Abstract
We investigate the extent, direction, and determinants of transitions in and out of self-employment across 31 European countries over the period 2014-2019. We focus on workers’ transitions from solo self-employment to paid employment and vice-versa, as well as on transitions within self-employment. We are particularly interested in exploring how exposure to a certain type of technology in one’s occupation affects the probability of moving in and out of solo self-employment, and how this effect differs across worker groups. To answer these questions, we match worker-level data from EU-SILC with indicators of exposures to AI advances and routine task-intensity at the occupational level. First, we use artificial intelligence (AI) indicators as a proxy for exposure to labour-augmenting technology. We find that employees who are more exposed to AI advances are more likely to move to solo self-employment. However, solo self-employment also seems to become more risky, as being exposed to AI advances also increases the probability of transitioning from solo selfemployment to paid employment. Certain workers, particularly the ones with low education, may not fully benefit from the labour-augmenting effects of AI, and therefore switch between employment status in search of better career prospects. Second, we use measures of routine task-intensity (RTI) as a proxy for the exposure to labour-saving technologies. We find that employees with higher RTI are less likely to enter self-employment, be it with or without employees. This is particularly the case for employees in manual occupations. This result may reflect that employees in these types of occupations tend to have limited access to financial resources and skills that are positively associated withthe odds of entering self-employment.