L E A D E R S H I PSTERNbusiness

S P R I N G / S U M M E R 2 0 0 5

L E A D E R S H I P

Creative Thinking:Entrepreneurship in the Digital Age

Corporate leaders speak: Larry Bossidy, Robert Rubin, Steven FlorioB i d d i n g f o r R e p u t a t i o n s o n E B a y A u t o m a t i o n ’ s N e x t W a v eW h y I n f o r m a l S t a t u s M a t t e r s F o r g e t A b o u t C o m m a n d a n d C o n t r o l

a l e t t e r f r o m t h e deanWhat does it mean

to lead? It’s a question

all executives – Wall

Street managing direc-

tors, retail store man-

agers, even business

school deans – must

answer.

To a degree, leader-

ship is a quality that is

demonstrated or learned through experience – not

taught. And that’s part of the reason we bring so

many leaders into our campus. Thanks to our loca-

tion in New York City, our faculty and students

have myriad opportunities to interact with a wide

range of people who have led large organizations.

Former Treasury Secretary Robert Rubin, ex-Allied

Signal CEO-turned best-selling author Larry

Bossidy, and Steve Florio, the former head of

Condé Nast Publications, all of whom are featured

in this issue of Sternbusiness, are just a few of the

many dynamic leaders who visited with us in

Washington Square last fall.

Institutions can lead, too. Management depart-

ments lie at the core of every business school. At

Stern, our robust management department is

distinguished by its particular focus on the behav-

ioral sciences. In two of the articles in this issue,

members of our faculty delve into concepts drawn

from psychology and sociology to highlight more

effective ways of managing and leading companies.

More broadly, business schools lead by carrying

out their core mission: sponsoring and conducting

innovative research, creating knowledge, and dis-

seminating knowledge to wide audiences. Moreso

than those in many other disciplines, scholars of

economics, finance, information systems, market-

ing communications, and management expect their

research to do work in the world. Stern faculty

lead by pushing knowledge beyond the confines of

our buildings. They offer expert opinion, consult to

businesses, testify before Congress, appear in the

media, and publish far and wide – all as part of an

overarching effort to place the insights they’ve

gleaned into the hands of others.

Part of our mission involves helping to develop

the next generation of business leaders – our

students. Ultimately, of course, business leaders

are forged in the workplace. But an excellent

business education that inculcates a grasp of the

fundamentals, an appreciation of the complexities

and challenges presented by the global workplace,

and a capacity for critical thinking is the sine qua

non for any leader.

These efforts lie at the core of what we do at

Stern. And I think you’ll find that they are embod-

ied in this issue of Sternbusiness.

Thomas F. CooleyDean

SternChiefExecutiveSeries

InterviewLarry Bossidy, former chairman and CEO HoneywellInternational and AlliedSignal Page 2

6 LeadingIndicatorsby Daniel Gross

Illustrations by:Ken OrvidasGordon StuderJuliette Borda Dave CutlerMichael CaswellRobert O’Hair

STERNbusinessA publication of the Stern School of Business, New York University

President, New York University � John E. SextonDean, Stern School of Business � Thomas F. CooleyChairman, Board of Overseers � William R. BerkleyChairman Emeritus, Board of Overseers � Henry KaufmanAssociate Dean, Marketing

and External Relations � Joanne HvalaEditor, STERNbusiness � Daniel GrossProject Manager � Lisette ZarnowskiDesign � Esposite Graphics

Letters to the Editor may be sent to:NYU Stern School of BusinessOffice of Public Affairs44 West Fourth Street, Suite 10-160New York, NY 10012www.stern.nyu.edu

contents S P R I N G / S U M M E R 2 0 0 5 L e a d e r s h i p

32 Command &ControlA better way to encourage employee cooperationby Steven L. Blader and Tom R. Tyler

38Hot Off the PressManaging media in a crisisby Irv Schenkler

44 Endpaperby Daniel Gross

8 Uncertain HoursA conversation with former Treasury Secretary Robert Rubin

12Media MessagesSTERNbusiness interviews Steven Florio, former president and CEO of Condé Nast Publications

14R-E-S-P-E-C-TInformal status can matter

as much as a job titleby Sandra E. Spataro and Cameron Anderson

20 Live Auction HeroesHow reputations are made on eBayby Luís Cabral and Ali Hortasçu

26Automation’s Next WaveOutsourcing may be just getting startedby Alexander Tuzhilin

GC: Larry, your book is called

Confronting Reality. Great idea,

but it's always been a great

idea, right? So what is the rea-

son for writing a book like this

now?

LB: We've always been asked

to confront reality. But the price

for not confronting reality is a lot

higher now because of three

mega-events. First, because

globalization has brought a lot

of excess capacity in some

industries, pricing has become

more difficult. Margins have

been compressed and the com-

moditization of products occurs

much faster now. That has

made business more competi-

tive. The second is that there's

been an enormous over-exten-

sion of credit. Usually, compa-

nies fail and go out of business

and it restores the supply-

demand balance. Now compa-

nies go bankrupt but they don't

go out of business, so the bal-

ance never gets restored. And

third, the arrival of mega-retail-

ers like Wal-Mart and Lowe's

creates a lot of disruption. So

the point of Confronting Reality

is that you’ve got to know where

you are, and if you wait too

long, it might be too late.

GC: This notion would seem to

apply equally to people who are

managing their careers. Aren’t

most people’s jobs threatened

by globalization today?

LB: Jobs will go to low-cost

locations, to the extent that's the

most efficient way to get them

done. People have to think

about how they can make a

competitive difference. Can

you, for example, do something

in information technology or in

science? And the way you keep

jobs in the United States is to

continue to pioneer things that

Larry Bossidy has led three Fortune 100 companies. After graduating from Colgate University in 1957 with aB.A. in Economics, he joined General Electric. In a 34-year career at GE, he served in a number of execu-tive and financial positions, and was named vice chairman and executive officer of General Electric Companyin 1984. In 1991, he became CEO of manufacturer AlliedSignal and engineered a transformation. AfterAlliedSignal and Honeywell merged in 1999, he became chairman of Honeywell International. He retired inApril 2000 but returned in July 2001 as chairman and CEO to stabilize the company following GeneralElectric's unsuccessful attempt to acquire Honeywell. Since stepping down from both positions in June 2002,he has spent time consulting and writing. Both his books, Execution: The Discipline of Getting Things Done(Crown Business, 2002) and Confronting Reality: Doing What Matters to Get Things Right (Crown Business,2004), co-written with consultant Ram Sharan, have been best-sellers. He serves on the Boards of J.P.Morgan Chase, Merck & Co. and Berkshire Hills Bancorp.

sternChiefExecutiveseries

Larry Bossidyformer chairman and ceo

Honeywell International andAlliedSignaland former vice chairman and executive officer

General Electric Company

2 Sternbusiness

are new to the world. We've

generated more jobs in this

country than anybody in the

history of the world, and we'll

continue to. But they'll be differ-

ent jobs.

GC: You mentioned a lot of

companies go bankrupt nowa-

days, but they don't go away.

And one can't help thinking

about the airline industry. You've

got some very pointed things to

say about the airline industry, as

well as a few other industries.

LB: There's a number of indus-

tries we say in the book that are

structurally defective: airlines,

steel, rubber and commodity

chemicals. If you're a big airline

– United, American, Delta – it

isn't clear how you're going to

compete with Northwest, or Jet

Blue. Delta got an enormous

wage concession from their

pilots just the other day, 32 per-

cent. They still are way over

JetBlue in terms of cost. You

look at the amounts General

Motors pays in health care and

pensions – that’s three thou-

sand dollars a car. The point is,

you've got to stand up to these

issues sooner, when you have

options. The longer you wait,

the fewer options you have.

GC: We can say they should

face reality, but it's still hard to

imagine what General Motors

can actually do?

LB: It isn't clear to me. We say

in the book that at some point in

time the government might

decide that it's in the nation's

best interests to have at least

one automotive manufacturer, or

maybe two. And they might be

able to help create a securities

offering that will take General

Motors over the hump. These

legacy costs do subside over

time. It seems to me that in the

absence of that, these compa-

nies are going to be in further

disrepair.

GC: You know, a big part of

what you did at AlliedSignal,

and at General Electric, was

evaluating people. You must

have formed some opinions

about what characterizes the

winners above all else?

LB: You learn to be humble

because you make mistakes. I

can remember promoting peo-

ple to a certain level, and being

concerned about the promotion,

and then seeing them blossom.

And I can also remember plac-

ing people in higher responsibili-

ties with a sure fire conviction

that they would succeed and

they didn’t. Executives at every

level have to continue to grow. I

always say that CEOs either

grow or they swell. You want to

stay away from the ones who

swell. I’ve got to continue to be

interested in education. I've got

to have a broader set of inter-

ests than just my job, because I

become a significant dullard if

that's all I do. And I've got to

expand my intellect in a way

that makes me valuable. At the

end of the day, the most difficult

decision you reach in terms of

who to select is not their intel-

lect or integrity, but how much

more will they grow?

GC: When people get evalua-

tions that aren’t good, and they

have to face that reality, what

should they do?

LB: When you get an apprais-

al that you disagree with, the

question is, who's right, you or

the person giving you the

appraisal? We started what we

call a 360 review about 10 years

ago, and they're quite prevalent

now. You’ve got to make sure

you’re considering the evalua-

tion in a way that allows you to

grow. And you have to get past

your disagreement. You've got

to do something about it. You

just can't accept it. You can say

I would like to have people do a

360 on me, and see what the

viewpoints are. Or that I'd like to

be assigned to another manager

to see what that manager's view

of me would be.

GC: Now what about the job

of the manager in this situation?

LB: When somebody came to

me and said others weren't

doing the job, I’d ask what have

you done to help this person do

the job? A manager's responsi-

bility is not just to hire but it's to

coach, it's to develop, it's to try

to make people better.

GC: When you came to

AlliedSignal, in 1991, it was a

company that needed a lot of

help. When you got in the door,

what needed to be done first?

LB: There was no self confi-

dence. People were disappoint-

ed in their careers. Because

when you don't do well, the

place doesn't expand, and new

jobs don't open up. I asked a lot

of questions; what do you think

we should do? As a conse-

quence of those discussions, we

put together a plan, and I said

I'm going to take this plan to

Wall Street now. Everybody who

finds that their knees are a little

bit weak, stand up, because

after tomorrow, we're going to

do this. And it got people excit-

ed. We told the Street what we

were going to do, and we deliv-

ered on it, and it was great to

see people’s self confidence

improve.

GC: Many managers say they

want to hear the straight unfil-

tered view from the people

working for them. But the peo-

ple might believe, perhaps

because of a corporate culture,

that there is nothing to be

gained and much to be lost by

putting their hands up in the

meeting and speaking honestly.

LB: Well, I think that's one of

the things that a 360 can help. If

people don't trust you, or they

think you have an ulterior

motive, then obviously people

are going to be careful. So it's

your job to make yourself trust-

worthy so you can get at some

of these issues.

Mr. Bossidy was interviewed by Stern alumnus Geoffrey Colvin, who is Senior Editor-at-Large of Fortune and co-anchor of “Wall Street Weekwith Fortune” on PBS.

Sternbusiness 3

“I always say that CEOs either grow or they swell. You want to stay away

from the ones who swell.”

suits because I didn’t have time

to go home before I went to

some ball game and was in the

dust pit. I’m not trying to

impress you about what I did. It

made me think in a broader

dimension than I otherwise

would have. I saw what was

going on in young people's

lives, and you know, I always

say to them, you're the best

thing that happened to me. I

still feel that way; I just feel like

I got so much more from them

than I gave.

GC: The company you were in

most of that time, General

Electric, is famously demanding.

And for 11 years your good

friend Welch was the CEO, and

he was about as demanding as

they come. How did it all work

out with your employer?

LB: I didn't ask for special dis-

pensations. With Welch, if you

GC: And now let's take the

point of view of the employee,

who wants to give his straight

unfiltered views to the manager,

but in fact believes, with good

reason, that that is not a career

enhancing thing to do.

LB: That's a harder question.

This employee can perhaps

express this viewpoint to an HR

person. But I think over time, if

that condition persists, you

ought to get another job. You

can't waste your time in an envi-

ronment that you know isn't the

way it should be, and where

there's no interest in changing it.

If you're not being allowed to

grow and flourish, go to an envi-

ronment where you can.

GC: When you look back on

your years as a CEO, is there

anything you wish somebody

had told you back when you

started?

LB: This question of evaluation

is an important one. When I

consult, I go to CEOs and I ask

to see the appraisals of their

direct reports – the ones I'm

consulting with. And invariably, I

see pages of circumlocution. In

other words, lots of words trying

hard not to say anything. It's

hard for people to think that

appraisals can be a constructive

process. It's supposed to identi-

fy the things you do well, and

it's supposed to identify the

4 Sternbusiness

GC: An issue for everybody

who works is balancing work

and family. Now, you had a

long career where you man-

aged these things, it would

appear, incredibly well. A,

you've been married to the

same person for a very long

time. B, you have nine children.

How did you think about these

issues as you went through

your career?

LB: While I picked a remark-

able, talented woman to

marry, the fact is you have to

make time too. I used to come

into work sometimes at 4:30 in

the morning, because I wanted

to be home for dinner. When

you have nine children, you

don't just interface with them,

you manage them. I used to

post a board as to who was

supposed to do what. I

coached the Little League

baseball team. I ruined more

things you should be doing bet-

ter. And it takes a long time to

get that through an organization.

At AlliedSignal people thought

that if they appraised people

accurately, it somehow will get

in a file and cause that person

everlasting harm. I said, no, this

is not an appraisal. This is the

beginning of a debate. It took a

series of years before it got

down in the organization where

people would finally be honest

enough to put down what they

believed, and then gave that

person a chance both to

improve and recover and go on.

It’s terrible to find a person at

mid-career with a series of defi-

ciencies that have never been

pointed out.

Now as far as myself is con-

cerned, I come from a small

town. I never knew anything

about corporate life. I was very

frank to the point of probably

being caustic. So somebody

finally took me aside, and I was

probably 28, and said, you

know, it isn't what you say but

how you say it, which was a

wonderful comment. And I

thought about it. And I tried to

not be withholding of viewpoints,

but to express them in a way

that was more positive and con-

structive. And some people still,

after having said all that, still

say I'm blunt.

“A manager's responsibility is not just to hire but it's to coach, it's to develop,

it's to try to make people better.”

got done what you had to get

done, if it takes you four hours a

day, you didn't have to stay for

the sake of staying. And he was

helpful and responsive. At

AlliedSignal among the first

things I did was put in a day

care center. We couldn't keep

women who wanted to have

children. Well, we put a 90-per-

son day care facility on site, and

we didn't lose a woman from

our employ for the next three

years.

GC: What was your attitude

towards people who had been

to business school?

LB: Well, first of all, not many

people who graduated in my

class went to business school.

GE had vaunted training pro-

grams, and you were told that

this was the same thing as busi-

ness school, but you get paid. I

thought it was a reasonable

proposition. If you look at our

educational system, you can

make the case that in grades

one through 12, there are sys-

tems that might be better than

ours, including the one in

Japan. But there is no system

that compares with the graduate

education that we have in the

States. I also think business

schools have worked harder to

try to stay contemporary. Ten

years ago business schools

always taught yesterday's war,

instead of trying to fight tomor-

row's war. Today, they’re better

in terms of preparing people for

what they might face.

