. 11
( 13)


and Gachter and Falk [2002]). In these experiments, there are two types
of subjects (employers and workers) and two stages of interaction.
During the ¬rst stage, each employer makes a wage offer, which is
either accepted or rejected by some worker. Acceptance leads to em-
ployment and to the second stage, where either the worker or the ex-
perimenter chooses an effort level. An employer can employ only one
worker, and a worker can work for only one employer. An employed
worker™s payoff is the wage minus a cost, which is increasing in the
effort level. The employer™s payoff increases in the effort level and,
of course, decreases with the wage.
Notice that the employer has no way to enforce the worker™s effort
choice. The two stages are repeated, usually 10 to 15 times. In some
experiments, one worker and one employer are paired for all the repe-
titions. In others, the experimenter changes the pairings after every
repetition. In still another version, the pairings are established at each
repetition by competitive bidding for workers and jobs. In such market
interactions, there are more workers than employers, so that market-
clearing wages should be little more than the workers™ reservation
level, which equals their cost of effort.
Experimenters consistently ¬nd that if workers choose the effort
level, the average wage is considerably higher than the reservation
level, even when competitive bidding should force wages down to it.
Furthermore, the workers™ average effort is higher than the minimum
allowed and increases with the wage offered. In addition, the wage is
little more than the reservation level if the experimenter chooses the
effort level and there is competitive bidding with an excess supply of
labor. These results hold even when the employer and worker interact
only once. That is, workers offer extra effort in exchange for a higher
than minimal wage, even though wages are agreed on before workers
choose effort levels and employers never have another opportunity to
reward or punish workers. Employers anticipate and exploit workers™
reciprocity by offering generous wages.
This series of experiments show that only some people reciprocate.
Others behave sel¬shly and offer the minimum amount of effort.
Some employers who would otherwise behave sel¬shly are probably
320 Bewley

induced to offer generous wages by the expectation that some workers
will react to them by offering liberal amounts of effort. Because wages
fall to minimal levels when the experimenter ¬xes the effort level, we
may tentatively conclude that employers™ behavior is driven mainly
by the expectation of reciprocation, not by a sense of fairness”that is,
by a desire to divide evenly the economic surplus generated by the
worker-employer interaction.
The tendency to reciprocate may be built into the human psyche.
Rilling et al. (2002) used magnetic resonance imaging to study the
reactions of the brain during repeated play of the prisoner™s dilemma
game and found that experiencing cooperative responses and deciding
to cooperate were both accompanied by patterns of brain activity nor-
mally associated with pleasure.2
All these ¬ndings support the explanation of wage rigidity proposed
by Akerlof (1982) in his gift exchange model. I pointed out earlier
that this theory does not seem to apply in a business context because
workers quickly grow to believe that they deserve whatever pay they
receive. Experiments do not continue for long enough to capture this
habituation effect.
What is important about the experiments is that they reveal that a
signi¬cant fraction of the population reciprocates. In addition, the
experimental ¬ndings do re¬‚ect some of the practices that managers
explained to me. When setting the pay of new hires, recruiters some-
times offer a little more than applicants expect in order to get the
relationship with them started off on the right footing and to create ex-
citement about the new job. One of the many reasons recruiters dislike
hiring overquali¬ed applicants is that they are likely to be disgruntled
because their pay disappoints their expectations.
Fehr and Falk (1999) performed interesting modi¬cations of the
experiments of Fehr, Kirchsteiger, and Riedl (1993) and others de-
scribed earlier in this chapter. Fehr and Falk make the bidding for
jobs and workers two-sided rather than one-sided in the situation with
competitive bidding and an excess supply of workers”that is, workers
as well as employers can make wage offers. The authors found that
when the experimenter determines the effort, level employers accept
only the lowest offers, and wages are forced down almost to the reser-
vation level. When the workers choose the effort level, however, the
wage is higher, just as in experiments where only employers make
offers. Workers make many low offers to try to obtain a job, but these
are refused, apparently because the employers hope to incite high ef-
Fairness, Reciprocity, and Wage Rigidity 321

fort by paying good wages. The experimental employers™ behavior cor-
responds to that of actual ¬rms that usually refuse workers™ offers to
work for very little.
Burda et al. (1998) have performed experiments involving wage
cuts. In their work, an employer and worker are matched for two peri-
ods, and in each of them the employer makes a wage offer, which the
worker may accept or reject. If the worker rejects the offer, the em-
ployer may (after paying a ¬xed training cost) hire a ¬ctitious worker
at a market wage, which the actual worker also receives (as if hired by
some other ¬ctitious ¬rm). The market wage is predetermined by the
experimenters and declines from the ¬rst to the second period. In the
experiments, there is little wage rigidity; the wages that employers
and employees agree on tend to decline along with the market wage.
The employer and worker in effect play two successive ultimatum
games, the bargaining position of the worker weakens from the ¬rst to
the second game, and as a result the wage declines. There is no recipro-
cation of effort for income that could give rise to an insult effect, and
the standard of living effect does not apply, since the workers do not
live from their earnings. The experiments, therefore, provide evidence
that without these two effects wages would be downwardly ¬‚exible.
Experimental evidence supports the view of businesspeople that ¬-
nancial incentives are effective, even when negative, provided they are
not presented in a hostile manner. For instance, Nagin et al. (1998) re-
port on a ¬eld experiment performed by a telemarketing ¬rm. In this
¬rm, the telemarketers™ pay increased with the number of successful
solicitations they claimed, and the company monitored these claims by
calling back a fraction of the people declared to be successes. The com-
pany secretly varied the fraction of bad calls reported to employees
while increasing the true call back rate. By analyzing the company™s
data, the authors found that cheating increased as the fraction of bad
calls reported declined, so that workers did respond to variation in the
negative incentive.
¨chter (1998a) and
Laboratory experimental work by Fehr and Ga
Brown, Falk, and Fehr (2002) shows that the possibility of negative
rewards does not keep reciprocation from being a powerful incentive.
Fehr and Gachter (1998a) performed the two stage experiments of
Fehr, Kirchsteiger, and Riedl (1993) with the modi¬cation that at stage
one, the employer requested an effort level. The authors compared the
results with experiments where in a third stage the employer could re-
ward or punish the worker. The amount of the reward or punishment
322 Bewley

was chosen by the employer and was not announced in advance. The
employer incurred a cost that increased with the absolute magnitude
of the reward or punishment. Despite the cost, many employers did
reward high effort and punish low effort, and workers on average
offered more effort and earned lower wages in the three stage than in
the two stage experiments.
Brown, Falk, and Fehr (2002) repeated the two stage experiments
of Fehr, Kirchsteiger, and Riedl (1993) 15 times under two condi-
tions. Under one, employers and workers could identify each other by
a number, and employers could make offers to a particular worker.
This arrangement made it possible for an employer and worker to
form a long-term relationship. In the other condition, the identifying
numbers were reassigned in every period, so that long-term relation-
ships were impossible. When identity numbers remained stable, in-
dividual workers and employers did form relationships that were
valuable to both, because they could establish a pattern of exchanging
high effort for high wages. Employers could and many did punish
workers for low effort by dismissing them”that is, by ceasing to
make them offers. Average wages and effort were considerably higher
when identity numbers were stable than when they were reassigned,
so workers were not discouraged from reciprocating by the threat
of dismissal. The fact that the negative incentives were not made ex-
plicit may have diminished any bad impression they made in the
experiments of Fehr and Gachter (1998a) and Brown, Falk, and Fehr
(2002). Another explanation for the effectiveness of the negative incen-
tives may have to do with the presence of both sel¬sh and reciprocat-
ing workers. Although the reciprocating workers might have been
offended by the possibility of punishment, sel¬sh ones might have
been induced to offer more effort by the prospect of reward and risk of
Other experiments that imitate the no shirking model provide addi-
tional evidence that punishments do not crush reciprocation and dis-
courage effort. These experiments are described in Fehr, Kirchsteiger,
and Riedl (1996); Fehr, Gachter, and Kirchsteiger (1997); Fehr, Klein,
and Schmidt (2001); and Fehr and Gachter (2002). The experiments
have the form of the two-stage experiments described in Fehr, Kirch-
steiger, and Riedl (1993), except that the employer requests a certain ef-
fort level and a worker is ¬ned with a ¬xed probability if the effort
level offered falls short of that demanded by the employer”that is, if
the worker shirks. In its offer, the employer speci¬es a wage, the ¬ne,
Fairness, Reciprocity, and Wage Rigidity 323

