Truth
Reciprocity and Selfishness: Society as the Interaction of Two Kinds of Human Beings
Abstract
© Dr M D Magee
Contents Updated: Tuesday, 17 November 2009
Reciprocity Versus Selfishness
Thomas Hobbes evoked an image of man, in the state of nature as being with no arts, no letters, no society, but continual fear and danger of violent death, and his life was “solitary, poor, nasty, brutish, and short”. Hobbes primeval man was therefore imagined as man before he became a social animal, a solitary animal that owed nothing to any other of its kind, except during the mating season. Its instincts are naturally selfish. But humans are social animals. Humanity and civilization require human beings to act in certain ways towards their fellow humans. It is society that has made humanity, and no one is human without it. Inhumanity is a reversion to the solitary brutish thing that went before.
Humanity in the past has been categorized as sociological man (Durkheim) and as economic man (Adam Smith). The two studies, sociology and economics, saw human beings as being motivated quite differently. Neoclassical (capitalist) economic theory suggests that economic man is driven by selfishness and personal aggrandizement (politely called “self regard”)—the economic self regarding preferences model (SRPM). Civilization depends upon socioeconomic institutions—property rights, codified law, and a strong state. But these institutions are recent in history—in the last 6000 years, probably too recently for evolution to have had any direct effect in their emergence. Human society and coöperation existed long before it developed its social institutions, so how was social order maintained beforehand?
Sociologists tend to take the view that people are socialized into being helpful to each other. Ever since Durkheim, before WW2, the internalization of social values that restrain people’s self interest has predominated in the sociological outlook. The sociological man lives according to the social norm of reciprocity or selfless coöperation. It has gone on so long that socialization has been internalized, and these internalized social norms constitute the aspects of human nature that civilize us. Thus, Talcott Parsons argued that people internalize social values, and want social approval to such a degree that there is little conflict between self interest and social values, except in a few social deviants. But how do self regard and social norms interact to determine individual behavior?
A self regarding actor in a social situation wants to maximize their own payoff. They care about the payoffs to other people only insofar as they influence their own payoff. Indeed, the economic SRPM often seemed successful, but these days, the economic assumption that humans are predominantly self regarding has been decisively rejected by controlled experiments in economics. This large body of experimental work has produced differences from the model’s predictions, notably when fairness or cooperation is a consideration.
In the ultimatum game, two people who do not know each other get $100 on condition that they agree how to share it among themselves, but with no interaction, or haggling. One of the two, chosen at random, is the one to propose the cut, and the other—who also knows the rules and the sum of money at stake—has to respond by accepting or rejecting it. When the proposal is accepted, the cut is made and each party gets their share, but when it is rejected, neither party gets anything. The games stops after one round, and is never repeated again with the same two people.
Many people offer 50 percent, on the grounds that it is the fairest cut, and so will be accepted. More selfish people might offer less than half expecting the responder to agree, because otherwise both get nothing. The responder can only say “yes” or “no” to the amount offered, but “no” means they get nothing whereas “yes” means they will get something, unless the amount offered is nil. Suppose the proposer did offer nothing, keeping the 100 bucks all to himself. The overwhelming response is to refuse the offer. The responder refuses to accept the offer of nothing, knowing they can be no worse off but that the proposer will lose the 100 bucks he hoped to keep to himself. The responder is punishing the proposer for being selfish.
Suppose the sum offered was $1. $1 is better than $0, but most people would consider it an insult, and would still refuse. They would lose the $1 but the selfish proposer, hoping to get $99, loses a lot more. An offer of $10 of the $100 is getting to be more serious. $10 is worth having, but the proposer will be getting $90, a cut of $100 that still looks unfair to most people. It is found that two thirds of offers are between 40% and 50%. Only four in 100 people risk offering less than 20% because too many people think it is unfair, and would rather punish the greedy proposer at the cost of the $20 to themselves. More than half of all responders reject offers that are less than 20%.