When I went to Allied, we had

a shortage of talent, and we

hired a lot of MBAs and they

saved my life. We were able to

give them more challenging

positions relatively early in their

career, and it was a major assist

to build the management team

that I was able to build.

GC: CEO pay is an issue that

was big in the headlines when

you got the job, and it's still big

in the headlines today and it

seems that nothing ever

changes despite all the talk

that goes on. What's going on?

LB: If your company does well

and your share price does well,

no one begrudges you to earn

a piece of that progress. On

the other hand, if your compa-

ny doesn't do well, or you get

fired and you leave and you

still get a lot of money, that is

going to be an ever growing

concern and I think it should

be. So, I think it's improved as

a consequence of this uprising.

You'll still see cases where it

gets abused, but by and large

it's better.

Audience Questions:

Q: Could you talk a little bit

about the challenges CEOs face

these days in terms of meeting

Wall Street's expectations in the

short-term, versus taking a long-

term management approach?

LB: You've got to deliver for the

current share owners, and

you've got to plant seeds for the

long-term. If you have people

who are unbalanced, who just

drive for the short-term, there's

a day of consequence. And if

you have a person who does

the opposite and doesn't plant

seeds, they run out of steam as

well. But you can’t get caught

up in what Wall Street thinks. I

mean, somebody comes out

and looks at your company for

20 minutes and tells you what

you’ve got to do, and people lis-

ten to them. I mean, come on. I

always felt that if I got fired, I

want to get fired on my own

mistakes, not on what some-

body told me to do.

Q: You're on the Board of

Merck. It must be very hard for

people without your stature to

be telling CEOs who are lead-

ing companies what to do from

a Board perspective. Are Boards

really strong enough to have an

effect on management?

LB: That's a great question.

You know, Sarbanes-Oxley has

done some good things, and it

has done some bad things too.

But it did call a lot more atten-

tion to the quality of Boards.

And it has made Boards more

introspective and increased the

Board's involvement with the

company. But a Director's role is

not to tell the CEO what to do.

Directors are there to listen to

various strategies and comment

on them in terms of their own

personal experience. So on the

Merck Board, for example, there

Sternbusiness 5

are a number of scientists who

speak very knowledgeably

about the science involved. I

probably speak more knowl-

edgeably about the business

aspects. I’m not trying to tell

the CEO what to do, but to

make sure that there's a good

dialogue that's inclusive. And I

don't think you have to be a

CEO to be a very good Board

member.

Q: A lot of the problems on

Boards seem to arise from

what some people have called

the Boardroom culture, and the

idea of not speaking up if it's

contrary to the prevailing view.

Even Warren Buffett has said

that he has failed to speak up

sometimes in a Boardroom

because it was just sort of

socially or culturally too difficult.

Have you observed this?

LB: I have, but I think one of

the corrections with Sarbanes-

Oxley is that there are now

mandatory Executive Sessions

of the Board, where the Board

convenes in the absence of the

CEO. A lot more gets said that

might not have been said. And

then somebody is appointed to

relate this to the CEO. I think

it's happening a lot less now, in

the presence of these

Executive Sessions. �

“You can't waste your time in an environment that you know isn't the way it should be, and where there's

no interest in changing it.”

n a way, leadership is a lot like hitting a 99-mile-per-hour fastball. It’s comparatively easy to describethe mechanics and tactics relating to the act. Butwhen it comes time for wood to hit leather, all thebooks and practice don’t really matter. Instinct and

natural ability matter more than preparation. A few peoplecan just hit a little white ball with red seams, and the vastmajority of us can’t. It’s tempting to conclude the sameabout leading an organization, a company, a country, or agroup of baseball players. The best leaders frequently seemto be born and not made, and many ofthem followed unorthodox, inimitablepaths to leadership.

Larry Bossidy is the rare personwho can write about and describe cor-porate leadership nearly as well as heactually does it. In his highly activeretirement, Bossidy, who served as CEO of not one butthree different Fortune 500 companies, has written twobest-selling books on leading businesses – Execution: TheDiscipline of Getting Things Done (Crown Business, 2002)and Confronting Reality: Doing What Matters to GetThings Right (Crown Business, 2004).

Last fall, Bossidy visited NYU Stern as part of our long-running CEO Series to discuss the art and science of run-ning large organizations (p. 2). And as he sees it, leader-ship is about more than a title, and hiring and firing.“When somebody came to me and said others weren'tdoing the job, I’d ask what have you done to help this per-son do the job?” he said. “A manager’s responsibility is notjust to hire but it's to coach, it's to develop, it's to try tomake people better.”

Bossidy’s point is backed up in large measure by theconclusions of a study by Sandra E. Spataro and CameronAnderson (“R-E-S-P-E-C-T,” p. 14). They conclude that

there’s a lot more to leadership than a title, which rep-resents formal status within organizations. Formalstatus matters. But informal status – how muchrespect and prestige a person enjoys by virtue of posi-tion, personality, demonstrated skills, and his or herability to connect with others – is a crucial componentof leadership.

One person who certainly enjoys both formal andinformal status is Robert Rubin, former TreasurySecretary, former co-chairman of Goldman Sachs,

and now chairman of the execu-tive committee at Citigroup.And while we expect leaders toproject certainty about theirdecisions and strategies, Rubin,who has worked with some ofthe most storied leaders of

recent decades, from Sandy Weill to PresidentClinton, takes the view that multi-dimensional leadersmust also embrace uncertainty. (“Uncertain Hours,”p. 8) In a conversation with students, Rubin offeredsome typically Rubin-esque thoughts on the burgeon-ing fiscal and trade deficits – it’s likely that “at somepoint, the markets will begin to look forward at theimmense projected deficits, and the markets willbegin to react by demanding sharply higher interestrates for providing long-term debt,” he said. We justdon’t know when.

In a course he taught at Stern last fall, “Leadershipin the Communications Industry” Steven Florio, theformer chief executive officer of Condé NastPublications, showed students a scene from the movie“Patton,” in which the actor George C. Scott rattleshis saber at the Soviet Union. “That scene fromPatton was about identifying who your competitors

6 Sternbusiness

“The best leaders frequently seem to be born and not made,

and many of them followedunorthodox, inimitable paths

to leadership.”

IL E A D I N G

are,” Florio said in an interview (“Media Messages,” p.12) in which he distilled the lessons he taught to – andlearned from – his class.

Patton represents a distinctive and old-fashioned modeof leadership, which relies on managing by commandingand using a series of carrots and sticks (in Patton’s case,mostly sticks) to motivate team members. But in today’sless hierarchical and more complicated workplaces, suchan approach may seem both outmoded and ineffective. Intheir article, (“Command and Control,” p. 32) Steven L.Blader and Tom R. Tyler argue that there’s a better way:Appeal to employees’ intrinsic desire to follow rules byconvincing them of the organization’s legitimacy. “We pre-dict that employees will be intrinsically motivated to fol-low their organization’s rules if they feel that those rulesdevelop from a system that is consistent with their own setof moral values,” they write.

oday, leading a company – especially a publiclyheld one – means dealing with the media.Whether it’s appearing on CNBC to discuss earn-ings or granting interviews to newspaper

reporters, executives must be media savvy – especially intimes of crisis. In an article excerpted from his book,Guide to Media Relations (Prentice Hall), Irv Schenklersays that while playing defense is important, there is a wayto feed the beast without getting bitten (“Hot Off thePress,” p. 38). “Whenever a company can position itsresponse as a meaningful effort to acknowledge and cor-rect the phenomenon that led to the crisis, media coveragewill become more favorable and stakeholder impressionswill in the long run not impugn the company’s reputa-tion,” he writes.

A great deal of press in the past year focused on thedanger of offshore outsourcing – the practice of corpora-tions moving jobs from America to distant locations where

wages are generally lower. Alex Tuzhilin argues that forall the hype surrounding the information-technologyinspired productivity revolution in the last two decades,we ain’t seen nothing yet (“Automation’s Next Wave,” p.26). The next wave of automation will affect not onlyroutine production workers, but also the better-paid andheretofore more secure group – engineers, office andknowledge workers, managers, educators, and othergroups of “mind workers.” Tasks that are “high onrepetitiveness, stability and structuredness – constitutethe primary candidates for automation,” he writes.

There’s one group of mind workers whose jobs arequite secure for the moment: the executives at eBay. Thegiant auction website has been one of the great corporatesuccess stories of the past decade. Millions of buyers andsellers of everything from old baseball cards to usediPods have come to appreciate eBay. And so too have agrowing number of economists, who see the site not as aplace to trade souvenirs but as a vast datamine. One ofeBay’s unique features is the ability of buyers to rate sell-ers, and thus potential bidders a highly public and trans-parent assessment. Luís Cabral and Ali Hortasçu haveexamined these feedback systems to determine how aseller’s reputation affects sales in a theoretically anony-mous marketplace (“Live Auction Heroes,” p. 20). Andit turns out that in the newfangled marketplace of eBay,the old-fashioned virtue of customer service still matters.

At Sternbusiness, we’ve long been the beneficiariesof some other old-fashioned virtues: innovative scholar-ship that challenges conventional wisdom, intelligentlycrafted writing, and attention-grabbing design. Leaders– and those who aspire to lead – will surely find plentyof useful ideas in this issue.

D A N I E L G R O S S is editor of STERNbusiness.

Sternbusiness 7

T

I N D I C A T O R S ILLU

STR

ATIO

N B

Y D

AVE

CU

TLE

R

8 Sternbusiness

Uncertain

HoursFor three decades, Robert Rubin has been at the center of high finance and public pol-

icy. A graduate of Harvard University and Yale Law School, Rubin spent 26 years at

Goldman Sachs, where he ultimately rose to co-chairman. In 1993, he went to

Washington to serve in the Clinton administration, first as director of the White House

National Economic Council and then as U.S. Treasury Secretary from 1995 to 1999.

During his tenure, Rubin was a key player in the debates over deficit-reduction and in

managing international financial crises

that cropped up in Mexico, Russia, and

Asia. Upon leaving Washington, he

joined Citigroup, where he is a director

and chairman of the Executive

Committee. On October 14, 2004, Rubin

appeared at NYU Stern’s Alumni Author

Lecture Series to discuss the economy,

the upcoming elections, and his

approach to decision-making.

Sternbusiness 9

JW: In the 1996 and 2000 elections people thought we had a very

strong economy. This time, you have a candidate arguing that it's

quite strong and another candidate arguing that it's weaker in some

ways. What do you think about both the health of the economy now,

and the prospects for the American economy and the world economy

going forward?

RR: It's been a complicated four years. On the one hand you've had

enormous stimulus, and on the other hand you've had job loss. It's

the first net job loss under any administration since 1932. You've had

declining price-adjusted median incomes in the United States. A cou-

ple of months ago, I was out in the Midwest and I had dinner with the

chief executive officer of one of America's largest companies. And he

said the company was doing well but noted that people are being

very cautious. Companies have an abundance of cash but they're not

spending it on investment or hiring. The reason is that there is a real

uncertainty due to an overhanging set of issues.

I've been involved with economic issues for a long time, and I think

this may be the most critical juncture for our economy in my lifetime.

And while the outlook is always complex and uncertain, I think this is

the most complex and the most uncertain in some number of

decades. That poses enormous difficulties and challenges for policy

The political people in the White House said in a debate between

paying down deficits versus tax cuts, tax cuts will win all the time

because tax cuts are something tangible, and the public simply

doesn’t understand the enormously dangerous long-term impacts

that our deficits can have. The way Clinton framed it was in terms of

protecting Social Security; that's something that resonates

politically. I just don’t think in this campaign there’s a way to

make this resonate in a public that is so underinformed on the

issues that are so critical to their future.

JW: In 1993, after Clinton was elected, when he decided to focus

on the deficit to the harm of some of the other things he had run on,

including a middle class tax cut and a health care plan, it was a very

good illustration of the way that you think probabilistically about

problems. As head of the National Economic Council, your role real-

ly was to set up this decision for Clinton. Explain a little bit to us how

that decision was made.

RR: One of the problems we have today is that we face these huge-

ly complex issues, but they're not being approached with the recog-

nition of that complexity. I know very little about Iraq. And I don't

know whether we should have gone in or not. But I read a book prior

to the invasion of Iraq, about Gertrude Bell, a British Arabist who is

responsible for the current borders of Iraq. It talked about the

Shiites, the Sunnis, the Kurds, the hundreds of years of friction

amongst all these people. When I saw that we were going in, it

seemed to me that there would be a plan that probabilistically took

into account the issues that we might face. Instead you had this

very, with all due respect, simplistic view that we would be wel-

comed. Well, I think the same approach is true for all these things,

because the issues are conceptually the same, even though the cir-

cumstances are different. There is an effort in this book to focus on

decision making. Larry Summers, who was my deputy at the time

and is now President of Harvard, at one point said, “Look, every-

body knows that issues are fundamentally about probabilities. But,

the difference is, when we were there, we had an internalized sense

of this. So when we got to actually

making decisions, it informed all

of our decisions.”

In the case of the deficit

reduction, on January 7th,

1993, during the transition, we

met with President-elect

Clinton. We said if

we don't make a

dramatic change

with respect to

these deficits, in

makers. But it also makes for a very difficult environment for

investors. On the one hand, we have a country with an enormous

comparative advantage in the global economy – our historical

embrace of change, flexible labor markets, and willingness to take

risk. On the other hand, we face hugely consequential and enor-

mously complex challenges and risks. And if we're going to realize

our potential, we have to deal with these challenges effectively. And

if we don't deal with them effectively, and in my judgment we are for

the most part on the wrong track right now, there is a real chance

that we could have a lot of trouble ahead some place.

Three years ago we had enormous projected surpluses. We now

have enormous projected deficits. We have very large current

account and trade deficits. Consumer debt as a percentage of GDP

is at historically high levels. We have an historic challenge, at least

in my judgment, with respect to our competitive position from China

and India, and the very large numbers of now well-educated work-

ers in low-wage environments connected to us by real-time com-

munications. These are challenges we can meet, but to do so we

have to act in a whole host of ways that we are not now acting, with

respect to policy.

JW: Let's start with the risk with which you are the most closely

identified, and most focused on: the deficit. I thought in 1992 Bill

Clinton successfully made a political issue out of the deficit. And it

seems to me that this time around, John Kerry for some reason is

not making an especially big deal out of it. Am I wrong?

RR: Well, just in terms of the numbers, in January 2001, the

Congressional Budget Office projected a $5.6 trillion 10-year sur-

plus. Goldman Sachs and most independent analysts are now pro-

jecting about a $5.5 trillion deficit. So that's actually a deterioration

of about $11 trillion, or $9 trillion after methodological adjustments.

Even President Clinton, with all of his enormous skills, had difficul-

ty communicating about fiscal matters to the American people. It

worked in the 1992 campaign because you had Paul Tsongas and

Ross Perot talking about it. And the American people associated

the tough economic times in some measure with the fiscal condi-

tions of the 1980s that led to a roughly quadrupling of the public

debt. In that context, the deficit had some traction politically.

