Abstract
This paper investigates the effects of robots on labour demand at the firm level in the Norwegian manufacturing industry from 2003 to 2016. Labour demand is proxied by the wage shares of different worker groups within a firm and access to robot technology is proxied by imports of industrial robots. The worker groups are based on their level of completed education; compulsory school (CS) workers, upper secondary school (US) workers and higher education (HE) workers. I use both dynamic and static two-way fixed effects (TWFE) regressions to examine the effects of robots on these worker groups wage share. The results show that robot firms generally have a higher wage share for HE workers before robot adoption and that there is a skill biased and robot-independent trend towards higher wage share for HE workers for all firms, although this trend is slightly stronger for robot firms. The opposite is true for the wage share of CS and US workers. However, after the firm adoptions robots, the wage share for US workers increases and the wage share for HE workers decreases. CS workers’ wage share is unaffected. The results are confirmed by robustness tests. Viewed together with the results from Barth et al. (2020), who examine the same firms and workers, these results suggest that robots lead to relatively higher wages for HE workers, while US workers are relatively more employed. The latter effect seems to dominate as the total effect, examined in this paper, is a relative increased demand for US workers and reduced relative demand for HE workers after Norwegian manufacturing firms adopt robots.