and the effort level demanded. The no shirking model of Shapiro and
Stiglitz (1984) also includes a probability of a worker™s being caught
shirking, and the ¬ne in the experiment corresponds to being ¬red.
One ¬nding is that the threat of being ¬ned elicits more than the mini-
mum possible level of effort. Also, some reciprocation exists, in that
employers obtain effort above the level they demand when they offer
generous wages. Probably because employers hope for reciprocation,
they often request effort levels that are too high to be enforced by the
¬ne. The average level of actual effort is reduced by a considerable
amount of shirking that may re¬‚ect reciprocation of the hostility per-
ceived in the possibility of being ¬ned.
The evidence is mixed on the degree to which the speci¬cation of
¬nes discourages reciprocity. Fehr, Klein, and Schmidt (2001) and
Fehr and Gachter (2002) compare experimental labor relations models
imitating the no shirking model (as in Fehr, Kirchsteiger, and Riedl
[1996]) with labor relations models that depend solely on reciprocity
or trust (as in Fehr, Kirchsteiger, and Riedl [1993]). In the trust model,
the employer offers a wage and makes a nonbinding effort request,
and the worker then offers an effort level. The no shirking model is as
described in the previous paragraph. The two papers report opposite
results. In Fehr and Gachter (2002), the trust model achieves higher
actual effort than the no shirking model.3 In Fehr, Klein, and Schmidt
(2001), the no shirking model achieves higher effort. I see no way of
explaining the discrepancy, as the payoffs are nearly the same in the
two experiments and the differences between them do not seem rele-
vant.4 Fehr and Gachter (2002) go on to make another comparison that
shows that the ¬ne may vex workers to some extent. Fehr and Gachter
compare the no shirking model with a mathematically equivalent bo-
nus model, in which the punishment is deprivation of a bonus rather
than a ¬ne. The bonus model gives rise to greater effort than the no
shirking model, but less than the trust model.
Further experimental evidence of the harmful effects of negative
incentives is contained in Fehr and Rockenbach (2002). In their experi-
ments, subjects play a game, in which an investor chooses a quantity of
money to give to a respondent and speci¬es the amount he or she
would like the respondent to return. The amount given is tripled by
the experimenter, so if the investor gives x the respondent receives 3x.
The respondent then chooses how much to return to the investor. In
another version of the game, the investor, when making the gift to the
respondent, may commit to imposing a ¬ne of a ¬xed magnitude on
324 Bewley

the respondent if he or she returns less than the amount requested by
the investor. On average, respondents were least generous when the
¬ne was imposed, more generous when there was no possibility of a
¬ne, and most generous when the investor could impose a ¬ne but
chose not to do so.
Two papers by Falk, Fehr, and Fischbacher (2000, 2003) provide ex-
perimental evidence that perceived intentions as well as the desire for
a fair division affect reciprocation. Falk, Fehr, and Fischbacher (2000)
report on experiments with a variant of the game ( just described) of
Fehr and Rockenbach (2002). On the ¬rst move, the investor may take
money away from or give money to the respondent, and the respon-
dent may then in turn give or take money away from the investor.
In another version of the game, the experimenter determines the
investor™s move according to a random distribution. In both versions,
respondents on average react by taking money back if it is taken from
them and give money back when it has been given to them. Their
responses are, however, of a larger magnitude when the ¬rst move is
chosen by the investor rather than by the experimenter.
This behavior shows that the respondents™ behavior was driven to
some extent by a desire to even the winnings from the game, but above
all by an urge to reciprocate the good or bad intentions of the investor.
Falk, Fehr, and Fischbacher (2003) reach the same conclusion from
experiments with various ultimatum games. Player A can propose one
of two possible splits of 10 monetary units to a respondent. One possi-
bility is always an °8; 2Þ split”8 for the proposer and 2 for the respon-
dent. Alternatives are °5; 5Þ, an even split, or °2; 8Þ, °10; 0Þ, or even
°8; 2Þ, the last of which means that there is really no alternative.
Respondents reject the °8; 2Þ split more frequently the less fair it
seems in comparison with the alternative. For instance, °8; 2Þ is re-
jected most often if °5; 5Þ is the alternative and least often if °10; 0Þ is
the alternative.
These results provide some”but not strong”support for managers™
assertions that using ¬ring systematically to stimulate effort would
dampen morale and depress productivity. I suspect that the effects
managers refer to are dif¬cult to capture experimentally, because ¬ring
is a much more severe punishment than can be imposed in the labora-
tory, and it is hard to reproduce in a laboratory the menacing atmo-
sphere that could be created in a workplace by frequent ¬rings or by
the threat of ¬ring.
Fairness, Reciprocity, and Wage Rigidity 325

11.5 Evidence from Organizational Psychology and Managerial

Although early investigations by managerial scientists and organiza-
tional psychologists of the relations between pay, morale, and produc-
tivity contradicted some of what managers say about these matters, the
subject has since evolved and now much of what managers say is
being corroborated by research. Recall that managers assert that pay
levels have little impact on motivation or performance, but that ¬n-
ancial incentives linked to performance can increase productivity con-
siderably. These conclusions have been supported by a large amount
of research by management scientists and psychologists, which I do
not describe. The relevant literature is reviewed in Vroom (1964, 252)
and Lawler (1971, 133).
The management intuitions that did not receive much support in
early research had to do with the link between morale and produc-
tivity. Morale was measured from questionnaire evidence on job sat-
isfaction, organizational commitment, and loyalty. Performance was
measured through direct observation or by supervisors™ evaluations.
There are many valuable reviews of the large amount of literature on
these management topics (Bray¬eld and Crockett 1955; Herzberg et al.
1957, chapter 4; Vroom 1964, 181“186; Locke 1976, 1330“1334; Iaffal-
dano and Muchinsky 1985; and Mathieu and Zajac 1990). The general
conclusion is that the correlations between the measures of morale and
performance are positive, but small. The measures of performance in-
clude those of both individuals and groups. In a way, these ¬ndings
con¬rm what managers say, because most of them assert that good
morale is not the same as happiness. There is a considerable amount
of evidence that job satisfaction is negatively related to quitting and
absences. The literature on this subject is reviewed in Bray¬eld and
Crockett (1955), Herzberg et al. (1957, 106“107), Vroom (1964, 175“
180), Locke (1976, 1331“1332), Price (1977, 79), Steers and Rhodes
(1978), Mobley (1982, 95“105), Staw (1984, 638“645), and Mathieu and
Zajac (1990).
There was interesting research in the 1950s that did support manage-
ment feelings about the importance of morale. The investigators made
experimental changes in management practices to determine the rela-
tion between work groups™ attitudes and performance (Viteles 1953,
chapter 8; Seashore 1954; Whyte et al. 1955 and 1961; and Likert 1961,
326 Bewley

chapter 3). A main conclusion was that performance is positively asso-
ciated with pride in the work group or ¬rm, but is not related to other
In response to the failure to ¬nd a signi¬cant relation between job
satisfaction and performance, researchers studied the link between job
attitudes and workers™ doing things for employers that are outside of
their normal duties. Contact with business may have led scholars to
look for such a connection, because managers claim that the impact of
good morale on productivity is felt mainly through employees™ will-
ingness to do more than the minimum required of them. Doing more
than the minimum has been given various names, such as spontaneous
behavior (Katz 1964), prosocial behavior (O™Reilly and Chatman 1986;
Brief and Motowidlo 1986), extra-role behavior (O™Reilly and Chatman
1986), and most commonly organizational citizenship behavior (Organ
1988). These concepts differ to some extent. Dennis Organ de¬nes
¬ve categories of organization citizenship behaviors: altruism (helping
other workers), conscientiousness (obeying company rules), sports-
manship (good humored toleration of inconveniences), courtesy (con-
siderate treatment of fellow workers), and civic virtue (participation in
the internal political life of the organization).
A ¬rst question is whether good morale increases organizational citi-
zenship behavior. Organizational psychologists have done most of the
research on this topic. They typically start with a number of loosely
de¬ned concepts, such as job satisfaction, perceptions of fairness in the
work place, and organizational citizenship behavior, and then try to
determine how these are related by analyzing responses to question-
naires from a sample of several hundred people. Each concept is usu-
ally broken into several components, such as Organ™s ¬ve categories of
organizational citizenship behavior, and a list of questions is associated
with each. Employees answer questions on job satisfaction and percep-
tions of fairness, and employees or their supervisors answer questions
on organizational citizenship behavior. Factor analysis is used to check
whether responses to the questions are such that those corresponding
to one conceptual component are highly correlated with each other
and have less correlation with responses to other questions. The rela-
tions among the concepts and their components are then estimated us-
ing regression analysis, which is used in nearly the same way that it is
in economics. The advantage of such surveys over laboratory experi-
ments is that they can investigate real-life situations where there are
long-term associations between workers and employers, whereas the
Fairness, Reciprocity, and Wage Rigidity 327

subjects in laboratory experiments are usually college students. The
disadvantage of surveys is that it is much harder to establish causation
than it is with experiments.
The ¬ndings of organizational psychologists do not all agree, but
their work supports the conclusion that typical measures of morale,
such as job satisfaction and organizational commitment, do have a pos-
itive relation with organizational citizenship behavior. What is more
important is that a perception of fairness within a business organiza-
tion has a positive relation with both job satisfaction and organiza-
tional citizenship behavior and may be the dominant factor affecting
both. Furthermore, procedural justice”especially the interactional
aspect of procedural justice”is more closely related to job satisfaction
and organizational citizenship behavior than is distributive justice. Dis-
tributive justice has to do with the actual allocation of rewards to
employees, whereas procedural justice has to do with the system used
to arrive at the allocation. Interactional justice has to do with the
consideration, politeness, and respect with which superiors treat their
Another conclusion is that organizational citizenship behavior de-
pends less on employees™ mood than on their conscious perceptions
about their jobs. The impact of fairness on organizational citizenship
behavior is discussed in Organ and Konovsky (1989); Moorman (1991,
1993); Folger (1993); Moorman, Niehoff, and Organ (1993); Niehoff and
Moorman (1993); Podsakoff and MacKenzie (1993); Organ and Ryan
(1995); Konovsky and Organ (1996); Netemeyer et al. (1997). Moorman
(1991) discusses the relative impact of the various forms of justice. The
impact of mood is discussed in Organ and Konovsky (1989), George
(1991), and Moorman (1993). The relative impacts of mood and cogni-
tive job satisfaction are discussed in Organ and Konovsky (1989) and
Moorman (1993). The impact of job satisfaction and commitment
on organizational citizenship behavior is discussed in O™Reilly and
Chatman (1986); Puffer (1987); Farh, Podsakoff, and Organ (1990);
Moorman (1991); Organ and Lingl (1995); Organ and Ryan (1995);
Konovsky and Organ (1996); Netemeyer et al. (1997); and MacKenzie,
Podsakoff, and Ahearne (1998). Good reviews of the impact of Fair-
ness, on organizational citizenship behavior are Organ (1988, 1990),
Schnake (1991), Greenberg (1993), and Organ and Moorman (1993).
Another connection between morale and organizational citizenship
behavior is made through studies of the impact of leadership style
on subordinates™ organizational citizenship behavior. A distinction
328 Bewley

