal to 5; F = 6.93, p = .01). iii They expected both a higher first year ROI (question b) (21.5% versus 6.5%; F = 67.4; p < .001) and a higher second year ROI (question I) (31% versus 9.6%, F = 89.5; p < .001). They were more likely to have entered the market (question g) (87% versus 52%, F = 6.77, p = .01) and were more willing to invest their own money (question f) (2.6 versus 3.5 on a scale from 1 to 5, with 1 indicating that they definitely would invest and 5 indicating they definitely would not invest; F = 4.96, p < .05).

High-outcome subjects rated the decision process utilized in the case (k) as better (3.2 versus 2.4, on a scale from 1 to 5 with 1 indicating a poor process and 5 an excellent process, F = 7.82, p < .01). They thought the timing of the entry decision (n) was about right whereas the low-outcome subjects thought entry was too early (2.8 versus 1.95, on a scale of 1 to 5 with 1 indicating too early an entry, 5 too late an entry and 3 indicating that the timing was about right; F = 19.5, p < .001). Finally, the high-outcome subjects seemed more favorably disposed to allowing managerial autonomy in future decision making (p) (2.65 versus 3.44 on a scale of 1 to 5 with 1 indicating more autonomy and 5 less autonomy; F = 11.6, p < .01).

In general, questions not showing hindsight effects are questions where the presence of such effects would not be expected. There were no differences between the high-and low-outcome subjects with respect to their perception of the adequacy of the information to make a decision about the project (c), whether additional information should be gathered for similar decisions in the future (r), or with respect to the confidence in their own decision to enter the Nigerian market (h).

These results are highly consistent with past evidence on hindsight bias in general, and with the Bukszar and Connolly study (1988) in particular, which show that decision makers are unable to ignore outcome information when assessing decisions.

Discussion Effects

Attention can focus on the question that motivated this study: would participation in a heterogeneous group discussion reduce hindsight bias? To further demonstrate the rationale for expected convergence, recall that subjects were asked to evaluate the suitability of the project for Hygeia. This allowed consideration of variables other than to performance to enter the evaluation.

77


Written case reports indicate that a number of subjects who received information suggesting poor performance (subjects in the 4% ROI treatment) favored project continuance by suggesting: a) its evaluation over a longer period, or b) organizational benefit from goodwill generated by the project and from synergies not likely to show up in the financial results of the chick and egg business, such as improved performance of related veterinary and pharmaceutical activities in Nigeria. A number of subjects with information indicating good performance (36% ROI) suggested the project detracted from the focus of the organization by shifting attention to activities outside of the organization’s core competence. They suggested “spinning-off” the activity given that the project was up and running. Under this argument, the project’s value was seen as a stimulant to sales of veterinary pharmaceutical products. The group discussion, which further illuminated these points, was expected to result in convergence on project continuance.

The reduction of hindsight effects was thought most likely to occur on the question of whether to continue in the market (question j). The continuation decision was forward looking. As such, subjects were no longer asked to ignore outcomes, but instead to determine, in light of their analysis and discussion, whether project continuance was advisable.

The results were somewhat unexpected. MANOVA results reveal no evidence of a main-effect for “Discussion” (F = .417, p = ns) but do show a significant interaction effect between “Outcome” and “Discussion” (F = 3.926, p = .001). Rather than convergence, evidence of polarization was found.

ANOVA shows a significant interaction effect on the bellwether question of project continuance (j). Following group discussion, high-outcome subjects were more willing to continue with the project while low-outcome subjects were less willing compared to their control counterparts (F = 4.68, p < .05).

Subjects appear to have had their initial positions reinforced by confirming evidence found in the discussion while simultaneously discrediting the contrary evidence, similar to the effect documented by Lord et. al. (1979). These results lead to the rejection of the hypothesis under investigation.

Experience and Gender Effects

Experience and gender covariates were utilized in a MANCOVA analysis. Results indicate no main or interaction effects for gender. For experience, there were no main effects and only a minor interaction effect between experience and discussion (F = 2.74, p < .05). ANCOVA indicates this interaction effect relates only to question “m.” Following discussion, experienced subjects were more willing to attribute the discrepancy between the estimated ROI and the actual ROI to the manager, whereas inexperienced counterparts were more willing to attribute the discrepancy to the decision process (F = 4.83, p < .05).

- 78

SUMMER 1999

The lack of effects for experience and gender is consistent with research results to date, which indicate that experienced and inexperienced decision makers are equally susceptible to hindsight bias, as are both men and women, and that debiasing efforts are unaffected by these variables (Fischhoff, 1982b).

DISCUSSION

This study suggests that informal attempts to have subjects “consider the opposite” may be insufficient to reduce hindsight bias, and might instead cause the bias to be exacerbated. Results also call into question whether the benefits of heterogeneous groups, as detailed in the strategy literature, are attainable by merely bringing heterogeneous groups together.

The heterogeneous discussion groups stimulated conversation that clearly illuminated opposing perspectives. Indeed, an observer of this process might well have been led to conclude that the vigorous discussion resulted in improved decision making as participants appeared to be considering different perspectives. This appearance could be particularly seductive, absent measurable results, in that it is consistent with conventional wisdom. However, results from this study suggest that exposure to opposing arguments may not sufficiently stimulate the cognitive involvement necessary for increased decision efficacy.

More formal processes of dialectic inquiry or devil’s advocacy force greater cognitive involvement and thus greater information processing. Such cognitive involvement with the opposing argument may be essential to improved decision making. With formal dialectic, devil’s advocacy and multiple scenario procedures, the purpose of the exercise is known by all and proceeds with the blessing of management, endowing the process with greater legitimacy. As such, decision makers may consciously try to assimilate new possibilities.

How can these results be reconciled with the findings in the strategy literature that heterogeneous groups lead to improvements in organizational performance in dynamic environments? The improvements that organizations witness, and that have been attributed to executive team heterogeneity, may be more the result of the processes these organizations use to take advantage of the heterogeneity than the heterogeneity itself. For example, executives who assemble heterogeneous groups for decision making may favor processes designed to enhance internalization of opposing viewpoints. Selection of heterogeneous teams may be a manifestation of their preferences, based on the belief that heterogeneity creates an ample supply of opposing views.

However, conveyance of a favorable attitude towards dialectic type processes could, in and of itself, legitimize multiple scenario procedures, and lead to enhanced performance for either heterogeneous or homogeneous teams. Heterogeneity of executive teams might be more the effect of predisposition than the cause of improved

79


performance. The degree of heterogeneity of the management team may be a relatively minor, or even spurious, factor in improvements. This raises an interesting question: Would heterogeneous groups, acting without processes designed to facilitate consideration of the opposite, outperform homogeneous groups acting with the help of said processes? The answer could have dramatic and far-reaching consequences for strategic management.

Homogeneous groups may offer organizations advantages in terms of decision speed and a sense of shared culture, which, when combined with “consider the opposite” processes, may enhance the effectiveness of strategy formulation and implementation. For example, Bourgeois and Brodwin (1985) suggest that shared culture enables strategy implementations to be conducted in a more decentralized and evolutionary manner, while minimizing risks from agency problems. Zajac, Golden and Shortell (1991) suggest that, while diversity amongst individuals may contribute to the generation of new ideas, it may prevent ideas from being successfully implemented. Regardless, results of the current study are an important reminder that executive team demographics are proxies and should not be considered as an end in themselves, since they do not illuminate actual strategic decision processes.

Limitations and Suggestions for Future Research

This study, while raising doubts about the value of informal devil’s advocacy processes, is by no means conclusive. Its primary value is to raise questions. Researchers should perhaps re-examine conventional wisdom surrounding studies of decision making that rely on demographic comparisons. This study does not invalidate those results.

The most important practical implication of this study is the suggestion that the benefits of group heterogeneity may be elusive if not augmented by processes designed to force internalization of opposing viewpoints. The nature of those processes is not yet clear. It may involve simple encouragement from a chief executive to take opposing viewpoints seriously, or the existence of an organizational culture which encourages such behavior, or more formal measures, particularly in organizations where opposing viewpoints have not historically been appreciated.

A good follow up to this study would be to compare the performance of heterogeneous groups without formal aids designed to get participants to internalize opposing viewpoints, with homogeneous groups utilizing such aids. By manipulating the degrees of heterogeneity and the formality of the decision aids, researchers could gain an understanding of what is necessary and sufficient to improve decision efficacy.

