On automation, labor reallocation and welfare

We develop an open-economy model of endogenous automation with heterogeneous firms and labor-market reallocation to quantify the contribution of various trends to the adoption of robots in the U.S. economy. The decline in the relative price of robots is the major trend leading to automation, but interacts with other trends that either hinder (rising entry costs, rising markups) or slightly foster (rising labor productivity, declining trade costs) the adoption of robots. Taken alone, the decline in the relative price of robots produces moderate welfare gains in the long run, but less than labor productivity growth. We then exploit our model to show that a decline in the relative price of robots (i) generates small positive cross-country automation spillovers and (ii) produces inefficient labor-market reallocation since a small subsidy on robots combined with a training subsidy can generate small welfare gains. Our main conclusion is that automation can not be simply modeled as an exogenous decline in the price of robots, and must be analyzed in a broader framework taking into account trends affecting firms, such as the decline in business dynamism and the rise in markups.✩ We would like to thank the editor B. Ravikumar and two referees for their invaluable comments. We also thank Basile Grassi, Sergio Rebelo, Ariel Resheff and Farid Toubal as well as participants in various seminars and conferences for interesting feedbacks. We acknowledge financial support from the French National Research Agency (ANR-20-CE26-0018-02).