Zusammenfassung:
This paper focuses on identifying determinants of 'automatability risk', namely the propensity of EU employees being in jobs with high risk of substitutability by machines, robots or other algorithmic processes, and uncovers its impact on labour market outcomes. Using relevant data on tasks and skill needs in jobs, collected by the European skills and jobs survey (ESJS), jobs are bundled according to their estimated risk of automation. The paper builds on the methodology of previous studies that estimate the latent relationship between 'true' automatability and job tasks (Frey and Osborne, 2013, 2017; Arntz et al., 2016; Nedelkoska and Quintini, 2018) but utilises highly disaggregated job descriptions provided by a subsample of the ESJS, as well as information on jobs' skill requirements. About 14% of EU adult workers are found to face a very high risk of automation. The distribution of high automatability across industries and occupations is also found to be skewed towards routine jobs with low demand for transversal and social skills. The risk of job displacement by machines is higher among males and lower-skilled workers, with little evidence of polarisation. It is prevalent in private sector jobs that fail to provide remedial training to employees, accentuating the vulnerability of at-risk-workers and highlighting the need for stronger lifelong learning policies at EU level.