Size-Dependent Regulations, Firm Size Distribution, and Reallocation

Abstract

In France, firms with 50 employees or more face substantially more regulation than firms with less than 50. As a result, the size distribution of firms is visibly distorted: there are many firms with exactly 49 employees. We model the regulation as the combination of a sunk cost that must be paid the first time the firm reaches 50 employees, and a payroll tax that is paid each period thereafter when the firm operates with more than 50 employees. We estimate the model using indirect inference by fitting the discontinuity of the size distribution. The key finding is that the regulation is equivalent to a combination of a sunk cost approximately equal to about one year of an average employee salary, and a small payroll tax of 0.04%. Our structural model fits well the discontinuity in the size distribution. Removing the regulation improves labor allocation across firms, leading in steady-state to an increase in output per worker slightly less than 0.3%, holding the number of firms fixed. However, if firm entry is elastic, the steady-state gains are an order of magnitude smaller.

The Firm Size Distribution across Countries and Skill-Biased Change in Entrepreneurial Technology

“…Because Gourio and Roys aggregate data across years, the figure plots the ratio of the number of firms at each integer level of employment from 20 to 100 to the number of firms with employment of exactly 100 for both sources. Close agreement of the two sources is obvious, including even the kink in the distribution at 50 employees that is the focus of Gourio and Roys (2014) and Garicano, Lelarge, and Van Reenen (2016). Below 50 employees, the accuracy of the size distribution obtained from Amadeus gradually deteriorates.…”

“…Amadeus thus provides an excellent picture of the size distribution of firms with 50 or more employees across the 19 countries under consideration. 12 Finally, I compare the firm size distribution in Amadeus to that in French official data using the extremely detailed data on firm sizes for France reported in Gourio and Roys (2014). 13 Figure 2 compares the size distribution for firms with 20 to 100 employees.…”

“…This also implies that for some large size categories, the number of firms in Amadeus could exceed that in Eurostat. 13 Thanks to Nicolas Roys for providing the detailed firm size counts underlying figure 1 in Gourio and Roys (2014). The data is complied by INSEE, the French statistical institute, based on a combination of administrative data and surveys.…”

“…Each of the two datasets thus provides a reliable image of part of the size distribution, with a substantial area of overlap in which both do a good job. Source: For the INSEE data: Gourio and Roys (2014) They also contain measures of fit, which are high for a bivariate relationship in cross-sectional data. Fact 1.…”

Growing Like Spain: 1995–2007

“…Another group of papers has been looking for examples of specific sources of misallocation that could help explain the large heterogeneity in firm productivity measured in the data. For instance, Gourio and Roys (2014) and Garicano et al (2016) looked at the labor regulation in France, and García-Santana and Pijoan-Mas (2014) assessed the Small Scale Reservation Laws in India. However, as Hopenhayn (2014) showed, it is hard to create large losses in aggregate productivity through misallocation without large rank reversals in firm sizes between the actual and the efficient economies, something that this type of size-dependent policies does not achieve.…”