Robots and wages: A meta-analysis

Given the mixed empirical evidence, our meta-study aims to uncover the effect of industrial robots on wages and to identify the drivers of heterogeneity in the primary literature. We collected 55 papers containing 2,468 estimates through systematic literature research. The overall effect of industrial robots on wages is close to zero and statistically insignificant. We observe little evidence of a publication selection bias in general. However, we find evidence of a preferential selection of negative results when authors focus on the USA. Our multivariate meta-regression analysis suggests that the heterogeneity among primary estimations is mainly driven by the selection of countries and control variables, aggregation level, and functional form. Nevertheless, we do not find any economically significant wage effect for specific country groups, aggregation levels, or subsamples of workers.