We propose a model with asymmetric firms where new technologies displace workers. We show that both leading (low-cost) firms and laggard (high-cost) firms increase productivity when automating but that only laggard firms hire more automation-susceptible workers. The reason for this asymmetry is that in laggard firms, the lower incentive to invest in new technologies implies a weaker displacement effect and thus that the output-expansion effect on labor demand dominates. Using novel firm-level automation workforce probabilities, which reveal the extent to which a firms’ workforce can be replaced by new AI and robotic technology and a new shiftshare instrument to address endogeneity, we find strong empirical evidence for these predictions in Swedish matched employer-employee data.
Reference:
Heyman Fredrik, Pehr-Johan Norbäck and Lars Persson (2021). "Artificial Intelligence, Robotics, Work and Productivity: The Role of Firm Heterogeneity". IFN Working Paper No. 1382. Stockholm: Research Institute of Industrial Economics (IFN).