Impact of innovation on employment in quantitative terms: review of empirical literature based on microdata

Because of the alarming discussion about robotisation and the potential loss of jobs, the effect of technological change on employment is back at the center of social debate. This study reviews the empirical evidence of 44 micro-level studies on the quantitative impact of innovation on employment, offering a systematic review of the methodologies used, (control) variables and data sets. It tries to explain their contradictory results and shortcomings and summarize their conclusions, offering a new taxonomy of the studies. Although we analyze all types of studies, the main part of our review is focused on two basic models. The first one is the static input-oriented model of Van Reenen (1997) and Bogliacino et al. (2008 and 2014), which shows a positive effect of R&D expenditures on employment. The second one is the dynamic output-oriented model of Harrison et al. (2008), which shows a positive effect of product innovation on employment and contradictory effects for process innovation. Most studies confirm the expected negative effect of process innovation. However, others (especially those which analyze low-income countries and low- tech sectors) reflect non-significant relationships. Further on, some conclusions, final critical remarks about detected limitations and contradictions and possible future lines of research will be presented.