The puzzle in this is that, according to the SRPM, everyone should accept something for nothing. No offer should be too small as long as it is more than nothing, and the sum taken by the other party ought not to affect the acceptance of a gain to oneself. The rational option is for a selfish person—and under SRPM we all are—to accept any offer, so a selfish proposer, who expects the responder to be similarly selfish, will make the smallest offer, keeping the rest, expecting it to be accepted as long as the responder got something, because $1 is better than nothing. So SRPM, which assumes that people are selfish and rational, predicts that the proposer would offer the smallest possible share and the responder would accept it.
Most people do not play the game that way, and so the SRPM must be wrong. Similarly, even without any legal enforceability, fair wage offers elicit a worker response above the shirking level expected. And again, under the the economic SRPM, contributions in “voluntary contributions public goods” experiments would be zero, but people often do make contributions, especially when free riders can be punished, even at a cost.
But experiment has not been as popular among sociologists, even though observational field data can rarely discriminate between contradicting sociological theories. Controlled experiments in sociology are needed to show to what extent society shapes each person’s preferences, and how? What is difficult is that a self denying motive observed in someone in a real social setting can be too easily attributed to the selfish motive of the agent desiring a good reputation, even if it is driven by genuine concern, or even altruism. In laboratory experiments, reputation can be ruled out as an incentive in a properly devised experiment.
People like Herbert Gintis and Ernst Fehr are gradually doing these experiments and are increasingly influencing sociological and economic thought. Gintis is keen to popularize what looks like important work. They not only reject the assumption of predominant selfishness routinely made in economics, but also support the sociologist’s view—reciprocity explains the deviations from the SRPM, albeit conditionally—the two contradictory motivations interact in subtle ways. Reciprocity means that people normally respond more cooperatively to friendly approaches, and more unpleasantly to hostile approaches, than predicted by the self interest model (Fehr and Gächter). So, strong reciprocity is the disposition:
- to coöperate with others on condition they coöperate with you
- to punish those who do not, even at a cost to the punisher.
Reciprocity differs from altruism in that altruism does not require any prior indications of reciprocal coöperation but is offered spontaneously with no obvious self interest at all.
In the experiments, as in the ultimate game, subjects were allowed to earn money, sometimes in substantial amounts, as an economic incentive, and to distribute it according to certain rules, a social response. To rule out reputation effects, interactions between the subjects were anonymous. In different cultural contexts from modern to older in various advanced and undeveloped countries, the results were essentially similar. Even when subjects could gain up to three months’ income (in Indonesia), in a two hour experiment by being self regarding, strongly reciprocal (coöperative, sharing) behavior dominated.
Herbert Gintis has developed an alternative model to the SRPM which he calls the “Belief, Preference and Constraint” (BPC) model, in which the interactions between strongly reciprocal and self regarding actors drive the patterns of coöperation and exchange that emerge in the experiments. Just how strongly reciprocal and self regarding people interact affects the emerging patterns of behavior, and are regulated by the social structure—the rules of the game. The BPC model takes people to have consistent preferences and beliefs about other people’s behavior and about the consequences of their choices:
People behave according to the choices that best satisfy their preferences, given their beliefs and their constraints.
Social Cooperation and Social Sanction
Any human population consists of a mixture of strong reciprocating people mixed with people who are self regarding.
Self interest and the social good can be distinguished clearly in public goods experiments. A public good is the product of coöperation. The incentive structure captured by the public goods experiment in coöperation is ubiquitous in reality, covering all communal institutions and benefits, like public security (ie the absence of violence and crime), military, communal health schemes, education, environmental pollution problems, and law and contract enforcement, as well as the enforcement of all kinds of rules that are beneficial for the common good, but costly for individuals to obey.
Each member of a group of N >= 2 people is endowed with $Y. Each group member can keep this money or invest up to $Y into a group account which represents the public good. The experimenter multiplies every dollar invested on the group account by a factor M—the beneficial effect of coöperating. When all group members have made their contributions to the group account, it is multiplied by M (> 1 but < N ie 1 < M < N), and the resultant amount is equally distributed among the group members.
