Is it feasible to estimate firm-level markups with commonly available datasets? Common methods to measure markups hinge on a production function estimation, but most datasets do not contain data on the quantity that firms produce. We use a tractable analytical framework, simulation from a quantitative model, and firm-level administrative production and pricing data to study the biases in markup estimates that may arise as a result. While the level of markup estimates from revenue data is biased, these estimates do correlate highly with true markups. They also display similar correlations with variables such as profitability and market share in our data. Finally, we show that imposing a Cobb-Douglas production function or simplifying the production function estimation may reduce the informativeness of markup estimates.
Basile Grassi, Giovanni Morzenti and Maarten de Ridder
20 December 2022 Paper Number POIDWP063
Download PDF - The Hitchhiker's guide to markup estimation
<! - KEYWORDS TAKEN OUT FROM HERE... AND PUT TO SIDE -->
This POID Working Paper is published under the Programme on Innovation and Diffusion.