is made between transactional and transformational leadership. The
transactional style asserts itself by means of praise and admonishment,
whereas the transformational style inspires people to go beyond their
personal interests and think of the interests of the company or task.
The transformational style attempts to entice people to identify with
the company, and the transactional style focuses on people™s self-
interest. The transformational style is intended to create the kind of
good morale that business people usually have in mind. Investigators
have found that transformational leadership has a strong positive im-
pact on both in-role job performance and on organizational citizenship
behavior, that its impact exceeds that of transactional leadership, and
that its impact is due in part to workers™ increased trust in the lead-
ership. The relevant studies are Podsakoff et al. (1990); Podsakoff,
MacKenzie, and Bommer (1996); and MacKenzie, Podsakoff, and Rich
An obvious question is whether organizational citizenship behavior
increases a company™s pro¬tability. Managers apparently think that it
does, because there is evidence that supervisors™ performance evalua-
tions of subordinates are strongly and positively in¬‚uenced by organi-
zational citizenship behavior. Papers that establish this connection are
MacKenzie, Podsakoff, and Fetter (1991, 1993) and Podsakoff, Mac-
Kenzie, and Hui (1993). A few studies have measured the impact of
organizational citizenship behavior on the performance of work
groups in various settings and have found the effects to be positive.
These studies include George and Bettenhausen (1990), Podsakoff and
MacKenzie (1994, 1997), Walz and Niehoff (1996), and Podsakoff,
Ahearne, and MacKenzie (1997). The observed correlations may be
spurious, however, because there is evidence from laboratory experi-
ments that the high performance of a work group may have a positive
in¬‚uence on perceptions within the group of organizational citizenship
behavior (Bachrach, Bendoly, and Podsakoff, 2001). The subject is
reviewed in Podsakoff et al. (2000).
Some interesting recent work has explored the connection between
identi¬cation with an organization on the one hand and quits and per-
formance (especially extra-role performance) on the other hand. Tom
Tyler has participated in much of this work. He thinks of identi¬cation
with a company as internalization of its goals and asserts that identi¬-
cation occurs as a result of judgments about organizational status
(which he calls pride), and about status within the organization (which
he calls respect). Pride has to do with a favorable view of the organiza-
Fairness, Reciprocity, and Wage Rigidity 329

tion as a whole, and respect has to do with being treated well within it.
Status judgments can be comparative or autonomous, where a compar-
ative judgment relates an organization or person to others and an
autonomous judgment is an absolute one about the overall organiza-
tion. Tyler believes that if people identify with an organization, they
will want it to succeed, because its success will strengthen their own
self-image. Identi¬cation with an organization is, in my opinion, a
much better interpretation of what managers mean by ˜˜good morale™™
than are job satisfaction and even organizational commitment. Tyler
and his co-authors ¬nd that identi¬cation is a dominant explanation of
voluntary cooperation with organizations. In the context of business
organizations, identi¬cation with the company is a much more impor-
tant explanatory factor than the ¬nancial rewards received from it.
These investigators ¬nd that identi¬cation has a greater impact on
organizational citizenship, extra-role, or discretionary behavior, as
opposed to in-role or mandatory behavior (that is, behavior required
by a job description). The primary impact of pride is on rule following
or conscientiousness, whereas the primary impact of respect is on help-
ing behavior (that is, assisting coworkers). Autonomous judgments of
status have a much bigger effect than comparative ones.
Tyler and his coauthors assert that perceptions of fairness and espe-
cially procedural justice have an important impact on judgments
about the status of an organization and hence on willingness to iden-
tify with it. Recall that management scientists cited earlier (Morris
Viteles [1953], Stanley Seashore [1954], William Whyte et al. [1955,
1961], and Rensis Likert [1961]), also found a connection between pride
in an organization and performance. The work of Tyler and his col-
leagues is reported in Tyler (1999) and Tyler and Blader (2000, 2001).
Abrams, Ando, and Hinkle (1998) observe a close association between
identi¬cation with an organization and intentions to quit. Much of the
work of Tyler and his co-authors on identi¬cation and cooperation
with organizations has been done in the context of political, social, and
educational institutions, but the recent studies just cited have to do
with businesses. This interesting work raises the question of why peo-
ple identify with organizations. Status is an incomplete explanation,
since the term status has little independent content and includes all
possible reasons for liking an organization. It is interesting that fairness
has a strong in¬‚uence on status and that people are proud of organiza-
tions that treat them and others fairly, but researchers have given no
explanation of why this is so.
330 Bewley

An obvious question is what evidence has been collected on the im-
pact of actual pay cuts or pay freezes on morale. The only works I
have found on the subject are Greenberg (1989, 1990) and Schaubroeck,
May, and Brown (1994). In the ¬rst paper, Greenberg ¬nds from a sur-
vey that workers did feel underpaid after a 6 percent pay cut, but job
satisfaction did not decline and employees instead paid more attention
to the non¬nancial advantages of their jobs. In the second paper,
Greenberg (1990) reports that theft of company property increased af-
ter a 15 percent pay cut. In this paper, he conducted an experiment in
which he gave employees a good explanation of the pay cut in one
plant where the pay cut occurred but not in another where it also
occurred. In the plant where the explanation was made, feelings of
pay inequity and pilferage were less than in the other plant. This evi-
dence supports the assertions managers make that employees tolerate
pay cuts more easily if they feel they are justi¬ed and that it is possible
to persuade workers that cuts are necessary.
These conclusions are further reinforced by the work of Schau-
broeck, May, and Brown (1994), who studied the reactions of salaried
employees to a pay freeze. These investigators also conducted an ex-
periment, giving a good explanation to some of the employees who
were affected by the pay freeze and not to others. The explanations of
the freeze diminished resentment. For those who did not receive the
explanation, job dissatisfaction increased with self-reported economic
hardship resulting from the freeze, and there was no such relation for
those who did receive the explanation.

11.6 Conclusion

Perhaps the outstanding conclusion to be drawn from the works dis-
cussed in this chapter is the importance of fairness to labor perfor-
mance. It is not easy to judge what fairness means. Fairness certainly
does not mean an equal distribution of the bene¬ts from a company™s
operations; pay levels within ¬rms are far from egalitarian. Even
workers doing the same job may receive very different pay because of
many factors, such as longevity with the company, skills acquired, and
productivity. Fairness is recognized in business as being inherently
ambiguous. For instance, judgments about the fairness of internal pay
structures are said to depend strongly on company tradition. Other
evidence that fairness does not mean equality of gains is evidence
from organizational psychology that procedural and interactive justice
Fairness, Reciprocity, and Wage Rigidity 331

are more important to an impression of fairness than is distributive
justice. A very signi¬cant ¬nding is that of Tyler and Blader (2000,
2001) that perceptions of procedural justice contribute to pride in an
We do not know why people so urgently desire fairness. Is it be-
cause it contributes to an atmosphere of positive reciprocation where
people like to exchange favors? Does fairness make people feel more
secure? Do people feel that fairness is right and want their surround-
ings to accord with their moral precepts? Do people simply want to
have a level playing ¬eld on which to compete? It is to be hoped that
further empirical work will give more insight into these questions.
An understanding of the need for fairness would contribute a great
deal to understanding how organizations obtain cooperation and to
the explanation of wage rigidity. A sense of fairness is probably the
most important determinant of good company morale. Other im-
portant factors are close ties among coworkers and the signi¬cance
attached to the ¬rm™s output. One reason pay cuts can be resented is
that they can dissolve the sense of fairness. Workers accept a pay cut if
they feel it is fair and they see it as fair when it saves a signi¬cant num-
ber of jobs.
Another important conclusion is that ¬rms try to gain the coopera-
tion of employees by getting them to identify with the company and
to internalize its objectives. As Tyler and Blader (2000, 2001) have
emphasized, an atmosphere of fairness makes workers more willing to
do these things. It would be useful to know why fairness promotes
identi¬cation with a company and why people identify with organiza-
tions at all. That they do is clear.


I am grateful to Professor Jennifer Smith of the University of Warwick for her comments.

