How does worker mobility affect business adoption of a new technology? The case of machine learning

Abstract Research Summary We investigate how worker mobility influences the adoption of a new technology using state‐level changes to the enforceability of noncompete agreements as an exogenous shock to worker mobility. Using data on over 153,000 establishments from 2010 and 2018, we find that changes that facilitate worker movements are associated with a significant decline in the likelihood of adoption of machine learning. Moreover, we find that the magnitude of decline depends upon the size of the establishment, the extent of predictive analytics adoption in its industry, and the number of large establishments in the same industry‐location. These results are consistent with the view that increases in outward worker mobility increase costs for adoption of a new technology that involves significant downstream investments in the early years of its diffusion. Managerial Summary Successful business adoption of new technologies such as machine learning requires skilled workers with experience in implementing those technologies. In the early years of technology diffusion workers in early adopting businesses typically acquire these skills through on‐the‐job learning that is paid for by the adopter. So, if such early adopters face an increased risk of those skilled workers quitting, then their incentives to adopt the technology decrease. We examine this possibility using changes in noncompete enforceability as a proxy for changes in worker mobility and find that the likelihood of adopting machine learning decreases as the risk of worker mobility increases, particularly for larger establishments, establishments in industries where adoption may be more beneficial and in locations with many large competing establishments.