Nonparametric Identification of Production Function, Total Factor Productivity, and Markup from Revenue Data
Professor Hiro Kasahara
Professor, Vancouver School of Economics
The University of British Columbia
Commonly used methods of production function and markup estimation assume that a
firm’s output quantity can be observed as data, but typical datasets contain only revenue,
not output quantity. We examine the nonparametric identification of production function
and markup from revenue data when a firm faces a general nonparametric demand function
under imperfect competition. Under standard assumptions, we provide the constructive
nonparametric identification of various firm-level objects: gross production function, total
factor productivity, price markups over marginal costs, output prices, output quantities,
a demand system, and a representative consumer’s utility function. We develop a simple
semiparametric estimator implementable with standard datasets. In simulation, our method
performs as well as standard methods using quantity data.
