Spatial externalities and wage distribution: The role of sorting

Recent literature has shown that sorting plays an important role when the impact of spatial externalities on disparities in average wages between locations is investigated. The aim of the paper is to show that sorting matters also when addressing the relationship between spatial externalities (employment density and specialization) and wage distribution, i.e. across workers located at different percentiles of the wage distribution. Previous empirical papers did not control for observed and unobserved heterogeneity since they used aggregate data (Wheeler, 2004, 2007, Moller and Haas, 2003). We can control for observed individual and firm heterogeneity since we make use of the Italian employer-employee panel data. By means of standard quantile estimates, we find an increasing impact of spatial externalities on wage distribution. Furthermore, in order to control also for unobserved individual heterogeneity, we perform quantile fixed effect estimates (Koenker, 2004). Results show that the sorting matters since it dampens all the spatial externality effects along the wage distribution, effects that remain positive only in the upper tail of the wage distribution. Moreover, the impact of sorting is even stronger when using the sample of displaced workers, which represents a random sample of the workforce