- 80

SUMMER 1999

NOTES

1. Subjects were informed that questionnaire responses would not affect their marks.2. Subjects conducted case analyses previously in their strategy seminar and learned to analyze cases without discussing outcomes. In this experiment, subjects were unaware that two outcomes were given. This was confirmed for each subject in a post-experiment debriefing. This manipulation is conservative in design. Had it broken down and subjects become aware that two outcomes were given, the significance of results would have been severely reduced.3. The mean of 2.7 for high outcome subjects in question ‘e’ is the weighted mean for the 6% / control subjects (2.8), and the 36% / discussion subjects (2.6), as depicted at the bottom Table 1. The remaining numbers reported in this section were calculated in like manner.

81


REFERENCES

Bantel, K. A. & Jackson, S. E. (1989). “Top Management and Innovations in Banking:Does the Composition of the Team Make a Difference.” Strategic Management Journal, 10, 107-124.

Bourgeois, L. J. III & Brodwin, D. R. (1985). “Strategic Implementation: Five Approaches to an Elusive Phenomenon.” Strategic Management Journal, 5, 241-264.

Bukszar, E. & Connolly, T. (1988). “Hindsight Bias and Strategic Choice: Some Problems in Learning from Experience.” Academy of Management Journal, 31(3), 628-641.

Chatterjee, S., Lubatkin, M. H., Schweiger, D. M. & Weber, Y. (1992). “Cultural Differences and Shareholder Value in Related Mergers: Linking Equity and Human Capital.” Strategic Management Journal, 13(5), 319-334.

Churchman, C.W. (1971). The Design of Inquiring Systems: Basic Concepts of Systems and Organization. New York: Basic Books.

Christensen-Szalanski, J.J. & Fobian Willham, C. F. (1991). “The Hindsight Bias: A Meta-Analysis.” Organizational Behavior and Human Decision Processes, 48, 147-168.

Connolly, T. & Bukszar, E. (1990). “Hindsight Bias: Self-Flattery or Cognitive Error?” Journal of Behavioral Decision Making, 3, 205-211.

Davies, M. F. (1987). “Reduction of Hindsight Bias by Restoration of Foresight Perspective: Effectiveness of Foresight-Encoding and Hindsight-Retrieval Strategies.” Organizational Behavior and Human Decision Processes, 40, 50-68.

Einhorn, H. J. & Hogarth, R. M. (1978). “Confidence in Judgment: Persistence of the Illusion of Validity.” Psychological Review, 55(5), 395-416.

Fischhoff, B. (1975). “Hindsight Foresight: The Effect of Outcome Knowledge on Judgment Under Uncertainty.” Journal of Experimental Psychology: Human Perception and Performance, 1, 288-299.

Fischhoff, B. (1977). “Perceived informativeness of facts.” Journal of Experimental Psychology: Human Perception and Performance, 3: 349-358.

Fischhoff, B. (1982a). “For Those Condemned to Study the Past: Heuristics and Biases in Hindsight.” In Kahneman, D., Slovic, P. & Tversky, A. (Eds.) ;Judgment Under Uncertainty: Heuristics and Biases. Cambridge: Cambridge University, 335-354.

Fischhoff, B. (1982b). “Debiasing.” In Kahneman, D., Slovic, P. & Tversky, A. (Eds.), Judgment Under Uncertainty: Heuristics and Biases. Cambridge: Cambridge University, 422-444.

Fischhoff, B. & Beyth, R. (1975). “I Knew It Would Happen: Remembered Probabilities of Once-Future Things.” Organizational Behavior and Human Performance, 13, 1-16.- 82

SUMMER 1999

Goodstein, J., Gautman, K. & Boeker, W. (1994). “The Effect of Board Size and Diversity on Strategic Change.” Strategic Management Journal, 15(3), 241-250.

Hambrick, D. C. (1994). “Top Management Groups: A Conceptual Integration and Reconsideration of the ‘Team’ Label.” In B. M. Staw & L. L. Cummings (Eds.), Research in Organizational Behavior, 16, 171-214. Greenwich, CT:JAI Press.

Hawkins, S. A. & Hastie, R. (1990). “Hindsight: Biased Judgments of Past Events After the Outcomes are Known.” Psychological Bulletin, 107(3), 311-327.

Herbert, T.T. & Estes, R. W. (1977). “Improving Executive Decisions by Formalizing Dissent: The Corporate Devil’s Advocate.” Academy of Management Review, 3, 663-667.

Lord, C., Lepper, M. R. & Ross, L. (1979). “Biased Assimilation and Attitude Polarization: The Effects of Prior Theories on Subsequently Considered Evidence.” Journal of Personality and Social Psychology, 37, 2098-2110.

Michel, J. G. & Hambrick, D. C. (1992). “Diversification Posture and Top Management Team Characteristics.” Academy of Management Journal, 35(1), 9-37.

Murray, A. I. (1989). “Top Management Group Heterogeneity and Firm Performance.” Strategic Management Journal, 10, 125-143.

Newman, W. & Logan, J. (1980). Strategy, Policy and Central Management. Cincinnati: South Western.

Priem, R. L. (1990). “Top Management Team Group Factors, Consensus, and Firm Performance.” Strategic Management Journal, 11(6), 469-478.

Schoemaker, P. J. H. (1993). “Multiple Scenario Development: Its Conceptual and Behavioral Foundation.” Strategic Management Journal, 14, 193-213.

Schwenk, C.R. (1990). “Effects of Devil’s Advocacy and Dialectical Inquiry on Decision Making: A Meta-Analysis.” Organizational Behavior and Human Decision Processes, 47, 161-176.

Slovic, P. & Fischhoff, B. (1977). “On the Psychology of Experimental Surprises.” Journal of Experimental Psychology: Human Perception and Performance, 3, 544-551.

Watson, W. E., Kumar, K. & Michaelsen, L. K. (1993). “Cultural Diversity’s Impact on Interaction Process and Performance: Comparing Homogeneous and Diverse Task Groups.” Academy of Management Journal, 36(3), 590-602.

Valacich, J. S. & Schwenk, C. (1995). “Devil’s Advocacy and Dialectic Inquiry Effects on Face-to-Face and Computer-Mediated Group Decision Making.” Organizational Behavior and Human Decision Processes, 63(2), 158-173.

Wiersma, M. F., & Bantel, K. A. (1992). “Top Management Demography and Corporate Strategic Change.” Academy of Management Journal, 35(1), 91-121.

Zajac, E. J., Golden, B. R. & Shortell, S. M. (1991). “New Organizational Forms for Enhancing Innovation: The case of Internal Corporate Ventures.” Management Science, 37(2), 170-184.

83


APPENDIX

Questionnaire Assume that the decision to enter agricultural production in Nigeria was made in

1990. Imagine that you are back at that point in time.a. Probability of success: With only the information you would have had in 1990,

state what you believe to be the probability of success and the probability of failure for the first full year of operations (1992) of the Nigerian chick-farming venture. (A successful operation would meet the 20% ROI goal, an unsuccessful operation would fall short of that goal). Be sure the two probabilities add to 100%.

b. Predicted first-year ROI: What would your best estimate have been of the ROI figure for 1992 (the first full year of operation)?

c. Adequacy of information: Was the available information adequate for making a decision of this significance? (completely inadequate = 1 and completely adequate = 5).

d. Prediction confidence: How confident do you feel making these ROI predictions from the information available in 1990? (not at all confident = 1 and completely confident = 5).

e. Perceived riskiness: How risky would you rate the decision to go ahead? (very low risk = 1 and very high risk = 5).

f. Readiness to invest: Would you have been prepared to invest your own money in this project? (definitely would invest = 1 and definitely would not invest = 5; reverse scored, so that high scores indicate greater readiness).

g. Would have entered: Putting yourself in Mr. Livingstone’s position at the time of the original decision to enter the chick business, would you have entered the market? (Yes = 1, No = 0).

h. Decision confidence: How confident would you be in your decision? (not confident = 1, confident = 5).

Using all of the information available to you today:i. Predicted second-year ROI: What is your best estimate of the ROI figure for 1993

(second full year of operation)?j. Recommend continuance?: Should Hygeia continue this project? (Yes = 1, No =

0). k. Rate decision process: Rate the strategic decision process utilized in the case.

(poor = 1 and excellent = 5). l. Credit/blame locus: Does the credit/blame for the decision in this case lie primarily

with the process or with the decision makers? (process = 1, 50/50 = 3, and decision makers = 5).

m. Error cause: As it turned out Murtala’s original estimate of ROI was substantially in error. Do you attribute this error to the difficulty of making estimates in situations like this or does the fault lie in Murtala’s estimating ability? (Murtala is the manager in Nigeria). (Error mainly due to Murtala = 1, 50/50 = 3, and situation = 5).