1. These data are cited in Akerlof, Dickens, and Perry (1996, 8).
2. I owe this reference to Angier (2002), who makes the connection with the experimental
work of Ernst Fehr.
3. See ¬gure 6 in Fehr and Fischbacher (2002).
4. In Fehr, Klein, and Schmidt (2001), the employer chooses the type of model used, there
is no excess supply of labor, and the experimenter matches one worker to one employer
in each period. In Fehr and Gachter (2002), the experimenter chooses the model, there is
an excess supply of labor, and the matching of workers to employers is determined by
market bidding.
332 Bewley


Abrams, Dominic, Kaori Ando, and Steve Hinkle. (1998), ˜˜Psychological Attachment to
the Group: Cross-Cultural Differences in Organizational Identi¬cation and Subjective
Norms as Predictors of Workers™ Turnover Intentions,™™ Personality and Social Psychology
Bulletin 24, 1027“1039.
Agell, Jonas, and Per Lundborg. (1995), ˜˜Theories of Pay and Unemployment: Survey
Evidence from Swedish Manufacturing Firms,™™ Scandinavian Journal of Economics 97, 295“
”””. (1999), ˜˜Survey Evidence on Wage Rigidity and Unemployment: Sweden in the
1990s,™™ Of¬ce of Labour Market Policy Evaluation, Uppsala, Sweden, Discussion Paper
1999, 2.
”””. (2003), ˜˜Survey Evidence on Wage Rigidity and Unemployment: Sweden in the
1990s,™™ Scandinavian Journal of Economics 105, 15“29.
Akerlof, George A. (1982), ˜˜Labor Contracts as Partial Gift Exchange,™™ Quarterly Journal of
Economics 97, 543“569.
Akerlof, George A., William T. Dickens, and George Perry. (1996), ˜˜The Macroeconomics
of Low In¬‚ation,™™ Brookings Papers on Economic Activity 1“76.
Akerlof, George A., and Janet Yellen. (1988), ˜˜Fairness and Unemployment,™™ American
Economic Association, Papers and Proceedings 78, 44“49.
”””. (1990), ˜˜The Fair Wage-Effort Hypothesis and Unemployment,™™ The Quarterly
Journal of Economics 105, 255“283.
Altonji, Joseph G., and Paul J. Devereux. (2000), ˜˜The Extent and Consequences of Down-
ward Nominal Wage Rigidity,™™ in Solomon W. Polachek (ed.), Worker Well-Being, vol. 19
of Research in Labor Economics 383“431.

Angier, Natalie. (2002), ˜˜Why We™re so Nice: We™re Wired to Cooperate,™™ New York
Times, Science Section, July 23, 2002.
Bachrach, Daniel G., Elliot Bendoly, and Philip M. Podsakoff. (2001), ˜˜Attributions of the
˜Causes™ of Group Performance as an Alternative Explanation of the Relationship be-
tween Organizational Citizenship Behavior and Organizational Performance,™™ Journal of
Applied Psychology 86(6), 1285“1293.
Baker, George, Michael Gibbs, and Bengt Holmstrom. (1994), ˜˜The Wage Policy of the
Firm,™™ Quarterly Journal of Economics 109, 921“955.
Bewley, Truman F. (1999), Why Wages Don™t Fall During a Recession. Cambridge, MA:
Harvard University Press.

Blinder, Alan S., and Don H. Choi. (1990), ˜˜A Shred of Evidence on Theories of Wage
Stickiness,™™ Quarterly Journal of Economics 105, 1003“1015.

Bray¬eld, A. H., and W. H. Crockett. (1955), ˜˜Employee Attitudes and Employee Perfor-
mance,™™ Psychological Bulletin 52, 396“424.
Brief, Arthur P., and Stephan J. Motowidlo. (1986), ˜˜Prosocial Organizational Behaviors,™™
Academy of Management Review 11(4), 710“725.
Fairness, Reciprocity, and Wage Rigidity 333

Brown, Martin, Armin Falk, and Ernst Fehr. (2002), ˜˜Contractual Incompleteness and the
Nature of Market Interactions,™™ Institute for Empirical Research in Economics, University
of Zurich, Working Paper No. 38.
Burda, Michael, Werner Guth, Georg Kirchsteiger, and Harald Uhlig. (1998), ˜˜Employ-
ment Duration and Resistance to Wage Reductions: Experimental Evidence,™™ Tilburg
University, Center for Economic Research Discussion Paper No. 9873.
Campbell, Carl, and Kunal Kamlani. (1997), ˜˜The Reasons for Wage Rigidity: Evidence
from Survey of Firms,™™ Quarterly Journal of Economics 112, 759“789.
Card, David, and Dean Hyslop. (1997), ˜˜Does In¬‚ation ˜Grease the Wheels of the Labor
Market™?™™ in Christina D. Romer and David H. Romer (eds.), Reducing In¬‚ation, Motiva-
tion and Strategy. Chicago: University of Chicago Press.
Chapple, Simon. (1996), ˜˜Money Wage Rigidity in New Zealand,™™ Labour Market Bulletin:
A Journal of New Zealand Labour Market Research 23“50.
Falk, Armin, Ernst Fehr, and Urs Fischbacher. (2000), ˜˜Testing Theories of Fairness”
Intentions Matter,™™ Institute for Empirical Research in Economics, University of Zurich,
Working Paper No. 63.
”””. (2003), ˜˜On the Nature of Fair Behavior,™™ Economic Inquiry 41, 20“26.
Farh, Jiing-Lih, Philip M. Podsakoff, and Dennis Organ. (1990), ˜˜Accounting for Organi-
zational Citizenship Behavior: Leader Fairness and Task Scope Versus Satisfaction,™™ Jour-
nal of Management 16(4), 705“721.

Fehr, Ernst, and Armin Falk. (1999), ˜˜Wage Rigidity in a Competitive Incomplete Con-
tract Market,™™ Journal of Political Economy 107, 106“134.

”””. (2002), ˜˜Psychological Foundations of Incentives,™™ European Economic Review 46,
Fehr, Ernst, and Urs Fischbacher. (2002), ˜˜Why Social Preferences Matter”The Impact of
Non-Sel¬sh Motives on Competition, Cooperation and Incentives,™™ The Economic Journal
112, C1“C33.
Fehr, Ernst, and Simon Gachter. (1998a), ˜˜How Effective Are Trust- and Reciprocity-
Based Incentives?™™ In A. Ben-Ner and L. Putterman (eds.), Economics, Values and Organi-
zations. Cambridge: Cambridge University Press.
”””. (1998b), ˜˜Reciprocity and Economics: The Economic Implications of Homo Recip-
rocans,™™ European Economic Review 42, 845“859.

”””. (2000), ˜˜Fairness and Retaliation: The Economics of Reciprocity,™™ Journal of Eco-
nomic Perspectives 14(3), 159“181.

”””. (2002), ˜˜Do Incentive Contracts Undermine Voluntary Cooperation?™™ Institute
for Empirical Research in Economics, University of Zurich, Working Paper No. 34.
Fehr, Ernst, Simon Gachter, and Georg Kirchsteiger. (1997), ˜˜Reciprocity as a Contract
Enforcement Device”Experimental Evidence,™™ Econometrica 65, 833“860.
Fehr, Ernst, and Lorenz Goette. (1999), ˜˜How Robust Are Nominal Wage Rigidities?™™
Discussion Paper, University of Zurich.
334 Bewley

Fehr, Ernst, Erich Kirchler, Andreas Weichbold, and Simon Gachter. (1998), ˜˜When Social
Norms Overpower Competition”Gift Exchange in Experimental Labor Markets,™™ Jour-
nal of Labor Economics 16, 324“351.
Fehr, Ernst, Georg Kirchsteiger, and Arno Riedl. (1993), ˜˜Does Fairness Prevent market
Clearing? An Experimental Investigation,™™ Quarterly Journal of Economics 108, 437“459.
”””. (1996), ˜˜Involuntary Unemployment and Non-Compensating Wage Differentials
in An Experimental Labour Market,™™ Economic Journal 106, 106“121.
”””. (1998), ˜˜Gift Exchange and Reciprocity in Competitive Experimental Markets,™™
European Economic Review 42, 1“34.
Fehr, Ernst, Alexander Klein, and Klaus M. Schmidt. (2001), ˜˜Fairness, Incentives and
Contractual Incompleteness,™™ Institute for Empirical Research in Economics, University
of Zurich, Working Paper No. 72.
Fehr, Ernst, and Bettina Rockenbach. (2002), ˜˜Detrimental Effects of Incentives on Hu-
man Altruism,™™ Working Paper, University of Zurich.
Folger, Robert. (1993), ˜˜Justice, Motivation, and Performance: Beyond Role Require-
ments,™™ Employee Responsibilities and Rights Journal 6(3), 239“248.
Fortin, Pierre. (1996), ˜˜The Great Canadian Slump,™™ Canadian Journal of Economics 29,
Gachter, Simon and Armin Falk. (2002), ˜˜Reputation and Reciprocity: Consequences for
the Labour Relation,™™ Scandinavian Journal of Economics 104, 1“26.
George, Jennifer M. (1991), ˜˜State or Trait: Effects of Positive Mood on Prosocial Behav-
iors at Work,™™ Journal of Applied Psychology 76(2), 299“307.
George, Jennifer M., and Kenneth Bettenhausen. (1990), ˜˜Understanding Prosocial Behav-
ior, Sales Performance, and Turnover: A Group-Level Analysis in a Service Context,™™
Journal of Applied Psychology 75(6), 698“709.

