This paper investigates how investment in automation-intensive goods impacts on worker flows at the firm level and, within firms, across occupational cate- gories. Resorting to an integrated dataset encompassing detailed information on firms, their imports, and employer-employee data for French manufacturing employers over 2002-2015, we identify ‘automation spikes’ using imports of intermediates embedding automation technologies and then test their impact on employment dynamics. We find that automation spikes are positively correlated with preceding and contemporaneous growth in employment, mainly due to lower separation rates of investing firms. These differential patterns of net and gross worker flows do not appear to change significantly across different types of workers (occupational categories, ‘techies’, routine-intensive vs. non routine-intensive jobs).
working paper 13/2019