- 84

SUMMER 1999

n. Decision timing: The timing of the Hygeia decision, like many business decisions, epresents a balance between moving too early, before enough information is available, and moving too late, after the good opportunities have been taken. Would you say Livingston (Hygeia’s U.S. manager) moved: (too early = 1, about right = 3, and too late = 5).

o. Personnel evaluation: Who do you think deserves the most credit/blame for the outcome in this case? (Murtala = 1, 50/50 = 3, and Livingstone = 5).

p. Autonomy shift: Should Murtala have more or less autonomy in future decisions? (more autonomy = 1 and less autonomy = 5).

q. Decision review: In the future, should someone in a position of authority over Livingstone review his decisions prior to implementation? (should not review = 1, should review = 5).

r. Information search: Imagine yourself in Mr. Livingstone’s position in the future. A decision of similar complexity arises. You have the option of gathering more information on market conditions prior to deciding but would do so knowing that your competitors may beat you into the market. Would you seek out additional information? (seek more information = 1 and make decision now = 5).

85


A COMPUTATIONAL PROCEDURE FOR SETTING

CUTOFF SCORES FOR MULTIPLE TESTS

Zvi Drezner *

George O. Wesolowsky**Willi Wiesner***

Organizations commonly use several different predictors such as ability tests, personality tests, interviews, and reference checks in selecting employee. In fact, the use of multiple predictors serves as a good way of triangulating on applicants’ abilities to do the job and generally provides better prediction of job performance than a single predictor. However, in using multiple predictors, organizational decision makers are faced with the challenge of how to make sense of the various, and sometimes conflicting, sources of information about applicants in order to make an informed decision.

Setting a cutoff on a single predictor is relatively straightforward. However, the subjective procedures used for setting cuttoffs on multiple predictors have not been satisfactory. This study demonstrates the feasibility of a computational procedure for setting multiple cutoffs. Such a procedure should be more cost-effective and accurate than the subjective methods currently used.

rganizations commonly use several different predictors such as ability tests, personality tests, interviews, and reference checks in selecting employees (Bureau

of National Affairs, 1988; Karren & Nkomo, 1988). In fact, the use of multiple predictors serves as a good way of triangulating on applicants’ abilities to do the job and generally provides better prediction of job performance than a single predictor (Gatewood & Feild, 1994). However, in using multiple predictors, organizational decision makers are faced with the challenge of how to make sense of the various, and sometimes conflicting, sources of information about applicants in order to make an informed decision. One way

O

*Zvi Drezner is affiliated with California State University, Fullerton, California. **George O. Wesolowsky is affiliated with McMaster University, Hamilton, Ont. Canada.*** Willi Wiesner is affiliated with McMaster University, Hamilton, Ont. Canada.

- 86

SUMMER 1999

of dealing with data from multiple predictors is to enter them into a regression equation. The applicant’s scores on each predictor (e.g., tests, interviews, reference checks) are

weighted and summed to yield a total score (e.g., predicted job performance). The appropriate regression weights or b values are determined through prior research where the unique contributions of each predictor score ( ) to predicting job performance (Y) are investigated.

The multiple regression approach assumes that a low score on one predictor can be compensated for by a high score on another predictor. Thus, an applicant could do very poorly in the interview (e.g., receive a score of zero) and still do well if he or she receives high scores on the tests and the reference check. However, this assumption made by the multiple regression model is not necessarily warranted. There might be a minimum level of competence required on each of the predictors for the individual to perform acceptably in the job. For example, a very low interview score might indicate that the applicant has such poor interpersonal and communication skills that he or she cannot function acceptably in retail sales, regardless of high cognitive ability and extraversion scores.

When the assumptions of the multiple regression approach cannot be met, an alternative is to use cutoff scores (whether in multiple cutoff, multiple hurdle, or combination methods. See Gatewood & Feild, 1994). Cutoff scores serve as criteria or thresholds in selection decisions. Applicants who score below the cutoff on any of the predictors are rejected. Thus, cutoff scores ensure that applicants meet some minimum level of ability or qualification to be considered for a job. The cutoff approach assumes that a minimum level is required on each of the attributes measured by the predictors for successful job performance (i.e., it does not assume a linear relationship among the predictors and job performance). The approach also assumes that the predictors are noncompensatory (i.e., it is not possible to compensate for a low score on one predictor with a high score on another predictor).

Setting cutoff scores is relatively straightforward when there is only one predictor. The usual method involves identifying the proportion of applicants who are to be hired and determining how stringent the cutoff score should be to select only the desired number of applicants. First, the expected selection ratio is calculated (number of individuals to be hired divided by the expected number of applicants). Next, the distribution of the applicants’ scores on the predictor is estimated by examining the predictor score distributions of past groups of applicants or of current employees (i.e., predictive or concurrent validation data). Finally, the cutoff score is established by applying the selection ratio to the predictor score distribution in order to determine the score that only the top applicants (the proportion to be hired) would attain. For example, if a fire department seeks to hire 5 firefighters and 150 people are expected to apply, the selection ratio will be .03 (5/150). About 3 percent of expected applicants will be accepted or, conversely, about 97 percent of expected applicants will have to be rejected. The cutoff score should therefore be set at the 97 th percentile of the distribution of predictor scores (plus or minus one standard error of measurement). That is, the cutoff score is set so that only 3% of applicants would be expected to meet or exceed the score (or 97% would fall below it). As noted above, this approach is limited to setting cutoffs for a single predictor.

87


When more than one predictor is used, setting cutoff scores is considerably more difficult (Buck, 1977; Cascio et al., 1988; and Cascio, 1991). Cascio (1991, p. 286) points out that “in general, no satisfactory solution has yet been developed for setting optimal cutoff scores in a multiple cutoff model.” Although it is not an ideal approach, one method that has been used to set cutoff scores is to use expert judges. There are several ways in which expert judges can be used to establish cutoffs but they differ only slightly in their methods. We will briefly consider the general approach and readers are encouraged to consult Cascio (1991), Gatewood and Feild (1994), or Schmitt and Klimoski (1991) for more detailed treatments of the various methods. Experienced employees, supervisors, or managers who know the job well or industrial psychologists typically serve as expert judges. Essentially, the expert judges are asked to rate the difficulty of test items (or interview questions) and to indicate what score on each item should be attained by a minimally competent applicant. These ratings are summed across items to yield a pass threshold or cutoff score. Cutoff scores can be established in this manner for each of the predictors used in the selection process.

Although expert judges are able to determine cut-off scores to ensure minimal competence on each of the predictors, they have tremendous difficulty setting optimal cut-off scores for several predictors so that the desired number of applicants is selected (Bazerman, 1986). That is, they must not only determine what scores would indicate minimal competence on each predictor but also what scores on each predictor would generate the desired number of selectees. Rank ordering scores for each predictor is inadequate for these purposes because the rank ordering of each applicant will differ across predictors. For example, an applicant who is ranked 1st on one predictor might be ranked 20th on another predictor and 59th on a third predictor. Clearly, a computational process is the only efficient means of setting multiple cut-off scores.

The purpose of this study is to explore the feasibility of a computational approach to setting multiple cutoff scores. Such an approach would be more objective and more efficient than using expert judges. Moreover, we expect that this approach will generate cutoff scores which will optimize the prediction of successful job performance.

FORMULATION

Assume a matrix of k columns and n rows. This matrix contains scores , representing the score of candidate i in test j. The

matrix also contains a vector indicating whether a candidate’s job performance is acceptable ( ) or unacceptable ( ). The problem is to find k cutoff scores, so that a candidate is classified as “acceptable” if he or she is above (or equal to) the cutoff score in at least k-p of the scores, and is classified as “unacceptable” if he or she misses the cutoff it more than p times. (The number p may be zero if a

- 88

SUMMER 1999

candidate is classified as “acceptable” if he meets the cutoff in all tests.) The objective is to minimize the number of incorrect classifications of candidates.

For each column j the are sorted in increasing order and distinct values

are obtained. because there may be ties in the scores. Each column

j defines a set of 0-1 variables If the score is below the cutoff point, then

, and if the score is above the cutoff point, then . In order to guarantee that the x’s define a cutoff score, we define the set of constraints for each j:

(1)

This set of inequalities guarantees that the vector starts with a sequence of uninterrupted zeroes and continues with a sequence of uninterrupted ones. It is possible to have either no zeroes at the beginning in which case all the ’s are equal to one, or no ones at the end of the vector in which case all the ’s are equal to zero. The cutoff point is between the two values where the zeroes end and the ones start (or the first

). If , then the score is below the cutoff point for column j, and if , the score is above the cutoff point. Once a solution is obtained, the cutoff points can be determined by examining the solution .