Greenberg, Jerald. (1989), ˜˜Injustice and Cognitive Reevaluation of the Work Environ-
ment,™™ Academy of Management Journal 32, 174“184.

”””. (1990), ˜˜Employee Theft as a Reaction to Underpayment Inequity: The Hidden
Cost of Pay Cuts,™™ Journal of Applied Psychology 75(5), 561“568.
”””. (1993), ˜˜Justice and Organizational Citizenship: A Commentary on the State of
the Science,™™ Employee Responsibilities and Rights Journal 6(3), 249“256.
Herzberg, Frederick, Bernard Mausner, Richard O. Peterson, and Dora Capwell. (1957),
Job Attitudes: Review of Research and Opinion. Pittsburgh: Psychological Service of
Iaffaldano, Michelle T., and Paul M. Muchinsky. (1985), ˜˜Job Satisfaction and Job Perfor-
mance: A Meta-Analysis,™™ Psychological Bulletin 97(2), 251“273.
Kahn, Shulamit. (1997), ˜˜Evidence of Nominal Wage Stickiness from Microdata,™™ Ameri-
can Economic Review 87(5), 993“1008.
Katz, Daniel. (1964), ˜˜The Motivational Basis of Organizational Behavior,™™ Behavioral
Science 9, 131“146.
Fairness, Reciprocity, and Wage Rigidity 335

Kaufman, Roger. (1984), ˜˜On Wage Stickiness in Britain™s Competitive Sector,™™ British
Journal of Industrial Relations 22, 101“112.
Keynes, John Maynard. (1936), The General Theory of Employment, Interest, and Money.
London: Macmillan.
Kirchler, Erich, Ernst Fehr, and Robert Evans. (1996), ˜˜Social Exchange in the Labor Mar-
ket: Reciprocity and Trust Versus Egoistic Money Maximization,™™ Journal of Economic Psy-
chology 17, 313“341.
Konovsky, Mary A., and Dennis W. Organ. (1996), ˜˜Dispositional and Contextual Deter-
minants of Organizational Citizenship Behavior,™™ Journal of Organizational Behavior 17,

Lawler, Edward E. III. (1971), Pay and Organizational Effectiveness: A Psychological View.
New York: McGraw-Hill.
Lebow, David E., Saks, Raven E., and Wilson, Beth Anne. (1999), ˜˜Downward Nominal
Wage Rigidity: Evidence from the Employment Cost Index,™™ Finance and Economics Dis-
cussion Series, Divisions of Research and Statistics and Monetary Affairs, Federal Re-
serve Board, Washington, DC, Paper No. 1999-31.
Lebow, David E., David J. Stockton, and William L. Wascher. (1995), ˜˜In¬‚ation, Nominal
Wage Rigidity, and the Ef¬ciency of Labor Markets,™™ Board of Governors of the Federal
Reserve System, Finance and Economic Discussion Series, 94-45.
Levine, David I. (1993), ˜˜Fairness, Markets, and Ability to Pay: Evidence from Compen-
sation Executives,™™ American Economic Review 83, 1241“1259.
Likert, Rensis. (1961), New Patterns of Management. New York: McGraw-Hill.

Lindbeck, Assar, and Dennis J. Snower. (1988), ˜˜Cooperation, Harassment, and Involun-
tary Unemployment: An Insider-Outsider Approach,™™ American Economic Review 78, 167“
Locke, Edwin A. (1976), ˜˜The Nature and Causes of Job Satisfaction,™™ in M. D. Dunnette
(ed.), Handbook of Industrial and Organizational Psychology. Chicago: Rand McNally, 1297“
MacKenzie, Scott B., Philip M. Podsakoff, and Michael Ahearne. (1998), ˜˜Some Possible
Antecedents and Consequences of In-Role and Extra-Role Salesperson Performance,™™
Journal of Marketing 62 ( July), 87“98.
MacKenzie, Scott B., Philip M. Podsakoff, and Richard Fetter. (1991), ˜˜Organizational Cit-
izenship Behavior and Objective Productivity as Determinants of Managerial Evaluations
of Salespersons™ Performance,™™ Organizational Behavior and Human Decision Processes 50,
”””. (1993), ˜˜The Impact of Organizational Citizenship Behavior on Evaluations of
Salesperson Performance,™™ Journal of Marketing 57, 70“80.
MacKenzie, Scott B., Philip M. Podsakoff, and Gregory A. Rich. (2001), ˜˜Transformational
and Transactional Leadership and Salesperson Performance,™™ Journal of the Academy of
Marketing Science 29(2), 115“134.
Mathieu, John E., and Dennis M. Zajac. (1990), ˜˜A Review and Meta-Analysis of the
Antecedents, Correlates, and Consequences of Organizational Commitment,™™ Psychologi-
cal Bulletin 108, 171“194.
336 Bewley

McLaughlin, Kenneth J. (1994), ˜˜Rigid Wages?™™ Journal of Monetary Economics 34, 383“
”””. (1999), ˜˜Are Nominal Wage Changes Skewed Away from Wage Cuts?™™ Federal
Reserve Bank of St. Louis Review 81(3), 117“132.
Mitchell, Daniel J. B. (1985), ˜˜Shifting Norms in Wage Determination,™™ Brookings Papers
on Economic Activity, 575“599.
Mobley, William H. (1982), Employee Turnover: Causes, Consequences, and Control. Reading,
MA: Addison-Wesley.
Moorman, Robert H. (1991), ˜˜Relationship between Organizational Justice and Organiza-
tional Citizenship Behaviors: Do Fairness and Perceptions In¬‚uence Employee Citizen-
ship?™™ Journal of Applied Psychology 76(6), 845“855.
”””. (1993), ˜˜The In¬‚uence of Cognitive and Affective Based Job Satisfaction Mea-
sures on the Relationship between Satisfaction and Organizational Citizenship Behavior,™™
Human Relations 46(6), 759“776.
Moorman, Robert H., Brian P. Niehoff, and Dennis W. Organ. (1993), ˜˜Treating Employ-
ees Fairly and Organizational Citizenship Behavior: sorting the Effects of Jobs Satisfac-
tion, Organizational Commitment, and Procedural Justice,™™ Employee Responsibilities and
Rights Journal 6(3), 209“225.
Nagin, Daniel, James Rebitzer, Seth Sanders, and Lowell Taylor. (1998), ˜˜Monitoring and
Motivation in an Employment Relationship: An Analysis of a Field Experiment,™™ Discus-
sion Paper, Carnegie-Mellon University.
Netemeyer, Richard G., James S. Boles, Daryle O. McKee, and Robert McMurrian. (1997),
˜˜An Investigation into Antecedents of Organizational Citizenship Behaviors in a Personal
Selling Context,™™ Journal of Marketing 61, ( July), 85“98.
Niehoff, Brian P., and Robert H. Moorman. (1993), ˜˜Justice as a Mediator of the Relation-
ship between Methods of Monitoring and Organizational Citizenship Behavior,™™ Acad-
emy of Management Journal 36(3), 527“556.

O™Reilly, Charles III, and Jennifer Chatman. (1986), ˜˜Organizational Commitment and
Psychological Attachment: The Effects of Compliance, Identi¬cation, and Internalization
on Prosocial Behavior,™™ Journal of Applied Psychology 71(3), 492“499.
Organ, Dennis W. (1988), Organizational Citizenship Behavior. Lexington, MA: Lexington
”””. (1990), ˜˜The Motivational Basis of Organizational Citizenship Behavior,™™ in Barry
M. Staw and L. L. Cummings (eds.), Research in Organizational Behavior 12, 43“72.
Organ, Dennis W., and Mary Konovsky. (1989), ˜˜Cognitive Versus Affective Determi-
nants of Organizational Citizenship Behavior,™™ Journal of Applied Psychology 74(1), 157“
Organ, Dennis W., and Andreas Lingl. (1995), ˜˜Personality, Satisfaction, and Organiza-
tional Citizenship Behavior,™™ The Journal of Social Psychology 135(3), 339“350.

