Wage Rigidity and Social Norms in Experimental Labour Markets

Bachelor Thesis 2006 41 Pages

Economics - Job market economics





1.1 The History of Wage Rigidity in Economic Thought
1.2 Layout of the Paper

2.1 The Neoclassical (Labour) Market
2.2 The Influence of Social Norms
2.2.1 Fairness in the Labour Market
2.2.2 Reciprocity in Wage Bargaining

3.1 The Gift Exchange Model
3.2 Fair Wage-Effort Hypothesis

4.1 The Labour Market in Experimental Economics
4.2 Testing the Model: Theories of Wage Rigidity in Experimental Labour Markets
4.2.1 Experimental Design of Gift-Exchange Games
4.2.2 Experimental Results
4.2.3 Introducing Wage References into the Game
4.2.4 Experimental Evidence
4.3 Social Norms in Experimental Labour Markets: Reciprocal Altruism and Unconditional Fairness vs. Reputation
4.4 The Effect of Third Party Intervention in an Experimental Labour Market
4.5 On the Robustness of Experimental Findings
4.5.1 Real Effort Experiments
4.5.2 The Effects of Social Framing in the Laboratory
4.5.3 Raising the Stakes: The Labour Market as a High Stake Game




Figure 1: The Neoclassical Labour Market. (Fischer and Heier, 1983, p. 51)

Figure 2: The Wage-Effort Relation. (Fehr et al., 1998a, p. 14)

Figure 3: The Wage-Profit Relation. (Fehr et al., 1998a, p. 16)

Figure 4: The Redemption Value-Wage Relation. (Fehr et al., 1996, p. 623)

Figure 5: Mean Punishment ( p < 1) or Reward ( p > 1) and Profit Differences. (Fehr et al., 1996, p. 617)


Table 1: Effort Cost Schedule (Fehr et al., 1998, p. 11)

Table 2: Punishment ( p < 1) and Reward Levels ( p > 1) and Associated Costs k(p) . (Fehr et al., 1996, p. 615)


1.1 The History of Wage Rigidity in Economic Thought

Are wages rigid, and if so, why? The question raised has attracted economists’ attention ever since Keynes (1936, p. 289) suggested that wages were rigid and could lead to involuntary unemployment (Gächter, 2001, p. 478). This wage rigidity can be defined as the tendency of wages to react slowly, if at all, to excess labour supply and/or demand (Wachtel, 1994, p. 482). Keynes’ theory is at odds with the neoclassical model of the labour market, in which wages are flexible and therefore full employment at a market-clearing wage will ultimately be reached (Fischer and Heier, 1983, p. 56).

Due to the clash between Keynes’ theory and the neoclassical model, a lively discussion among economists arose as to whether wage rigidity existed or not. Some economists, such as Lucas and Rapping (1969, p. 748), claimed that wage rigidity was an illusion and that existing unemployment was voluntary, i.e. real wages were below workers’ reservation wages. Others claimed that wages were rigid and started to implant sociological findings into economic models, which gave further explanations as to why wage rigidity existed.

As a consequence of this debate it became clear that evidence was needed as to which of the models and theories actually applied to real world labour markets. Some economists went about this by conducting surveys in the labour market (see, e.g., Bewley, 1999, and Campbell III. and Kamlani, 1997). Others used experimental methods to simulate labour markets and test theories of wage rigidity for their robustness. Their findings, which generally confirm the sociological approaches to wage rigidity, will be the basis of this paper.

1.2 Layout of the Paper

This paper discusses the role of social norms in the labour market and the resulting effects on wage bargaining. It investigates whether social norms can explain the phenomenon of wage rigidity sufficiently by reviewing sociological, theoretical and experimental findings.

As described above, a turn in economic thinking came about when economists first started to implement sociological findings into their theories. For the first time, they questioned the neoclassical rational choice model, which is equivalent to the model of homo economicus (Kaufman, 1999, p. 361). This homo economicus is summarized by Ockenfels and Weimann as a “strictly rational fellow without sex, age, or cultural identity” (Ockenfels and Weimann, 1999, p. 275). The second chapter of this paper presents the boundaries of this rational choice model and therefore of the neoclassical concept of the labour market. Further, the importance of social norms particularly in the labour market is investigated.

