Estimating Production Functions When Output Prices and Quality Are Unobserved

In this work, we propose a technique for consistently estimating production functions when output prices and quality are unobserved. Unobserved output prices and product quality are correlated with inputs and will lead to bias when we estimate production functions with traditional approaches. We show that the markup can serve as a control function for unobserved prices and quality and address the problems caused by them. The markup can be computed with candidate parameters and data, so our approach does not need more data than traditional approaches. We implement our method as an extension of the proxy-variable approach to estimating production functions. The empirical results give strong and consistent support to our method.