DOES MUTUAL KNOWLEDGE OF PREFERENCES LEAD TO MORE
NASH EQUILIBRIUM PLAY? EXPERIMENTAL EVIDENCE
EUROPEAN ECONOMIC REVIEW (ACCEPTED FOR PUBLICATION)
(JOINT WITH CHRISTOPH BRUNNER AND FLORIAN KAUFFELDT)
Nash equilibrium often does not seem to accurately predict behavior. In experimental game theory, it is usually assumed that the monetary payoffs in the game represent subjects' utilities. However, subjects may actually play a very different game. In this case, mutual knowledge of preferences may not be satisfied. In our experiment, we first elicit subjects' preferences over the monetary payos for all players. This allows us to identify equilibria in the games that subjects actually are playing. We then examine whether revealing other subjects' preferences leads to more equilibrium play and find that this information indeed has a significant effect. Furthermore, it turns out that subjects are more likely to play maxmin and maxmax strategies than Nash equilibrium strategies. This indicates that subjects strongly rely on heuristics when selecting a strategy.
INEQUALITY, FAIRNESS AND SOCIAL CAPITAL
EUROPEAN ECONOMIC REVIEW, 2020, VOL. 129, ARTICLE 103566
(JOINT WITH DIETMAR FEHR, STEFAN TRAUTMANN AND YILONG XU)
We study the impact of unjust inequality on social trust and trustworthiness, and how it interacts with economic status in a large-scale controlled experiment. We document that unfair economic inequality is detrimental for social interactions, resulting in a significant decline in trust and trustworthiness. Probing the boundaries of this effect, we demonstrate that this erosion of social capital critically depends on the context: if an economically successful person is not directly responsible for the outcome of the unsuccessful person, we observe no negative effects on trust and trustworthiness in the aggregate. Finally, our data do not support the view that higher status or wealth leads to an erosion of pro-social attitudes: the successful are always more generous, whereas unsuccessful persons display the least efficient and generous behavior.
HEDGING, AMBIGUITY AND THE REVERSAL OF ORDER AXIOM
GAMES AND ECONOMIC BEHAVIOR, 2019, VOL. 117, PP. 380-387
(JOINT WITH JÖRG OECHSSLER AND ALEX ROOMETS)
We ran experiments that gave subjects a straight-forward and simple opportunity to hedge away ambiguity in an Ellsberg-style experiment. Subjects had to make bets on the combined outcomes of a fair coin and a draw from an ambiguous urn. By modifying the timing of the draw, coin flip, and decision, we are able to test the reversal-of-order axiom. Our main result is that the reversal-of-order axiom seems to hold. We also confirm low levels of ambiguity hedging despite the relative obviousness of the opportunity.
PREFERENCED-BASED GAMES AND NASH EQUILIBRIUM PLAY
RESEARCH PROPOSAL, 2020 (UNDER REVIEW AT DFG)
The Nash equilibrium is a widely used concept in applied game theory and experimental economics. However, in several situations it fails to accurately predict behavior. Sometimes players even violate the criterion of strict dominance. As this happens in very simple games as well, it seems unlikely that a lack of rationality is the driving force behind behavior. In many applications, the Nash prediction is based solely on players’ own (material) payoffs. These are assumed to represent the utility of the decision makers. However, it is a well-established fact that for many players their utility does not only depend on their own payoff. Instead, they frequently take additional factors into account, such as the payoffs of others or the type of interaction (“social preferences”). As a result, the actual strategic situation may be different from the one implied by the material payoffs of the game. People even might play a completely different kind of game compared to what the analysts think they do.
In this project, we intend to explore the latter aspect by conducting a comprehensive analysis how the nature of a strategic situation changes when incorporating players’ (social) preferences into the decision problem. We plan to first examine simple one-shot 2x2 games and subsequently extend the analysis to more complex categories of games. We plan to use experimental methods to elicit people’s evaluations of the outcomes of games, which in our context usually correspond to payoff tuples. We first start with an ordinal preference ranking and later on extend it to a cardinal measure. These evaluations can then be used to infer the strategic properties of the game the agents are playing according to their preferences (the “preference-based game”). Empirical frequencies of different types of preferences yield information how often specific pairings of types occur. From that may determine the likelihood by which games transform into other classes of games when accounting for the social preferences of the involved players. Because the Nash-prediction that is based solely on players’ own payoffs is biased towards selfish behavior, through this approach more cooperative outcomes can be justified as an equilibrium.
