Measuring Markups with Revenue Data

Abstract

Standard production-based markup estimators are biased and inconsistent when output prices are unobserved. Absent additional assumptions, researchers cannot disentangle whether firms have higher revenues because of higher productivity or higher markups. We propose a method that generates unbiased and consistent estimates by modifying physical productivity process assumptions into revenue productivity process assumptions, and by flexibly modeling markups as a specified function of observables and fixed effects. Our method builds on past work designed to estimate production functions in competitive environments. Our suggested two-step estimator is simple in concept and implementation, requiring only common regression techniques.