Quantifying the Sources of Firm Heterogeneity

Abstract

We develop and structurally estimate a model of heterogeneous multiproduct firms that can be used to decompose the firm-size distribution into the contributions of costs, "appeal" (quality or taste), markups, and product scope. Using Nielsen bar-code data on prices and sales, we find that variation in firm appeal and product scope explains at least four fifths of the variation in firm sales. We show that the imperfect substitutability of products within firms, and the fact that larger firms supply more products than smaller firms, implies that standard productivity measures are highly dependent on implicit demand system assumptions and probably dramatically understate the relative productivity of the largest firms. Although most firms are well approximated by the monopolistic competition benchmark of constant markups, we find that the largest firms that account for most of aggregate sales depart substantially from this benchmark, and exhibit both variable markups and substantial cannibalization effects.

Estimating Unequal Gains across U.S. Consumers with Supplier Trade Data

“…42 Hottman et al (2016), using an oligopoly model and U.S. scanner data, find similar results in the sense that large firms are found to quantitatively deviate from the CES benchmark while small firms do not. differenced out.…”

“…To address this potential issue, we pursue an instrumental variables approach as in Hottman et al (2016). Note that, as in section 3.2.5, the change in the log of the sectoral price index can be linearly decomposed into four terms as follows:…”

“…A useful property that we will exploit is that the CES subcomponent of our aggregate import price index, following Hottman et al (2016), can be linearly decomposed as below:…”

“…There is some potential for these shocks to be correlated, however, whereby an increase in a variety's quality relative to another ( k;t ' vt ") leads to an increase in the intercept of the cost function for that variety relative to the same comparison ( k;t ı vt "). Hottman et al (2016) …”

“…In the spirit of Redding and Weinstein (2016), we normalize the demand shifter for the geometric average firm-product to be one for all time periods.…”