Tomasz Michalski at the RITM Economics Seminar

Tomasz Michalski (HEC Paris) will present his paper “Returns to Experience Across and Within Cities,” co-authored with Donald R. Davis, Eric Mengus and Elie Vidal-Naquet.

Abstract: A growing literature in regional, urban, and labor economics considers the origin of urban and skill premia in wages that rises with city size and density. An area of contest has been the extent to which these premia are explained by the sorting of firms and skills versus location effects in both levels and dynamics. We argue that resolving this dispute requires a focus on the spatial resolution of the inquiry. The traditional AKM approach focuses on the commuting zone as the appropriate spatial unit, since this defines the labor market for matching. However, a rich literature in urban economics focuses on spatial heterogeneity within cities, especially the role of job-dense central business districts. These may have a quite distinct character that promotes knowledge spillovers, with consequences for wages. In this paper, we pursue this insight by using geocoded French administrative data that allows us to distinguish between high and low job-density locations and so to consider both within and between city differences. We also partition jobs along an axis with high versus low analytical and interpersonal skills that may matter differently in different locations. We estimate both localized and occupation-specific returns to experience. We find that a small number of highly dense areas within large cities are at the apex of both the dynamic and static components of the urban wage premium. After ten years of experience, workers in analytical and interpersonal occupations in the 50 densest km² of the Paris metropolitan area earn dynamic wage gains 3.4 times higher than similar workers elsewhere in Paris (+88pp vs. +26pp), relative to workers outside large cities. Occupations involving high computer use display particularly strong gains, suggesting that task content and local job density jointly shape wage trajectories. Controlling for establishment fixed effects or the time evolution of job type wages at the area level does not qualitatively change our results.