Inequality and Procedural Justice in Social Dilemmas

Abstract

This study investigates the influence of resource inequality and the fairness of the allocation procedure of unequal resources on cooperative behavior in social dilemmas. We propose a simple formal behavioral model that incorporates conflicting selfish and social motivations. This model allows us to predict how inequality influences cooperative behavior. Allocation of resources is manipulated by three treatments that vary in terms of procedural justice: allocating resources randomly, based on merit, and based on ascription. As predicted, procedural justice influences cooperation significantly. Moreover, gender is found to be an important factor interacting with the association between procedural justice and cooperative behavior.

Keywords:

• game theory
• procedural justice
• social dilemmas

ACKNOWLEDGMENTS

We thank Vincent Buskens, Manuela Vieth, and Werner Raub for their comments on earlier drafts and their help in conducting the experiment. We also acknowledge helpful comments from anonymous JMS reviewers.

Notes

Crawford and Ostrom (1995) further differentiate between internal and external sources of utility that result from obeying/breaking the norm of cooperation. In our design, there is no external source which imposes the norm of cooperation, thus this differentiation is not necessary. We thank an anonymous reviewer who reminded us of this equivalence.

One can extend the list of possible alternative social utility models even further by including many other utility models proposed in the literature. An overview of these social utility models can be found in Fehr and Schmidt (2006).

Notes: “Equality condition” is a dummy for the symmetric game. “Period” is the period in the experiment when the decision is made. The last two columns include differences in coefficients between men and women and their standard errors (SE).

∗p < 0.1;∗∗p < 0.05;∗∗∗p < 0.01 for two sided tests.

Fehr and Schmidt's (1999) utility function for an actor with an outcome allocation (x, y) for the self (x) and the other (y) is defined as U = x − βmax (0, x − y) − αmax (0, y − x), where β and α represent sensitivities to advantaged and disadvantaged inequality, respectively. Fehr and Schmidt further assume that β ∊ [0, 1] and α > β. With the assumption of α > β, defection is a dominant strategy for at least one actor in the asymmetric games that we use, and the single Nash equilibrium is mutual defection unless the share parameters of actors are equal, i.e., λ₁ = λ₂ = 0.5. Cooperation can be an equilibrium only if the share parameters of actors are equal, and even in this case, mutual defection still remains an equilibrium.

One reviewer offered the fear-greed hypothesis (Simpson, 2003; Kuwabara, 2005) as a possible explanation. This hypothesis claims that men defect in the PD in response to greed, that is, to earn more; whereas women defect in response to fear, that is, to avoid being exploited by cheaters. It is possible that under the ascription condition the advantaged women fear that their partners will be defective because of the illegitimacy, and due to this fear do not cooperate themselves. Although this explanation seems plausible, Theorem 1 shows that the probability of cooperation does not depend on the expectation about the behavior of the opponent in the asymmetric investment game. Moreover, in our case, the greed component, (T-R), is always the same as the fear component, (P-S).