But once President Clinton got elect-

ed, and he had to turn the concept into

reality, we proposed a program that

passed by two votes in the House. It

was a tie in the Senate, which Vice

President Gore broke. In 1998, we

were beginning to see these

large projected surplus-

es. President Clinton’s

view was that they

should go largely to

paying down the debt.

10 Sternbusiness

Mr Rubin was interviewed by Jacob Weisberg, editor of Slate, and the co-authorof Rubin’s best-selling memoir, In An Uncertain World: Tough Choices from WallStreet to Washington (Random House).

From left to right:Jacob Weisberg, Dean Thomas Cooley,and Robert Rubin.

Sternbusiness 11

our judgment, the probability is very high that any time the econo-

my begins to do well it will get choked off by higher interest rates.

On the other hand, if we do make a dramatic change in fiscal poli-

cy, while we think the probability is high that that will create an inter-

est rate regime low enough, it is also possible that the fiscal con-

traction will be the dominant effect, and there'll never be the confi-

dence you need to have lower interest rates. There are no guaran-

tees, but probabilistically we think this is the right way to go. And he

related to that.

And you'll see this underlying phenomenon on all the policy deci-

sions that we discussed in the book, but also investments. If you

thought with discipline, probabilistically, about markets and about

valuation in 1999, that was not an investment environment in which

one would have participated.

JW: The International Monetary Fund has now issued two different

reports suggesting that the deficit in particular could be a problem

that plays into a global economic crisis of some kind, or at least an

economic crisis in America. What kind of crisis are we talking

about?

RR: Jacob always wanted me to use the word "crisis" in the book,

and we never did. The two most important challenges are geopolit-

ical issues and also these immense fiscal imbalances. I think it's

impossible to predict when it will come. And I think one reason mar-

kets don't reflect that is simply because there's no way to quantify

it, there's no way to fit it into models. It could be six months off, and

it could be six years off. What I think at least is that the probability

of serious economic difficulty is very high. If you have large fiscal

deficits that absorb part of the savings pool, you have less savings

available for private investment, and therefore low rates of produc-

tivity and low rates of growth. That is a long-run problem. But I actu-

ally think there's a much more serious risk. At some point, the mar-

kets will begin to look forward at the immense projected deficits,

and the markets will begin to react by demanding sharply higher

interest rates for providing long-term debt. On the other hand, none

of this may happen. We may muddle through one way or the other.

It's also possible that our political system may rise to the challenge.

JW: When I was watching the second presidential debate, the line

that almost knocked me out of the chair was when President Bush

said that "Bob Rubin says that Kerry's anti-outsourcing plan won't

work." I was under the impression that you had been one of the

people who had helped in some sense put that plan together. So

what was the deal with that?

RR: What I said is that if you look at outsourcing, it is a part of a

larger phenomenon of trade liberalization. I think trade liberalization

contributed substantially to our well-being in the 1990s, and I think

it's the right path going forward. But trade liberalization has to be

intertwined with programs that will deal with those that are dislocat-

ed by trade. And we must have a much more effective program to

promote competitiveness in our economy. We have got to have a

world-class public education system. We've got to invest far more

substantially and effectively in basic research.

Audience Questions:Q: You mentioned the need for more investment in our schools and

the need for more fiscal discipline. Is there a theoretical place at

which taxation reaches a tipping point, where we have so much that

we're taking care of in that regard that we collapse the economy

from the other end?

RR: I think the answer is probably yes. But I don't think we're any-

where remotely near that today. My recollection is that federal rev-

enues are something like 16 percent of GDP and that's the lowest

percentage of GDP since the 1950s.

Q: When you look at the numbers you cited, they're astounding.

And I wonder if you have an opinion on why you don't hear more

from Federal Reserve Board Chairman Alan Greenspan or why

some other prominent economic business leaders aren't out there

talking about potential crises?

RR: Greenspan at various times has expressed great concern

about fiscal matters, even when he was supporting the 2001 tax

cut. Why business leaders don't speak out more is a very interest-

ing question. Back in the mid 1990s, the Business Roundtable was

taking the very strong position on deficits. John Snow, today’s

Treasury Secretary, was at that time chairman of the Business

Roundtable and in favor of fiscal discipline. I meet with a lot of busi-

ness people, and almost always they talk about this as a very seri-

ous problem. But it's longer term, it's out there, it's intangible, it's not

quantifiable, and it’s laden with politics.

Q: There's been a lot of talk and pressure about China and having

them float their currency. And I know eventually that day is coming,

and it could be soon. What do you think the net impact of that would

be on our economy?

RR: I was in China three weeks ago, and met with Premier Wen,

and he talked about the currency. I think China would benefit from

being on a floating exchange rate system, and I think that's ulti-

mately where they will wind up. But they have always had great

concern about trying to minimize the risk of instability in a country

of 1.3 billion people. My instinct is that they will continue to very

heavily weight stability and until their financial system is in materi-

ally better shape than it is today, I suspect they're going to lag on

moving ahead with such matters as exchange rate liberalization. �

“While the outlook is always complexand uncertain, I think this is the mostcomplex and the most uncertain in somenumber of decades.”

M E D I AMESSAGESSB: What does it mean to lead a company?SF: The only way you can really drive a company in any industry isthrough effective leadership, not just being the boss. And, that startswith top-line revenue. You can have brilliant financial people, butunless you have someone who understands how to build the top-line,it won’t go anywhere. In the case of Condé Nast, it was putting themagazines editorial up front. Look at the quality of Vogue. It’s about[editor] Anna Wintour and her staff. At the end of the day, she is ademanding, tough boss. She will not compromise the quality of themagazine that she was charged with. That’s leadership. Then thepublisher takes the product and sells it to advertisers and readers.

SB: Demanding bosses often get tagged in the media with certainlabels.SF: When I did it as CEO I was, "a street tough Italian guy fromQueens who rose to the top, an effective charismatic leader." Whena woman does it, people write novels about her, like The Devil WearsPrada. Martha Stewart, who is a good friend of mine, is anotherexample of that. She’s no less tough on driving her business than Iwas. They label her with unflattering labels too.

SB: You showed a clip from the film "Patton" in one session on lead-ership. What did that illustrate?SF: The reason I showed that clip is to show that sometimes you

need different leaders for different types of companies. I don’t knowif you’d want Patton in the White House after the war was over. Hewas determined to march into Russia, for God’s sakes. But for histime and his assignment, he was the perfect leader. In the clip, he’ssaying that the object is not to die for your country, the object is tohave the other poor guy die for his country. That scene from Pattonwas about identifying who your competitors are, and then to havelaser focus.

SB: William Lauder, the CEO of Estée Lauder, was one of yourguest speakers. What did he bring to the discussion?SF: We were trying to talk about the difference between being aboss and being a true leader. From the time he could breathe,William Lauder knew his challenge would be to develop into an effec-tive leader. He had to fill the shoes of his grandmother, Estée Lauder,and his Dad, Leonard Lauder, and keep the company moving for-ward. We talked about family-run companies, and how some of themessentially die with the third generation, which too frequently oper-ates with a sense of entitlement. I wanted William in there becausehe didn’t have a sense of entitlement. He once said that he workstwice as hard for half the credit. He’s a good human being. Peoplelike working for him because he’s a leader. He never felt thatbecause his family had started the company he was entitled torespect. He earned it.

12 Sternbusiness

A graduate of NYU Stern, Steven T. Florio has spent more than three decades in New York’s publishingworld. After starting out as an advertising executive at Esquire, he served as publisher of CondéNast’s GQ, and as president and chief executive officer of The New Yorker in the 1980s. When TheNew Yorker was acquired by Advance Publications, he rejoined Condé Nast. In 1994, he was namedpresident of Condé Nast Publications, Inc., the largest privately held U.S. magazine company, whichincludes such tit les as Vogue, Vanity Fair, Glamour, and GQ. Two years later, he was named CEO. Inearly 2004, after suffering health setbacks, Mr. Florio stepped down from this position, and wasnamed vice chairman of Advance Magazine Group. Last fall, he returned to his alma mater to teachan MBA course entit led “Leadership in the Communications Industry.” The course featured guestappearances by CEOs including Donald Trump and Will iam Lauder of Estée Lauder.

I n a n i n t e r v i e w w i t h S t e r n b u s i n e s s , M r. F l o r i od i s c u s s e d h i s e x p e r i e n c e s t e a c h i n g t h e c o u r s e ,a n d t h e l e s s o n s i t t a u g h t h i m a b o u t l e a d e r s h i p .

SB: Does leadership have a different meaning when it comes to run-ning a public and a private company?SF: The answer is yes. If you work for a privately held company,you’re held to the same standard, but you’re not so quick to burn thefurniture on a quarterly basis. S.I. Newhouse was my boss for 25years, and the lesson I learned from him was to build for the long-term. Was he interested in profit? You bet! But he wasn’t willing tomake short-term, quick decisions based on monthly earnings.Unfortunately, Wall Street has put the pressure on modern-day CEOsof public companies to hit earnings figures on a short-term basis. I’veoften said to people, look at the great job that Michael Eisner did atDisney. If you had invested $1,000 the day he took the job, you’dhave $50,000 now. And yet he had eight, nine, 10 quarters in a rowthat weren’t particularly good, and they’re trying to paint him as a ter-rible CEO. He did a brilliant job. He always took the long-term view.

SB: Donald Trump was another guest lecturer. What did he bring tothe class?SF: I told everybody there were three Donald Trumps: (1) the per-sonality, the character you see on television; (2) the real estate devel-oper, who is brilliant; and (3) Donald Trump the guy, the friend. Whenyou’re really his friend, he’ll take a bullet for you. When I got reallysick and was in the hospital, he was there. But I asked him to beDonald Trump, the MBA from Wharton. And he gave them a rousinglecture about building a business and a brand.

SB: Who were some of the others?SF: Michael Wolf, the head of McKinsey’s media and entertainmentpractice, came and talked about what you look for as you examinethe CEO for leadership. I had Michelle James, who runs the execu-tive search firm James & Co., who talked about what you look for intrying to match an executive with the culture of the company.

SB: It sounds like you had a bunch of living case studies.SF: You can read case studies anywhere. Someone once asked mewhy I went to NYU. I said if I wanted to learn how to ski, I would havegone to Aspen. I wanted a career in business, so I came to NewYork. The benefit of going to school at NYU is going to a classroomand learning from a guy that ran a $3 billion business, listening to awoman who hires seven-figure executives, talking to the president ofMcKinsey, and shooting your hand up and asking a question. That’sthe great benefit of NYU — it’s not just the quality of the teaching, butthe universe that it operates in.

SB: What did you learn from this course?SF: It really forced me to think about what kind of a leader I am.Preparing three-hour lectures makes you do that. And it forced me tothink now about what kind of leader I will be in the future. I have totell you very selfishly that a big part of my healing from heart surgerytook place in that classroom. �

Sternbusiness 13

ERIn the modern corporation,

positions confer power. But

informal status — the prestige

a worker or executive can

build up among co-workers

— can depend on much more

than a job title.

14 Sternbusiness

ILLU

STR

ATIO

NS

BY

GO

RD

ON

STU

DE

R

Snyone who has everworked in a corpora-tion knows that titlesand positions can give

employees leverage over one another.But researchers have also concludedthat individuals with higher informalstatus – the level of respect and pres-tige they enjoy among coworkers – canbe at a distinct advantage. Theyare given better opportunities,get more support when needed,and are awarded more creditwhen they succeed. Indeed, indi-viduals sometimes strive harderfor informal status than they dofor formal promotions or higherfinancial compensation. And asinformal collaboration andteamwork have become moreimportant in the 21st century corpora-tion, it’s likely that informal status isgrowing even more significant.

In an effort to understand andquantify how informal status emerges,

we set out to develop and test a theo-retical model of informal status inorganizations. We did so by puttingforth several hypotheses and then test-ing them against data gathered fromthree different types of organizations.And the results – some of them quiteunexpected – shed some interestinglight.

Power can be defined as the abilityto influence others. But people withpower are not always respected andheld in high regard. In other words,they don’t always have status com-

mensurate with their rung in the orga-nizational ladder. In any group, mem-bers assess others’ personal attributesand behaviors and assign greatervalue – and ultimately informal status– to those they deem more likely tomake greater contributions to sharedobjectives. And when individualsbehave in ways that contradict shared

objectives, they can be pun-ished with social neglect, oreven ridicule.

Informal status emergesnaturally from social interac-tions and can be distinguishedfrom a person’s formal positionor rank. And while informalstatus might be correlated withsocial connectedness, one caneasily imagine individuals who

have high informal status but do nothold advantaged positions in socialnetworks – such as highly esteemedyet reclusive academics.

Sternbusiness 15

A“Individuals with higher informal status – the level of respect and

prestige they enjoy among coworkers –can be at a distinct advantage.

They are given better opportunities, get more support when needed,

and are awarded more credit when they succeed.”

TEP C By Sandra E. Spataro andCameron Anderson

16 Sternbusiness

Status PerceptionsAchieving high consensus on the

informal status order is likely acomplex and difficult task in organ-izations. People in organizationswork on multiple tasks, value differ-ent types of work differently, andmay have little contact with oneanother. But such social interaction is

nonetheless set within a socialsystem of agreed-upon ideals.

Further, the richness ofinformation coworkers

have about oneanother – gleanedthrough workingwith them directlyor hearing aboutthem via second-hand sources –should contribute

to its reliability. Finally,the advantages that areassociated with highinformal status tend to behighly visible. People withhigh informal status aregiven more control overinteraction patterns andresources, more opportu-

nities, and more social andmaterial support. We therefore startby predicting (Hypothesis 1) that:Individuals in organizations will con-sensually perceive informal status dif-ferences among their coworkers.

The characteristics and behaviorsthat lead to high status in any givengroup are derived from the group’sspecific goals. A research group willvalue behaviors that contribute to thegroup’s success, such as new discov-eries. But being an effectiveresearcher might not lead to status ina group of salespeople. What’s more,organizations prioritize not only thecompletion of any single task, butalso the success and future vitality ofthe organization itself. As a result,(Hypothesis 2) holds that: Informalstatus will likely be accorded to thosewho both represent and reinforce the

organization’s distinctive values.Since values are prioritized differ-

ently across organizations, it followsthat the characteristics that lead tohigh informal status should also differacross organizations. But there shouldalso be some characteristics andbehaviors that uniformly lead to high-er informal status across organiza-tions. These include an individual’sskills and experiences that contributeto the organization’s core technologies,which rise with a person’s familiarity

with the organization’s history. Wetherefore suppose (Hypothesis 3) that:Individuals’ tenure in the organizationwill be positively related to their infor-mal status. Similarly, across organiza-tions, individuals’ contributions to theshared technical objectives will likelybe assessed, in part, by their ability tofacilitate the completion of importanttasks. Therefore we posit (Hypothesis4) that: Individuals’ job performancewill be positively related to their infor-mal status.

ndividuals with high formal ranktypically have responsibilitiesmore important to the organiza-tion’s success. And fulfilling more

important duties is generally construedby peers as making a stronger contri-bution to the group. So we also propose(Hypothesis 5) that: Individuals’ rankin an organization’s formal hierarchywill be positively related to their infor-mal status. Finally, once individualsattain high levels of respect and esteemamong peers, they become more cen-tral in the flow of communication.People seek them out for informationor advice, to pass along ideas, or sim-ply to seek their companionship. Andso we speculated (Hypothesis 6) that:The amount individuals interact withothers will be positively related to theirinformal status.

