a Promised Land for economic models of fairness

Get Complete Project Material File(s) Now! »

Fairness in the economic field

Markets show the characteristics that are the least favorable to the emergence of fair behaviors. First, norms associated to markets are traditionally those of profit making therefore inducing individuals to focus on their material self-interest. The experiment conducted by Hoﬀman et al. (1994) is very illustrative of this point. They show that adding a market frame in a ultimatum game so that participants are called buyers and sellers significantly reduces oﬀers. More-over, markets typically involve large stakes. Finally, individuals contracting on a market are rarely informed about the valuations of buyers and about the costs or reservation prices of sellers. In other words, computing the impact of their actions on others’ surplus seems out of individuals’ reach. Yet, there exist configurations of market transactions which could oﬀer some room for individuals to express their concerns for fairness. More specifically, market transactions can be split up into two groups: transactions that involve complete contracts17 and transactions that involve incomplete contracts18. In the following, we investigate whether concerns for fairness could modify the outcome on markets involving complete and incomplete contracts respectively.

Selling and buying under complete contracts

Transactions that involve complete contracts are those whose characteristics are observable by both contracting parties. We consider two types of markets in-volving complete contracts. The first type encompasses competitive markets where the lack of information on others’ surplus is likely to be an insurmount-able obstacle to the emergence of concerns for fairness. The second type includes monopolistic markets where even a vague information on the monopoly’s unde-servingness is likely to induce desert-sensitive altruistic behaviors on the part of individuals belonging to the dominated side of the market.

Competitive markets

Smith (1965) was the first to compare the competitive final price of a double-auction19 repeated market game with the competitive price predicted by the homo oeconomicus postulate. Smith (1965) finds out an amazing convergence of the experimental exchange price to the competitive price. Many subsequent market game experiments confirmed this finding. In other words, certainly due to individuals’ lack of information on others’ surplus, concerns for fairness do not aﬀect the outcome on markets involving complete contracts in a laboratory setting20. Given that the outcome of market games is not impacted by concerns for fairness while we have shown that experiments conducted in a laboratory setting have generally a fairness inducing eﬀect, we feel confident in concluding that, a fortiori, concerns for fairness should not influence the outcome on real life competitive markets.

Monopolistic markets

We rely on evidence provided by market games conducted in a laboratory setting where the bargaining power is concentrated on one side of the market. We derive the relevant implications of such evidence for real life monopolistic markets. Market power concentrated on the selling side Roth et al. (1991) imple-ment a competitive ultimatum game where all the nine dictators simultaneously propose an amount si of an initial endowment normalized to 1 to a receiver that either accepts or rejects the highest oﬀer s = max si. In case several dictators propose s, one of them is randomly selected with equal probability.
This experimental setting can be interpreted as a market game with buyers’ competition where the seller enjoys a monopolistic power. More precisely, the nine dictators could stand for nine buyers willing to buy one unit of commodity that each values at a price of 1, while the responder could be viewed as a seller whose reservation price is equal to 0. s stands for the highest bid to buy proposed by the buyers to the seller. In case the seller accepts s, the buyer’s surplus is equal to 1 −s, while the seller’s surplus is equal to s. In case the seller rejects this bid to buy, both the seller and the buyer obtain a surplus equal to 0. The computation of the unique subgame perfect equilibrium in case of purely self-interested buyers is straightforward. The competition among dictators in-duces each of them to increase his bid to buy si up to 1. In other words, one side of the market (the seller) reaps all the buyers’ surplus. This prediction is confirmed by Roth et al. (1991) who show that, in each of the countries where the experiment was conducted (Israel, Japan, Slovenia and United States), the market game converges to the competitive price equilibrium predicted by the Market power concentrated on the buying side In the reverse ultimatum game where there is competition among the receivers and were the proposer enjoys a monopolistic power, G¨uth et al. (1997) also show that the experimental equilibrium price converges to the competitive equilibrium price of 0 predicted by the postulate of homo oeconomicus. More precisely, in the final round of their repeated market game (encompassing five rounds overall), 71% and 9% of the receivers have an acceptance threshold of 0 and 0.02 respectively. A summarizing experiment Cason and Williams (1990) analyse the results of market games in which the competitive equilibrium model predicts that all exchanges profits (surpluses) are received by one side of the market only. More precisely, they assume that the reservation price PS is the same for all suppliers and that all buyers share the same valuation PD . In a first treatment, the quantities demanded D are lower than the quantities supplied S which should lead, under the homo oeconomicus postulate, to a price equilibrium equal to PS . In a second treatment, the quantities demanded D are greater than the quantities supplied S which should lead to a price equilibrium equal to PD .
The market game is organized according to a ‘posted-oﬀer’ procedure. In con-trast to the double auction procedure, the posted-oﬀer procedure does not allow exchange of both bids to buy and oﬀers to sell. It is particularly characteristic of retail markets where sellers post a price that the buyers may accept or not on a simple ‘take-it-or-leave-it’ basis. In the experimental design, each period begins with all sellers making their oﬀers to the market (each seller is shown the oﬀer prices entered by other sellers). The computer then chooses randomly the order that buyers enter the marketplace.
Cason and Williams (1990) show that in either treatment, the price converges to the competitive price, even though the convergence is slower in the first treatment than in the second one which is easily understandable. In the sec-ond treatment, sellers quickly learn that their objective is to avoid being the low-priced seller. The upward price trend triggered oﬀ by this learning eﬀect is not easily hampered by desert-sensitive altruistic buyers who know that the undeserving oﬀer will certainly be accepted by another purely self-interested buyer if they refuse it. Conversely, in the first treatment, sellers quickly learn that their objective is to avoid being the high-priced seller. This learning ef-fet induces a downward price trend that some desert-sensitive altruistic sellers try to moderate through costly signaling consisting in oﬀering purposely high prices that do not meet any demand. However, due to the presence of purely self-interested sellers, this attempt to make the price diverge from the compet-itive equilibrium is not sustainable and the price ultimately converges to PS . Conducting the same experiment, but in a double auction framework, Smith and Williams (1990) also find a convergence of the exchange price to the com-petitive price which is symmetric between both treatments due to the double auction procedure.
Implications for real-life monopolistic markets Experimental results ob-tained in a laboratory setting therefore suggest that desert-sensitive altruistic behaviors on monopolistic markets with complete contracts are enforceable only under the unrealistic assumption that a suﬃciently large number of individuals agree to bear the costs of punishment. Anticipating that punition won’t be successful, fairness-concerned agents themselves tend to renounce of showing desert-sensitivity.

