John Rehbeck

Ph.D. Candidate – Department of Economics

Job Market Paper

Revealed Preference Analysis of Characteristics in Discrete Choice (with Roy Allen)

Abstract: This paper studies a descriptive model of stochastic choice that includes information about observable characteristics. The model can describe behavioral phenomena observed in datasets while retaining properties similar to the standard consumer demand problem. Necessary and sufficient conditions for the model to describe a dataset are formalized using a system of linear inequalities. We perform an empirical analysis and find that the model can describe individual stated preference data on flight choice from Louviere et al. (2013). After performing a power correction, a model with linear utility over characteristics often provides the most powerful description of individual datasets.

Published and Forthcoming Papers

Menu-Dependent Stochastic Feasibility (with Richard Brady) (Supplemental Material)

Econometrica, May 2016

Abstract: We examine the role of stochastic feasibility in consumer choice using a random conditional choice set rule (RCCSR) and uniquely characterize the model from conditions on stochastic choice data. Feasibility is modeled to permit correlation in availability of alternatives. This provides a natural way to examine substitutability/complementarity. We show that an RCCSR generalizes the random consideration set rule of Manzini and Mariotti (2014). We then relate this model to existing literature. In particular, an RCCSR is not a random utility model.

Every Choice Correspondence is Backwards-Induction Rationalizable

Games and Economic Behavior, November 2014

Abstract: I extend the result from Bossert and Sprumont (2013) that any single valued choice function with strict preferences is backwards induction-rationalizable via strict preferences to the case of choice correspondences via weak preferences.

Working Papers

Nonseparable Costly Information Acquisition and Revealed Preference (with Christopher P. Chambers and Ce Liu )

Abstract: We provide revealed preference characterizations for choices made under costly information acquisition. In particular, we examine nonseparable, multiplicative, and constrained costly information acquisition. The revealed preference characterization is performed using primitives from Caplin and Dean (2015). Each model has an analogous model in the literature on decisions under uncertainty. The techniques we use parallel the duality properties in the standard consumer problem.

Note on Unique Nash Equilibrium in Continuous Games

Abstract: This note studies the following question: If a researcher is given a finitely supported mixed strategy for each player over a continuous strategy space, does there exist a game whose unique Nash equilibrium is given by these mixed strategies? This note finds that any set of finitely supported mixed strategies on metric spaces can be represented as the unique Nash equilibrium to a separable game. If the strategy spaces are subsets of Euclidean space, then the mixed strategies can be represented as the unique Nash equilibrium to a polynomial game.

Note on Symmetric Utility (with Christopher P. Chambers)

Abstract: This note studies necessary and sufficient conditions for consumer demand data to be generated by a symmetric utility function. We find that a dataset of prices and consumption decisions can be rationalized by a symmetric utility function if and only if the symmetrized dataset satisfies the generalized axiom of revealed preference.

Complementarity in Perturbed Utility Models (with Roy Allen)

Abstract: This paper extends the Hicksian definition of complementarity to settings outside of the standard consumer problem. Using the structure of perturbed utility models, we show how to identify complementarity using variation in observable characteristics. This does not require price variation. We propose a derivative ratio as a measure of the degree of complementarity, which can be estimated with standard instrumental variables techniques. Turning to specific settings we show (i) the assumption of no income effects is not needed to identify complementarity in the analysis of Gentzkow (2007) and (ii) complementarity is ruled out by many, but not all, discrete choice models.

Identification of Average Demand Models (with Roy Allen)

Abstract: This paper studies the nonparametric identification of a model of average demand for multiple goods, once unobservable heterogeneity has been integrated out. The model can be used for bundles, decisions under uncertainty, stochastic choice, and other examples. We establish nonparametric identification of the model using exclusion restrictions. Our technique uses an analogue of Slutsky symmetry, and our key results do not rely on special regressors or identification at infinity. As a special case we provide new conditions for identification of additive random utility models (ARUM). These conditions also apply to a stochastic choice model allowing bounded rationality. In an illustrative application, we refute ARUM in favor of this more general model.

Testable Implications of Constrained Walrasian Equilibrium

Abstract: I provide a non-parametric characterization of the testable implications of locally nonsatiated preferences in a constrained Walrasian equilibrium using market data. I show that a market data set is consistent with a constrained Walrasian equilibrium with nonsatiated preferences if and only if it is consistent with a constrained Walrasian equilibrium with monotone and continuous preferences. Using the motivating example from Brown and Matzkin (1996), I demonstrate that the difference between the testable implications of constrained and unconstrained Walrasian equilibrium with locally nonsatiated preferences is captured by a concave utility function.