Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
How do cities' average wages adjust to a shock, e.g. trade-induced or technology-induced, shifting the industrial and/or occupational composition of employment? This question is commonly answered using a shift-share analysis, which imputes the effects of composition changes on cities' average wages in a purely mechanical way. This paper argues that a shift-share approach can be misleading if compositional shifts spill over into industrial and occupational wages. Using data on Germany, this paper finds that the spillover effects from a shift in industrial and occupational composition are large and imply a total effect on cities' average wages of 2.8 times the effect that is predicted by a shift-share exercise. In addition, results indicate that in the case of Germany, accounting for adjustments along the occupational mix is crucial to identify these spillovers, even if one is only interested in the effect of an industry-specific shock.