Cognitive Resource Depletion, Choice Consistency, and Risk Preferences

We investigate the consistency and stability of individual risk preferences by slightly manipulating the cognitive resources of subjects through sleepiness. Participants are recruited and randomly assigned to an experiment session at a preferred time of day relative to their diurnal preference (circadian matched) or at a non-­‐preferred time of day (circadian mismatched). For the decision task, subjects and are asked to choose how much to allocate between two state-­‐dependent assets (using the Choi et al., 2007, design). We have two main findings. First, the consistency of behavior for circadian matched and mismatched subjects is statistically the same. This is true whether it is (nonparametrically) defined as consistency with GARP, payoff dominance, expected utility, disappointment aversion or cumulative prospect theory. Second, while our cognitive resource manipulation yields no difference in consistency of behavior, it results in an increased tendency to take risk. Our experiment confirms theoretical predictions that preferences are consistent yet state-­‐dependent.