Organizational BehaviorTo test these hypotheses, we exam-

ined three organizations: (1) an engi-neering department within a telecom-munications firm. Comprising prima-rily engineers and technicians, itincluded some general managers andadministrative support individuals;(2) a family medicine department of aresearch hospital, comprising fourseparate health clinics, each managedby medical doctors and populated bymedical and administrative staff, as

well as a centralized staff of adminis-trators and faculty; and (3) a consult-ing firm that specialized in brand andimage development, comprisingemployees in three different regionaloffices. The staff includes artisticdesigners, computer programmers,business professionals, and adminis-trative and general managementsupport.

We collected data in informalinterviews, from human resourcedepartments, and through surveys ofall members of each departmentinvolved. The surveys allowed us toobtain peer-ratings of informal statusand social connectedness, self-reportsof tenure with the company, stablebehavioral patterns, race, socioeco-nomic status, education level, andratings of the organization’s values.The total sample included 427 partic-ipants, divided roughly between thethree organizations.

We used peer-ratings to measureinformal status. Each participant wasasked to rate 10 randomly selectedcoworkers on how much status he orshe had, from 1 = (“low”) to 7 =(“high”), where status was defined asthe amount of “prestige or socialstanding” each member had.Participants’ jobs were assigned theformal rank their occupation held rel-

�

“Once individuals attain high levels of respect and esteem among peers, they become more central in

the flow of communication.”

I

R-E-S

Sternbusiness 17

ative to the other jobs in theirorganization. In the engineeringfirm, support staff received a “1,”while managers received a “4.”Similar gradations were made in thehospital department and consultingfirm.

e assessed social con-nectedness using peer-ratings. Participantsrated the same 10

coworkers on how much they interact-ed with each of them, on a scale from1 = (“Never”) to 7 = (“All the time”).We were able to obtain job perform-ance data in the consulting firm in theform of manager evaluations of per-formance. Ratings were on a scale of1 (“Rarely or never meets expecta-tions/Red flag”) to 5 (“Consistentlyexceeds expectations/Top performer/Leader”). In the surveys, participantsreported tenure in their organization,coded in number of months.

To measure the extent towhich individuals behaved inways valued by their organiza-tion, we assessed the values ofeach organization through infor-mal interviews and then assessedindividuals’ stable patterns ofbehavior through self-reportmeasures of personality. We observeddifferences between the three organi-zations in how much they emphasizedtangible results, or the “bottom line,”and in their focus on teamwork andcollaboration. These dimensions close-ly mirrored what researchers CharlesO'Reilly, Jennifer Chatman, and DavidCaldwell in 1991 labeled “outcome-orientation” and “team-orientation.”

To measure their firm’s outcome-orientation, participants rated theextent to which each of six dimensionscharacterized their organization: resultor outcome-oriented, detail-oriented,reward-oriented, decisive, aggressive,and competitive. For team-orientation,participants rated the extent to whicheach of five dimensions characterizedtheir organization: team-oriented,

group-oriented, people-oriented, sup-portive, and cooperative. Eachdimension was rated on a scale from1 (“Extremely uncharacteristic”) to 7(“Extremely characteristic”). Theengineering department was rated asvaluing outcome-oriented dimensionsof culture more than the hospitaldepartment, which was in turn fol-lowed by the consulting firm. The con-sulting firm was rated as highest onteam-orientation, followed by the hos-pital department, followed by theengineering department.

Based on these assessments, wefocused on two personality dimensionsthat provided the best content matchesto the organizational value-dimensionsof outcome- and team-orientation.First, Conscientiousness, which is apersonality dimension that “facilitatestask- and goal-directed behavior;”conscientious individuals are dutiful,

hard-working, and organized. Second,we focused on Extraversion, which is apersonality trait that involves an“energetic approach to the social andmaterial world and includes traits suchas sociability, activity, assertiveness,and positive emotionality.” In theworkplace, Conscientiousness hasbeen linked with diligence in tasks,whereas Extraversion has been shownto predict job performance in occupa-tions that require interpersonal skills,such as in sales positions.

A rich literature on status charac-teristics has shown that demographiccharacteristics such as sex, race, orsocioeconomic status can becomesalient status characteristics. As aresult, we assessed and controlled forsex, ethnicity, socio-economic status,

and education in our analyses of ori-gins of informal status.

ResultsWhat did our results find?

Hypothesis 1 predicted that organiza-tional members would reliably per-ceive informal status differencesamong the coworkers in their depart-ment. And indeed, we found that in allthree organizations, a consensus hademerged among coworkers as to whohad high and who had low informalstatus.

ased on our assessments ofthe three organization’s val-ues, we conducted two testsof Hypothesis 2, which held

that informal status will likely beaccorded to those who both representand reinforce the organization’s dis-tinctive values. First, we hypothesizedConscientiousness would most strongly

relate to informal status in theengineering department, fol-lowed by the hospital depart-ment, then by the consultingfirm. Second, we hypothesizedExtraversion would most strong-ly relate to informal status in theconsulting firm, followed by thehospital department, then by the

engineering department. As shown in Table 1, the

strength of the relation betweenConscientiousness and informal statuswas highest in the engineering depart-ment, followed by the hospital depart-ment, followed by the consulting firm.This provides support for Hypothesis2. Figure 1a illustrates graphicallythe difference between the engineeringdepartment and the consulting firm.

As shown in Table 1, the strengthof the relation between Extraversionand informal status was highest in theconsulting firm, followed by the hos-pital department, followed by theengineering department. That meansExtraversion was a stronger predictorof informal status in the consultingfirm than it was in the engineering

“Understanding the political landscape in an organization

requires knowing the individuals that are respected and admired in addition to understanding the

formal organizational chart.”

WB

department and that it was in the hos-pital department. Figure 1b illustratesgraphically the difference between theconsulting firm and the engineeringdepartment.

Hypotheses 3 and 4 predicted thattenure and job performance, respec-tively, would be related to informalstatus. But tenure turned out to be asignificant contributor to informal sta-tus only in the consulting firm. Andjob performance, which could begauged only in the consulting firm,proved to be significantly related toinformal status.

In Hypotheses 5 and 6, we pre-dicted informal status would be relat-ed to structural position within theorganization’s formal hierarchy, andto social connectedness, respectively.In all three organizations, in fact,

informal status was significantlyrelated to formal rank. Similarly,informal status was independentlyand directly related to social connect-edness in all three organizations. Butin our study, informal status differ-

ences were moderately related to, butclearly distinct from, the formal orga-nizational hierarchy and from pat-terns of social connectedness.Implications

What are the implications of these

18 Sternbusiness

�P-E-C-T

Equation 1

EngineeringDepartment

.290*(.133).069(.092)

.001(.001)

.352**(.131).592**(.071)

-.004(.168).574(.359).585(.388).499(.375).397(.385).163*(.073)-.141(.123)-.138(.143)-.015(.110)-.237(.130)8.90**0.53

Equation 2

Hospital Department

.075(.176).063(.128)

.002(.002)

.278**(.080).543**(.104)

-.093(.220)-.738(.674).245(1.059)-1.332(.836)

-.195(.172).086(.181).060(.210)-.293(.166)-.129(.195)3.00**0.29

Equation 3

Consulting Firm(without performance)

-.177(.262).569**(.204)

.015**(.005)

.381**(.112).391**(.110)

-.127(.308).218(.486).763(.777).391(.600)1.541(1.027).254(.179).373(.242).204(.258).107(.269)-.122(.286)4.66**0.52

Equation 4

Consulting Firm(with performance)

.073(.261).551**(.198)

.013**(.005).262*(.147)

.324**(.144).382**(.107)

-.021(.305).278(.473).738(.754).507(.587)1.380(1.001).218(.175).343(.235)-.002(.276)-.149(.298)-.209(.282)4.83**0.55

Table 1 Regressions Predicting Informal Status in Each Organization

Independent variable

Organization-specific origins: Embodiment of Org’l ValuesConscientiousness

Extraversion

Origins related to the technical coreTenure with organization

Job performance

Structural correlatesFormal rank in organization

Social connectedness

Control variablesSex (1 = female)

Caucasian (1 = Caucasian)

African-American (1 = African-Amer.)

Asian-American (1 = Asian-American)

Hispanic/Latino (1 = Hispanic)

Socioeconomic status

Education level

Agreeableness

Neuroticism

Openness

F-statisticAdjusted R-squared

*p < .05; **p < .01.Entries are unstandardized beta coefficients. Standard errors are in parentheses.

Info

rmal

Sta

tus

(z-s

core

d)

More outcome-oriented organization(Engineering Department)

Low ConscientiousnessHigh Conscientiousness

Less outcome-oriented organization(Consulting Firm)

0.2

0.15

0.1

0.05

0

-0.05

-0.1

-0.15

-0.2

Figure 1A

findings? First, if informal status islargely distinct from formal rank, itsuggests that individuals who lack for-mal authority in the organization canstill wield considerable influence overothers. Thus, understanding the polit-ical landscape in an organizationrequires knowing the individuals thatare respected and admired in additionto understanding the formal organiza-tional chart.

Second, individuals’ behaviormight be shaped by their strivings forinformal status as much or more thanby their aspirations for formal promo-tion or higher financial compensation.On an organizational level, this canbecome a problem if the behaviorsrewarded and punished by the infor-mal status structure differ from thosethat managers seek to instill in theirworkers. For example, in their classic1939 study, Management and theWorker, Fritz Julius Roethlisbergerand William J. Dickson observed howthe informal status structure in a fac-tory punished workers who performedtoo efficiently.

It is also clear that emergent infor-mal status differences are associatedwith organizations’ unique values. Inthe engineering department, whichvalued task-oriented over socially-ori-ented behavior, being diligent and

task-focused led to status but beingsociable and talkative did not. In con-trast, in the consulting firm, whichvalued socially-oriented over task-ori-ented behavior, being talkative andsociable led to status but being diligentand task-oriented did not. Our studyshowed that the presumed linkbetween task abilities and domain-specific status characteristics is not sotight. Rather, we found that upholdingand projecting the values of the organ-ization, independent from experiencein the organization or even job per-formance, was an important source ofinformal status.

rior research has docu-mented other benefits indi-viduals enjoy when theypossess values that are con-sistent with their organiza-

tion: higher satisfaction, lowerturnover, and increased organiza-tional commitment. Our research sug-gests that those who possess such val-ues could also achieve positions ofpower and influence in organizationsthrough higher informal status. Inother words, individuals who attainhigh levels of respect and esteemamong their peers garner influenceover their coworkers’ attitudes, workhabits, performance levels, and satis-faction with the organization.

Organizations appear to promote cul-turally appropriate behavior not onlythrough formal means, such as theallocation of bonuses, bigger offices,and promotions but also throughinformal means – namely, the alloca-tion of high informal status amongpeers.

In contrast to a long tradition ofresearch on ad hoc groups, the demo-graphic variables included in ouranalyses as controls did not emerge asindependent contributors to informalstatus. Sex and ethnicity did not havean effect on informal status in any ofthe three organizations. This mayreflect the “relevancy” principle ofExpectation States theory, whichholds that diffuse characteristics suchas sex and race are less relevant whenother more relevant information –such as behavior, expertise, and jobperformance to gauge their status – isavailable.

Due to some of the limitations ofthe current research, a number ofquestions need further examination.First, given the small number oforganizations, we could not statistical-ly test the relation between group val-ues and the traits that lead to status.Future work should focus on theimportance of status to individualorganizational members and examinethe importance of informal status tothe dynamics of organizations at thecollective level. As a system of socialrewards and punishments, informalstatus hierarchies might provide arobust and powerful way to controlemployees’ behavior. It is thereforecrucial to understand whether thesesocial rewards and punishments areshaping employee behavior in waysthat managers wish them to.

SANDRA E. SPATARO is assistant pro-fessor of organizational behavior atthe Yale School of Management.CAMERON ANDERSON is assistantprofessor of management at NYU Stern.

Sternbusiness 19

Figure 1B

Low ConscientiousnessHigh Conscientiousness

Info

rmal

Sta

tus

(z-s

core

d)

More team-oriented organization(Consulting Firm)

Less team-oriented organization(Engineering Department)

0.4

0.3

0.2

0.1

0

-0.1

-0.2

-0.3

-0.4

P

arkets rely significantly on the trustcreated by repeated interaction andpersonal relationships. Is it possible to

obtain the same level of trust and efficiency in moreanonymous electronic markets? One possibility is tocreate reputation mechanisms that allow traders toidentify and monitor each other. And that’s preciselywhat has happened on eBay.

Since its launch in 1995, eBay has become the dom-inant online auction site, with millions of items chang-ing hands every day. In 2003, more than $21 billion insales were transacted on eBay by 69 million users. EBayacts purely as an intermediary, and collects revenuefrom seller fees upon successfully completed auctions.But the company doesn’t just process orders. To enablereputation mechanisms to regulate trade, eBay hasdeveloped an innovative feedback system.

After an auction is completed, buyers and sellers cangive one another grades of +1 (positive), 0 (neutral), or-1 (negative), and add textual comments. EBay thendisplays several aggregates of the grades: the overallrating, the sum of positives minus negatives received bya seller; the percent of positives; the date when the sell-er registered; a summary of recent reviews from the pastweek, month, and six months; and the entire feedbackrecord, an exhaustive list of reviews left for the seller.With its well-defined rules and mass of available infor-mation, eBay thus presents the researcher with a fairly

20 Sternbusiness

MEBay may seem to be a vast,

anonymous electronic market-

place. But the website’s

mechanisms that allow buyers

and sellers to rate one another

ensure that reputation matters.

By Luís Cabral and Ali Hortasçu

LIVE AUCTIONHEROES

ILLU

STR

ATIO

N B

Y J

ULI

ETT

E B

OR

DA

Sternbusiness 21

controlled environment for theorytesting. So we decided to use thisdata to investigate how feedbackcomments affect reputation, futuresales, and the willingness of eBayparticipants to continue buying andselling goods.

e collected data fromeBay’s website atmonthly intervalsbetween October 24,2002 and March 16,

2003 on three products. First, wechose collectible coins, one of themost active segments on eBay. Weexamined activity in 1/16 oz. fivedollar gold coins of 2002 vintage(gold American Eagle) and 2001 sil-ver proof sets, a set of 10 coins ofdifferent denominations. Both itemsare produced by the U.S. mint. Theaverage sale price for the gold coinin our data set was $50, and theproof sets sold on average for $78.Second, we chose IBM ThinkpadT23 PIII notebook computers.Notebook computers, according tothe Federal Bureau of Investigation’sonline fraud unit, give rise to the

not feasible, since users can’t accesstransaction level information that ismore than 30 days old, and many ofthe sellers in our sample have beenusing eBay for much longer thanthat. However, assuming that a con-stant fraction of transactions israted by bidders, the total numberof feedback points is a good proxyfor the total number of transactionsconducted by the seller. The averageseller in our sample has 1,625 totalfeedback responses (See Table 1).The median seller has 397. Thelargest seller has 52,298 feedbackresponses, and the smallest has 0.Sellers were largest in the marketfor Thinkpads, followed by teddies,gold coins and the proof sets.