For each score there exists an index such that . The variable

determines whether this score is above the cutoff for that column or not.

Therefore, the number of tests which are above the cutoff point for row i is .

If the candidate is classified as “acceptable”, otherwise the

candidate is classified as “unacceptable”. Consider the expression (recall that means an acceptable candidate and means an unacceptable one):

(2)

1. If candidate i is acceptable ( ), the second term in (2) is zero. If he or she is classified as “acceptable” the expression (2) is non-positive If he or she is classified as “unacceptable” the expression (2) is positive (is at

least one)2. If candidate i is unacceptable ( ), the first term in (2) is zero.

If he is classified as “unacceptable” the expression (2) is non-positive

89


If he is classified as “acceptable” the expression (2) is positive (is at least one)

3. Therefore, if the classification is it correct, expression (2) is non-positive, and if the classification is wrong, expression (2) is at least one.

4. The maximum possible value for expression (2) is k-p if and p+1 if . This is the same as

n additional 0-1 variables are added to the formulation. indicates whether the classification is correct ( ), or incorrect ( ). The objective is to

maximize

By the above discussion, is defined by the constraint:

Rearranging terms leads to:

(3)

The complete formulation is:

(4)

DISCUSSION

The Case p=0

In our investigation we found that for p=0: when only are required to be 0-1 variables and may be real numbers, we usually obtain integer values for the x’s as well. We did not find an example where this is not true, but it cannot generally be proven. This makes the formulation “integer friendly” for the values of the x’s and only n integer

- 90

SUMMER 1999

variables are required. We propose to solve the problem with only the requirement that the n y’s are integer, and only if non-integer values are obtained for the x’s, to resolve the problem requiring that it all variables are integers. The reason for this property is as follows. Consider the second constraint in (4). There are four possibilities for the values of and summarized in Table 1.

Examining Table 1 shows that when the constraint is superfluous because it is always satisfied whether the x’s are fractions or integers. This should be expected because can only be 0 or 1 and since we maximize the sum of the y’s, will be equal to 1 at the maximum unless the constraint involving it is not satisfied. In this case, if the constraint for is not satisfied, we have an infeasible solution. On the other hand, if

and , then all the x’s on the left hand side of this constraint must all equal to 1. All the constraints for which force many x’s to be equal to 1. Therefore, a constraint for which , will usually have some of its x’s on the left hand side equal to one because of other constraints. If k-1 of them are equal to 1, then the remaining one must be equal to zero. By this reasoning many of the x’s are forced to be zero or one (mostly one). The only case where we may get non-integer x’s is when for a constraint of the type , fewer than k-1 x’s are fixed to zero or one by the above reasoning by other constraints. And even then it is possible that the solution will be zero or one for these “free” x’s. We would also like to note that for practical problems (and especially near the optimum) there are many because means that we predict correctly candidate i and we expect to predict correctly many of them. That means that the majority of the constraints will be for which yields many “non-free” x’s that are forced to get the value zero or 1.

91


Table 1: Possible Constraints

Constraint

0 0

0 1

1 0

1 1

This argumentation is not effective for p>0 because even for a constraint of the type no x’s are surely fixed to 1.

Computational Experience

The integer programming formulation was programmed in AMPL (Fourer et al.,1993) and given in the appendix. The program is available from the first author.

We constructed two illustrative problems: one problem with n=30 candidates and k=5 tests, and the other with n=100 candidates and k=3 tests. The first problem was investigated more extensively. The raw data for the first problem have been already sorted for each test and common scores grouped into one variable. We had nine different scores among the 30 candidates for each test. The data used are given in Table 2.

The first problem is based on 45 x-variables and 30 y-variables. There are 40 constraints of the first type and 30 constraints of the second type.

We solved the problem for p=0, 1, 2, 3, 4 by linear programming (no integrality constraints), only 0-1 y’s and all integer variables. The solution for this particular problem was obtained in a couple of seconds of computer time on a 486 33MHz computer. The response was so fast that we did not notice a difference in run time between the linear programming solution, only 0-1 y’s solution, and all integer solution. The results are summarized in Table 3.

- 92

SUMMER 1999

Table 2: The First Example Problem

i Test1 Test2 Test3 Test4 Test51 0 2 4 6 3 12 1 4 5 7 9 43 0 3 7 4 5 14 0 5 2 2 1 95 1 7 3 6 8 7

6 0 6 7 2 7 27 0 3 7 7 2 18 0 1 7 1 5 39 1 6 1 2 9 410 0 3 7 3 2 8

11 1 8 7 5 3 512 1 5 6 3 7 613 1 1 7 4 8 914 0 5 9 5 4 215 1 8 8 6 6 1

16 1 9 3 9 6 317 1 9 1 9 8 618 0 2 6 7 1 419 0 3 7 8 3 820 0 3 4 2 5 9

21 0 5 8 1 6 222 1 9 9 4 6 723 0 4 3 3 8 424 1 5 3 8 9 625 0 1 6 7 1 3

26 0 4 5 4 2 927 0 9 1 9 2 628 0 3 6 1 5 929 1 8 7 5 7 130 0 5 8 6 3 1

For p=0, 27 candidates were categorized correctly (the objective function is 27, and all except 9, 11, 23). The cutoff points were 1, 1, 3, 6, 1, respectively, for the five

ii

iii

93


tests. These cutoff scores were determined by checking the solutions for the x’s. For j=1, 2, and 5 all the x’s were equal to 1 and therefore the cutoff score is the lowest value. For

j=3 the first two x’s were equal to 0 and the rest equal to 1, and for j=4 the first five x’s were equal to 0. A candidate is classified “acceptable” if he scored at least these values in all five tests and is classified “unacceptable” if he failed to get at least the particular score on any of the five tests. Note that tests #1, #2 and #5 can be discarded from consideration because the “passing” score is the minimum possible score and all candidates (whether “acceptable” or “unacceptable”) passed these tests successfully. The testing procedure can be simplified to include only tests #3 and #4 because the other three tests do not provide any information that helps to distinguish between acceptable and unacceptable candidates. They are dominated by tests #3 and #4.

The same solution was obtained when we require only the y’s to be integer for all p’s except for p=1. For p=1 we obtained the same y’s but some of the x’s were non-integer. The linear programming solution was non-integer in all cases.

Table 3: The First Example Problem Solved by Linear and Integer Programming

p=0 p=1 p=2 p=3 p=4L.P. Solution (z) 29.4 29.75 29.667 29.5 29.40-1 y Solution 27 29 29 28 27Are x’s Integer? yes no yes yes yes

All integer 27 29 29 28 279,11,23 9 27 12,27 12,13,27

Cutoff Points 1,1,3,6,1 4,6,4,6,3 5,10,4,6,4 6,10,4,8,10 7,10,9,9,10

The second problem with 100 candidates and three tests is about the largest problem that can be solved by the students’ version of AMPL which can handle problems with up to 300 variables and 300 constraints. The full-size version of AMPL can efficiently handle much larger problems. The data for the second problem are given in Table 4. In order to prepare the data for AMPL, the scores for each test were sorted, and the lowest score was entered as “1”, the second lowest as a “2” and so on.

The student version of AMPL (Fourer et al., 1993) found a solution to the second problem in a couple of seconds. The cut-off points are 99, 28, and 52.Ninety seven of the applicants were classified correctly. Only applicants #6, #11, and #20 were classified incorrectly. Both cases when only the y’s are integer or all the variables are integer were solved in about the same computer time and yielded the same solution. The linear programming relaxation of the problem resulted in a solution with an objective function of 99.

- 94

SUMMER 1999

CONCLUSION

Our program was able to quickly generate multiple cutoff scores with a very high degree of accuracy. This computational method requires considerably less time and effort than the usual expert judgement approach and has the potential of providing substantial savings to the hiring organization. Moreover, this approach is more likely to provide accurate cutoff scores because it is keyed to job performance scores rather than subjective judgements made by job experts.

