Production and learning in teams
Kyle Herkenhoff, University of Minnesota
Jeremy Lise, University of Minnesota
Guido Menzio, New York University
Gordon Phillips, Dartmouth College
The effect of coworkers on the learning and the productivity of an individual is measured combining theory and data. The theory is a frictional equilibrium model of the labor market in which production and the accumulation of human capital of an individual are allowed to depend on the human capital of coworkers. The data is a matched employer-employee dataset of US firms and workers. The measured production function is supermodular. The measured human capital function is non-linear: Workers catch-up to more knowledgeable coworkers, but are not dragged-down by less knowledgeable ones. The market equilibrium features a pattern of sorting of coworkers across teams that is inefficiently positive. This inefficiency results in low human capital individuals having too few chances to learn from more knowledgeable coworkers and, in turn, in a stock of human capital and a flow of output that are inefficiently low.