The average seller in our samplehas 4.9 negative feedback points,corresponding to 0.9 percent of allcomments. The maximum numberof negative feedbacks received by aseller is 819, but this seller took partin 52,298 transactions. The medianseller in our sample has only onenegative feedback, and more than aquarter of the sellers have none.Our subjective impression, afterbrowsing through eBay communitychatboards, is that the informationcontained by a neutral rating isperceived by users to be muchcloser to negative feedback thanpositive. Given this, we decided to

most customer complaints aboutauction fraud. The average saleprice of the Thinkpad T23’s was$580. Finally, we chose the 1998Holiday Teddy Beanie Babies, ahugely popular item. With an aver-age sale price of $10.70, the BeanieBabies were the least expensive itemwe examined.

Along with transaction-leveldata, we also downloaded each sell-er’s “feedback summary” page, asshown in Figure 1. The constructionof entire transaction histories formany of the sellers in our sample is

22 Sternbusiness

LIVE AUCTION HEROES

Table 1: Distribution of feedback aggregates across sellers.

Mean

Standard Dev.

Minimum

Maximum

Number ofPositives

1,625

3,840

0

52,298

Number ofNegatives

4.9

25.1

0

651

Number ofNeutrals

7.2

33.5

0

654

N/(N + P)(entire history)

0.009

0.038

0

1

W

Figure 1

lump negative and neutral com-ments together when talking about“negative” comments.

Negative Feedback andSales

Having assembled the data, weset out to test several assumptions.We assume that generally speak-ing, the higher the seller’s reputa-tion, the more he has to gain fromputting an object up for auction,and thus the more often he will doso. Conversely, the worse the repu-tation of the seller, the less he hasto gain from doing so. We thushypothesized that after the firstnegative feedback experience,there would be a drop in the rate atwhich the seller puts objects up forsale and manages to sell them.

n order to test this hypothe-sis, we constructed a proxyfor weekly sales totals byadding the total number ofsales-related feedback com-

ments received by a seller in a givenweek. We then marked theweeks in which a sellerreceived his first, second, andthird negatives. We averagedthe weekly sales rates over afour-week window before andafter the week in which theseller got his first (or second,or third) negative. We also calculat-ed the sellers’ “before” and “after”weekly growth rates by averaginggrowth rates over these two four-week windows. The results, report-ed in Table 2, are striking: For allfour object categories, the impactof the first negative is to slowgrowth by 14 percent a week, andthis difference is highly statisticallysignificant. By and large, the sec-ond and third negatives did nothave statistically significant affectson growth rates.

impact on the frequency of negativefeedback.

At the very least, it appears fromthese results that there is somethingspecial about the first negative thata seller receives: Once the first neg-ative arrives, the second one arrivesfaster. Given the significance of thisresult, both in statistical and in eco-nomic terms, we set out to find pos-sible explanations.

One explanation is that buyershave a threshold of dissatisfactionabove which they give a negative,

and that this threshold dropsafter the first negative. Thereare several behavioral mecha-nisms through which this canhappen. There could be adecrease in the cost of writinga negative comment. Manynegative comments from buy-

ers are followed by a retaliatorynegative comment given by the sell-er; and seller retaliation mightimpose an economic cost on thecomplaining buyer, especially ifthe buyer is also a seller. Such aneffect would confound our resultsif the probability of retaliation by aseller in reaction to her first negativeis higher than retaliation to hersecond negative, an explanationproposed by several eBay users wetalked to.

In order to investigate this possi-bility, we checked whether each

Frequency of NegativeFeedback

We next examined whether thearrival of the first negative ratinghas an impact on the frequency withwhich subsequent negative ratingsare given. We measured time in twoways: number of sales transactionsand calendar time (number of days).For the Thinkpad, it takes on aver-age 129 transactions before a sellerreceives his first negative, but only60 additional transactions before thesecond arrives. Similar results are

obtained for the other three objects.When we replicated this analysiswith time measured in days, the dif-ference between the interarrivaltimes of the first vs. the second neg-ative is again quite striking. In theThinkpad market, for example, ittakes on average 300 days for thefirst negative to arrive, but only 66days for the second one. In all caseswe considered, the increase in fre-quency after the first negative is sta-tistically significant. By contrast, thearrival of the second, third, up tofifth negative seems to have no

Sternbusiness 23

Table 2: Impact of first negative on sales growth (%).

OBJECTAverage WeekGrowth Rate Thinkpad

7.12

–6.76

–13.88***

4.88

66

Proof set

6.85

–7.51

–14.36***

3.45

130

G. Eagle

9.04

–3.89

–12.92***

3.58

95

B. Baby

14.19

–4.28

–18.47***

3.69

136

“With its well defined rules andmass of available information, eBaythus presents the researcher with a fairly controlled environment for

theory testing.”

I

Before

After

Difference

Standard Error

Number of Observations

particular negative comment by abuyer was accompanied by a retal-iatory negative left by the seller.The result was striking: Of thealmost 10,000 negative/neutralinstances in our data, 2,462 result-ed in a retaliatory comment by theseller. However, our data indicatesthat sellers are not more likely toretaliate upon their first negative,as opposed to subsequent nega-tives. So it does not appear that“fear of retaliation” is a significantdriver of the difference in interar-rival times of negative comments.

Next, we considered the possi-bility that buyers are influenced byother buyers’ behavior. In particu-lar, faced with poor performance bya seller with a perfect record, abuyer might be inclined to thinkthat there is no ground for a nega-tive feedback. For example, ifthere is a communication problembetween buyer and seller, the for-mer may attribute this to a problemwith him or herself. However, if theseller has already received a nega-tive feedback, especially regardingthe same problem that the buyer isnow facing, then the buyer mayhave a greater inclination to attrib-ute this to a problem with the seller.

To consider this possibility, weclassified the first and second nega-tive remarks according to theirnature. The buyer-influence storyshould imply an increase in the rel-ative importance of subjectiveproblems in second negatives.However, the results suggest a verysimilar pattern for first and secondnegatives. Moreover, “item neversent,” arguably the most objectivereason for negative feedback, actu-ally increases in relative impor-tance (though by a small amount).At the opposite extreme, “bad com-munication,” arguably the mostsubjective reason for negative feed-

changing his identity and starting anew reputation history. Intuitively,we would expect the seller’s tenden-cy to do so to be decreasing in theseller’s reputation. We supplement-ed our data set by revisiting oursample of sellers in the first weekof January 2004, and checkingwhether they were still in business.Of the 819 sellers originally sam-pled, we found that 152 had notconducted any transactions withinthe last 45 days and 61 sellers hadnot sold anything within the last 45days, but had bought an item. Wealso could not locate the feedbackrecords for 104 sellers in our sam-ple, since eBay’s database claimedthat these seller IDs were no longervalid.

hen we ran regressionson these data, theresults implied that aten-fold increase in the

total number of positives (as of May2003) translates into a decline inexit probability (in January 2004)of between 14 to 21 percent. Also, a1 percent level increase in the per-centage of negatives in a seller'srecord translates into an increase inexit probability of 1.6 to 2.1 per-cent. For Beanie Babies, the magni-tude of the coefficient estimateimplies that an increase from 1 per-cent to 2 percent of negatives in aseller’s record translates into 12.5percent higher exit probability.

We also investigated whether the“exits” we see in our data set areaccompanied by opportunistic prof-it-taking by sellers, and whetherreputational variables can predictsuch behavior. We collected data onthe last 25 sale transactions con-ducted by exiting sellers, andcounted the number of negativecomments. Some cases were quitestriking: One of the sellers in oursample, who had 22,755 positives,

back, also increases in importance(though by an even smalleramount).

Finally, if the “threshold” storyholds true, we would expect thecomments accompanying first nega-tives to be nastier than the com-ments accompanying the secondand subsequent negatives. In orderto test this possibility, we createdpairs of comments corresponding toeach seller’s first and second nega-tive. We then asked a third party (astudent) to make a subjective evalu-ation as to which of the two remarkswas more negative. (We randomlymixed the order of the comments sothat the student could not tell whichwas the first and which was the sec-ond negative). The results show that51 percent of the second negativeswere considered nastier then the cor-responding first negative, a split thatis not statistically different from50/50.

In sum, the empirical evidencesuggests that the behavioral changefrom the first to the second negativeis not due to changes in buyerbehavior, but rather to changes inthe seller behavior. Our interpreta-tion is that, once the first negativearrives, a seller’s reputation is worthless and the value of protectingsuch reputation is also lower.Accordingly, the seller makes lesseffort to guarantee a good transac-tion and as a result more negativefeedback experiences take place.

Reputation and ExitNext we considered the possibili-

ty of a seller “exiting,” i.e., secretly

24 Sternbusiness

LIVE AUCTION HEROES

“For all four object categories, the impact of the first negative is

to slow growth by 14 percent a week.”

W

racked up 11 negatives inher last 25 transactions;whereas she had a total of 54negatives in her previoustransactions. On average,the percentage of negativesin the last 25 comments ofexiting sellers we examinedwas 4.38 percent, asopposed to an average 1.61percent over their entire his-tories. The results of theseregressions indicate that, forthe entire sample of sellers, aten-fold increase in a seller’scount of negatives is corre-lated with a 5 percentincrease in “opportunistic”exit as defined above.

In summary, our data is consis-tent with the possibility of oppor-

tunistic profit-taking and exitbehavior by sellers. There are, how-ever, alternative stories consistentwith the data. For example, itmight be that some unexpectedexogenous event leads the seller tooffer poor service for a period oftime, which results in an increase innegative feedback, which in turnresults in the seller’s decision to exit(given such a poor record).

Buying a ReputationCasual observation of feedback

histories suggests that many sellersappear to start out as buyers, com-pleting a string of purchases beforeattempting their first sale. As canbe seen from Figure 2, bearsylva-

from being a buyer to a seller. Soalthough we do find indisputableevidence for the existence of switch-ing behavior on eBay, our evidencefor a clear economic incentive to doso is weak.

ConclusionThe marketplace can be quite

efficient in meting out punishmentfor those who don’t adhere toexpected norms. On eBay, a vastmarketplace itself, the reputationmechanism plays the role of punish-ing poor performance and behavior.Indeed, it is clear from our researchthat eBay’s reputation system givesway to noticeable strategic respons-es from both buyers and sellers.That is, the mechanism has “bite.”Of course, this does not imply thatthe current structure of the system isoptimal. In fact, we believe an excit-ing area for future research is pre-cisely the design of an efficient rep-utation mechanism.

L U Í S C A B R A L is pro fessor o feconomics at NYU Stern.ALI HORTASÇU is assistant professor ofeconomics at the University of Chicago.

nia – a Beanie Baby dealer – startedout as a buyer first, and quicklychanged the pattern of his transac-tions from purchases to sales.

We then defined a seller as havingswitched from being a buyer to beinga seller if more than 50 percent of thefirst 20 comments referred to pur-chases, and more than 70 percent ofthe last 20 comments referred tosales. We found that 38 percent ofBeanie Baby sellers, 22 percent oflaptop sellers, 31 percent of gold coinsellers, and 31 percent of proof setsellers followed the “buy first, selllater” strategy. We also found that,on average, 81 percent of a seller’slast 20 transactions were sales, com-pared to 46 percent of the first 20transactions. These results show that“buying first and selling later” is awidespread phenomenon on eBay.

Next, we investigated the correla-tion of the “buy first, sell later” indi-cator variable with the percentage ofnegatives in a seller’s record, as wellas the length of the seller’s record.This regression suggests that a 1 per-cent level increase from the meanvalue of 0.7 percent of negatives to1.7 percent of negatives is correlatedwith a 6.4 percent decrease in theprobability that the seller switched

Sternbusiness 25

1

0.8

0.6

0.4

0.2

00 10 20 30 40 50 60

No. of weeks on eBay

Figure 2: How “bearsylvania” became a seller

Perc

ent o

f tra

nsac

tions

con

duct

ed a

s sa

les

“On average, the percent-age of negatives in the last

25 comments of exitingsellers we examined was4.38 percent, as opposed

to an average 1.61 percentover their entire histories.”

great deal of journalisticand academic attentionhas been focused on thestrong growth in produc-

tivity in the U.S. economy. Between1996 and 2003, productivity roseat a 3 percent annual rate, doublethe pace of the first half of the1990s. Automation, frequentlydriven by advances in informationtechnology (IT), has been one of thesources of this productivity growth.To take but one recent example,Atmos Energy, a Dallas, TX-basedgas company, is automating its gasmeter reading capabilities by using

growth over the next 10 to 15 years. Industrial automation goes back

to the Industrial Revolution of the18th century, when machinesreplaced physical labor on a massivescale. From the advent of the steamengine to the assembly line, workpreviously done by human handscame to be done by machines thatcould harness the power of water,steam, and, eventually electricity. Inthe past 25 years, automation trans-formed manufacturing as industrialrobots replaced manual jobs inindustries such as automobiles,computers, and telecommunication

wireless technologies, a move thatwill allow it to reduce its staff by225 employees over the next fiveyears and thus attain significantincreases in productivity.

It is natural to wonder if suchautomation-driven productivityenhancements can be sustained.After all, it seems like so many tasksand components of jobs are nowautomated. And yet there’s an argu-ment to be made that we are still inthe early stages of a new wave ofautomation, which will profoundlyaffect the economy and significantlycontribute to the productivity

A

26 Sternbusiness

A U T O M AT I O N ’ SN E X T AV EIn recent years, productivity-enhancing information technology has wrought signifi-cant changes in global labor markets. But the process may just be getting started.

By Alexander Tuzhilin

ILLU

STR

ATIO

NS

BY

KE

N O

RV

IDA

S

Sternbusiness 27

equipment. More recently, automa-tion has been primarily driven byIT. The toll booth collectors whohave lost their jobs to EZ-Passtechnologies may be a harbinger offuture trends. It is possible, forexample, that many cashiers indepartment stores and supermar-kets will soon lose their jobsbecause of the advancements of theRadio Frequency Identification(RFID) tag technologies.

Most of the jobs lost toautomation have been rou-tine production jobs,according to the job classifi-cation proposed by formerLabor Secretary RobertReich in his 1991 book TheWork of Nations. Examplesof these jobs, which arecharacterized by repetitiveness andstructuredness, include assemblyline workers, foremen, data proces-sors, and toll collectors.

The next wave of automationwill affect not only routine produc-tion workers, but also the better-paid and heretofore more securegroup that Reich called symbolic-analytic workers – engineers, officeand knowledge workers, managers,educators, and other groups ofmind workers. Although few ofthese jobs will be eliminated com-pletely, many of the more routinetasks in these jobs will be delegatedto smart machines within the next10-15 years, leading to majorrestructuring and consolidation.

Symbolic AnalystsI recently taught a course on

Advanced Technologies for BusinessApplications at NYU Stern. Thestudents, who were predominatelypart-time MBA students, wereasked to describe what parts (ifany) of their jobs or the jobs of theirclosest colleagues, could be auto-

are high on repetitiveness, stability,and structuredness – constitute theprimary candidates for automation.For example, the task of a salesper-son meeting with the same clientover and over again and interactingwith the client in a structured man-ner, asking the same set of questionsand offering a simple array of serv-ices based on the answers, is a goodcandidate for automation by an

intelligent software agent.Moreover, most of the rou-tine production jobs thathave been lost to automa-tion rate highly on all ofthe three dimensions. Incontrast, the tasks that areaway from the origins onall three dimensions are thehardest to automate. For

example, the task of a salespersonmeeting with a different and ever-changing clientele and havingunstructured open-ended discus-sions with them is very hard toautomate.