Organ, Dennis W., and Robert H. Moorman. (1993), ˜˜Fairness and Organizational Citi-
zenship Behavior: What Are the Connections?™™ Social Justice Research 6(1), 5“18.
Fairness, Reciprocity, and Wage Rigidity 337

Organ, Dennis W., and Katherine Ryan. (1995), ˜˜A Meta-analytic Review of Attitudinal
and Dispositional Predictors of Organizational Citizenship Behavior,™™ Personnel Psychol-
ogy 48, 775“802.
Podsakoff, Philip M., Michael Ahearne, and Scott B. MacKenzie. (1997), ˜˜Organizational
Citizenship Behavior and the Quantity and Quality of Work Group Performance,™™ Journal
of Applied Psychology 82(2), 262“270.
Podsakoff, Philip M., and Scott B. MacKenzie. (1993), ˜˜Citizenship Behavior and Fairness
in Organizations: Issues and Directions for Future Research,™™ Employee Responsibilities and
Rights Journal 6(3), 257“269.
”””. (1994), ˜˜Organizational Citizenship Behaviors and Sales Unit Effectiveness,™™ Jour-
nal of Marketing Research 31, (August), 351“363.
”””. (1997), ˜˜Impact of Organizational Citizenship Behavior on Organizational Per-
formance: A Review and Suggestions for Future Research,™™ Human Performance 10(2),
Podsakoff, Philip M., Scott B. MacKenzie, and William H. Bommer. (1996), ˜˜Transforma-
tional Leader Behaviors and Substitutes for Leadership as Determinants of Employee Sat-
isfaction, Commitment, Trust, and Organizational Citizenship Behaviors,™™ Journal of
Management 22(2), 259“298.
Podsakoff, Philip M., Scott B. MacKenzie, and Chun Hui. (1993), ˜˜Organizational Citizen-
ship Behaviors and Managerial Evaluations of Employee Performance: A Review and
Suggestions for Future Research,™™ in Ferris, Gerald R. (ed.), Research in Personnel and Hu-
man Resources Management 11, 1“40.
Podsakoff, Philip M., Scott B. MacKenzie, Robert H. Moorman, and Richard Fetter.
(1990), ˜˜Transformational Leader Behaviors and Their Effects on Followers™ Trust in
Leader, Satisfaction, and Organizational Citizenship Behaviors,™™ Leadership Quarterly
1(2), 107“142.
Podsakoff, Philip M., Scott B. MacKenzie, Julie Beth Paine, and Daniel G. Bachrach.
(2000), ˜˜Organizational Citizenship Behaviors: A Critical Review of the Theoretical and
Empirical Literature and Suggestions for Future Research,™™ Journal of Management 26(3),
Price, James L. (1977), The Study of Turnover. Ames, IA: Iowa State University Press.

Puffer, Sheila M. (1987), ˜˜Prosocial Behavior, Noncompliant Behavior, and Work Perfor-
mance Among Commission Salespeople,™™ Journal of Applied Psychology 72(4), 615“621.
Rilling, James K., David A. Gutman, Thorsten R. Zeh, Giuseppe Pagnoni, Gregory S.
Berns, and Clinton D. Kits. (2002), ˜˜A Neural Basis for Social Cooperation,™™ Neuron 35(2),
Schaubroeck, John, Douglas R. May, and R. William Brown. (1994), ˜˜Procedural Justice
Explanations and Employee Reactions to Economic Hardship: A Field Experiment,™™ Jour-
nal of Applied Psychology 79(3), 455“460.
Schnake, Mel. (1991), ˜˜Organizational Citizenship: A Review, Proposed Model, and Re-
search Agenda,™™ Human Relations 44(7), 735“759.

Seashore, Stanley. (1954), Group Cohesiveness in the Industrial Work Group. Ann Arbor:
University of Michigan Press.
338 Bewley

Shapiro, Carl, and Joseph E. Stiglitz. (1984), ˜˜Equilibrium Unemployment as a Worker
Discipline Device,™™ American Economic Review 74, 433“444.
Smith, Jennifer C. (2000), ˜˜Nominal Wage Rigidity in the United Kingdom,™™ The Economic
Journal 110, C176“C195.
”””. (2002), ˜˜Pay Cuts and Morale: A Test of Downward Nominal Rigidity,™™ Warwick
Economic Research Paper No. 649, Department of Economics, University of Warwick.
Solow, Robert M. (1979), ˜˜Another Possible Source of Wage Stickiness,™™ Journal of Macro-
economics 1, 79“82.
”””. (1990), The Labor Market as a Social Institution. Cambridge, MA: Basil Blackwell.
Staw, Barry M. (1984), ˜˜Organizational Behavior: A Review and Reformulation of the
Field™s Outcome Variables,™™ Annual Review of Psychology 35, 627“666.
Steers, Richard M., and Susan R. Rhodes. (1978), ˜˜Major In¬‚uences on Employee Atten-
dance: A Process Model,™™ Journal of Applied Psychology 63, 391“407.
Tyler, Tom R. (1999), ˜˜Why People Cooperate with Organizations: An Identity-Based
Perspective,™™ Research in Organizational Behavior 21, 201“246.
Tyler, Tom R., and Steven L. Blader. (2000), Cooperation in Groups: Procedural Justice, Social
Identity, and Behavioral Engagement. Philadelphia, PA: Psychology Press.
”””. (2001), ˜˜Identity and Cooperative Behavior in Groups,™™ Group Processes and Inter-
group Relations 4(3), 207“226.
Viteles, Morris S. (1953), Motivation and Morale in Industry. New York: W. W. Norton.
Vroom, Victor H. (1964), Work and Motivation. New York: John Wiley.
Walz, Sandra M., and Brian P. Niehoff. (1996), ˜˜Organizational Citizenship Behaviors
and Their Relationship with Indicators of Organizational Effectiveness in Limited Menu
Restaurants,™™ in J. B. Keys and L. N. Dosier (eds.), Academy of Management Best Papers Pro-
ceedings. Statesboro, GA: George Southern University, 307“311.

Weiss, Andrew. (1980), ˜˜Job Queues and Layoffs in Labor Markets with Flexible Wages,™™
Journal of Political Economy 88, 526“538.
”””. (1990), Ef¬ciency Wages, Models of Unemployment, Layoffs, and Wage Dispersion.
Princeton, NJ: Princeton University Press.
Whyte, William Foote, et al. (1955), Money and Motivation: An Analysis of Incentives in
Industry. New York: Harper and Brothers.
”””. (1961), Men at Work. Homewood, IL: Corsey Press, Inc.
Wilson, Beth Ann. (1996), ˜˜Movement of Wages over the Business Cycle: an Intra-Firm
View,™™ Federal Reserve Board, Discussion Paper.
12 The Logic of Reciprocity:
Trust, Collective Action,
and Law

Dan M. Kahan

12.1 Introduction

The Logic of Collective Action has for decades supplied the logic of
public policy analysis.1 In this pioneering application of public choice
theory, Mancur Olson elegantly punctured the premise”shared by a
diverse variety of political theories”that individuals can be expected
to act consistently with the interest of the groups to which they belong.
Absent externally imposed incentives, wealth-maximizing individuals,
he argued, will rarely ¬nd it in their interest to contribute to goods that
bene¬t the group as a whole, but rather will ˜˜free ride™™ on the contri-
butions that other group members make. As a result, too few individu-
als will contribute suf¬ciently, and the well-being of the group will
suffer.2 These are the assumptions that currently dominate public pol-
icy analysis and ultimately public policy across a host of regulatory
domains”from tax collection to environmental conservation, from
street-level policing to policing of the internet.
But as a wealth of social science evidence (much of it appearing
elsewhere in this volume) now makes clear, Olson™s Logic is false. In
collective action settings, individuals adopt not a materially calculating
posture, but rather a richer, more emotionally nuanced reciprocal one.
When they perceive that others are behaving cooperatively, individ-
uals are moved by honor, altruism, and like dispositions to contribute
to public goods even without the inducement of material incentives.
When, in contrast, they perceive that others are shirking or otherwise
taking advantage of them, individuals are moved by resentment and
pride to retaliate. In that circumstance, they will withhold bene¬cial
forms of cooperation even if doing so exposes them to signi¬cant mate-
rial disadvantage.3
340 Kahan

This behaviorally realistic picture of human motivation suggests not
only an alternative account of when collective action problems will
arise, but also an alternative program for solving (or simply avoiding)
them through law. Whereas the conventional logic of collective action
counsels the creation of appropriate external incentives, the new logic
of reciprocity suggests the importance of promoting trust. Individuals
who have faith in the willingness of others to contribute their fair share
will voluntarily respond in kind. Spontaneous cooperation of this sort,
moreover, breeds even more of the same, as individuals observe others
contributing to public goods and are moved to reciprocate. In this self-
sustaining atmosphere of trust, reliance on costly incentive schemes
becomes less necessary. By the same token, individuals who lack faith
in others can be expected to resist contributing to public goods, in-
ducing still more persons to withhold their cooperation as a means of
retaliating. In this self-sustaining atmosphere of distrust, even strong
(and costly) regulatory incentives are likely to be ineffective in promot-
ing desirable behavior.
Indeed, such incentives may well undermine the conditions of trust
necessary to hold collective action problems in check. Conspicuous
rewards and punishments can imply that others aren™t inclined to
cooperate voluntarily, a message that predictably weakens individuals™
commitment to contributing to public goods. In addition, incentive
schemes tend to mask the extent to which individuals are inclined to
contribute to public goods voluntarily, thereby weakening the ten-
dency of observable cooperation to generate reciprocal cooperation by
others. In short, manipulating material incentives may not only be an
inef¬cient regulatory strategy for solving collective action problems; it
may often be a self-defeating one.
This chapter will elaborate upon and apply these claims. It begins by
distilling from the reciprocity literature a set of behavioral dynamics
pertinent societal collective action problems. It then shows how these
dynamics can be used to analyze and improve policymaking in various
regulatory ¬elds, with a particular emphasis on tax compliance, the sit-
ing of noxious facilities, and the policing of street crime.