Furthermore, we are able to test if the equilibrium prediction can substantially be improved when taking players’ preferences into account. This can be done by comparing empirical frequencies of Nash equilibrium play, once based on the predictions from the original payoff structure with those from the corresponding preference-based games. If the latter frequencies are significantly higher, then it is not the concept of the Nash equilibrium itself to be blamed for the imprecise prediction of behavior. Rather the material payoff structure of the game (the “game-protocol”) does not represent the strategic situation adequately.
DO PEOPLE EXHIBIT MORE ANTISOCIAL BEHAVIOR IF THE INCOME ALLOCATING PROCESS HAS BEEN UNFAIR?
We examine whether an unfair process of income allocation leads to a higher degree of antisocial behavior. In order to test this hypothesis, we run an experiment where we vary the way players’ endowments are determined: by a fair, random, or unfair process. The initial distribution has a certain degree of inequality, which is held constant across treatments. After receiving their income, subjects can anonymously reduce the income of another player at a cost. The overall frequency and percentage of destruction is similar and not significantly different across treatments. Surprisingly, even if money is allocated in an unfair manner, subjects do not destroy more. We furthermore elicit subjects’ perceptions about the fairness of the income-generating process. They are in line with the intended treatment effect, but we find almost no correlation between subjects’ fairness evaluation and the propensity to burn money. The findings indicate that the degree of antisocial behavior is rather constant in this context and independent of the fairness of the income-allocating process. Subjects’ justifications of their decision and insights of related studies suggest that the decision to destroy other’s income depends a lot on whether other subjects can be held responsible for the initial (unfair) distribution.
GENDER DIFFERENCES IN BARGAINING EXPERIMENTAL EVIDENCE
(JOINT WITH GRACIELA KUECHLE)
We study gender differences and gender pairing effects in a laboratory experiment with alternating-offers bargaining for a fixed pie, framed as an employer-employee interaction. We vary the degree of asymmetry in bargaining power between roles, as well as the disclosure of genders of the negotiating partners. With low asymmetry, we find no gender effects in behavior or outcomes. When there is high asymmetry and gender is known, men achieve more favorable deals than women in both roles, especially in mixed gender pairings. These differences are significant, despite gender information being transmitted only in a very subtle way. However, differences disappear, when no gender information about the bargaining partner is provided. Additionally, we examine the bargaining strategies of the players in detail. In presence of high asymmetry, men behave more aggressively in mixed gender parings, while the opposite effect is true for women.
FAIRNESS PROPERTIES OF COMPENSATION SCHEMES
(JOINT WITH DIETMAR FEHR, STEFAN TRAUTMANN AND YILONG XU)
We study the effect of different characteristics of merit-based compensation schemes on their perceived fairness in a large-scale online survey. Respondents rate these compensation schemes for a real effort task as an impartial observer. Three main categories of pay-for-performance schemes are considered here: continuous piece rate incentives, step level incentives that pay a bonus when surpassing a threshold performance level, and tournaments with two workers pitted against each other. Within each class, we vary the degree of earnings inequality (steepness of the merit-related pay) and consider time bonuses and handicaps as a function of earlier performance in the step-level and tournament incentive scheme. The specific properties strongly affect the fairness perceptions of the payment schemes. We find that the induced level of inequality has a clear negative effect on perceived fairness. Given the same degree of inequality, people consider piece rate schemes fair than those with a step-level or tournament design. Furthermore, adding elements of time bonus or handicap generally makes a compensation scheme perceived as less fair. Bonusses are especially aversive under tournament schemes, while handicaps are perceived as unfair in general. Given the handicaps’ positive effect on equality, this demonstrates the importance of merit for fairness judgments.