Po Hyun Sung

We study stochastic choice across decision problems, each represented as a menu of action labels paired with observable outcome vectors. We propose a consistency condition for behavior in decision problems composed of two separable components: choice probabilities must agree with those obtained when each component is considered in isolation. Together with monotonicity and continuity, this separability requirement characterizes the family of random coefficients logit rules.

We study solution concepts for normal-form games. We obtain a characterization of Nash equilibria and logit quantal response equilibria, as well as generalizations capturing non-expected utility. Our axioms reflect that players are responsive to payoffs induced by the play of others and, whenever several games are played simultaneously, players may consider each separately.

This manuscript identifies a largely unrecognized connection between, and distinction in results for,  Ellsberg's three-color paradox and Allais' common-consequence paradox. Both phenomena derive from reducing a common state payment to zero, but the reactions to this reduction are directionally opposed across the two problems. Using a common experimental framework, we develop intermediate tasks to identify the source of this difference. Key to the difference is the reversal of Allais' common consequence paradox when problem parameters are made more similar to those for Ellsberg.

We theoretically examine connections between choices and valuations when preferences are stable, but measurements are noisy and individuals are heterogeneous. Even under strong assumptions governing noise and heterogeneity, stability does not generally imply identical measurements, challenging the inferences of the ``preference reversal’’ literature. We derive the attainable set of joint measurements under strengthening assumptions for noise and heterogeneity. We propose corresponding statistical tests for stability and techniques quantifying instability. Re-examining the preference reversal literature  and datasets connecting choices and valuations, we find that the data are, in fact, largely consistent with stable preferences and deviations from stability are not systematic. 

We study the FOSD independent random utility (FIRU) model, in which random utilities are independent and ranked by first-order dominance. This model generalizes the widely used i.i.d. noise model, which includes logit and probit. We provide necessary conditions on binary choice data for it to be rationalized by the FIRU model. When there are three alternatives, these restrictions are sufficient to characterize the model. In this case, the set of binary choice data rationalized by the FIRU model is convex. Thus, the restrictions of the FIRU model extend to the i.i.d. mixture noise model, which includes mixed logit. 

Draft coming soon!

Draft coming soon!