95


Table 4: The Data for the Second Problem

# Scores # Scores # Scores1 130 51 56 1 34 95 33 60 0 67 114 43 69 12 85 40 71 0 35 95 48 67 0 68 120 40 54 13 100 37 77 1 36 92 42 39 0 69 122 46 56 14 90 52 58 0 37 131 43 63 1 70 99 43 53 15 105 42 69 1 38 110 38 63 1 71 131 45 59 16 91 39 79 1 39 103 39 73 1 72 134 35 66 17 114 39 75 1 40 121 34 61 1 73 114 45 65 18 129 43 82 1 41 86 22 52 0 74 111 38 60 19 120 33 54 1 42 100 42 51 0 75 119 38 63 110 122 44 52 1 43 102 50 76 1 76 136 41 61 111 94 28 66 1 44 103 36 75 1 77 113 37 60 112 120 53 72 1 45 115 40 66 1 78 117 39 61 113 122 32 73 1 46 90 40 80 0 79 101 34 56 114 103 39 63 1 47 99 29 69 1 80 88 32 73 015 118 39 73 1 48 124 43 62 1 81 125 43 72 116 100 44 43 0 49 131 46 68 1 82 127 38 76 117 92 44 47 0 50 128 30 59 1 83 117 30 57 118 99 34 62 1 51 91 38 75 0 84 132 37 85 1

19102 46 66 1 52 131 40 72 1 85 112 30 81 1

20 92 40 71 1 53 105 43 77 1 86 111 40 68 121 101 49 62 1 54 90 33 68 0 87 122 41 42 022 89 31 74 0 55 83 35 71 0 88 94 35 54 023 111 33 69 1 56 143 39 47 0 89 123 40 73 124 120 35 54 1 57 136 49 70 1 90 84 45 54 025 90 50 63 0 58 113 37 75 1 91 101 28 71 126 119 38 76 1 59 114 46 72 1 92 131 40 76 127 113 50 65 1 60 115 38 62 1 93 102 37 64 128 103 41 64 1 61 103 34 65 1 94 110 40 71 129 125 46 73 1 62 85 39 77 0 95 140 38 55 130 105 43 53 1 63 117 40 61 1 96 98 34 52 031 102 35 50 0 64 107 34 52 1 97 116 42 67 132 118 45 68 1 65 117 48 73 1 98 112 38 78 133 90 41 46 0 66 119 40 66 1 99 113 54 87 1

100 119 47 65 1

- 96

SUMMER 1999

APPENDIX: THE AMPL PROGRAM AND DATA FILE

The AMPL program:

option solver cplex;param n;param k;param p;set K := {1..k};set I := {1..n};param g {i in I};param s {i in I, j in K};param N {j in K};set M {j in K} := {1..N[j]};var x {j in K, i in M[j]} integer >=0 <=1;var y {i in I} integer >=0 <=1;maximize objective: sum {i in I} y[i];subject toconst1 {j in K, i in M[j]: i<N[j]}: x[j,i] <= x[j,i+1];const2 {i in I}: (1-2*g[i])*(sum {j in K} x[j, s[i,j]])+((k-p)*g[i]+(p+1)*(1-g[i]))*y[i] <= k*(1-g[i]);

Note that if one is interested in a solution with continuous x variables, then one should just remove the definition “integer” from the variable definition of x.

The data file for the first example problem is:

param n=30;param k=5;param p=0;param g :=1 02 129 130 0;param s :=1 1 22 1 43 1 328 5 929 5 130 5 1;param N :=1 92 93 94 95 9;

97


REFERENCES

Bazerman, M. H. (1986). Judgement in Managerial Decision Making. New York: Wiley.

Buck, L. S. (1977). Guide to the Setting of Appropriate Cutting Scores for Written Tests: A Summary of the Concerns and Procedures (Report No. TM-77-4). Washington, DC: U.S. Civil Service Commission.

Bureau of National Affairs. (1988). Recruiting and Selection Procedures (PFF Survey No. 146). Washington, DC: Bureau of National Affairs, Inc.

Cascio, W. F. (1991). Applied Psychology in Personnel Management (4th ed.). Engelwood Cliffs, NJ: Prentice-Hall.

Cascio, W. F., Alexander, R. A., & Barrett, G. V. (1988). “Setting Cutoff Scores: Legal, Psychometric, and Professional Issues and Guidelines.” Personnel Psychology, 44, 1-24.

Fourer, R., D.M. Gay, and B.W. Kernighan (1993). AMPL: A Modeling Language for Mathematical Programming. San-Francisco: The Scientific Press

Gatewood, R. D., & Feild, H. S. (1994). Human Resource Selection (3rd ed.). Fort Worth, TX: The Dryden Press.

Karren, R. J., & Nkomo, S. M. (1988). “So you want to work for us.” Personnel Administrator, 38, 88-92.

Schmitt, N. W., & Klimoski, R. J. (1991). Research Methods in Human Resources Management. Cincinnati OH: South-Western Publishing.

- 98

SUMMER 1999

99


SIX-SIGMA QUALITY IN SMALL BUSINESSES:

A GENESIS FOR GROWTH

Mustafa R. Yilmaz*

Sangit Chatterjee**

ABSTRACT

Six-sigma quality has become a common term for describing a production process of extremely robust quality; one that has very low variability compared with the tolerance limits specified for the product being manufactured. In addition to ensuring dependable quality, such a process also makes efficient use of resources and eliminates waste. Although it is commonly associated with large scale production systems, six-sigma quality is also a metaphor for excellence in managerial thought. In this paper, we discuss this metaphor from the viewpoint of small but growing businesses, and argue that six-sigma way of thinking is not a luxury for a small business, but rather, a genesis for growth. This argument brings forth the importance of taking a systems view of a business regardless of its size, products, or services.

INTRODUCTION

he United States Small Business Administration (SBA) defines any business that has between 1 and 499 employees as a small business. Although exact numbers are

elusive, SBA statistics show that there are approximately 25 million small businesses in the United States. They represent a dynamic segment of the economy where many new businesses are born each year, and many cease to exist. This segment creates virtually all of the net new jobs each year, employs more than half of the private workforce, represents 99.7 percent of all employers, provides 55 percent of all innovations, and accounts for more than half of the private sector output (see http://www.sba.gov).

T

A new business is born to offer something that did not exist before, be it a new service or product, lower cost, better quality, just about anything to fit a niche. The business could be as ordinary as a new neighborhood grocery store, an auto repair shop, a dental practice, a house painting business, or as exotic as a high-technology product or service resulting from many years of development. While the novelty of its products or services are often critical to get a business off the ground, it cannot survive long without meeting the demands and expectations of its customers, that is, without achieving “perceived quality” (Garvin, 1988). Thus, quality is of paramount importance for the *Mustafa Yilmaz is affiliated with Northeastern University.** Sanjit Chaterjee is affiliated with Northeastern University.

- 100

SUMMER 1999

survival of a business, and this is keenly recognized by many small business owners. This is illustrated by the excerpts in Table 1 which are taken from the best business tips provided by recent winners of the SBA outstanding business owner awards (see http://www.sba.gov/advice.html and http://smallbusinesssuccess.sba.gov/best.html).

It will be observed that each comment in Table 1 has a direct connection to quality, and a content analysis of all tips shows that quality is one of the main themes woven into the success stories. In summary, quality is perceived to be an important factor in the survival and growth of a business (Ehresman, 1995).

In the 1980s, the business climate in the United States was swept by a quality revolution. On the heels of that revolution came the term “six-sigma quality,” which became well-known after the highly successful quality improvement efforts at Motorola Inc.. This new term was used to describe manufacturing processes of highly robust quality, that is, product quality characteristics exhibiting extremely low variability compared with the tolerance limits for the product. It was realized that the development of this capability would ensure both dependable quality and the efficiency of the manufacturing processes. With a real example at Motorola, this idea has led other large companies to undertake similar efforts.

The objective of this paper is to discuss six sigma as a concept which is applicable not only in large organizations, but small, growing ones as well. Here, we use six-sigma as a metaphor for excellence-oriented management thought rather than the specific details of its use in particular processes. Even without the technical details, we argue, six-sigma thinking is a useful concept for robust management; one that is able to confront the challenges of a highly competitive environment. These challenges include the rapid spread of technology, immediate and worldwide accessibility information, growth of electronic commerce, and the continual downward pressure on prices and profit margins.

In the next section, we begin our discussion with a brief summary of the basic idea of six-sigma quality. We then turn to the question of whether six-sigma quality is a luxury or a necessity in the growth of a small business. This question is relevant because attainment of very high quality requires managerial commitment as well as the commitment of scarce resources at the outset. We then describe a managerial pathway toward the adoption of six-sigma quality thinking in small businesses. Our conclusion is that six-sigma quality thinking is not a luxury for a small business, but a genesis for its growth and long-term survival.