If all N group members contribute their personal wealth of $Y, each of them will receive N × $Y × M/N = $Y × M. If M = 2, each person can double their income by contributing everything to the public good because their coöperation, the public good, doubles the value of the joint pot. But, if no one knows how much you actually put into the public good, there is an incentive to free ride, by putting in less than the maximum, in the expectation that others will put in all they own, and after the share out, you will benefit as a free rider by the sum held back.
If a subject contributes, say, $5 to the group account, the group as a whole earns $5 × M from this contribution. The doner of the $5 will get back a portion of the public good fund, but, of his $5 contribution, they will get only $5 × M/N—the sum donated multiplied up, and divided by the size of the group sharing it. As M < N, the personal investment of $5 exceeds the return of $5 × M/N from that investment alone. So, the person investing always gets a smaller payoff from their own investment alone than they contribute to the public good. Besides this the donor gets similarly calculated contributions from all the others, though some selfish participants might not have put in anything! Someone so thoroughly self interested that they want a positive gain on any buck invested will never contribute to the public good. Yet the potential return, from the public good fund, if they are not so constrained, is $Y x M!
Now when the experimenter runs the experiment several times with the same sets of people, and they always know who contributed what at the end of each run, coöperation falls, typically to a low level.
This figure (Fehr and Gächter) shows data from six groups with N = 4, Y = 20, M = 1.6 and ten runs initially, followed by another ten under different rules. In period ten, roughly 55% of the subjects contribute nothing to the public good and the remaining subjects contribute little. Towards the end of the first ten runs, individual self interest dominates behavior, and the sociological view that internalization of coöperative values suffices for social coöperation is wrong. Humans who obey social norms without regard to self interest seem uncommon.
After period 10, subjects could punish each other based on their knowledge of each other’s personal contribution. They could assign a financial punishment to any free rider, decreasing their payoff, but only at a cost to themselves as a punisher. Each punishment point costs a punisher $1 (or money unit) and penalizes the easy rider by 10 percent (roughly 3 money units). Purely self interested people would not punish anyone, because it is a cost to them, so the punishment system ought not to alter the motives of purely selfish people, especially in the last few runs.
Given that punishment can be expected after run 10, behavior converges to almost full coöperation. In period 20, the average coöperation rate reaches almost 100%. 83% of the subjects contribute their whole endowment to the public good, and most of the remaining subjects contribute close to the maximum. The same subjects whose coöperation reduced in the first ten runs, were highly coöperative in the second ten. In some groups, every group member contributed nothing to the public good in run 10 and everything in run 20. The same subjects showed almost diametrically opposite behavioral patterns.
In the first ten runs, people were self interested, but in the second ten, the Durkheim/Parsons notion of internalized social values seems to obtain. As the subjects knew the experiment ended at run 20, there were no future economic benefits to be had, yet a cost was still incurred by punishing, so self interest inhibits punishing in run 20. Why, then, should self interested subjects coöperate in the last run in the expectation that anyone with self interest in mind would be punishing? Moreover, a rate of coöperation of roughly 50% was maintained for the first few runs even in the absence of punishment.
The experiments showed a majority of subjects prefered strong reciprocity, but a substantial minority were completely self regarding. So strong reciprocity amounts to a propensity to coöperate when other members coöperate, and to punish easy riders even though it is costly for the punisher. The behavior of strong reciprocators is not primarily shaped by their self interest but presumably by internalized coöperative social values, and conditional coöperation and punishment motives too. Strong reciprocity is not simply long term, enlightened self interest, because the strong reciprocator coöperates and punishes even in anonymous one shot interactions where the coöperative and punishing acts obviously reduce their economic net gain.
Reciprocal Fairness and Inequity Aversion
Two of the most prominent social forces have been termed “reciprocal fairness” and “inequity aversion”. A “reciprocally fair” subject is motivated by the desire to respond to kind acts with kindness and to hostile acts with hostility. An “inequity averse” subject is motivated by the desire to avoid inequity and to implement equitable outcomes. Inequity averse and reciprocally fair subjects are no saints who resist unfair outcomes and punish unfair behavior under all circumstances. Rather, these subjects value equity and reciprocal fairness in addition to their economic self interest, implying that if the costs of maintaining equity or of reciprocally fair behaviors increase they are less likely to engage in these behaviors. Models of inequity aversion and reciprocal fairness can, in particular, explain why we observe little coöperation in the absence of a punishment opportunity, and why coöperation flourishes when punishment is possible.
