Are Small Firms more cyclically Sensitive than Large Ones? National, Regional and Sectoral Evidence from Brazil

An important issue facing policymakers is the degree to which fluctuations in economic activity affect employment in large and small businesses across sectors and regions. These issues are particularly relevant for developing countries, as they matter for the understanding of the labour market dynamics, and for devising regional, sectoral, and national labour policies. The unique data used in this paper was constructed using the CAGED database, which is a comprehensive administrative census dataset collected monthly by the Ministry of Labour in Brazil covering the formal sectors of the economy. Thus, monthly employment data for small and large establishments across regions and industries were constructed from 2000:1 to 2009:6 for each state and industry across the two-digit sectoral classification. This paper draws on the work of Moscarini and Postel-Vinay (2009) who analyse the correlations between measures of relative growth rate of employment by size class and business cycle conditions. As in their work, this paper uses detrended difference in employment growth rates between large and small firms as a measure of the relative performance of firms in different size bins. The evidence suggests that in Brazil small firms are more sensitive cycles, a result that contradicts Moscarini and Postel-Vinay (2009). The differential growth of employment between large and small establishments is negatively correlated with measures of business cycles, indicating that SMEs shed proportionally more jobs in recessions and gain more in booms. This pattern is also observed in most of the Brazilian States; however, there is a substantial variation in the manner the difference in employment growth rates correlates with business cycles at regional level. Besides, the sectoral analysis supports the evidence that formal small businesses are more sensitive than large ones in all sectors but in the commerce. This finding is important and might be related with the fact that the commerce sector relies heavily on informal workers that are the first ones to be hired or made redundant over the business cycles. Therefore, the evidence from this paper suggests that in a developing country context small establishments are more sensitive than large ones to business cycle conditions.