WHAT IS SIX-SIGMA QUALITY THINKING?

The term sigma, borrowed from statistics, is a measure indicating the “typical deviation” from the average of a measured characteristic. A six-sigma process has very low variability compared with the tolerance limits established for a measured quality characteristic, as illustrated in Figure 1. The bell-shaped curve in the middle represents the variability in the measured characteristic, with its mean located exactly at the center of the tolerance range. Relative to the tolerance range, variation in the quality characteristic is so small that, even if the process mean varies up or down by as much as 1.5 standard deviations (dotted curves in Figure 1), no more than 3.4 per million items

101

http://www.sba.gov/advice.html


Table 1. Excerpts from the best business tip provided by SBA small business winners.

-

“Trust your own talents/instincts, and surround yourself with good people.”

“Hire the very best professionals you can find, i.e., CPA. Lawyer, computer consultant, whatever.

“If one’s guiding light is simply doing what you say you are going to do ‘at all costs and whatever it takes’ you will not be surprised at the end of the day when your accounts/customers/clients reward you with the realizations of your expectations.”

“Take the time to listen to people. Thank people and care about them.”

“Important as it is to choose the right business, create the right strategy and develop the appropriate support systems, it is even more paramount to get the right people and liberate them to perform at their full potential.”

“Quality goods and service is the most important part of business.”

“Follow up and make sure it is being done, whatever IT is.”

“Treat your customers and employees the way you want to be treated.”

“ALWAYS remember the customer first, even if financially it may not be the wisest.”

“Do not put profits before customer and employee relations.”

“A successful business is based upon all people within the organization working towards a common goal.”

“Give your customers what they want, not what you want them to have.”

“Identify and document your core competence. Ensure each employee understands your core competence.”

“Have excellent staff and give them the freedom to do their job.”

“From the beginning to end, do it right; otherwise all your efforts will be worthless.”

“If you don’t inspect it, don’t expect it.”

“Build your business on a solid foundation of customer focus and employee esprit de corps. Always respect your customers’ freedom of choice and value their loyalty above all else.”

“Keep the quality of your product and your relationships high. People respond well to something that is well made, well designed ... and to being treated with respect.”

“Find the best people for the job and compensate them well.”

“Surround yourself with the best people you can find.”

“Employees are the key to making your business successful, yet they are also the most expensive. So choose them wisely, and then treat them with respect and caring.”

“Never become arrogant. You can always improve upon the quality of service you provide to your customers.”

“Build in quality from day one. Make a commitment to continuous improvement and respect for others.”

102

SUMMER 1999

would fall outside the tolerance limits. If the process average is maintained exactly at the center (solid curve in the middle), there would be only 2 defectives per billion items (these numbers are for a single quality characteristic with a normal distribution). A six-sigma process is therefore capable of yielding virtually no defects, even though it is not free of variability.

To a small business that has a relatively small number of customers, six-sigma quality thinking is tantamount to the goal of not losing even a single customer because of bad quality. Striving towards this goal is an excellent guide for ensuring customer satisfaction and loyalty, even though it may not be entirely practical. This is because small businesses must build their reputation mainly by word of mouth, one customer at a time, rather than by mass advertising or other means available to large businesses. Though it may seem ironic, pursuing this goal can also be less expensive in the long run than compromising it. This is because six-sigma quality can only be achieved by improving the capability of production processes. This in turn leads to improved efficiency, elimination of waste, and reduction of costs. Thus, thinking in terms of six-sigma quality is desirable for small as well as large businesses. We shall further argue that achieving six-sigma quality can in fact be easier in small businesses.

103


In a recent investigation, Ahire and Golhar (1996), conducted a survey of 499 plant managers in 275 small and 224 large businesses in the motor vehicle parts industry. For each business, whether it had implemented a quality management program was recorded, and firms with fewer than 250 employees were classified as small businesses. Comparisons were made along the following quality dimensions: (1) Top management involvement, (2) Customer focus, (3) Supplies quality management, (4) Design quality management, (5) Benchmarking, (6) Statistical process control (SPC) usage, (7) Internal quality information usage, (8) Employee involvement, (9) Employee training, and (10) Employee empowerment. Comparisons based on the implementation of a quality management program showed significant differences in most of the dimensions, as well as in product quality, but comparisons based on size did not show significant differences. Differences that did show up were in SPC usage and employee involvement, and they actually favored small firms which had higher scores. Thus, the practice of quality management or attainment of high quality is apparently not constrained by size. Though small firms are limited in access to capital, market clout and other resources, they can nevertheless achieve high quality just as well as large firms. Small firms have better potential for high employee involvement, multi-functional and versatile employee skills, and encouragement of employee innovation, all of which contribute to high quality (Sonfield, 1984).

In another recent survey, Wijewardena and Cooray (1995) collected data from 52 small but growing manufacturing firms in Japan. Their focus was on the factors related to sales growth (measured over a ten-year period), and considered such factors as the age of the firm, size, advertising, research and development expenses, capital intensity, export orientation, market competition, and type of industry. Except for a type of industry variable, the only other significant growth factor was the ratio of skilled workers to total employees, a major determinant of quality. In a different context, Kaldenberg and Gobeli (1995) found that business outcomes such as increased revenue, lower cost, attracting new customers, etc. are positively correlated with quality management practices in dental services.

The primary means to achieving six-sigma quality is to eliminate the causes of quality problems before they lead to defects. Once a defect occurs, it can consume additional resources for repairs, or worse, it can be delivered to a customer without being discovered. If a customer is dissatisfied, the chance of losing that customer increases dramatically, and a small business can hardly afford the loss of any customers. To put it differently, the effect of losing a customer is greater in a small business than a large one. Six-sigma quality aims to prevent this from happening in the first place, and in doing so, it gives a high-quality image to the business. As new customers begin purchasing from a company known for its high quality product or service, market share and revenues also increase. This chain reaction set off by quality improvement is shown in Figure 2 (adapted from Evans and Lindsay, 1996).

In its common usage, six-sigma quality is a technical term based on statistical concepts. Used literally, it would mean that there is practically no chance of a defective

- 104

SUMMER 1999

product or service reaching a customer. Here, we use six-sigma quality as a metaphor for management attitude that aims towards this state of affairs. It would be prudent for management to think of it as a way of minimizing risk, of losing customers or incurring waste that could have been prevented. Without six-sigma quality thinking, any business would be exposed to greater levels of these risks.

It is generally perceived that small entrepreneurial businesses do not internalize total quality practices, and six-sigma quality in particular, to the same extent as large companies. This is due to the notion that very high quality is expensive to achieve. We argue to the contrary that the management of a growing business cannot afford not to pursue six-sigma thinking. At the root of the argument is the observation that, almost without exception, small businesses that thrive also achieve high quality and customer satisfaction. Indeed, some important quality innovations have been pioneered by successful small businesses, such as providing personal service to each individual customer, and they have later been adopted by large organizations. Another example is employee involvement and empowerment which are characteristics of a well-managed small business, and which were later adopted by large businesses in the form of cross-functional teams and quality circles.

105


Figure 2. Effect of quality on the creation of value (adapted from Evans and Lindsay, 1996)

It is worthwhile to re-emphasize that quality by itself does not ensure success, but that it is a necessary success factor in the long run. Businesses can exist and grow temporarily without focusing on high quality, provided that they offer unique products or services that have no competitors. In a competitive environment, such unique advantages can disappear quickly with the entry of competitors into the market, or the appearance of substitute products. In this context, Murphy’s law may state, “You show me a profitable business, and I will show you a competitor.” Today, the anticipated duration of almost any competitive advantage is rather short, and it is getting shorter with the spread of technology, electronic commerce, and rapid worldwide availability of information.

-

Improvement in quality of

product or service

Higher profitability and value

Increased market share

Higher or stable revenues

Lower production and service costs

Higher prices or higher

resistance to price cutting

Higher perceived value

106

SUMMER 1999

Quality At The Start

Most small businesses are born with one or two people as their founders who may also constitute their entire workforce. The founders are often very skilled in what they do, and usually, this is why the business is born in the first place. They do most of the work themselves, and when they hire new employees, they keep a close watch over what they do and how they do it. With fewer people making the products or providing the services, there are fewer sources of variability. Thus, high quality is easier to achieve in a very small startup business, and without requiring dedicated systems or specific resources for quality control.