In the absence of a punishment opportunity, strongly reciprocal subjects will initially coöperate if they believe that others will also coöperate. However, they notice over time that other group members—the self regarding ones—free ride. As strong reciprocators are only willing to coöperate if most others also coöperate, they cease to coöperate in the later periods of the experiment. Strong reciprocators also have a desire to punish free riders because they perceive free riding on their coöperation to be unfair. However, stopping coöperation is the only way to punish other group members in the absence of a direct punishment opportunity that enables the subjects to target the punishment on the free riders. Thus, both the free riders and the strong reciprocators contribute little or nothing to the public good towards the end, albeit for different reasons. The self regarding subjects contribute nothing because this maximizes their economic payoff. The strong reciprocators contribute nothing because they are only willing to coöperate if sufficiently many others also coöperate, and because free riding is the only way available to punish the self regarding subjects. The self regarding subjects ultimately induce the reciprocators to free ride as well in the absence of a direct punishment opportunity.
In the presence of a direct punishment opportunity, virtually all subjects ultimately coöperate despite the fact that there is a substantial share of self regarding subjects. The reason is that the strong reciprocators can now punish the defectors directly, creating an economic incentive for the self regarding subjects to coöperate. Moreover, the strong reciprocators will also coöperate because they need not fear others’ defection, as the self regarding people are disciplined. Thus, the strong reciprocators induce the self regarding subjects to coöperate in the presence of a direct punishment opportunity.
The figure is based on the same public goods incentives, without a direct punishment opportunity, as in the first figure. The squares show the conditional coöperator’s average behavior. These subjects increase their coöperation level if they believe that the other group members will do the same. 50% of the subjects behaved in this way. On average—the conditional coöperators do not completely match the others’ expected average contribution, staying slightly below the 45 degree line. Self interest must affect conditional coöperators’ choices. The circles in the Figure show the large fraction of subjects, 30% of participants, who are purely self regarding regardless of what the others do—the perpetual free riders. The remaining 20% of the subjects exhibited other behavioral patterns such as an increasing response to others’ contributions over the first 10 expected contribution units while afterwards they show a decreasing response.
The figure explains why the maintenance of a high coöperation level is not possible in the absence of a direct punishment opportunity. Suppose all subjects initially believe other group members will contribute their whole endowment of $Y = $20 to the public good, then the response curve of the conditional coöperators (the squares) show that they would contribute approximately 14 units for this belief, while the self regarding subjects would contribute nothing. The information feedback at the end of the period would reveal that the actual average contribution is considerably below the expected average contribution of $20. The subjects are therefore likely to revise their beliefs about others’ expected contribution downwards which would induce the conditional coöperators to reduce their contribution level in the next period. In this way, the heterogeneous mix of self regarding and strongly reciprocal subjects generates a downwards trend in coöperation levels.
If the self regarding subjects rationally anticipate the response of the reciprocal subjects, they will also contribute during the initial periods of a finitely repeated game. By contributing in period t they can induce higher contribution levels of the reciprocal subjects in the subsequent periods. However, towards the end, these future (selfish) returns from current contributions decline so that rational egoists will cease to coöperate, which then induces the reciprocal subjects to also stop coöperating.
Another important lesson from this is that social order sometimes breaks down—for example, after natural disasters or at the end of a war (as in Iraq)—but it does not prove that all people are self regarding. The social order can break down even if a large share of people have internalized coöperative social values and are, in principal, willing to coöperate. But, if free riders go unpunished, coöperative people tend to stop coöperating. In theory, a relatively small minority of self regarding people suffices to generate a breakdown of coöperation in the public goods game, if the free riders cannot be punished individually. It shows that social rules, laws and such, though they need not be ruinous, must be applied, just as Beccaria said two centuries back. Society will begin to break down when people realize that people, whether criminals, the rich, or our Representatives, can act antisocially and get away with it. Likewise, a relatively small minority of inequity averse people will suffice to generate a fully coöperative outcome when punishment can be directly targeted at the individual free rider.
