Journal Article

Nonparametric estimation of a nonseparable demand function under the Slutsky inequality restriction

Date: 02 May 2017
Authors:
Publisher: MIT Press Journals
Published in: Review of Economics and Statistics
JEL classification: C14, C21, D12
DOI: 10.1162/REST_a_00636

We derive conditions under which a demand function with nonseparable unobserved heterogeneity in tastes can be estimated consistently by nonparametric quantile regression subject to the shape restriction from the Slutsky inequality. We consider nonparametric estimation of the nonseparable demand for gasoline in the U.S. The estimated function detects differences in behavior between heavy and moderate gasoline users, and reveals systematic variation in the responsiveness of demand to plausible changes in prices across the income distribution. We test for exogeneity of prices and develop a new method for estimating quantile instrumental variables to allow for endogeneity of prices. The empirical results illustrate the improvements in finite-sample performance of a nonparametric estimator from imposing shape restrictions based on economic theory.

This article is forthcoming.