In new small businesses, the founders are driven by a desire to succeed and to survive, rather than by quarterly rates of return as it is often the case in large companies. This leads to a constancy of purpose and commitment to quality that may not be true of the management of a large organization. Small businesses also have close relationships with their customers, neighborhoods and communities. They contribute directly to the economy in their local communities, and to the improvement of social welfare. All of these factors are conducive to the pursuit and attainment of high quality and customer satisfaction, and a resulting minimization of the “cost to society” that is considered to be a measure of quality by Taguchi (1986). Thus, high quality is usually a given for startup businesses.

Quality Can Become Elusive With Growth

A growing business necessarily becomes more complex as it grows. Certain tasks that the founders performed easily and skillfully must now be performed by others. As more people are hired, it also becomes necessary to delegate responsibilities. New people have to learn what they are supposed to do, sometimes down to the smallest details. The founders’ knowledge and skills are no longer sufficient, and other people must acquire some of these skills. It therefore becomes necessary to train people to perform their duties, and to do this adequately, it is necessary to define, describe, and document the procedures to be followed. These can be daunting tasks for unprepared business owners, who find themselves in need of new organizational and managerial skills. Unless the founders are prepared for the increased complexity that comes with growth, the business can stumble or fall at this stage.

A case in point was recently reported by Yilmaz and Chatterjee (1997). The small company in question was a contract software developer for other companies. The founder of the company was a good programmer himself, and successfully developed several software products. As the demand for his software began to increase, he began hiring other programmers to help with development. Each new programmer would be given the assignment of developing some part of the project, along with a deadline for completion. There were no established procedures for programmers to follow or standards for testing

107


the programs, and programmers worked independently without knowing what other programmers were doing or how their work would fit together. Trial-and-error (or code-and-fix) was the common approach used by programmers. No project data was collected during development, and when completed modules were tested together, they would often fail to work as a whole. A great deal of effort was spent fixing bugs, and getting parts to work together. Deliveries were delayed and the quality of what was delivered frequently did not meet contract specifications. The business fell into disarray, and it began to lose customers as well as employees. The business never achieved the growth potential it had at the beginning.

Costs of Quality

Achieving high quality in a growing business entails added costs because of the necessity of defining and documenting procedures, training, record keeping, data collection, etc. The standard manual published by the American Society for Quality Control (ASQC, 1971) explicitly identifies several categories of costs associated with the pursuit of quality. They include costs incurred for the prevention of defects, appraisal of quality in production, internal failure of defects produced, and external failure resulting from defects that reach the customers, such as warranty repairs or lost future sales.

It is a common perception that small, growing businesses cannot afford these added costs, at least not as easily as large companies can. While this seems true on the surface, it reflects a lack of understanding about the growth process and the added complexity that accompanies it. Costs will rise as the business grows, simply because there will be more and new things to do, and new systems will be necessary to support these activities. The need for these systems and their costs must be anticipated and planned for. When viewed in this light, there are no unexpected additional costs in the pursuit of quality, only the costs associated with a larger business.

In the process of growth, a small business becomes a more complex organization, or even more generally, a system of interrelated subsystems. This is a new way of thinking for most entrepreneurs. Instead of focusing on individuals, it will become necessary to think of subsystems and their functions, such as production, marketing, accounting, human resource management, etc., and their relationships with each other. How will these functions be carried out? What methods and procedures will be used for each? How will the quality of outputs be evaluated? What data will be collected? How will defects be detected, and how will they be handled?

- 108

SUMMER 1999

Ironically, questions like these are more difficult and costlier to resolve in large organizations than in small businesses which have the advantages of lower complexity higher potential for employee involvement, and collaboration. Failure to utilize these advantages at the growth stage can impede growth, and the issues can become even more difficult and costly to tackle later on.

ROADWAY TO SIX-SIGMA QUALITY AND GROWTH

It is natural that, in the pursuit of achieving high quality and growth, businesses differ from one another in various ways, but they also share some common characteristics. One of the shared aspects is the importance of an overarching management view, for which the theory of W. Edwards Deming is eminently appropriate, and another is a set of management tools that have been used successfully by many growing businesses.

Deming’s Theory

In his last book titled The New Economics (1994), quality pioneer W. Edwards Deming emphatically argued that success in the world of modern business requires a management transformation to quality in all aspects of a business. His term for the means to this transformation was “the system of profound knowledge” which must be acquired and adopted by management. This system consists of four interrelated parts:

Appreciation for a system; Knowledge about variation; Theory of knowledge; Psychology.

First in the system of profound knowledge is the recognition that an organization is a system of interdependent subsystems that work together to accomplish the system’s aim. Subsystems include logistics, production, marketing and sales, customer service, as well as administrative subsystems such as accounting, finance, and human resource management. The system’s aim is to stay in business and create value for owners, employees, customers, suppliers, community, and the environment. As Deming stated in the first his fourteen points, “constancy of purpose: to stay in business ever and forever ... to create jobs, and more jobs…”

The second part is the recognition that there is always variability in any process: among people, in service, in product. It is important for management to learn about the reasons for variability in a business process, develop measures to quantify variability whenever possible, and to use these measures to track performance. Strict deadlines, numeric goals, or quotas are not consistent with a thorough appreciation of variation, even if they are based on historical data. It is better to work on reducing variability to produce the desired results. Rather than numeric goals or quotas, Deming believed that

109


management must provide the methods and training that are needed to achieve those results. As desirable as it is to hire good people, it is just as important train and develop them, and to create an atmosphere that allows them to perform at a high level. Penalties for failing to meet quotas, or monetary rewards for a exceeding them simply avoid the fundamental question, “How is this person supposed to do what he/she is asked to do?” or as Deming put it, “By what method?”

The third part, theory of knowledge, entails a management vision of how the business will “get there from here.” It involves a predictions of the future of the business based on information about the past, current observations, and an insight or vision that links them to the future. By itself, information is not knowledge; a dictionary contains a lot of information but not knowledge. Knowledge requires a theory which can explain the past and allows predictions of the future, even though the predictions may not be accurate (due to chance variability, for example). Management of growth requires a theory of knowledge beyond ambition and entrepreneurship. To gain that knowledge, founders and managers need to continually educate themselves.

Clearly, knowledge about variation and its sources is an essential part of a theory of knowledge. A bell-shaped curve to represent the distribution of a measured quality characteristic is a theory about the process that generates the characteristic. Such a curve allows us to predict the proportion of defects that can be expected to occur if the factors influencing the system remain stable. Theory of knowledge requires management to study and learn about those factors.

The fourth part, psychology, helps management understand people. As individuals, people are naturally different from one another. They learn in different ways, they have different sources of motivation, and they also work in different ways in doing a given task. Management must be aware of people’s differences, and use this awareness to optimize their abilities and inclinations. Performance ratings of people according to fixed numeric criteria are not consistent with psychology or knowledge about variation. A common pitfall for managers is to reward someone (including themselves) who performed above a numeric goal, or to chastise someone who fell short. This is indicative of a failure to understand variability as well as psychology.

TOOLS FOR SIX-SIGMA QUALITY

A number of tools are available to the management of diverse businesses in leading their quality efforts. Without attempting to give a comprehensive discussion or a mere listing of these tools, we mention a few conceptual tools that can help small business management in leading the attainment of high quality. Our purpose in doing so is to raise management’s awareness of quality issues and methods, rather than to provide expositions of the tools. Ehresman (1995) provides a detailed manual for the various methods small businesses can use in implementing a total quality management program. Voehl, Jackson, and Ashton (1994) provide guidelines for small to mid-size businesses to

- 110

SUMMER 1999

obtain ISO 9000 certification, an internationally recognized symbol of the ability of a business to manage and achieve quality.

Two basic aspects of quality are the quality of design and the quality of conformance to design. Clearly, building quality into the products or services at the design stage can improve the likelihood of satisfying customer requirements, result in fewer changes in design, and reduce or eliminate defectives. For manufacturing processes, the idea of introducing quality at the design stage was developed by Taguchi for ensuring what he called robust quality (Taguchi and Clausing, 1990). This idea has since been adopted for a variety of products as well as services, and it culminated in the development of the Quality Function Deployment (QFD) method that has been used successfully by many organizations (Sullivan, 1986). This idea is also the foundation of the “house of quality” matrix developed by Hauser and Clausing (1988). In both developments, the basic approach to a high-quality design is to begin with the solicitation of customer requirements, to analyze and prioritize these requirements, and finally, to translate them to design specifications. The end result of this process is a set of measurable technical product specifications as well as their priorities. Involvement of the customer in the design stage ensures that the ultimate product or service is what the customer wants, or as stated in one of the tips in Table 1, “Give your customers what they want, not what you want them to have.” Although many small businesses do not have the resources to hire experts to do this, help may be available from a variety of sources such as the SBA. An interesting business-government-university collaboration in the state of New Hampshire is described by Gaudard, Schoof, and Paterno (1996).

