Artificial intelligence and outward foreign direct investment: Evidence from China

We investigate the impacts of robot adoption (a representative form of artificial intelligence) on outward foreign direct investment (OFDI) by firms in China. First, we construct a unique firmlevel dataset of A-share listed firms and conduct empirical analysis by adopting a staggered difference-in-differences (DID) strategy. Second, we find that changes in robot adoption on both the extensive and intensive margins encourage firms to conduct more OFDIs. Specifically, robot adoption has positive effects on multidimensional categories of OFDI but has no significant impact on buying OFDI. Third, we explore the underlying mechanism to show that the rise in output and growth of total factor productivity (TFP) matter. Fourth, we conduct a series of empirical tests to check robustness and obtain consistent results. Finally, we analyze the heterogeneous effects and find that adoption firms with state-owned ownership or higher profit tend to send more OFDI projects, especially to the Organization for Economic Co-operation and Development (OECD)-rich countries. Our findings elucidate the impact of robot adoption on OFDI at the firm level in developing and transition countries.