Stable and Transient Employment
This figure illustrates the punishment pattern in the experiment with a punishment opportunity (black bars). The deviation of the punished subjects’ contributions from the average contribution of the other group members is an important determinant of punishment. The more they target individual free riders relative to the group average, the higher is the punishment. If the free rider’s contribution deviates between 20 and 14 from others’ average contribution, the free rider receives 7 punishment points which reduces his income by 70%. Thus, free riding is strongly punished when the group composition is stable (partner treatment) and, as a consequence, potential free riders have a strong incentive to coöperate.
The black bars in the chart do not yet demonstrate that strong reciprocity drives punishment because if the same group of people interact for 10 periods, there may be a selfish incentive to punish free riders. After all, a free rider who is punished increases contribution levels in the next few periods which provides also a benefit for the punisher. This problem is exacerbated if—as in the work of Yamagishi (1986) and Ostrom, et al, (1992)—the subjects do not know how many periods they will interact together in a group. In this earlier work, the researcher deliberately did not tell the subjects the number of periods they would interact together because they wanted to implement a truly repeated game in which there are also selfish incentives to coöperate and to punish.
In other experiments, the group composition changed randomly from run to run in such a way that no subject ever encountered another subject more than once. In this setting (stranger in the Figure), a punisher could not reap any benefit from the future coöperation of the punished free rider because the free rider was never in the punisher’s group in future periods. Therefore, there cannot be any selfish benefit from punishing a free rider in the stranger setting. Nevertheless, punishment is still very high (white bars), indicating the existence of strong reciprocators. The differences in the punishment pattern between the partner and the stranger setting are not significant, so other regarding motives drive the bulk of punishment choices.
The Influence of Social Exchange
Different social structures, such as the degree to which targeted punishment opportunities are available, may generate completely different aggregate patterns of interaction. Because the same people cause these divergent patterns, we cannot attribute the divergence solely to individual characteristics. The social structure is causally involved in this emergent pattern. Some researchers think this is a reason to take care in uncritically accepting these studies until the sensitivity of the conditions are fully understood. If they are useful, they might be at the edge of chaos, and small changes can have unpredictable effects.
If only self regarding people existed, the two social structures would generally produce the same outcomes. The social structure generates non obvious, perhaps counterintuitive, patterns. In the absence of a targeted punishment opportunity, even a large majority of strong reciprocators will not be able to sustain coöperation, whereas even a relatively small number of strong reciprocators can enforce a fully coöperative outcome in the presence of a punishment opportunity.
A social exchange involves the mutual transfer of benefits such as goods, ideas, aid, or social approval under conditions of incompletely specified obligations (Blau 1964). The incomplete specification of the exchange partners’ obligations implies that such a social arrangement cannot be enforced through binding contracts serviced by third parties (say the judiciary). Sociologists noted long ago that social exchanges permeate most human interactions and are of relevance for many economic relations such as the employment relationship or the provision of complex professional services. Social exchanges are likely to affect almost all long term relationships between economic actors. This has led to a rich theoretical literature.
Economics largely ignored social exchange. How are social exchanges enforced? After all, the partners’ obligations are ill specified. So why should the exchange partners obey such ill specified “terms of trade”? How do social exchanges affect social relations between the exchange partners? How do successful social exchanges get started? What is their impact on aggregate entities like prices and efficiency? How do social exchanges affect competition between actors and vice versa?
In the gift exchange game, one subject takes the role of an employer who offers a wage W, the other subject takes the role of an employee who provides effort E in exchange for W. First the employer offers a binding wage payment W, then the employee observes W and responds with a choice of E. Because the wage offer is binding, W has to be paid regardless of how large the employee’s effort E will be. This feature simplifies the situation but it is not essential for our main arguments. The employer’s monetary payoff, denoted by p , is given by p = A × E - W where A is a positive constant determined by the experimenter. Higher effort increases, and a higher wage reduces p.
