Labor pooling as an agglomeration factor : evidence from the Brazilian Northeast in the 2002-2014 period

This is a pioneering study of Brazil of the importance of labor pooling to explain industrial agglomeration in the Northeast of Brazil, employing firm-level microdata. We applied the theoretical model proposed by Krugman (1991) (labor market pooling model) with the adaptations by Overman and Puga (2010) to examine how firms react to shocks in the labor market that influence their productivity. For this purpose, we applied regression models in which we regressed the Ellison and Glaeser (1997) index as a function of a proxy for labor pooling, to capture exogenous shocks in the labor market while controlling for observed sector characteristics that vary in time and sector fixed effects. The results are consistent with a reduction in the level of industrial concentration in the period from 2002 to 2014. For labor pooling, as predicted, the role of the labor pooling variable is positive and significant.Thus, industries where, on average, plants face more idiosyncratic shocks relative to their industry are more spatially concentrated.