If all the tasks of a given job canbe automated, then the entire jobcan be eliminated. However, this isunlikely to occur for most of thesymbolic-analytic jobs since most ofthem have some tasks that areranked high along at least onedimension in Figure 1. Therefore,most of the symbolic-analytic jobscan be automated only partially (ifat all) within the next 10-15 years.

Extent and ScopeOne of the surprising outcomes

of the student projects was theextent and scope of possibleautomations they identified for dif-ferent types of jobs in diverse indus-tries, including accounting, finance,healthcare, human resources, IT,marketing and sales.

For example, one type of a job astudent described as already auto-

mated within the next 10 to 15years. All the students were symbol-ic-analytic workers according toReich’s classification, and some ofthem worked in managerial posi-tions. Based on about 30 studentreports, an interesting pictureemerged about the types of jobs thatcan be automated and the extentand scope of this automation.

In general, jobs can be classified

along three dimensions. First, repet-itiveness – for example, a salesper-son repeatedly meeting with clients.Second, stability – a job that doesnot change over time. For example,a salesperson meeting with the sameclient, as opposed to meeting differ-ent clients. Third, structuredness – ajob that can be described with aclear procedure, perhaps evenexpressed as an algorithm. Forexample, a salesperson can have astructured interaction with the clientasking several standard questionsand making several standard offer-ings of products. Alternatively, theinteraction can be unstructured andopen-ended.

Many jobs consist of severaltasks, with each task characterizedby the three dimensions of repeti-tiveness, stability and structured-ness. Graphically, a job can be rep-resented with a set of points in thethree-dimensional space shown inFigure 1, where each point consti-tutes a particular task of the job.

The tasks that are closer to theorigin in Figure 1 – i.e., those that

28 Sternbusiness

“There’s an argument to be madethat we are still in the early stages

of a new wave of automation, which will profoundly affect the

economy and significantly contributeto the productivity growth over the

next 10 to 15 years.”

mated is that of the ClientAccountant. This job is responsiblefor ensuring that all the client’stransactions settle properly, allfunds are transferred, and all theaccount balances are reconciledwith various parties involved in atransaction. It is a very routine andpaper intensive job that rates veryhigh on all three dimensions inFigure 1 (the point is close to theorigin). Over the past few years,this job has been automated in thefinancial services and other indus-tries. A single client accountant cannow monitor the transaction activi-

process involving running variousreports, cutting and pasting infor-mation from Excel and Word docu-ments, and eventually building aPowerPoint presentation. In manyapplications this process is struc-tured, straightforward, and doesnot require much creativity. It alsorates high on all three dimensionsin Figure 1, and is a good candi-date for automation. Some com-panies are currently trying toautomate this task. However,that does not mean that the jobof a Marketing Associate willbe eliminated, since it alsoinvolves other tasks that areless routine and structured.Instead, Marketing Associatejobs are more likely to be consol-

idated and restructured byautomating the tasks of respond-

ing to RFPs and RFIs and lettingMarketing Associates focus on themore human-oriented parts of theirjobs.

These two examples representthe simplest types of symbolic-ana-lytic service jobs that are currentlythe primary targets for automation.The students also provided numer-ous examples of more advancedautomation tasks. Currently, manybusiness processes have alreadybeen partially automated by dele-gating some parts to machines andother parts to humans. Examples ofsuch human-centered tasks includemoving information from one sys-tem to another or checking theresults returned from one part ofthe business process before initiat-ing another. These human activitiesare often required because varioussystems may not “talk” to eachother or may return questionableresults that need to be inspectedbefore the business process can con-tinue. These activities usually con-stitute the leftovers from previous

ties of 10 times more accounts thanwas feasible in the past.

Another example of a job cur-rently being partially automated insome companies is that of aMarketing Associate, who helps cre-ate a company’s responses to variousRequests for Proposals (RFPs) orRequest for Information (RFIs). Oneof the tasks for which MarketingAssociates are responsible for is thecollection, reviewing, and compilingof the account-related information(such as performance figures, mar-ket values, etc.) into a presentableformat. It is a laborious, manual

Sternbusiness 29

Figure 1 Job representation as a set of tasks in 3D space(each dimension is measured in terms of High and Low, where High is at the origin and Low at the end of the axis)

Task 4

Task 2Task 3

Sta

bilit

y

Repet

itiven

ess

Structuredness

Task 1

automation projects and comprisethe hardest parts of these projectsthat were left un-automated for thereasons mentioned above.Naturally, they are primary candi-dates for new automation attemptsusing more recently developedinformation technologies.

The students also explored vari-ous other jobs that are significantlyharder to automate, such as newproduct development, sales sup-port, systems analysis, and projectmanagement, which all require sig-nificant advances in technologiesbefore smart machines can performthese jobs. Although they claimedthat such unstructured, non-repeti-tive, and evolving jobs are impossi-ble to eliminate, the students iden-tified various tasks within thosejobs that could be automated with-in the next 10 to 15 years.

More Deep BluesAlthough many findings in the

student reports were quite unusual,they should not be very surprisingupon further reflection. Considerthe chess program Deep Blue,developed by IBM, which defeatedthe world champion Gary Kasparov

in 1997. Or the projects attemptingto automate the art of painting,writing poetry, and composingmusic, such as robotic painterAaron, music-generating softwareEMI, and Kurzweil’s CyberneticPoet, that are described in Ray

and marketing associate functionsdescribed earlier. As another exam-ple, Lehman Brothers Inc. is cur-rently automating payroll and otheradministrative functions. The mainquestion is: How far will the ITindustry be able to advance alongthe three dimensions of Figure 1within the next 10 to 15 years?

The scope and extent of possibleautomation of the symbolic-analyt-ic jobs described is possible onlybecause of the development ofadvanced technologies that canenable these automation processes.It is these technologies that will pro-pel the continued productivityenhancements in the comingdecade. Many smart devices andtechnologies have been developedover the past few years, includingsmart homes, refrigerators, laundrymachines, even tires. These areenabled by so-called smart software

Kurzweil’s book The Age of SpiritualMachines. Although these efforts arestill in their infancy, it is quite possi-ble that significant progress can beachieved in the next 10 to 15 years.And if such highly creative, unstruc-

tured, non-repetitive, andevolving tasksas playingchess, painting,c o m p o s i n gmusic, andwriting poetrycan be auto-mated, thensignificant por-

tions of the work currently per-formed by symbolic-analytic work-ers can also be.

Moreover, the low-hanging fruitsare being picked right now, as is evi-denced from such activities asautomation of the client accounting

30 Sternbusiness

“If such highly creative, unstructured,non-repetitive and evolving tasks asplaying chess, painting, composing

music and writing poetry, can be automated, then significant

portions of the work currently performed by symbolic-analytic

workers can also be.”

that monitors their behavior anddrives and guides these devices.Meanwhile, numerous informationtechnologies help remove the userfrom the loop from various businessprocesses and thus make theseprocesses more automated. Theseinclude Web services that help dis-tributed computer systems interactamong themselves and understandone another without any humanintervention, workflow automation,and document analysis and pro-cessing technologies. Much humaneffort in the knowledge economypertains to the processing of variousmultimedia documents con-taining text, images, video,and audio information. Thislabor-intensive activity isvery difficult to automatebecause it involves naturallanguage understandingand/or computer vision,which constitute two veryhard areas of computer science.However, significant progress hasbeen made in both of these areasover the past several years, and cer-tain types of specialized textualdocuments and images can be ana-lyzed by machines now.

ecent advancements innetworking and wirelesstechnologies will enablethe development of new

automation methods and new waysto redesign business processes. Forexample, RFID tag technologiesmight allow for the elimination ofthe check-out lines in the depart-ment stores and elimination ofmany cashier jobs. The tags wouldalso enable automation of the busi-ness processes in the supply chainsresulting in numerous efficiencyimprovements. And EZ-Pass-liketechnologies are certainly not limit-ed to toll collection applications.The EZ-Pass concept – a person

computing power can be doubledessentially every 18 months. Thatmakes computation-intensiveautomation solutions more feasible.It is important for some of theautomation activities that Moore’sLaw continues to be followed in thefuture.

It is easy to be sanguine aboutthe promise of new technologies,and frequently IT advocates andindustry representatives paint a pic-ture of unblemished progress whendiscussing innovation. However, thenext wave of automation will haveboth positive and negative out-

comes. It will have signif-icant effects on produc-tivity in terms ofimproved efficiencies andincreased productionspeeds which will reducecosts. But these produc-tivity improvements willhave profound effects on

the labor market, with many jobsand job categories being restruc-tured, significantly reduced, oreliminated. Of course, job restruc-turing and elimination in someparts of the economy will result injob creations in other parts of theeconomy. Companies have learnedover the last few decades thatinformation technology can be apowerful competitive weapon thatcan significantly affect the econo-my and the society at large. To beable to respond properly to thiscoming wave of automation thatwill change not only routine pro-duction but also symbolic-analyticjobs, it is crucial to study and dis-cuss the effects of this wave ofautomation before it affects us inprofound ways.

ALEXANDER TUZHILIN is associateprofessor of information systems atNYU Stern.

walking through a monitoringdevice that recognizes the providedservice and automatically bills thisperson – could find numerous appli-cations in all spheres of businesswithin the next several years.

The integration of wireless, loca-tion-based (e.g., Global PositioningSystems, [GPS]) and Web servicestechnologies constitutes a powerfulcombination that would enablenumerous automation applicationswithin the next 10 to 15 years. As anexample, there should be no need forparking meter inspectors in thefuture. When a parking meter

expires, and the car is still located inthe parking spot, computer visiontechnologies could read the licenseplate of the car and the pertinentinformation for issuing a parkingviolation ticket could be wirelesslysent to the central office using Webservices technologies.

Machine-to-machine interactiontechnologies, which facilitate directinteractions between the machines,currently include distributed sys-tems, networking, Web services, andworkflow technologies. However,more complicated and smartermachine interactions will be possiblein the future by integrating othertypes of technologies into the mix,including some of the artificial intel-ligence-based technologies.

Some automation applicationsrequire formidable computingpower. So far, the IT industry hasmet this challenge and continues tofollow Moore’s Law – the notion that

Sternbusiness 31

“These productivity improvementswill have profound effects on thelabor market, with many jobs and job categories being eliminated

or significantly reduced.”

R

32 Sternbusiness

he viability of any com-pany depends on ensur-ing that employees fol-low the organization’sformal rules and proce-

dures. Historically, the dominantapproach to gaining such adher-ence has been a top-down, com-mand-and-control approach.This approach regards employ-ees as rational actors whose pri-mary concern is the maximiza-tion of the outcomes theyreceive from their organization.From this perspective, an orga-nization’s best hopes for realiz-ing employee adherence to rulesand policies lie with linking employ-ee rule-following to the outcomesthey receive. In many companiesthis translates into the developmentand implementation of systems thatprovide incentives (to encouragerule-following) and sanctions (to

clocks, and ordinary managerialmonitoring of employees – is testa-ment to just how prevalent suchsystems are in the modern work-place. In addition to tangible costs,such systems have intangible socialcosts. The command-and-controlapproach typically communicates a

message to employees that theyare not trusted and that theorganization is their adversary(and vice versa), leading to thebreakdown of trust betweenemployees and their organiza-tions.

Evidence also suggests thatthe success of this approach is limit-ed. Employees intent on breakingrules often find ways to do so unde-tected. This often leads organiza-tions to devote even more resourcesto surveillance, further adding tothe tangible and intangible costs ofthose mechanisms. And even then,

discourage rule-breaking). Theintention is that these incentives andsanctions will shape employees’workplace behavior.

However, this approach comeswith significant costs. To implementsuch systems appropriately, organi-zations must be able (and willing) to

devote substantial resources to sur-veillance methods, so that employeesfeel that their rule-following orbreaking behavior will be detected.Indeed, the extensive use of such sur-veillance techniques – cameras, themonitoring of telephone calls andcomputer usage, drug testing, time

COMMAND AND CONTROL?

In most companies, management tries to shape employee behavior through a system ofincentives and sanctions. But creating an environment that encourages self-regulation of

behavior may be a more effective way to ensure that everybody follows the rules.

By Steven L. Blader and Tom R. Tyler

T

Sternbusiness 33

“We predict that employees will be intrinsically motivated to follow

their organization’s rules if they feelthat those rules develop from a

system that is consistent with theirown set of moral values.”

ILLU

STR

ATIO

NS

BY

KE

N O

RV

IDA

S

employees often find new ways tocircumvent detection. The downwardspiral of this approach can be debili-tating to an organization.

Such problems raise the questionof whether there might be an alter-nate strategy for gaining employeeadherence to organizational rules. Wehave been conducting researchexploring just such an alternative: theself-regulatory approach. Rooted insocial psychological research, theself-regulatory approach argues thatemployee rule-following can be bestachieved by activating an intrinsicdesire by employees to follow organi-zational rules. Rather than relying onextrinsic factors such as incentivesand sanctions, our self-regulatorymodel argues that the key toemployee rule-following is to designorganizations that lead employees tointrinsically want to follow organiza-tional rules. In such an environment,

organization are entitled to such con-trol. We predict that employees willalso be intrinsically motivated to fol-low the organization’s rules if theyfeel that those rules are developedand enforced by authorities theyregard as legitimate.

In sum, our research on the self-regulatory approach is based on theprediction that these two judgmentsfoster an intrinsic desire on the partof employees to follow organizationalrules, and, furthermore, that theintrinsic desire they inspire outweighsthe influence of command-and-con-trol mechanisms. We set out to testthis prediction by creating studiesthat compare the relative efficacy ofthese two approaches for promotingthree forms of employee rule-follow-ing: compliance, deference, and rule-breaking. Compliance refers toemployee’s straightforward followingof their organization’s rules.Deference refers to rule-followingthat is specifically discretionary innature – i.e., do employees followrules even when no one, includingtheir bosses, will know that they didso? On the other hand, we also setout to consider the flip side of com-pliance and deference, which isrule-breaking. Rule-breaking refersto conscious decisions by employeesto ignore or violate organizationalrules or policies.

Testing PredictionsThe first study we conducted was

based on confidential questionnairesdistributed to the employees at a divi-sion of a multinational banking firm.The respondents held positions rang-ing from clerical to managerial, withthe bulk of the employees involved indirectly providing banking services tohigh-profile customers. A total of 540surveys were returned. For thoseresponding to the survey, the meantenure with the firm was 13 years, themean age was 42 years, and there wasan average salary of $84,000. Thesecharacteristics mirrored those of the

employees do not need to be coercedinto following rules through the provi-sion of incentives and sanctions.