The employee’s monetary payoff, denoted by U, is given by U = W - C(E) where C(E) represents the cost of effort. As higher effort levels are (by the construction of the experiment) associated with higher cost levels, a higher effort reduces U. We also assume, for the moment, that the two parties play the game only once. This game captures the key element of a simple social exchange situation because the employee is completely free to choose whatever effort level he likes if there is no third party who can enforce a contractually specified effort level. The absence of third party enforcement requires that one of the following two conditions be met:
- the employer cannot stipulate a contract with a well specified effort level, so third party enforcement cannot function—the third party has nothing certain to enforce
- the third party cannot verify the actual effort level the employee chooses.
Let the gift exchange game takes place only once. What pattern of behaviors should we expect? When both the employer and the employee are purely self regarding, caring only for their own economic payoff, the employee will always choose the lowest possible E, because this maximizes U, and the employer will choose the lowest possible W, because this maximizes p. However, when, among the employees, there are enough strong reciprocators who respond to a kind act—a fair wage—with fair (non minimal) effort levels, the employer may have an incentive to pay fair wages. The possibility of this kind of reciprocal fairness gave this game its name because if third parties cannot enforce the effort the exchange partners essentially can only exchange “gifts”.
Indeed, reciprocally fair exchange patterns have frequently been observed in the gift exchange game. If the employer pays higher wages, the employees generally chose to work harder. When strong reciprocators are present among the workforce, people’s internalized social values act to enforce fairness. The existence of sufficiently many strong reciprocators among the employees may even transform the interactions in labor markets in a profound way—profit maximizing employers have an incentive to pay “fair” wages, because higher wages generally bring a higher effort among the workforce in response. It means wages in labor markets may not be solely determined by the laws of supply and demand. The perception of employees that the wages paid by the employers is fair motivates them to work harder because of their internal social values.
Fehr and Falk (1999) compared a bilateral gift exchange experiment with a competitive gift exchange experiment. In both, an employer could only employ one worker, but there was an excess supply of workers in the competitive condition, whereas the number of workers equalled the number of employers in the bilateral condition. In the competitive case, employees had to compete to get an offer because there were fewer jobs than workers. According to standard economic principles, the excess supply of workers should drive wages in the competitive conditions below those in the bilateral condition. However, Fehr and Falk found no difference in wages between the experiments because the existence of strong reciprocators in both meant firms in both cases had an incentive to pay “fair” wages. Fairness norms affect price formation in social exchanges by weakening the impact of supply and demand. Fairness norms shape price formation because the internalized social values of strong reciprocators induce them to respond to fair wages with fair effort levels, which is an economic incentive to employers to pay fair wages.
And social exchanges may transform the pattern of market exchanges in a more fundamental way than simply affecting wages or prices. They may change the way in which trading partners interact with each other. The enforcement problem inherent in social exchanges causes long term relationships between the trading partners. If third parties enforce contractual obligations, ie if the possibility for social exchanges is absent, there is no necessity for the traders to trust each other and they do not care about their partner’s identity. As a consequence, trade predominantly takes place in one shot interactions. However, because traders need to trust each other in the absence of third party enforcement, so long term relationships emerge.
As traders care a lot about the trustworthiness of their partner under social exchange conditions, it is essential for them to know with whom they are trading while the partners’ identity plays no role under third party enforcement of contracts. To initiate trade under social exchange conditions, offers should be made to specific potential trading partners who have a reputation for being trustworthy. Usually, offers are made to the whole group of potential trading partners under third party enforcement because this maximizes the chances that one of the potential partners accepts the offer. Trust enhancing strategies of the trading partners are associated with an enormous increase in effort enforcement relative to a situation where only one shot interactions are possible.
The experiment lasted 15 runs in both the “one shot” and the “relations” treatment, and the subjects in the role of an employer could make wage offers to the workers in each period. There was always an excess supply of workers, ie some workers were unemployed in every period. Wage offers could either be targeted to specific workers or to the whole group of workers. Once a worker had accepted a wage offer he chose an effort level between 1 (the minimum effort) and 10 (the maximum effort). Employers and workers were assigned identification numbers in both treatments, but they were randomly assigned to the subjects in every run in the “one shot” treatment, and they remained the same throughout the 15 runs in the “relations” treatment.