The acknowledgment of the role of social norms in the labour market led to a series of wage rigidity theories based on sociological findings (Yellen, 1984, p. 204). Two of the most prominent papers concerning social norms in the labour market are the gift exchange model (Akerlof, 1982) and the fair-wage effort hypothesis (Akerlof and Yellen, 1990), which will be presented in the third chapter of this paper.

The fourth chapter of this paper presents experimental findings on the topic of wage rigidity. Experimental evidence on the robustness of the theories of wage rigidity presented in the third chapter of this paper will be discussed, as well as further findings as to the role of social norms in the labour market.

The last chapter of this paper summarizes the theoretical and experimental findings presented in the preceding chapters, as well as their compatibility, and draws a conclusion.


2.1 The Neoclassical (Labour) Market

In neoclassical economics, the labour market is modelled as any other goods market. It is therefore based on the same assumptions as the neoclassical goods market, including representative actors (Wagner and Jahn, 1997, p. 31) who behave according to the rational choice model. This implies purely selfish behaviour of the actors in terms of maximization of their own utility. The concept of representative actors assumes perfect homogeneity of supply and demand, which allows the aggregation of individual demand and supply preferences to market supply and demand curves.

illustration not visible in this excerpt

Figure 1: The Neoclassical Labour Market. (Fischer and Heier, 1983, p. 51)

The above figure shows a diagram of the neoclassical labour market with an equilibrium and market-clearing wage W0. At the market-clearing wage the employment equals E0. At any wage above W0 involuntary unemployment occurs, e.g. at a wage equal to W1 excess supply and hence involuntary unemployment amounting to E1E2 occurs. At the going wage, these people would be willing to work but will not get employed. In the neoclassical labour market, these wage rigidities can only exist in the short-term, as wages are perfectly flexible (Wagner and Jahn, 1997, p. 38) and the price mechanism will ultimately induce the market-clearing wage.

Permanent wage rigidity therefore contradicts the neoclassical model, yet the empirical fact of wage rigidity is largely undisputed (Gibbons and Katz, 1992, p. 515). An explanation for this is needed, and most economists base their criticism on one particular neoclassical assumption: contractual completeness and perfect information. A complete contract states exactly what each of the parties has to do and how the payoffs are distributed in every possible situation, including the one in which the contract is broken. This completeness, however, is rarely the case in labour relationships (Fehr and Falk, 1999, p. 109). Usually the employee has some discretion over her effort during her working hours without the employer noticing. This leads to a contractual incompleteness in which the employer has to rely on the integrity of her employee.

2.2 The Influence of Social Norms

The assumption of incompleteness of contracts in the labour market gives rise to the influence of social norms in the labour relationship (Fehr and Gächter, 2002, p. 3). The above mentioned integrity of the employee needs to be based on some sort of social norm, as it contradicts the rational choice model. A rational, self-contained employee who knows that she will not be checked on during her time at work will do whatever gives her the greatest positive utility, and that will rarely be the task she is assigned. Still, even with the awareness of the incompleteness of their contracts, most employees do not shirk just because they can, and an explanation for this behaviour given by many sociologists are social norms.

The main criticism of the neoclassical model is therefore the assumption of the representative actor who behaves as if atomized from the influence of others and their behaviour (Granovetter, 1988, p. 187). Economic subjects are social beings who live and can only survive in groups. Within these groups, social norms arise which regulate the behaviour of the group members. Complying to these rules is necessary for the individual in order to survive, and therefore economic activity of the individual is influenced by the social norms of his or her group (Bolton and Ockenfels, 2000, p. 189).

This implies that social norms need to be considered in any market. In the labour market, however, social norms play a particularly important role, as this market is determined by social interaction more than any other market and, in particular, by long-term relationships (Charness, 1999, p. 2). The nature of this social interaction is therefore the main difference between the labour market and any other goods market, or, as Solow put it: “There is a difference between a long-term relationship and a one-night stand, and acceptable behavior in one context may be unacceptable in another” (Solow, 1980, p. 9).