QFD is an explication of the “Plan” phase in the famous Plan-Do-Check-Act (PDCA) cycle. The PDCA cycle was originally proposed by Shewhart in the 1930’s, but it was popularized by Deming in the 1950s in Japan (where it became known as the Deming Cycle, and later exported back to the West). It is a never-ending cycle of continuous improvement consisting of planning, doing (experimenting with the plans), checking (studying the results of experimentation), and acting (on the results of the previous steps). This is shown in Figure 3 as a ball which is being rolled uphill and raising the level of quality in the process (Deming, 1982). The energy and direction for rolling the ball uphill must be provided by management, and six-sigma thinking can provide the impetus in this effort.

111


Failure Modes and Effects Analysis (FMEA) is a useful technique that can be utilized to identify high-risk failure mechanisms a business process can experience, together with possible “counter-measures” to mitigate those risks. For some key internal business processes, common failure modes and counter-measures are given in Table 2 for illustration. Used with FMEA, six-sigma thinking can thus lead to more dependable, reasonably error-free processes. It is noted that a great majority of the counter-measures given in Table 2 involve the collection and tracking of accurate data about the process in each case. Along with the quality of performance in a process, this brings forth the importance of data collection, and the necessity of a computerized information system. Even in a low-technology business, use of an adequate information system is essential for growth.

While it is desirable for management to study sophisticated quality tools and practices, some of the most useful tools are relatively simple and inexpensive to implement. After the sale of a product or service, for example, it is easy to obtain feedback from customers concerning their satisfaction, suggestions, or complaints (e.g., a brief postage-free questionnaire included with product shipments). Responses to these questionnaires can be very informative since the customers have fresh memories of their

- 112

SUMMER 1999

experiences. It should be noted, however, that customer feedback can be useful only if the management is willing to track and act upon the opinions received. Similar to the lawyers’ rule for examining witnesses at a trial, “do not ask a question if you do not know the answer,” it can be said in the business context, “do not ask a question if you will not do anything about it.”

Table 2. Failure mode and effects analysis (FMEA) for some key internal business processes.

Process Failure Modes Effects Counter MeasuresCosting Inaccurate cost data

Inaccurate cost allocation

Lack of timeliness

Inaccurate cost estimates

Cost information not available

Inaccurate report of income

Collect and track accurate cost data

Activity based cost allocation

Pricing Wrong pricing decisions

Wrong timing

Lost sales Customer

dissatisfaction

Collect and track accurate cost data

Track market price trends

Budgeting Wrong budgets Lack of timeliness

Loss of control Collect and track of budget data

Accounting Needed data not collected

Inaccuracies in data collected

Inaccurate report of income

Inaccurate report of cash flow

Improve accuracy of data collection and entry

Improve accounting methods

Capital Planning

Lack of need assessment

Lack of planning

Impact on current profitability

Impact on current profitability

Establish criteria for capital decisions

Collect and analyze needs- related data (e.g. capacity usage)

Payroll Inaccurate payroll data

Mismatching of wages with work

Employee dissatisfaction

Impact on shareholder concerns

Wage/salary data collection

Track and review wages and salaries

Accounts Receivable

Invoicing errors Lack of timeliness

Customer dissatisfaction

Impact on cash flow

Collect and track accuracy of invoicing data

Accounts Payable

Payment errors Lack of timeliness

Supplier/vendor dissatisfaction

Collect and track accuracy of voucher payments

Inventory Inaccurate data on inventory levels

Errors in inventory valuation

Overstocks or stockouts;

Impact on profitability

Collect and track inventory data

113


CONCLUSION

The main focus of this paper has been to argue how six-sigma thinking can be a genesis for growth for small businesses. Our argument is that the adoption of six-sigma thinking can enhance the chance of success through improved customer satisfaction and efficient use of resources. While six-sigma thinking does not ensure long-term success and growth, there is no single controllable factor that can guarantee these outcomes.

Clearly, a small business can fail for a variety of reasons other than lack of quality. According to SBA statistics, as many as one out of four new businesses cease to exist within the first two years, and a vast majority of these are failures rather than buyouts by other firms. Common reasons for failure include lack of adequate financing, lack of demand for the products, and inability to hire and keep people with the necessary skills. Some family owned businesses fall apart simply because of infighting among family members. Unlike the pursuit of quality, such failures are often beyond the control of managers. Pursuit of quality is a controllable way of reducing the risk of failure and enhancing the prospect for growth.

The essence of six-sigma quality is the reduction of variability in the outputs of business activities. To appreciate this, management must anticipate the inevitable variability in all kinds of business processes, and be committed to study, understand, and control it as much as possible. Contrary to common thought, small businesses are as able to achieve high quality as large businesses. While large organizations have advantages in terms of access to resources and mass marketing, small businesses have advantages of their own in terms of “people” factors and lower complexity.

More than monetary resources, the pursuit if six-sigma quality requires an attitude favoring long-term commitment to quality. Costs range from insignificant to none when the gains and losses associated with quality are taken into account, and some even feel that it is completely free (Crosby, 1976). In an era of unprecedented accessibility and immediate availability of information in the marketplace, management’s commitment to quality is a basic necessity for the survival of any business. The system of profound knowledge combined with six-sigma quality thinking can help the management of a small business as a rudder on the path to growth.

- 114

SUMMER 1999

REFERENCES

Ahire, S. L. and Golhar, D. Y. (1996). “Quality Management in Large vs Small Firms- An Empirical Investigation.” Journal of Small Business Management, 34 (2), 1-13.

ASQC (1971). Quality Costs – What and How (2nd ed.). Milwaukee, WI: American Society for Quality Control.

Crosby, P. B. (1976). Quality Is Free. New York, NY: McGraw-Hill.Deming, W. E. (1982). Out of the Crisis. Cambridge, MA: Massachusetts Institute of

Technology, Center for Advanced Engineering Study.Deming, W. E. (1994). The New Economics. Cambridge, MA: Massachusetts Institute of

Technology, Center for Advanced Engineering Study.Ehresman, T. (1995). Small Business Success Through TQM. Milwaukee, WI: ASQC

Quality Press.Evans, J. R. and Lindsay, W. M. (1996). The Management and Control of Quality (3rd ed.).

Minneapolis/St.Paul, MN: West Publishing.Gaudard, M, Schoof, J. and Paterno, J. J. (1996). “Small Companies Learn How to Design in

Quality.” Quality Progress, 29(2), 51-56.Hauser, J. R. and Clausing, D. (1988). “The House of Quality.” Harvard Business Review,

66(3), 63-73.Garvin, D. A. (1988). Managing Quality. New York, NY: The Free Press.Kaldenberg, D. O. and Gobeli, D. H. (1995). “Total Quality Management Practices and

Business Outcomes: Evidence from Dental Practices.” Journal of Small Business Management, 33(1), 21-33.

Sonfield, M. (1984). “Can Japanese Management Techniques Be Applied to American Small Businesses?” Journal of Small Business Management, 22(3), 18-23.

Sullivan, L. P. (1986). “Quality Function Deployment.” Quality Progress, 19(6), 39-50.Taguchi, G. (1986). Introduction to Quality Engineering–Designing Quality into Products

and Processes. Tokyo: Asian Productivity Organization.Taguchi, G. and Clausing, D. (1990). “Robust Quality.” Harvard Business Review, 68(1),

65-75.Voehl, F., Jackson, P. and Ashton, D. (1994). ISO 9000: An Implementation Guide for

Small to Mid-Sized Businesses. Delray Beach, FL: St. Lucie Press.Wijewardena, H. and Cooray, S. (1995). “Determinants of Growth in Small Manufacturing

Firms: The Japanese Experience.” Journal of Small Business Management, 33(4), 87-92.

Yilmaz, M. R. and Chatterjee, S. (1997). “Deming and the Quality of Software Development.” Business Horizons, 40(6), 51-58.

i

115

Date post:	24-Nov-2020
Category:	Documents
Upload:	others
View:	0 times
Download:	0 times

Fall 1998jbm.johogo.com/pdf/volume/0602/JBM-0602.doc · Web viewPublished jointly by the Western...

Documents