Panel and Pseudo-Panel Estimation of Cross-Sectional and Time Series Elasticities of Food Consumption: The Case of American and Polish Data

TL;DR

This paper investigates biases in estimating food consumption elasticities using pseudo-panel data, highlighting the effects of unobserved heterogeneity and measurement error through empirical comparisons with true panel data.

Contribution

It provides an empirical assessment of biases in elasticity estimates from pseudo-panels and evaluates the effectiveness of different estimators in correcting these biases.

Findings

Unobserved heterogeneity biases income elasticity estimates downward for at-home food.

Unobserved heterogeneity biases income elasticity estimates upward for away-from-home food.

Instrumental variables reduce bias in within and first-difference estimators.

Abstract

The problem addressed in this article is the bias to income and expenditure elasticities estimated on pseudo-panel data caused by measurement error and unobserved heterogeneity. We gauge empirically these biases by comparing cross-sectional, pseudo-panel and true panel data from both Polish and American expenditure surveys. Our results suggest that unobserved heterogeneity imparts a downward bias to cross-section estimates of income elasticities of at-home food expenditures and an upward bias to estimates of income elasticities of away-from-home food expenditures. "Within" and first-difference estimators suffer less bias, but only if the effects of measurement error are accounted for with instrumental variables.