So, when employer 5 traded with worker 8 in period t in the relations treatment, and the employer was satisfied with the performance of worker 8, he could make another wage offer to the same person (worker 8) in the next run. In the one shot treatment, worker 8 was a different person in every run. Effort enforcement is limited when long term relations between the trading parties are ruled out, but high levels of effort can be enforced when long term interactions are possible. The difference again lies in the interaction between self regarding and reciprocal types. If no long term relations are possible, the self regarding people will never choose effort levels above the minimum. Only the strong reciprocators will provide non minimal effort depending on the wage they are offered.
However, when long term relations are possible, the presence of strong reciprocators among the workers disciplines the self regarding workers. Due to the existence of strong reciprocators, it is profitable for employers to pay fair wages—in excess of the income of being unemployed—even in the final period of the experiment. The employer can implicitly threaten not to renew the contract (relationship) in the next period to punish the worker in any run other than the last one. The refusal to renew a contract always imposes a loss on the worker owing to a strong possibility of unemployment due to the excess supply of workers, so self regarding workers have strong incentives to provide non minimal effort levels in the “relations” treatment, incentives based on the presence of the strong reciprocators.
Fair Wages and Social Exchange
Social exchanges not only transform the interaction patterns between the trading partners in radical ways, but also cause a very different distribution of the earnings from trade, as can be shown by comparing the relations treatment with a treatment in which third parties enforce the effort level. The two treatments are identical in all other respects, particularly that excess workers in both treatments put them in a weak bargaining position. As the Figure shows, the workers’ weak bargaining position causes a very uneven distribution of earnings when third parties enforce the contract, whereas the distribution is much more equal under social exchange conditions (in the relations treatment). The possibility of varying effort according to the workers’ fairness preferences, counteracts the weak bargaining position that is generated by their excess supply.
The chart also nicely illustrates the efficiency gains that accrue in relationships with a longer duration. The longer a relationship ultimately lasts in the relations treatment, the higher are the earnings of both the employer and the employee. These efficiency gains are a result of the fact that the effort level is higher in longer lasting relationships. Employers did not renew the relation with those workers who did not provide sufficient effort, so, relationships which survive longer generally have a higher effort, which also brings higher total earnings. In relationships that lasted longest (11-15 runs), the sum of the earnings is as high as under third party enforcement.
The Impact Of Society On People’s Preferences
“Society” shapes individual preferences through socialization, but other behavioral sciences often ignore and implicitly reject these observations, which are part of the core of sociological theory. Experimenters conducted ultimatum games cross culturally in 15 small scale societies, and public goods and other experimental games in some of them. In an ultimatum game, two players—a proposer and a responder—bargain about the distribution of a given sum of money (or some other valuable resource). The proposer makes a proposal on how the money should be distributed among the two, and the responder accepts or rejects it. When accepted, the money will be split according to the proposal. When rejected, both players receive nothing.
If both players are purely self regarding, the responder will accept any positive amount, however small it is, so the proposer who anticipates the responder’s behavior will offer the smallest positive amount to the responder. However, responders frequently reject a sizeable fraction of the available money even if the stakes are as high as three months income. Self regarding preferences cannot explain a rejection of a sum of money in an anonymous one shot ultimatum game, but rejections most likely reflect internalized fairness norms. Due to the high rate of rejections, even self regarding proposers have an incentive to make relatively fair offers.
The 15 societies were rank ordered in five categories—“market integration” (how often do people buy and sell, or work for a wage), “coöperation in production” (is production collective or individual), “anonymity” (how prevalent are anonymous roles and transactions), “privacy” (how easily can people keep their activities secret), and “complexity” (how much centralized decision making occurs above the level of the household). Using statistical regression analyses, only market integration and coöperation in production were significant, accounting for 66% of the variation among societies in mean ultimatum game offers. So differences in societies in “market integration” and “coöperation in production” explain a substantial portion of the behavioral variation between groups.