READ Long range energy transfer in self assembled nanoplatelets

Selling and buying under incomplete contracts

Transactions that involve incomplete contracts are those whose some character-istics are not observable by all contracting parties, like the employer-employee relationship where the eﬀort of the employee is not fully observable by the em-ployer. Within incomplete contracts, individuals having a discretionary power on one of the contract characteristics can fully express desert-sensitive altruism. Concerns for fairness are therefore likely to impact the aggregate outcome on markets involving incomplete contracts. We start reviewing experimental evi-dence in a laboratory setting and then extend our analysis to naturally occurring field experiments.

Experiments in a laboratory setting

Fehr et al. (1993) were the first to propose an experimental test of the ‘fair’ wage-eﬀort hypothesis developed by Akerlof (1982) and by Akerlof and Yellen (1990). The ‘fair’ wage-eﬀort hypothesis goes a step further compared to the ‘eﬃciency’ wage hypothesis theorized by Shapiro and Stiglitz (1984). While the latter stipulates that high wage levels reduce workers’ temptation to shirk by raising the cost of being fired, the former assumes that wage increases do not only induce workers to better comply with the eﬀort standard imposed by the firm, but possibly to increase their eﬀort level above this standard. Akerlof (1982) bases the ‘fair’ wage-eﬀort hypothesis on the classic anthropological literature on the gift and notably on the essay of Marcel Mauss (1954) who emphasizes the obligatory nature of reciprocity in archaic societies as soon as one receives a gift from someone else21.
Description of the experimental setting The experimental gift exchange game initiated by Fehr et al. (1993) consists in two stages. The first stage is a one-sided oral auction where employers make wage proposals. As soon as one worker accepts the bid, a binding contract is concluded between the employer and the worker. At the second stage, workers choose their eﬀort level (which is revealed only to the employer with whom they trade).
More precisely, the usual design consists for the worker in choosing eﬀort e from the interval [e, e], 0 < e < e. With possibly some variants from an experimental setting to the other, the firm’s payoﬀ is summarized by xF = ve − w where v denotes the marginal product of eﬀort. The worker’s payoﬀ is given by xW = w − c(e) where c(e) denotes the eﬀort cost (c(e) = c′(e) = 0 and c′ > 0, c′′ > 0 for e > e). Moreover, it is generally assumed that v > c′(e) so that e = e is the eﬀort level consistent with utilitarian altruism since it maximizes the total surplus. An excess supply of workers is deliberately created by experimenters to ensure that workers are willing to accept virtually any oﬀer greater or equal to the market-clearing one.

Table of contents :

1. Summary and avenues for future research
1. General motivation
2. Research questions… and preliminary answers
2.1. Paper 1
2.2. Paper 2
2.3. Paper 3
3. Avenues for future research
References
2. Politics: a Promised Land for economic models of fairness
1. Motivation
2. Individuals’ concerns for fairness
2.1. Utilitarian altruism
2.2. `Rawlsian’ altruism
2.3. From responsibility-sensitivity to desert-sensitivity
2.3.1. The responsibility cut
2.3.2. From responsibility-sensitivity..
2.3.3. …to desert-sensitivity
2.4. A summarizing `fair utility function’
3. Fairness in the economic ¯eld
3.1. Selling and buying under complete contracts
3.1.1. Competitive markets
3.1.2. Monopolistic markets
3.2. Selling and buying under incomplete contracts
3.2.1. Experiments in a laboratory setting
3.2.2. Naturally occurring ¯eld experiments
4. Fairness in the political ¯eld
4.1. Preliminary insights
4.1.1. Norms
4.1.2. Stakes
4.1.3. Information
4.2. Retrospective voting
4.2.1. Studies based on aggregate data
4.2.2. Studies based on individual data
4.2.3. Evidence of `truly’ fair sociotropic voting
4.3. Spatial voting
4.3.1. The impact of political ideology..
4.3.2. … and of ethnic prejudice
5. Concluding remarks
References
3. Voting for redistribution under desert-sensitive altruism
1. Motivation
2. The model
2.1. Individual characteristics
2.2. Private preferences for consumption and leisure
2.3. Altruistic preferences for redistribution
2.4. Di®erent scenarios of altruism
3. Political equilibrium
4. Desert-sensitive altruism in practice
5. Concluding remarks
Appendix A: impact of incomplete information
Appendix B: descriptive summary
References
4. Containing ethnic con°icts through ethical voting? Evidence from Ethiopia
1. Motivation
2. The `ethnicization’ of Ethiopian politics
2.1. The four main ethnic groups in Ethiopia
2.2. The adoption of ethnic federalism in the 1990s
2.3. The division of political parties along ethnic lines
3. Data, econometric method and descriptive statistics
3.1. Survey
3.2. Econometric method
3.3. Descriptive statistics
4. Results
4.1. The characterization of political parties by university students
4.2. Voting results
4.3. Logit regression results
5. Concluding remarks
Appendix
References