What might breed an intrinsicdesire among employees to adhere toorganizational policies? In our work,we have focused on the influence oftwo judgments employees make abouttheir work organizations. The first isthe assessment by employees thatthere is congruence between their ownmoral values and those of the organi-zation. That is, we predict thatemployees will be intrinsically moti-vated to follow their organization’srules if they feel that those rules devel-op from a system that is consistentwith their own set of moral values.The second judgment is the assess-ment by employees regarding thelegitimacy of those with power intheir organizations – i.e., whetheremployees believe that those whowield control over the rules of the

34 Sternbusiness

COMMAND AND CONTROL?

broader set of employees working inthis division.

Respondents were asked torespond to a series of questions abouttheir rule-following. Sample ques-tions, with responses on a scale of (1)never to (6) very often, included:“How often do you follow the policiesestablished by your supervisor?” (forcompliance); “How often do you fol-low organizational policies evenwhen you do not need to do sobecause no one will know whetheryou did or not?” (for deference); and“How frequently do you neglect tofollow work rules or the instructionsof your supervisor?” (for rule-break-ing). To assess the two judgmentsrelated to the self-regulatoryapproach, respondents indicatedtheir agreement with statements suchas “Disobeying one’s supervisor isseldom justified” (for legitimacy)and “I find that my values and thevalues of my company are very simi-lar” (for moral value congruence).

Finally, we asked two types ofquestions to assess the perceivedincentives and sanctions for rule-following and behavior (i.e., toassess the command-and-controlapproach). First, to meas-ure the expectancy thatrule-following would bedetected in the first place,we asked respondentsquestions such as “Howmuch attention does yoursupervisor pay to whetheror not you follow workrules?” Second, to meas-ure the value employeesplaced on the incentives orsanctions for rule-follow-ing or breaking, we askedthem questions such as “Ifyou were caught breakinga work rule, how muchwould it hurt you?” Sinceit is possible that thesetypes of judgments (i.e.,judgments regardingexpectancy and value)

ing. Specifically, judgments about thelegitimacy of organizational authori-ties significantly shaped compliance,deference, and employee rule-break-ing. Moral value congruence also sig-nificantly shaped all three forms ofemployee rule-following. Central toour predictions, the analysis found nosignificant unique impact of any ofthe command-and-control judgmentson these forms of employee rule-fol-lowing. The results therefore providestrong support for the hypothesizedimportance of the self-regulatoryvariables.

We also utilized an additional sta-tistical analysis called structuralequation modeling (SEM) to furtherinvestigate the relative influence ofthe self-regulatory and command-and-control approaches. Rather thanexamining the variables includedunder each approach individually,this analysis allows direct examina-tion and comparison of the variablescomprising each approach as a set.The results of these analyses, alsopresented in Table 1, confirm thatthe self-regulatory approach is a moreeffective way of fostering employeerule-following.

interact with one another, such thatpunishments or rewards only matter ifthey seem likely, we also included aninteraction term in all our analyses tocapture any such effects.

e used respondents’ratings to statisticallytest our key predic-tion regarding therelative efficacy of

the command-and-control and self-regulatory approaches for encourag-ing rule-following. Specifically, weperformed regression analyses inwhich we included the command-and-control variables (expectancy regard-ing detection of their behavior, valuelinked to the reaction to their behav-ior, and their interaction) and the twoself-regulatory variables (legitimacyand moral value congruence) andexamined their relationship with eachof the three types of employee rule-fol-lowing. The results of these analysesare presented in Table 1.

Confirming our predictions, theresults indicate that self-regulatoryvariables such as employee viewsabout legitimacy and moral value con-gruence had a unique impact on allthree types of employee rule-follow-

Sternbusiness 35

TABLE 1Study 1: Antecedents of employee adherence to organizational policy

Self-regulatory variables .60*** .58*** -.45***

Legitimacy .29*** .29*** -.10*

Moral value congruence .14** .19*** -.31***

Command & control variables .03 .01 -.02

Detection of behavior (expectancy) .08 .00 .05

Reaction to behavior (value) .16^ -.11 .03

Expectancy X Value -.14 .14 -.18

Variable Compliance withorganizational policy

Deference toorganizational policy

Rule-breaking

Regressionanalysis

SEM1 Regressionanalysis

SEM1 Regressionanalysis

SEM1

1 Two latent variables reflecting: (1) the two self-regulatory variables and (2) the three command-and-control variableswere created. Chi-sq = 373.33, d.f. = 35, CFI = 0.98, NFI = .98, IFI = .98.

Note: ^p < .10; * p < .05; ** p < .01; *** p < .001. n = 540Results control for respondent age, gender, and tenure with the organization

W

COMMAND AND CONTROL?

Additional EvidenceThe first study did not contain

independent ratings of rule-followingbehavior, and thus raises issues relat-ed to self-report of behavior. Forinstance, perhaps respondents werenot accurately reporting their ownbehavior. We conducted a secondstudy to address this limitation. Thisstudy was based on a large sample ofemployees from around the UnitedStates, who worked in a wide varietyof companies and industries.Importantly, we obtained supervisorratings of employee rule-following fora significant subset of respondents toaddress this potential criticism ofthe first study.

tudy Two was basedon responses to aquestionnaire pre-sented to a nationalpanel of respondentsvia WebTV. We

screened to ensure that allrespondents worked at least 20 hoursa week, had a primary supervisor,and had worked at their currentemployer for at least three months.Our final sample included 4,430employees from a variety of organi-zations: 24 percent worked forsmall businesses, 20 percent forlarge companies in one location, 36percent for large multi-city Americancompanies, and 20 percent in multi-national companies. Respondentsanswered questions similar to thosedescribed for Study One. In addition,supervisors of these employeesanswered a survey in which we askedthem about the respondent’s actualrule-following behavior on the job.Supervisors provided ratings ofrespondent’s behavior at work for allthree types of rule-following that weexamine: compliance, deference, andrule-breaking.

First, we examined respondents’self-report of their behavior, to deter-mine whether the effects we found inStudy One were replicated in StudyTwo. As in the first study, variables

ducted study two was to test ourpredictions using supervisor ratingsof employee rule-following.Regression analyses and structuralequation models similar to those con-ducted on the self-report data weretherefore conducted using the super-visor ratings of behavior that we col-lected. The results of these analysesare presented in Table 2.

These regression analyses indicatesignificant effects of legitimacy oncompliance and deference and effectsof moral value congruence on defer-ence and rule-breaking. The onlysignificant effect among the com-

mand-and-control variableswas for the effect of perceivedreactions to behavior (value)on rule-breaking. Results of thestructural equation modelslikewise indicate that, as pre-dicted, the self-regulatoryapproach outpaced the influ-ence of the command-and-con-

trol approach on supervisor-ratedcompliance, on supervisor-rated def-erence to organizational policies, andon supervisor-rated rule-breaking.

n summary, using thesesupervisor reports ofemployee behavior, weagain found that the self-regulatory model provided asuperior account of the fac-tors that shape employeerule-following than the

command-and-control approach.This confirms that employees followorganizational rules, and are per-ceived by their supervisors as follow-ing those rules, when they hold anintrinsic desire to do so. Although thecommand-and-control approach toencouraging rule-following (whichemphasizes threatening employeeswith punishments for breaking rulesand rewarding them for followingthem) may also influence rule-follow-ing, its relative utility is secondarywhen compared with the self-regula-tory approach.

representing the self-regulatory (legit-imacy and moral value congruence)and command-and-control (detectionof behavior, reaction to behavior, andtheir interaction) strategies wereregressed on each of the three types ofrule-following.

By and large, the results confirmedthe findings from the first study.Specifically, legitimacy and moralvalue congruence were related to allthree forms of rule-following, withevidence suggesting that legitimacyhad a somewhat stronger effect thanvalue congruence. However, in StudyTwo the command-and-control vari-

ables had effects not found in StudyOne. In particular, the belief that one’swork behavior would be detected(expectancy) was significantly relatedto all three forms of rule-following,and expectations about the reactionsto detected behavior (value) also hada significant, though small, effect oncompliance and rule-breaking.Structural equation modeling wasagain used as an additional way ofexploring the relative impact of theself-regulatory and command-and-control approaches to employee rule-following. The results of this analysisconfirmed that, as predicted, the self-regulatory approach prevailed overthe command-and-control approachin facilitating deference and for pre-venting rule-breaking. Contrary topredictions, however, the influence ofthe command-and-control approachactually exceeded that of the self-reg-ulatory approach in shaping compli-ance. This was the only finding whichdid not replicate the support for ourpredictions found in study one.

However, the key reason we con-

36 Sternbusiness

“Not only is the self-regulatoryapproach to employee rule-following

more effective, it is also more efficient, since employees take

the responsibility of following ruleson themselves.”S

I

ImplicationsTraditionally, organizations have

subscribed to a belief that the onlyway to get employees to follow orga-nizational rules is to monitor themand then reward or punish them,depending on whether they didindeed follow the rules or not. Suchan approach can get employees to fallin line with organizational expecta-tions, at least to the extent that mon-itoring systems are extensive andreward/punishments sufficientlylarge. However, our studies show thatthis approach generally has a weakerimpact on rule-following than theself-regulatory approach we haveoutlined and tested.

Companies may thus have a greatdeal to gain by going beyond conven-tional instrumental strategies ofsocial control. Further, not only is theself-regulatory approach to employeerule-following more effective, it isalso more efficient, since employeestake the responsibility of followingrules on themselves. This leads to areduction in the command-and-con-trol strategy’s tangible costs as well

strive to engage inactions that supportjudgments amongemployees that theirauthority is legitimate.Additional research wehave conducted alsoemphasizes the impor-tance of fostering organi-zational cultures domi-nated by fairness andtreatment of employeeswith respect. Such fair-ness and respect likewisefoster an intrinsic moti-vation among employeesto follow rules and towork in pursuit of theorganization’s success.

The more generalpoint is that the devel-opment of intrinsicmotivations amongemployees begins at the

top, with the leadership of theorganization. When upper manage-ment does not itself conform to ethi-cal codes of conduct, as appears tohave been the case in several recentcorporate scandals, the legitimacy ofthose authorities is eroded and theperceived congruence of valuesbetween the employee and the organ-ization is diminished. But when theleaders of the organization appeal toemployees’ value systems and presentthemselves as deserving of the powerthey hold, new approaches to foster-ing employee cooperation becomeviable and superior routes to organi-zational success emerge.

STEVEN L. BLADER is assistantprofessor of management andorganizations at NYU Stern.TOM R. TYLER is professor of psy-chology at New York University.

The research discussed in this articleis drawn from an article that willappear in an upcoming issue ofAcademy of Management Journal.

as the intangible toll of polluting theemployee/employer relationship. Inturn, the self-regulatory approachallows organizations to devote greaterorganizational resources to uses thatare more central to the achievement oforganizational goals. It also enablesorganizations to more easily gain theloyalty and commitment of employees.

he results therefore suggestthat one promising way tobring the behavior ofemployees into line withcorporate codes of conduct

is to design organizations in ways thatactivate an intrinsic desire amongemployees to follow rules. The chal-lenge, of course, is to know how todesign organizations in ways thatbreed an intrinsic desire amongemployees to follow organizationalrules. The current findings suggesttwo important ways to foster such adesire. First, organizations shouldstrive to bring organizational practicesinto line with employees’ moral valuesand to make such congruence in val-ues apparent to employees. Second,organizational authorities should

Sternbusiness 37

TABLE 2Study 2: Antecedents of supervisor-rated employee adherence to organizational policy

Self-regulatory variables .81*** .90*** -.67***

Legitimacy .08* .08* -.01

Moral value congruence .04 .08* -.13***

Command & control variables .04 .09 .02

Detection of behavior (expectancy) .10 .12 -.11

Reaction to behavior (value) .19 .22 -.34*

Expectancy X Value -.19 -.19 .31

Variable Compliance withorganizational policy

Deference toorganizational policy

Rule-breaking

Regressionanalysis

SEM1 Regressionanalysis

SEM1 Regressionanalysis

SEM1

1 Two latent variables reflecting: (1) the two self-regulatory variables and (2) the three command-and-control vari-ables were created. Chi-sq = 1,288, d.f. = 41, CFI = 0.82, NFI= 0.82, IFI = 0.82.

Note: ^p < .10; * p < .05; ** p < .01; *** p < .001. n = 4,430. Results control for respondent age, gender, and tenure with the organization

T

orporate leadership fre-quently finds its greatesttests during crises.Characterized by fast

moving developments and an ele-ment of danger, crises can over-whelm executives and present themwith often contradictory tasks andresponsibilities. And it’s difficult toact strategically when minute-by-minute demands require a host ofimmediate decisions. The mediaadds yet another degree of com-plexity to such situations. It can act

affects a company’s financial well-being and image and reputation inthe eyes of critical constituencies –it’s likely a crisis.

The U.S. National WeatherService ranks hurricanes in severityfrom Category 1 to Category 5.Organizational crises can beassessed in a similar way. And justas a Category 1 or 2 hurricane cansuddenly mutate into a Category 4or 5 monster, so organizationalcrises can spiral into unexpectedproportions. And the media’s pres-

as a conduit to important audiencesduring a crisis, or become an obsta-cle to the delivery of messages.

Of course, not every organiza-tional difficulty attracts medianotice. And not every problem is acrisis. When a new product roll-outgets caught in an unexpected patentinfringement suit, it’s definitely aproblem. A crisis? Not necessarily.But when a business problem threat-ens to severely affect the organiza-tion’s normal workflow, when it dis-tracts senior management, when it

Hot OffThe Press

D u r i n g a c o r p o r a t e c r i s i s , t h e m e d i a c a n a g g r a v a t e a n a l r e a d y d i f f i c u l t s i t u a t i o n . B u t b y e m p l o y i n g e f f e c t i v e m e d i a r e l a t i o n s ,

l e a d e r s c a n g u i d e t h e i r c o m p a n i e s s a f e l y t h r o u g h t h e s t o r m .

By Irv Schenkler

38 Sternbusiness

CIL

LUS

TRAT

ION

S B

Y G

OR

DO

N S

TUD

ER

Sternbusiness 39

ence during a crisis raises thepotential for greater organizationalharm.

Understanding which categoryof crisis a company faces can helpdetermine the media sources bestsuited to deliver a message. Andbecause crises are by definitionvolatile, the nature of a crisis mayshift from one category to another.These shifts occur when a new setof stakeholders become affected orconcerned and when the mediatracks this new-found interest.

Varieties of CrisesCrises come in several varieties.

If a company is facing bankruptcy,hostile take-over, a strike, or mas-sive employee lay-offs, the crisisoriginates in the financial arena,and financial media and businessreporters will naturally have themost initial interest. Examples ofnotable financial crises include theSalomon Brothers Treasury Bill cri-sis of 1991-92 and Enron. Whenthe courts form the battleground –in litigation crises – the adversarialsystem lends itself to vocal, antago-nistic claims from plaintive parties,which can find their way onto thefront page. Examples of such crisesinclude the long-running Dow-Corning breast implant crisis of the1990s and the Microsoft-U.S. anti-trust battles.

Some business issues becomepopular crises, which can gainstrength over time and can fomentthe need for legislative action.Frequently, these crises arise whenthe issue’s advocacy group canenlist media to publicize its sense offear, or when an issue speaks to aconstituency’s sense of moral orlegal rights. Examples includeNike’s use of Third World factoryproduction and Nestle’s marketingof infant formula in Third Worldcountries in the 1970s and 1980s.

journalism’s excesses. The goal oftitillation has plenty of critics. Butnews directors at competing sta-tions will still risk the wrath ofpolice chiefs by flying so many newschoppers above a breaking storythat they impede the police.Assigning blame is a consistent goalfor the press – and something com-panies want to avoid. Finally, thenational exposure that can accom-pany an individual journalist’s cov-erage of a crisis can be a spring-board for job offers and careeradvancement.