Moreover, experimental behavior mirrors patterns of interactions found in everyday life in that society. Among the Papua New Guinean tribes of the Au and the Gnau, many proposers offered more than half of the available sum of money, but many of these offers were rejected! It reflects the Melanesian culture of status seeking through gift giving. Making a large gift is a bid for social dominance in everyday life in these societies. Rejecting the bid is a rejection of being subordinate.
The Hazda, a tribe located in Tanzania, made low offers and had high rejection rates, which mirrors the tendency of these small scale foragers to share meat, but with a high level of conflict and frequent attempts of hunters to hide their catch from the group.
Almost all offers made by the Ache, a tribe located in Peru, are close to 50% and no rejections occurred. In daily life, the Ache regularly share meat, which is distributed equally among all the households, irrespective of which hunter made the catch.
The impact of piece rate incentive schemes, hourly payment schemes, and team based pay on the degree of conditional coöperation of workers in several bike messenger companies in Switzerland and the US were tested in a sequentially played prisoners’ dilemma (PD) game. Conditional coöperation is the frequency with which the second movers in the sequential PD coöperate in response to first mover coöperation. Conditional coöperation is significantly more frequent in firms that pay hourly wages or where the revenue is shared among the workers than in firms with piece rate incentives. The lower degree of conditional coöperation in firms with piece rate incentives could be due to the possibility that self regarding workers self select into these firms.
Yet, it could also be caused by the incentive system itself. Piece rate incentives renders mutual help costly for the workers and may make workers even more self regarding. The latter is the more likely reason because firms with piece rate incentives are not located in the same cities as firms that pay on an hourly basis, or share the revenue among the workers. Because the cities are relatively far from each other, the bike messengers typically cannot chose to work in firms with different incentive schemes, rendering the self selection explanation unlikely.
Women in the US are competition averse. When given the choice to enter a piece rate incentive scheme (that rewards absolute experience) and a tournament incentive scheme (that rewards only the winner of the tournament) women predominantly prefer the piece rate scheme. The vast majority of US men prefer the tournament scheme. This preference pattern across men and women even prevails when they control for risk preferences and equity preferences.
To test whether women’s competition aversion had to do with the fact the US culture is still dominated by males, the same experiments were carried out in a paternalistic culture, among the Maasei in Tanzania, and in a matrilineal society, the Khasi in India. Among the paternalistic Maasei, men were more likely to prefer the tournament incentive than were the women, but in the matrilineal society of the Khasi, the gender results were the opposite—the majority of the women (54%) preferred the tournament while a majority of the men (61%) preferred the piece rate incentive. These results are consistent with the notion of culturally determined gender roles as an important determinant of preferences for or against competition.
Summary
A minority of selfish people may make sustaining coöperation in the absence of a punishment mechanism impossible, whereas a minority of strong reciprocators may permit coöperation to flourish when a punishment option is available. The models show:
- how a mixed population of selfish actors and strong reciprocators enforces and transforms anonymous trading relationships in rich social exchanges
- the differences in the distribution of the gains from trade between both standard neoclassical markets with third party enforcement, and social exchange markets with endogenous enforcement facilitated by strong reciprocators
- why fairness plays a prominent role in bilateral bargaining whereas its role is limited in competitive markets with third party enforcement
- why the prisoner’s dilemma game may be better viewed as an assurance game with multiple equilibria rather than as a game about a prisoner’s dilemma
- why employers may deliberately generate social exchanges by offering incomplete contracts to their employees rather than complete ones.
Confining attention to equilibrium behavior is not always justified. People sometimes hold irrational beliefs and the understanding of belief formation constitutes a difficult problem. To understand how beliefs and preferences in BPC form, and to better predict and understand individual behaviors and aggregate level phenomena, the distribution of preferences and beliefs in the population needs to be examined empirically. This empirical knowledge can then be fed back into the model yielding a unification of sociological and economic behaviour theory based on carefully controlled experiments and the BPC model.
Further Reading
- More about morality, and justice as fairness, and more, the Principle of Humanity
- A lot more on religious origins—five linked webpages
- More on the death of God and secular Christianity
