Customer Capital, Markup Cyclicality, and Amplification

Abstract

This paper studies the importance of firm-level price-markup dynamics for business cycle fluctuations. Using state-of-the-art IO techniques to measure the behavior of markups over the business cycle at the firm level, I find that markups are countercyclical with an average elasticity of -1.1 with respect to real GDP. Importantly, I find substantial heterogeneity in markup cyclicality across firms, with small firms having significantly more counter-cyclical markups than large firms. Then, I develop a general equilibrium model by embedding customer capital (due to deep habits as in Ravn, Schmitt-Grohé, and Uribe, 2006) into a standard Hopenhayn (1992) model of firm dynamics with entry and exit. The calibrated model replicates these empirical facts and produces counter-cyclical firm sales dispersions consistent with the data. The resulting input misallocation amplifies both the volatility and persistence of the aggregate productivity shocks driving the business cycle. (JEL Codes: E31, E32, L11, L13.)

Markups and Financial Shocks

“…In particular, the authors find that markups of smaller firms tend to be counter-cyclical, while those of large firms tend to be pro-cyclical, such that the cyclicality of the aggregate markup depends, among other factors, on the firm size distribution in the sector or economy. Empirical findings by Hong (2019) confirm that the cyclicality of the aggregate markup varies by firm size. He rationalizes this finding in a model featuring customer markets, where small firms have a higher exit probability, implying that they become more myopic during recessions.…”

“…Second, our results are relevant with respect to the debate on the role of firm heterogeneity for the cyclicality of markups, as recently raised by Burstein, Grassi, and Carvalho (2020) and Hong (2019). This strand of the literature tries to explore sources of firm heterogeneity -in particular, in relation to firm size -to shed light on the long-standing debate on the cyclical properties of the aggregate markup.…”