An organization under intensemedia scrutiny can respond with avariety of overall strategies (SeeFigure 1). At root, leaders have tomake decisions in two dimensions:how aggressive or passive a compa-ny responds, and how free a compa-ny feels about its options.

Based on those decisions, fourinitial strategies emerge: “Free toattack” is the strategy that can workonly when a company has a highlycredible story to tell. The underlyingfacts must support the messageand the tactics employed must take

Finally, crises can be caused byphysical events such as fires, earth-quakes, hurricanes, contamination,or criminal actions. Examples ofsuch physical crises include ThreeMile Island in 1979 or the Exxon-Valdez oil spill of 1990. In mostcrises, a primary objective is to keepthe situation anchored in one of thecategories and reduce the likelihoodthat it will shift to another.

If handled properly, the mediacan be an important ally. But onehas to understand the five key driv-ers that spur the media’s crisis cov-erage in order to do so.Fundamentally, informing is themedia’s business – who, what,when, where, and why. Failing torespond to questions is thus a pre-scription for confrontation withreporters. In physical crises – suchas natural disasters and chemicalspills that threaten communities –the public and government rely onthe media, especially electronicmedia, to convey information andaid to the public. The downside ofmedia coverage is the press’ tenden-cy to play to the worst of tabloid

40 Sternbusiness

Figure 1 CRISIS RESPONSE STRATEGIES

Freeto

Choose

Forcedto

React

Forced toDefend

Forced toAvoid

Free toAvoid

Free toAttack

Aggressive

Passive

into account the culpabilitylurking beneath the accu-sations. In the 1990s, thisstrategy was adopted withvarying degrees of success.In 1993, NBC’s Datelinebroadcasted footage osten-sibly showing a Chevroletpick-up explode as a resultof gas tanks positionedbeneath the driving cab.General Motors learnedthat the sequence had beenstaged, using incendiarydevices to set off the explo-sions. GM obtained out-takes, called a news confer-ence, and exposed thedeception. NBC apologizedpublicly and fired thoseresponsible.

When choices are limit-ed and a companynonetheless feels compelledto react, it is “forced to defend.”For such an approach to work, thecompany needs to receive a recep-tive hearing from the media or, atleast, get its version of events deliv-ered in a credible context. Accusedof underpaying and exploitinglabor in the Third World, Nike inthe 1990s defended its businessstrategy by claiming financial ben-efits for the shoe workers. Later,Nike defended its actions by claim-ing it did not directly run thesneaker factory. Eventually, thecompany adopted a problem-solv-ing strategy – to be discussed laterin this article – and instituted aseries of remedies, including over-seers and improved accountability.

The “forced to avoid” approachpositions the company in passivemode, in which conditions force akind of silence. Because of extenu-ating circumstances or lack ofinformation, this strategy is mostfrequently framed by the message,“We are unable to comment at this

Free to IgnoreIf a company believes it

cannot be harmed by mediacoverage, then it may opt notto make spokespersons avail-able. This strategy has oftenbeen adopted by industrialsuppliers or non-consumerbased companies. Privatelyheld companies also oftenbelieve that this response tomedia inquiry is the best. Andinternational companies withU.S. based subsidiaries some-times adopt this approach;their cultural misreading ofthe media’s role in the U.S.business scene leads them tobelieve that what works in thehome country – silence – willwork well here.

The danger behind thisstrategy lies in its assumptionsof insularity and strength.

Most companies in fact do need toconsider the secondary effects thatresult from the media’s coverage ofits actions. And privately held com-panies ultimately have customerswho are susceptible to media influ-ence. During the 1980s, companiesfelt particularly free to ignore mediainquiry. Hooker Chemical, whichadopted a no-comment policy in thewake of the Love Canal environ-mental crisis, was a prime example.By the time the company respondedto allegations and lawsuits, its cred-ibility was profoundly undermined.Unanswered, third-party commentswill accumulate, potentially affect-ing image and reputation.

Overlooked StrategyA fifth strategy is frequently over-

looked: problem-solving. Whenevera company can position its responseas a meaningful effort to acknowl-edge and correct the phenomenonthat led to the crisis, media cover-age will become more favorable and

time.” Legal departments will oftencounsel using this approach since itoffers apparent protection from mis-cues and misquotes that could beused against the firm. Even whenjustified legally, however, this strate-gy puts the company in a vulnerableposition. In the court of public opin-ion, silence or the refusal to defendoneself is equated with guilt.Meanwhile, any number of otherinterested parties will be sure tovoice their reactions. In 1999, whenthe Bank of New York was chargedwith money laundering for organ-ized crime, the firm adopted thisstrategy. However, information wasleaked to the press, apparently fromthose close to the investigation.Coverage increased as reporterssearched for the names of responsi-ble parties to blame. The Bank ofNew York, with its patrician reputa-tion in the crucible, suffered frommedia over-exposure and tabloid-like investigation.

Sternbusiness 41

“In the court of public opinion,silence or the refusal to defendoneself is equated with guilt.”

stakeholder impressions will in thelong run not impugn the company’sreputation.

Figure Two illustrates how thisapproach can be placed on the orig-inal model’s horizontal axis span-ning “free to respond” to “forced torespond.” Thus, a company can beproactive and “free to solve.” Forexample, a company can appoint anoutsider to oversee new personnelpolicies in the face of federal dis-crimination suits. Or it can beapologetic – “forced to solve” – aswhen it acknowledges system orworker error as the cause of anaccident.

problem-solving strategy,however, rarely comes tomind for most managers.Reasons range from con-

cerns about admitting legal liabilityto the cultural reluctance of com-petitive, success-oriented businessexecutives to admit error. But whena company can position its responseas a meaningful effort to acknowl-edge and correct the phenomenonthat has led to the crisis, media cov-

months, however, the results werean increase in negative media cov-erage, a U.S. boycott, and protestsat Tokyo corporate headquarters.The company changed course,replaced the management of theU.S. subsidiary, and moved into the“mutual problem-solving” mode.An outside overseer was appointedto examine and change employmentpolicies.

By contrast, A.H. Robins, whichmade the Dalkon Shield, employeda “free to ignore” strategy whenconfronted with scores of reportsthat the controversial birth controldevice was malfunctioning andcausing harm. When media atten-tion heightened and lawsuits wereabout to be filed, it shifted to “freeto attack” and attempted to vilifythe women who brought suit,alleging an assortment of unsavorypersonal behaviors as the cause ofthe malfunctions. This damagedRobins’ credibility and reputationimmensely. The company wentbankrupt and was later acquired bya competitor.

Effective TacticsOnce a strategy is chosen, com-

panies must assess which tacticswill best help achieve their desiredobjectives. Regardless of the strate-gy, all corporate leaders should fol-low one inviolate rule: Do not lie tothe press. Lying to the press is likethrowing blood into the shark tank.Lies are always found out. Beechnutfound out the hard way when its“100%” apple juice was found tocontain a cocktail of sugar andwater and very little real fruit juice.The company was fined $2 millionand its president pleaded guilty tofelony charges.

Companies must also assessreporters’ motivations. Reporterscan lie, too. It is incumbent uponexecutives to conduct due diligence

erage will usually become morefavorable and stakeholder impres-sions will not ultimately impugn thecompany’s reputation.

A notable example of the prob-lem-solving approach came in 1996.When the oil giant Texaco wasaccused of racial discrimination bythe Equal Employment OpportunityCommission, Texaco CEO PeterBudjar pledged an impartial investi-gation, brought in a respected out-side jurist to conduct it, and publiclyacknowledged the need for improve-ment in hiring and promotion proce-dures. Media coverage, which wasintense, waned.

While each of these five strategiesmay represent an appropriateresponse to crisis, it’s also importantto realize that shifts in strategy mayoccur – or become necessary. Mostcommonly, strategic shifts will mili-tate towards “mutual problem-solv-ing.” For example, in 1996, theEEOC sued Mitsubishi over sexualdiscrimination in the workplace.The company first responded in a“free to attack” mode. After several

42 Sternbusiness

Figure 2 CRISIS RESPONSE STRATEGIES: PROBLEM-SOLVING

Freeto

Choose

Forcedto

React

Forced toDefend

Forced toAvoid

Free toAvoid

Free toAttack

Aggressive

Passive

Mutual Problem Solving

A

on the media, to know withwhom they’re dealing. Andleaders should never committo put themselves or othermembers of their organizationon the phone or in front of thecamera unless they are trainedand capable. A more personaltouch from the top mightensure the best possible stories inone or two key business and trademedia. Executives should alsoavoid saying “no comment.” Thephrase has come to be associatedwith admission of guilt. It’s farbetter to provide a sense of theprocess involved. Finally, theyshould avoid the blame trap. Formany organizations, the instinctiveresponse to accusations of blame isfirst to deny and then later find ascapegoat. Yet nothing sets off afeeding frenzy among the pressmore quickly than an attempt toshift blame that doesn’t stand up toscrutiny.

very crisis response mustbalance responsiblebehavior with protectingreputation. Those are not

mutually exclusive concepts.Accepting responsibility can be dif-ferent from taking the blame. It canalso be the best way to move for-ward to address the real crisis, andat the same time develop supportfrom the general public, the media,and other key audiences.

Let’s contrast the handling oftwo similar crises by two oil com-panies. In 1989, the Exxon Valdezspilled tons of oil into pristinePrince William Sound in Alaska.More than a decade later, Exxonstill is vilified by many for its mis-handling of that crisis. Seeminglyat every turn, Exxon’s response washostile and combative. First, thecompany tried to assign blame tothe single individual – the “drunk-en” boat captain. This tactic

Avoiding the media maywork sometimes, but in a timeof crisis, it won’t work forlong. The press can find toomany other sources of infor-mation – disgruntled employ-ees, state environmental offi-cials, competitors – many ofwhom may be more than

happy for the media exposure.However, occasionally avoidingcalls may be an appropriate short-term strategy – and the best way totemporarily delay media contact isto issue a “holding statement” thatdoesn’t misrepresent current cir-cumstances and provides enoughinformation to fend off additionalquestions.

good rule of thumbagainst which to measurecrisis response is to take“the 60 Minutes Test,”

named after the grand-daddy of allinvestigative television programs.Executives should answer threequestions: What did you know?When did you know about it? Whatdid you do once you knew about it?Acknowledging an appropriate levelof responsibility and helping drivetoward solutions is the best way topass this test and win acquittal inthe Court of Public Opinion. Whenit comes to reminding the public ofalleged or actual corporate errors,missteps or misdeeds, the mediasuffers no amnesia. As a NativeAmerican proverb instructs: “Don’tshoot an arrow that will returnagainst you.”

I R V S C H E N K L E R is c l in ica lassociate professor of managementcommunication at NYU Stern.

This article is adapted from Guide toMedia Relations, by Irv Schenkler andTony Herrling (Prentice Hall, 2004).

begged the question as to whetherExxon had put too much responsi-bility in one set of shaky hands, andwithout adequate backup systems.Then, the company appeared tofight cleanup efforts andbesmirched those with concernsabout the pollution of the sound,creating fresh enemies at everyturn. Each of those strategiesensured that Exxon’s name wouldforever be associated with a well-covered disaster.

In contrast, few today rememberthat Ashland Oil experienced its owndisastrous spill. In 1988, 700,000gallons of diesel fuel poured from aruptured Ashland tank and was car-ried by the Monongahela River intothe Ohio River, threatening thedrinking water of Pittsburgh and anestimated one million people inPennsylvania, West Virginia, andOhio. Because Ashland’s CEO insist-ed that local media be apprisedimmediately of the situation andwhat the company was doing aboutit, the story remained under control.The company signaled that it wasmore important to accept responsi-bility and do something about thecrisis, rather than stop to figure outwhether real blame lay with thebuilder of the storage tank, or themanufacturer of the steel fromwhich the tank was constructed. Bykeeping the media informed,Ashland was able to limit conjectureand rumor and to reduce the ava-lanche of criticism that such a sig-nificant oil spill would normallyproduce.

Sternbusiness 43

“Whenever a company can position its response as a

meaningful effort to acknowledgeand correct the phenomenon that led to the crisis, media

coverage will become more favorable.”

EA

250,000 units. George Parker, thenin his late 60s, turned to one of hisyouthful rules: “Bet heavily whenthe odds are long in your favor.”Parker in 1936 developed six dif-ferent editions of the game, rangingin price from $2 to $25. In 1936,1.81 million copies were sold.

George Parker died in 1952 atthe age of 86. In the post-waryears, Parker Brothers introduced asuccession of smash hits: Risk,Careers, and Trivial Pursuit. Thecompany, now a unit of Hasbro,has maintained its leadership byadhering to Parker’s simple rulesabout business – and games. Themost universal, perhaps, is RuleNo. 4: “Learn from failure; buildupon success.”

DANIEL GROSS is editor of STERNbusiness.

enerally speaking, chil-dren’s games aren’tseen as having muchbearing on the busi-

ness world. Sure, most companieswould like to play Follow theLeader, or King of the Mountain, orCapture the Flag. But as any man-agement consultant worth hishourly fee will tell you, business isnever that simple.

Nonetheless, for 120 years, oneboard-game company – ParkerBrothers – has managed to makemarket leadership look like child’splay. And at the center of ParkerBrothers’ unlikely story, told by for-mer executive Philip E. Orbanes inThe Game Makers (HarvardBusiness School Press, 2003), liesseveral crucial insights on the busi-ness of games, and on the game ofbusiness.

In 1883, George Parker, wholived near Boston, modified a simplecard game to make a game he called“Banking.” When big-city publish-ers turned down his ideas, he print-ed up 500 copies and sold Bankingout of a suitcase. In the 1880s, asOrbanes noted, Parker “came tosee a relationship between thestrategies that guided success inparlor games and the principles thatenhanced success in the ‘game’ ofbusiness.” In fact, he codified themin 12 rules, many of which sound asif they were ripped from present-day management tomes: “Knowyour goal and reach for it;” and

endpaper

G

44 Sternbusiness

Follow The Leader By Daniel Gross

“When faced with a choice, makethe move with the most potentialbenefit versus risk.”

Parker Brothers’ largest and mostlasting breakthrough came withanother business-oriented game. In1934, when the company’s sales hadfallen by two-thirds from the 1929level, George Parker was offered theopportunity to purchase a game inwhich players could amass riches bytrading properties and railroads, andby building hotels and houses.Parker politely passed. But after itsinventor, failed plumbing salesmanCharles Brace Darrow, publishedMonopoly himself and it caught on,Parker Brothers was offered a secondchance.

The company acquired Monopoly,and launched it nationwide duringthe 1935 Christmas shopping season.It became an instant hit, selling

ILLU

STR

ATIO

N B

Y K

EN

OR

VID

AS

Office of Public Affairs44 West Fourth Street, Suite 10-160 New York, NY 10012

NON-PROFIT ORG.U.S. POSTAGE

PaidNew York, NYpermit no.7931