12.2 The Logic of Reciprocity

Accepted for decades on a combination of faith and anecdote, the
premises of the conventional theory of collective action have only
recently been subjected to sustained and rigorous empirical examina-
The Logic of Reciprocity 341

Conventional Theory Reciprocity Theory

Agents Wealth Maximizers Reciprocators

Collective Behavior Unique Equlibrium Multiple Equilibria

Promoting Cooperation Incentive Trust

Variability of Preferences Homogeneous Heterogeneous

Figure 12.1
Agents: Wealth maximizers vs. Emotional/Moral reciprocators.

tion. This research suggests an alternative ˜˜strong reciprocity theory™™
that differs from conventional collective action theory in four impor-
tant respects, as re¬‚ected in ¬gure 12.1. Each of the contrasts between
the conventional theory and the strong reciprocity theory merits spe-
ci¬c attention.

12.2.1 Agents: Wealth Maximizers versus Reciprocators
The ¬rst pair of contrasting elements relates to the nature of individu-
als™ utility functions. The conventional theory assumes that individuals
in collective action settings”ones that take the form of a standard Pris-
oners™ Dilemma”behave like wealth maximizers. That is, they refuse
to contribute to collective goods and instead free-ride on the contri-
butions made by others, who, as wealth maximizers, also contribute
nothing. The strong reciprocity model, in contrast, sees individuals as
moral and emotional reciprocators. Most persons think of themselves
and want to be understood by others as cooperative and trustworthy
and are thus perfectly willing to contribute their fair share to securing
collective goods. By the same token, however, most individuals hate
being taken advantage of. Accordingly, if they perceive that most other
individuals are shirking, they hold back too to avoid feeling exploited.
Individuals who care only about maximizing their wealth are at best
weak reciprocators. If a rational wealth maximizer anticipates that she
will be engaged in recurring transactions with another identi¬able
agent over a suf¬ciently long period of time under circumstances
342 Kahan

where both can observe and keep track of one another™s actions,
then her best strategy is to reward cooperation with cooperation and
defection with defection in a ˜˜tit-for-tat™™ pattern.4 Emotional and
moral reciprocators, in contrast, are strong reciprocators. They will con-
dition their contributions to collective goods on the contributions of
others even in ¬‚eeting transactions with multiple actors whose be-
havior they cannot keep track of and whose identities they can™t even
The prevalence of this sort of strong reciprocity is supported by a
vast body of evidence. Much of it is experimental in nature. So-called
˜˜public goods™™ experiments”laboratory constructs designed to simu-
late collective action problems”have consistently shown that the will-
ingness of individuals to make costly contributions to collective goods
is highly conditional on their perception that others are willing to do
so.5 Empirical studies of real-world behavior corroborate this ¬nding.
Individuals have been shown, for example, to reciprocate the disposi-
tion of others to give (or not) to charity,6 to refrain (or not) from litter-
ing,7 and to wait their turn (or not) in lines.8 Indeed, individuals
behave like reciprocators even in markets: econometric and other forms
of ¬eld research, for example, suggest that when ¬rms compensate
their workers more generously, workers reciprocate by voluntarily
working harder.9

12.2.2 Collective Behavior: Unique versus Multiple Equilibria
The strong reciprocity theory also takes issue with the conventional
theory™s view of collective behavior. In typical collective action settings,
the conventional theory treats defection or free-riding as the dominant
strategy for every individual. Accordingly, the theory predicts a single
collective behavioral equilibrium: universal noncooperation.
Under the strong reciprocity theory, in contrast, there is no ˜˜domi-
nant™™ individual strategy. Individuals prefer to contribute if they be-
lieve others are inclined to contribute, but to free-ride if they believe
that others are inclined to free-ride.
Such interdependencies tend to generate patterns of collective behav-
ior characterized by multiple equilibria punctuated by tipping points.10 If,
for whatever reason, some individuals conclude that those around
them are inclined to contribute, they™ll respond by contributing in
kind, prompting still others to contribute, and so forth and so on until
a highly cooperative state of affairs takes root. But if some individuals
conclude that others are free-riding, then they will respond by free-
The Logic of Reciprocity 343

riding, too, spurring others to do the same, and so forth and so on until
a condition of mass noncooperation becomes the norm.
This dynamic has also been empirically documented. In multi-round
public goods experiments, for example, contribution levels tend to mi-
grate steadily toward or away from the social optimum depending on
whether subjects behaved relatively cooperatively or noncooperatively
early on.11 Scholars have also documented that the incidence of litter-
ing, recycling, smoking in public, safe sex, and other types of behavior
that affect collective welfare are likewise subject to feedback effects and
multiple equilibria”generating dramatic variations in their incidence
across space and over time.12

12.2.3 Promoting Cooperation: Incentives versus Trust
The strong reciprocity theory and the conventional theory also dis-
agree about policy prescriptions. The conventional theory sees incen-
tives as the solution to collective action problems: Because wealth
maximizers cannot be counted on to contribute to public goods, they
must be prodded to do so with either rewards or punishments that
bring their individual interests into alignment with their collective
The strong reciprocity theory suggests an alternative policy”the
promotion of trust. If individuals can be made to believe that others
are inclined to contribute to public goods, they can be induced to con-
tribute in turn, even without recourse to incentives. When permitted to
communicate during in multi-round public goods experiments, for
example, subjects tend to assure one another that they will contribute
rather than free-ride. Although unenforceable, such assurances do in
fact prompt larger contributions, which subjects quickly increase to-
ward the social optimum as they observe others doing the same.13
Face-to-face assurance-giving, in sum, promotes trust, which in turn
generates reciprocal cooperation.
Indeed, ¬eld and laboratory research suggests that incentives, far
from solving collective action problems, can sometimes actually mag-
nify them by dissipating trust. The simple existence of an incentive
scheme can be seen as a cue that other individuals are not inclined
to cooperate voluntarily: if they were, incentives would be unneces-
sary. This inference can in turn trigger a reciprocal disposition to
withhold voluntary cooperation, thereby undercutting”if not wholly
displacing”the force of the incentive. In addition, the existence of
incentives can mask voluntary contributions to public goods, thereby
344 Kahan

diluting the power of such contributions to trigger reciprocal coopera-
tion. Relatedly, incentives can crowd out dispositions such as altruism
by extinguishing the opportunity of individuals to demonstrate (to
themselves and to others) that they are willing to sacri¬ce material
gain for the public good. If for any of these reasons, the advent of a
material incentive induces even a few individuals to contribute less to
a public good, moreover, reciprocity dynamics will induce still others
to contribute less, thereby inducing others to do the same. This new
noncooperative equilibrium that results is likely to be impervious to
the subsequent removal of material incentives.14
It would be a mistake, however, to conclude that material incentives
invariably diminish trust. They are most likely to have that effect,
research suggests, when individuals start out with the belief that
most other individuals are inclined to contribute to some public good
voluntarily”when the advent of material incentives creates the great-
est risk of adverse cueing, masking, and crowding out. But things are
likely to be different if individuals start out with the belief that most
other individuals are inclined to shirk or free-ride. In that case, the ad-
vent of a credible reward or penalty can work”not just by changing
individuals material incentives, but by changing (in a positive way)
their impression of the willingness of other individuals to behave coop-
eratively rather than noncooperatively in a collective action setting.
An example is the power of higher-than-average wages to elicit
higher-than-average productivity in the workplace. Workers naturally
suspect their ¬rms of being unwilling to share a fair portion of the sur-
plus generated by the workers™ labor. But when a ¬rm offers workers
a wage that exceeds the industry average, workers are likely to infer
that that particular ¬rm is willing to divide the surplus fairly. They
therefore respond by voluntarily working more productively, which
inclines the ¬rm to maintain or even raise their wages. The result is a
self-sustaining form of reciprocal cooperation that obviates the need
for costly performance monitoring regimes.

12.2.4 Variability of Preferences: Homogeneous versus
Finally, the conventional theory and the strong reciprocity theory differ
on the variability of preferences across individuals. The conventional
theory imagines that the disposition to free-ride in collective action
settings is relatively uniform. In contrast, the evidence on which the
strong reciprocity theory rests suggests that the disposition to cooper-
The Logic of Reciprocity 345

% Contribuing in tn+1




0 40 50 60 100
% Contribuing in tn
Figure 12.2
Heterogeneity of collective action dispositions.

