Heterogeneous Causal Impacts of Highway on Local Employment Growth

Abstract

This study examines the heterogeneity in the causal impact of highway development on local employment growth. To this end, we apply causal forests, which is a machine learning algorithm for causal inference, to employment data for the manufacturing sector in Japanese municipalities from 1971 to 2011. We then find evidence that an opening of highway interchange increased local employment by 8.5 ％ on average between 1971 and 1991, while there was no significant impact between 1991 and 2011. We also identify regional characteristics that affected the heterogeneity in the causal effects for each period, and reveal that those differed between periods of economic growth and stagnation.