2.2.1 Fairness in the Labour Market

Fairness norms in the labour market are usually concerned with the allocation of payoffs. Even though there is no generally accepted notion of fairness, Fehr, Klein and Schmidt emphasize that a common feature of all fairness definitions is that equals should be treated equally (Fehr et al., 2004, p. 33). In a homogeneous labour market, which implies that all subjects are equal, a “fair” wage therefore resembles an equal split of payoffs between the firm and all its employees. Based on this notion of fairness, Fehr and Schmidt introduced a theory of inequity aversion which assumes that people do not only care about their own monetary payoffs, but also about that of others and that they dislike inequity (Fehr and Schmidt, 1999, p. 2).

Without a doubt, the assumption of perferctly homogeneous subjects in the work force and among the employers does lack realism. An even split between employees and employers therefore does not necessarily represent a reasonable distribution in real world labour markets. A more realistic approach to the level of a “fair” wage is presented by Akerlof (1982, p. 551-557): in his view, each employee has a reference group and compares her own wage with that of others within her reference group. Another factor that builds into the fair wage is the level of past wages earned. The perceived “fairness” of wages according to Akerlof is therefore influenced by the wages of others as well as unemployment benefits and past wages earned, not only the payoff of the employer.

Other perceptions of fairness are concerned with the treatment of workers, such as the implementation of minimum standards. Sociological findings show, for example, that workers’ work morale is influenced by the way they themselves and their co-workers are treated.

2.2.2 Reciprocity in Wage Bargaining

Norms of reciprocity generally regulate reactions to actions, i.e. they regulate how to behave “in return”. Gift giving, for example, is regulated by these norms. Even though there may not be a norm which states to give a Christmas present to each of your cousins, once your cousin starts giving you Christmas presents you do have to return it (Elster, 1989, p. 101). However, if your cousin should fail to give you a gift one Christmas, you are not obliged to give her a Christmas present ever again. Good and bad behaviour is therefore rewarded and punished respectively, regulated by social norms.

This gift giving can also be applied to the labour market, as Akerlof showed in his 1982 gift-exchange model, which is presented in the next chapter. In his view, workers and employers exchange gifts in the sense that employers pay higher than market-clearing wages and workers exhibit greater work effort than necessary in return (Akerlof, 1982, p. 543). According to Dufwenberg and Kirchsteiger, reciprocal norms are particularly applicable to the labour market. This results from the fact that both the employer and the worker know the possible range of wages which could be paid by the employer, as well as the possible range of effort levels. Therefore, the kindness of any action in the form of a particular wage paid or an effort level chosen can be judged by both parties and rewarded or punished accordingly (Dufwenberg and Kirchsteiger, 1999, p. 3).

Hence, the awareness of each of the parties’ intentions in wage bargaining and the ability to reward and punish due to incomplete contracts make norms of reciprocity a strong influence on the labour market.


In this chapter two prominent theories of wage rigidity will be presented which take into account the importance of social norms in the labour market. The first theory presented is the gift exchange model, which explains wage rigidity with the existence of reciprocal behaviour among actors in the labour market. The second theory presented is the fair-wage effort hypothesis, which explains wage rigidity with the existence of fairness norms in the labour market.

3.1 The Gift Exchange Model

The gift exchange model by Akerlof (1982) is the first explicitly sociological model of the efficiency wage theories. Efficiency wage theories explain wage rigidities by assuming that higher wages will lead to an increase in output and therefore induce higher profits. Profit-maximising firms will therefore resent from paying market-clearing wages (Thaler, 1989, p. 186).

Akerlof’s intention was to design a model which was more readily applicable to real world labour markets than is the neoclassical model. He noticed that norm-gift-exchange models had been used successfully by sociologists to explain workers’ behaviour (Akerlof, 1982, p. 544), yet they had to a great extent been ignored by economists up to that date. Akerlof’s approach was to model norms of behaviour, i.e. reciprocal gift-exchange, as a major determinant of work relationships, and therefore of wage bargaining.

Akerlof’s main criticism of the neoclassical model is the assumption that firms can hire and use (or abuse) labour as a factor of production just as they hire and use capital (Akerlof, 1982, p. 545). This, however, does not seem to be an appropriate model of the relationship between firms and workers. Apart from legal restrictions which rule out the possibility of the firm to abuse the work force as it desires, firms also depend on the willingness of the work force to cooperate (Akerlof, 1982, p. 545), which implies the necessity to consider this social interaction in some way in order to explain labour markets and which is the reason why the labour market cannot be modelled in the same way any goods market can be.

