Electronic Thesis and Dissertation Repository

Thesis Format

Integrated Article

Degree

Doctor of Philosophy

Program

Economics

Supervisor

Aguiar, Victor H.

Abstract

This dissertation consists of three chapters studying revealed preference theory and its applications to testing and inference. The first chapter develops a semiparametric revealed preference methodology to analyze the effects of price search on prices paid across income groups. The second chapter derives a novel representation of the exponential discounting model to make inference on the discount factor. The third chapter derives an axiomatization of the exponential discounting model and uses it to propose a nonparametric test of dynamic discrete random utility models.

The first chapter develops a novel semiparametric approach to estimate the impacts of price search on prices paid. My methodology allows for nonparametric preferences, rich heterogeneity, and measurement error in prices. I derive the implications of the model and use the resulting shape constraints to gain additional identification power. Using data on shopping expenditures from the Nielsen Homescan Dataset, I show that the data are consistent with the model and that a doubling of shopping frequency decreases prices paid by about 19%. Moreover, I find that heterogeneity in price search generates within-group consumption inequality and mixed impacts on between-group consumption inequality.

The second chapter derives a novel representation of the exponential discounting model that allows me set identify the discount factor while accounting for measurement error. My approach uses a revealed preference methodology that makes no parametric assumption on the utility function and allows for unrestricted heterogeneity. Using longitudinal data from checkout scanners, I bound household-specific discount factors and assess their sensitivity to measurement error. I find that unobserved heterogeneity is important as observable characteristics fail to capture differences in discounting.

The third chapter generalizes the previous representation and provides an axiomatic characterization of the exponential discounting model. This axiom is used to propose a nonparametric test for dynamic discrete random utility models. The methodology can be used to understand the welfare implications of price changes in demand data when there is no uncertainty for consumers that have preferences over discrete product characteristics or quantities of products.

Summary for Lay Audience

Revealed preference analysis is an approach that makes minimal assumptions on the consumer or firm behavior. It can be used to test various theories and to study the behavioral response to various policies. This thesis consists of three chapters studying revealed preference theory and its applications to testing and inference.

The first chapter develops a semiparametric revealed preference methodology to analyze the effects of price search on prices paid across income groups, where price search describes the process whereby buyers actively seek to gauge the most favorable prices. Using data on shopping expenditures, I show that the data are consistent with the model and that a doubling of shopping frequency decreases prices paid by about 19%. Moreover, I find that heterogeneity in price search generates within-group consumption inequality and leads to mixed impacts on between-group consumption inequality.

The second chapter derives a novel revealed preference representation of the exponential discounting model to make inference on the discount factor. This endeavor is motivated by the fact that assumptions on the consumer preferences may lead to erroneous conclusions about his time preferences and misrepresent a decision maker's intertemporal choices in a vast range of applications. Using longitudinal data from checkout scanners, I bound household-specific discount factors and assess their sensitivity to measurement error.

The third chapter derives an axiomatization of the exponential discounting model and uses it to propose a nonparametric test of dynamic discrete random utility models. This methodology can be used to understand the welfare implications of a change in prices and to predict the change in the distribution of demand from a change in prices.

Download

Included in

Economics Commons

Share

COinS