Inter-firm Heterogeneity in Production

TL;DR

This paper investigates inter-firm heterogeneity in production by estimating diverse Cobb-Douglas functions across firms using an innovative Empirical Bayes approach, revealing significant heterogeneity and correlations impacting productivity analysis.

Contribution

It introduces a novel Empirical Bayes methodology to estimate heterogeneous production functions, emphasizing differences in output elasticities and productivity across firms.

Findings

Substantial heterogeneity in productivity and elasticities

Negative correlation between productivity and elasticities

Robust results across multiple datasets and specifications

Abstract

This paper studies inter-firm heterogeneity in production. Unlike much of the existing research, which primarily addresses heterogeneous production through unobserved fixed effects, our approach also focuses on differences in factors' output elasticities. Using manufacturing data from Chile, Colombia, and Japan, we apply an innovative Empirical Bayes methodology to estimate heterogeneous Cobb-Douglas production functions. We uncover substantial heterogeneity in both factor neutral productivity and factor elasticities, with a strong negative correlation between them. These findings are consistently observed across datasets and remain robust when using CES and intensive Cobb-Douglas specifications. We show that accounting for these features has significant implications for issues such as markup estimation, firms' technology adoption, and productivity measurement.