This paper analyses the influence of country-level education mismatch on the individual-level relationship between education and the probability of being unemployed or staying in alternative labour statuses, for young people aged 15-34 in 2006, 2008 and 2010, living in 21 EU countries. We assume that young people may fall in five labour market statuses: 1) Employee; 2) Self-employed; 3) Unemployed; 4) In Education; 5) Inactive, and perform a multinomial logit model to study the effects of years of education on relative probability of being in labour statuses 2, 3, 4, or 5, compared to the base category (Employee). Afterwards, we interact the individual-level years of education with a country-level indicator of education mismatch in order to identify the heterogeneous effects of the aggregate mismatch among people with different educational attainments. Results show that more years of education: i) reduce the relative probability of being unemployed; ii) have a cumulative effect by extending the period of education; iii) slightly raise the relative probability to be self-employed. As regards country-level education mismatch, we found that only after 2008 it produces an additional effect on better educated young people by further reducing their relative unemployment risk when it is compared to that of low educated youngsters. This outcome tells us that improving access to university degrees remains the main road to tackle youth unemployment caused by education mismatch, even after the outburst of the current financial and economic crisis.
youth unemployment education mismatch multinomial logit model