Essays on decision making in social environments.

Abstract

Individuals who respond to those that affect them may be influenced primarily by the actions or intentions of those they are responding to or to the outcomes their actions produce. Recent experimental economic studies and predictions adapted from behavioral economic theories of fairness suggest that individuals respond primarily to intentions. In Chapter 1 we present an experiment called the “accountability game”. Using the data from the experiment and an alternative framework, we demonstrate and rationalize the opposite conclusion, i.e. outcomes loom larger than actions or intentions in accountability judgments, even when they do not provide information about actions. The experimental design consists of one subject choosing between a lottery and a certain amount on behalf of another subject who responds after the outcome of the choice is revealed. We find that lottery outcomes and not actions or intentions are important in determining the nature of the responses. We investigate a form of fairness termed the Control Principle which asserts that we should hold others responsible only for the events which they can control. In our environment first movers cannot control the outcome of the lottery, yet second movers will assign a higher payment to first movers when the lottery outcome is favorable. We find that this “unfair” sensitivity to outcomes is cut in half when second mover responses are elicited contingently before the first mover acts. The pattern of responses in each experimental manipulation can be rationalized by a framework that first considers what is termed as a salient pertur- bation of a decision environment (Myerson 1991) and then applies principal-agent theory to the perturbed environment. In Chapter 2 we apply the ideas developed in Chapter 1 to data generated in an experimental environment we call the “voting game”. In the voting game individuals make a decisions in the context of a group where the payoffs are perfectly correlated. In this context we find that individuals make fewer risky decisions in the context of a group when it is salient that the choice is attributable to them. The data cannot be rationalized by a popular model of social decision theory (Eliaz, Ray, and Razin 2006). The data can be explained by incorporating well known experimental results on social norms and our results on accountability judgments from Chapter 1. In Chapter 3 we investigate important gender differences in risk-taking and accountability judgments. We find that women are more risk-averse than men and tend to hold others accountable according to an absolute measure of performance. Men on the other hand tend to hold others accountable according to benchmarks less related to performance.