Automation, Globalization and Vanishing Jobs: A Labor Market Sorting View

We show, theoretically and empirically, that the effects of technological change associated with automation and offshoring on the labor market can substantially deviate from standard neoclassical conclusions when search frictions hinder efficient matching between firms with heterogenous tasks and workers with heterogenous skills. Our key hypothesis is that better matches enjoy a comparative advantage in exploiting new technologies. It implies that technological change promotes employment when initial productivity is low so that firms and workers are not very selective in matching, whereas it hampers employment when initial productivity is high enough to make firms and workers sufficiently selective. It also implies that, conditional on high enough selectivity, the negative impact of technological change on employment eventually gets weaker as firms and workers become increasingly selective. Adding an endogenous choice to offshore to the model predicts that higher offshorability reduces employment holding constant the state of technology and high offshorability amplifies the negative effect of technological change the more the lower the selectivity. Capturing task heterogeneity at the sectoral level and skill heterogeneity at the occupational level, we find empirical support to both predictions in a dataset covering 21 occupations and 21 sectors in 13 countries from 1995 to 2010. Joint with Sébastien Laffitte, Gianmarco Ottaviano and Maximilian Mayer.