Estimating the Relationship between Economic Preferences: A Testing Ground for Unified Theories of Behavior

We estimate a variety of different economic preferences in the same group of subjects: risk aversion, discounting, ambiguity aversion, compound lottery aversion, the common ratio and common consequence effects, present bias, loss aversion, the endowment effect, the reflection effect, and behavior in trust games. This allows us to estimate the empirical relationships between these behaviors across subjects. We compare our results to both formal and informal theories that make predictions for these correlations. Our preliminary results show that, broadly speaking, there are five major factors that explain behavior, relating to (1) violations of EU in risky choice, (2) ambiguity aversion and failure of reduction of compound lotteries (3) loss aversion and the endowment effect (4) time preferences and (5) social preferences. Underlying this broad picture, there are, however, many additional subtleties, some of which have been predicted by previous theory. For example, ambiguity/compound lottery aversion is related to the common ratio effect, while present bias is related to the common ration effect. These relationships remain after intelligence is controlled for. We also find some more unexpected links: for example between between overplacement (a form of overconfidence), loss aversion and ambiguity aversion, and between lost aversion and behavior in the trust game.