An experimental study of asymmetric reciprocity

When deviating from best responses, do people have a stronger propensity to increase or decrease other people's payoffs? Offerman (2002) finds that negative intentions are more likely to induce payoff decreases than positive intentions are to induce payoff increases. Using the Falk and Fischbacher (2006) model, we approach the same question as Offerman, but from a structural angle. This requires measuring what a subject predicts that other subjects predict that he will do (known as a subject's second-order expectations). This permits us to interpret any asymmetry in the propensities to increase and decrease payoffs in terms of the determinants of payoff increases and decreases. Our results are largely consonant with Offerman (2002). We also find that in situations with exogenously created inequity (rather than inequity that is the result of an intentional action by an actor), payoff increase is more likely to be used to diminish the inequity than is payoff decrease. Our results exhibit some sensitivity to whether we obtain second-order expectations by eliciting them directly from the subject making the payoff increase/decrease decision or by inducing them by reporting to the decider the expectations of the target of the payoff increase/decrease.