The behaviour of a utilitity company in New England and its employees serve Akerlof as an example which proves the inability of the neoclassical model to explain the labour market. This company was observed by Homan in his 1954 study “The Cash Posters”. The employees considered in his studies were female cash posters whose job was to record customers’ payments on ledger cards at the time of receipt (Akerlof, 1982, p. 546). The company had set a standard of 300 cash postings per hour which had to be achieved by each of the women. Not complying to this standard resulted in a “mild rebuke” (Akerlof, 1982, p. 546), and therefore did not have severe consequences.

In this situation, the neoclassical theory of contracts predicts that either the cash poster will reduce their speed to the standard (as long as they have marginal disutility for work at that point) or, if they do have positive utility for work at the standard level, that the firm will increase the standard up to that point where the workers’ marginal disutility equals zero (Akerlof, 1982, p. 547). This, however, was not observed by Homan. Instead, the average number of postings exceeded the standard by 17.7 percent (Akerlof, 1982, p. 546). The deviation of the standard varied significantly among the workers, ranging from 6 to 139 excess cash postings per hour. Yet, wages and standards were identical for all cash posters, and even the possible explanation of this increased effort by an implicit contract stating the possibility of promotion or increased pay was belied. Promotion hardly ever occured, and when it did occur, it was not followed by an increase in pay. Therefore, implicit contracts do not seem to explain the observations of increased effort.

The observations made can, according to Akerlof, only allow for two possible explanations: either the firm maximizes something other than profits, or workers have a different utility function to the neoclassical one, influenced by the interactions among themselves and with the firm (Akerlof, 1982, p. 548). This induces the recognition of social norms in labour contracts. They lead Akerlof to explain the observations by means of reciprocal gift exchange: the workers give a gift to the firm in the form of excess effort and the firm therefore gives the gift of reduced pressure in return (Akerlof, 1982, p. 549). Instead of reduced pressure, it is also possible to imagine a wage paid above the market-clearing level to be the gift given by the firm.

Whereas this gift-giving between persons and an institution seems surprising at first sight, an easy sociological explanation for it can be found. According to Akerlof (1982, p. 550), people who work for a firm tend to anthropomorphize this institution, i.e. the firm has human features to its workers. Therefore, just as people derive positive utility from giving gifts to other people they like, they do so when giving a gift to an institution they have sentiments for. They might also have sentiment for their coworkers and derive positive utility from reduced pressure on them as well. This explains why the same standard was set for all workers in the Homan study: increasing the minimum standard to a level which would put pressure on the slowest workers might be considered by all workers as a failure by the firm to reciprocate to the gift of the entire group, which amounted to 17.7 percent excess effort (Akerlof, 1982, p. 551). This in turn could lead to reduced gift giving by all workers which could severly harm the firm. It can therefore be rational for the profit-maximizing firm to set the work standard below the level that could be implemented without the workers actually leaving the firm, as workers reward this behaviour with higher effort.

The model set up by Akerlof includes these preferences and can therefore simultaneously explain the workers’ and the firm’s behaviour observed by Homan. In his model, the wage is determined by norms of effort, the workers’ utility function and the firm’s profit maximization. The firm’s profits depend on the workers’ effort, as does the workers’ utility. The workers’ utility depends further on the wage paid. The norms for effort in turn depend on the work rules of the firm, the average wage paid by the firm and the utility of the co-workers in the firm. Because of this interdependency, an interior solution with a higher wage than the market-clearing one and involuntary unemployment occurs (Akerlof, 1982, p. 557-560). Therefore, according to Akerlof’s model of gift-exchange, wage rigidity and involuntary unemployment occur due to norms of reciprocity in the labour market.



ISBN (eBook)
File size
564 KB
Catalog Number
Institution / College
University of Göttingen
Wage Rigidity Social Norms Experimental Labour Markets



Title: Wage Rigidity and Social Norms in Experimental Labour Markets