Mismatch between jobs and skills in the EU

In order to study the mismatch between the available skills of labour force and the skills required by labour market, we developed the concept of Labour Market Attractiveness. This concept consisted of the combination of a set of variables from 6 Eurostats datasets on different subjects (i.e. Demographics; Earnings structure; Education and training; Life conditions; Employment and unemployment; and National accounts). The impact of this combined dataset on Skills Mismatch, as well as on Labour Market Mobility and Emigration, was assessed using various data mining techniques, particularly, clustering analysis, model selection analysis using multivariate regression and weighted network correlation analyses. We showed that Labour Market Attractiveness is able to form consistent clusters at country-level, which can be well defined using only the variables “Youth Unemployment” and “GDP”. Furthermore, from this combined dataset we defined 6 Eigenvariables, namely, “Unemployment”, “Poverty”, “Ageing Population”, “Secondary Education (Adults)”, “Employment” and “Earnings structure”. Considering these Eigenvariables, we found that: Skills Mismatch is negatively associated to “Employment” and “Secondary Education (Adults)”, while being positively associated to “Poverty” and “Unemployment”; Labour Market Mobility is associated to “Earnings structure” and “Employment”; and Emigration is negatively associated to “Secondary Education (Adults)” and “Ageing Population”. From model selection, we showed that: Skills Mismatch is best explained by “Proportion of employed youth” and “Proportion of employed youth working at NACE M-N”; Labour Market Mobility is best explained by “Proportion of employed adults working at NACE L and K”; and Emigration is best explained by “Proportion of employed adults working at NACE K”, “Proportion of employed youth with Higher Education”, “Proportion of employed youth working at NACE B-E and O-Q” and “Total Population size”.