Seminar 13 October 2014

The Aggregate Significance of Labor Reallocation

Abstract

We analyze the impact of workers switching between firms on GDP. The decomposition in Basu and Fernald (2002, European Economic Review) identifies the contributions of firm-level productivity growth, worker and capital flows and technological change on aggregate productivity growth and subsequently GDP. This paper quantifies the importance of the worker reallocation term on aggregate productivity growth. In a nutshell, worker flows boost GDP if high-productivity firms expand and low-productivity firms shrink. Our data spans the population of workers and firms in Sweden between 1997 and 2011, and allows us to match workers to firms. Separating out the firm-specific component of wages and under the assumption of cost minimization, we show that there is no systematic flow of workers towards high-productivity firms, resulting in a near-zero contribution of worker reallocation to aggregate productivity growth. This result is in contrast with previous studies of the manufacturing sector, where between-firm reallocation is an important component of aggregate growth. We study the divergence between these results, in terms of the methods used and data sources. Finally, we make two independent contributions. First, we test for and reject an efficient bargaining model in determining wages. Second, we show that the estimated firm-specific component of wages is highly correlated with both measures of nominal and real labor productivity, suggesting that the output price bias may be partly eliminated.

Contact: Battista Severgnini and Cédric Schneider