Replacing Workers: Is It a Boon or a Bane for Firm Productivity?

Using a uniquely rich longitudinal matched employer-employee data set, this paper is the first to investigate the impact of replacing workers, as measured by excess worker turnover, on firm productivity. Using a modified version of the method proposed by Ackerberg et al. (2006), that allows to take into account unobserved heterogeneity, an augmented production function with excess worker turnover entering as the regressor of interest is estimated. The main result is that replacing workers is beneficial to firm productivity. A 1 standard deviation increase in the excess worker turnover rate is estimated to increase productivity by 0.81%. The possibility of finding more suitable employer-employee matches and the presence of knowledge spillover effects are seen as the main determinants of the impact. Robustness checks indicate that the impact has an inverted U-shape, suggesting that, beyond a certain point, replacing workers ends up being harmful. However, since about 90% of firms lie before this point, increases in excess worker turnover are beneficial for the vast majority of them. They also suggest that the effect is diversified across different categories of firms. High-tech firms and firms belonging to industrial districts benefit the most from excess worker turnover. On the contrary, young and very small firms seem to even suffer from it.