ate varies. In public goods experiments that generate multiple equilib-
ria, for example, neither universal cooperation nor universal defection
is the ¬nal resting point.
It makes more sense, then, to envision a distribution of cooperative
dispositions across the population.
A relatively small fraction of the population (consisting, perhaps, of
those who™ve been trained in neoclassical economics) consists of com-
mitted free-riders, who shirk no matter what anyone else does, and
another small fraction (consisting maybe those who™ve read too much
Kantian moral philosophy) of dedicated cooperators, who contribute
no matter what. But most individuals are reciprocators who cooperate
conditionally on the willingness of others to contribute. Moreover,
some reciprocators are relatively intolerant: they bolt as soon as they
observe anyone else free-riding. Others are relatively tolerant, continu-
ing to contribute even in the face of what they see as a relatively mod-
est degree of defection. And a great many more”call them neutral
reciprocators”fall somewhere in between.
Under these circumstances, individuals are unlikely fully to over-
come collective action problems through reciprocity dynamics alone.
No matter how cooperative the behavior of others, the committed free-
riders will always free-ride if they can get away with it. Indeed, their
shirking could easily provoke noncooperative behavior by the less tol-
erant reciprocators, whose defection in turn risks inducing the neutral
reciprocators to abandon ship, thereby prompting even the tolerant
346 Kahan

reciprocators to throw in the towel, and so forth and so on. If this
unfortunate chain reaction takes place, a state of affairs once character-
ized by a reasonably high degree of cooperation could tip decisively
toward a noncooperative equilibrium in which only the angelic uncon-
ditional cooperators are left contributing (probably futilely) to the rele-
vant public good.
Maximum cooperation, then, probably requires that reciprocity dy-
namics be supplemented with appropriately tailored incentives”most
likely in the form of penalties aimed speci¬cally at persistent free-
riders. Although trust and reciprocity elicit cooperation from most
players, some coercive mechanism remains necessary for the small
population of dedicated free-riders, who continue to hold out in the
face of widespread spontaneous cooperation, thereby depressing the
contributions made by relatively intolerant reciprocators. In the face of
a credible penalty, however, the committed free-riders fall into line.
The existence of such penalties in turn assures the less tolerant recipro-
cators that their cooperation won™t make them chumps; they thus con-
tinue to cooperate, less out of material interest than out of positive
reciprocal motivations. And because the less tolerant reciprocators
contribute, so do the neutral and tolerant reciprocators, generating
an equilibrium of near-universal cooperation. Again, these dynamics
are borne out by empirical evidence, particularly from public goods
experiments where subjects can retaliate against defectors.15
The uneven effect of penalties in promoting and dissipating trust
calls attention to the expressive dimension of incentives. Incentives do
more than affect individuals™ calculations of the costs and bene¬ts of
particular forms of conduct; they also shape their impressions of the
attitudes and intentions of those around them.16 Laboratory and real-
world schemes that use generally applicable incentives convey the mes-
sage that noncooperation is the norm, and thus sti¬‚e the reciprocal
motivations of even neutral reciprocators, whose defection predictably
spills over onto even the most forgiving ones. Targeted retaliation, in
contrast, conveys a very different message. Because all individuals are
aware from social experience that there are some committed free-riders
out there, no one is surprised or disappointed to see penalties aimed at
those types. Accordingly, such penalties don™t create the cueing, mask-
ing, or crowding out effects associated with more generalized incentive
regimes. On the contrary, penalties understood to be necessitated only
by the existence of committed free-riders have a trust-enhancing effect,
for they imply that most individuals are not inclined to shirk. Targeted
The Logic of Reciprocity 347

retaliation works, in sum, because it simultaneously coerces dedicated
free-riders, calms unforgiving reciprocators, and avoids confusing or
demoralizing neutral and forgiving reciprocators.

12.3 Tax Compliance

Tax compliance is the consummate collective action problem from a
public policy point of view. Society collects taxes to ¬nance a variety
of goods”from education to highways to national defense”that
bene¬t its members collectively. Nevertheless, it is in the individual
material interest of every citizen to free-ride on her fellow citizens™ con-
tributions to these goods while withholding any contribution of her
own. Accordingly, the conventional theory predicts that individuals,
as wealth maximizers, will evade their taxes unless furnished with
incentives”in the form of threatened penalties”that make the ex-
pected return from evasion smaller than the expected return from
This account of tax evasion is embarrassingly ill-supported by em-
pirical evidence. Econometric studies have concluded that the expected
penalty for evasion explains little if any of the variation in compli-
ance across space or over time.18 Survey measures also ¬nd only very
modest correlation between reported compliance and individuals™ sub-
jective perception of the expected penalty for evasion.19 Finally, labora-
tory experiments that simulate the decision to evade taxes suggest that
probability and severity of detection can in¬‚uence individual decisions
to evade, but only when they are set at levels far in excess of those
associated with actual policies.20
Substantially more important, empirical research suggests, is a com-
plex of factual beliefs and emotional dispositions. Thus, an individual™s
perception of the extent of evasion is a powerful predictor of com-
pliance behavior: the higher an individual believes the rate of tax-
cheating to be, the more likely he or she is to cheat as well.21 The
prospect of shame (or potential stigma) and guilt have a similar effect.
The more likely an individual believes it is that she will be condemned
by others should she be caught, the more likely she is to refrain from
evading. By the same token, the more regret or remorse that an indi-
vidual believes she™d experience for engaging in evasion, the less likely
she is to engage in that crime.22
These are exactly the factors one would expect to in¬‚uence tax com-
pliance if individuals behave like moral and emotional reciprocators. A
348 Kahan

strong reciprocator wants to understand herself and be understood by
others as fair, but she loathes being taken advantage of. With tax col-
lection as with other collective action settings, the extent to which
others appear to be contributing to the good in question determines
which of these sensibilities comes into play. If most other individuals
seem to be paying their taxes, then evasion will provoke either guilt,
shame, or both in the reciprocator who covets the respect of others
and of herself. If, in contrast, most individuals appear to be evading
taxes, then complying will not make her feel guilty or ashamed at all;
it will make her feel like a sucker.
This interpretation of the data is con¬rmed by an experiment that
tested how the 1986 Tax Reform Act affected compliance levels.23 One
hypothesis, suggested by the conventional theory, was that individuals
would become more or less willing to evade depending on whether the
Act had increased or decreased their relative tax burden. The study
found no such correlation. What did shift patterns of compliance, the
researchers found, were the types of interactions that individuals had
with other taxpayers in the months leading up to the reform: those
who encountered others who expressed a positive attitude toward,
and commitment to complying with, the Tax Reform Act displayed
greater commitment to complying with it themselves, whereas those
who encountered others who expressed negative attitudes displayed
less commitment.24 This effect, moreover, was explained completely
by variation in the shame and guilt that the two groups of taxpayers
anticipated for failing to pay their taxes.25 In other words, as moral
and emotional reciprocators, these individuals naturally felt guilt
and shame for failing to contribute to the public good of tax compli-
ance in proportion to their perception that others were or were not
The conventional theory of collective action is just as weak at ex-
plaining variance in tax compliance across nations as it is in explaining
compliance across individuals. Tax compliance rates vary dramati-
cally across nations. Essentially none of this variance, however, can be
explained by differences in the expected penalty for evasion. More
important, researchers have concluded, are differences in public atti-
tudes toward tax laws. In some nations (including the United States),
individuals tend to view paying their taxes as an important civic obli-
gation and are highly motivated to pay for that reason. In other nations
(including many in Western Europe), individuals regard tax obliga-
tions much more casually (akin, say, to traf¬c regulations in the United
The Logic of Reciprocity 349

States) and display no particular moral aversion to evading them if
they feel they can safely do so.26
Varying national ˜˜tax cultures™™ of this sort are perfectly under-
standable under the strong reciprocity theory. Because individuals
are reciprocators, their decisions in a collective action setting feed on
each other, generating multiple high- and low-cooperation equilibria
independent of the material payoffs associated with cooperating or
defecting. If individuals believe that those around them are inclined
to pay their taxes, they will (as a result of guilt, shame, pride and
the like) be more likely to comply, thereby strengthening the per-
ception that individuals are generally inclined to pay. If, in contrast,
individuals believe that those around them are inclined to evade, re-
sentment will inhibit them from complying, strengthening the percep-
tion that most individuals are inclined to cheat. In other words, what
we should expect to see under the strong reciprocity theory is exactly
what we do see”competing and relatively durable norms toward tax
The empirical evidence also bears out the strong reciprocity theory™s
anxiety about the effect of self-defeating material incentives. Experi-
mental evidence suggests that when taxpayers are exposed to informa-
tion highlighting the penalties for evasion, they respond in much the
same way that subjects in public goods experiments do when fur-
nished with generalized material incentives to contribute”namely, by
contributing less.28 Researchers have also found that highly politicized
auditing campaigns tend to provoke a higher incidence of tax cheating
rather than a lower one.29
The mechanism for these effects appears to be social cueing. When
government engages in dramatic gestures to make individuals aware
that the penalties for tax evasion are being increased, it also causes
individuals to infer that more taxpayers than they thought are choos-
ing to cheat. This distrust of one™s neighbors triggers a reciprocal
motive to evade, which dominates the greater material incentive to
comply associated with the higher than expected penalty.30
Is there a way for tax enforcers to bolster taxpayers™ trust in one an-
other? One policy that seems to do that is simply to advise citizens
that the vast majority of taxpayers are in fact complying. In a study
sponsored by the Minnesota Department of Revenue, researchers sent
letters to a group of individuals stating that tax compliance rates were
in fact much higher than what public opinion polls suggested citi-
zens believed them to be. The individuals who received these letters
350 Kahan

subsequently reported more income and claimed fewer deductions
than did individuals in a control group. This is exactly what the
phenomenon of strong reciprocity would predict: when they learn
that others are in fact disposed to contribute their fair share, indi-
vidual taxpayers, just like individuals in public good experiments,
cooperatively respond in kind. Likewise, consistent with the strong
reciprocity theory”and at odds with the conventional economic one
”the Minnesota study found that individuals advised of high compli-
ance rates paid more taxes than did individuals who received letters
advising them that their returns would be subject to a greater rate of
Another policy that appears to promote trust and hence bolster re-
ciprocal cooperation is the enactment of popular reforms. As the study
of the 1986 Tax Reform Act demonstrates, such reforms promote the
expression of positive views toward the law. When exposed to these
views, individuals infer that others are inclined to comply. This con-


. 11
( 13)