Unemployment risks in four EU countries: a validation study of the ESeC

The aim of the paper is three-fold. First, it tests the construct validity of ESeC, that is to say a nine-fold class scheme usable in every EU country and based on the classification of all the occupations listed in ISCO.08 according to the substantive criteria underlying the EGP class scheme. Second, it seeks to falsify the hypothesis that the globalization process, changes in the labour market regulations, and retrenchment of welfare state provisions have produced an individualisation of the risks of unemployment. Third, it is intended to corroborate the hypothesis that the likelihood of becoming unemployed is still structured on the basis of social classes. To perform the above tasks, data coming from the Austrian, British, Danish and Italian sections of the eight (1994-2001) ECHP annual waves were used. The annual incidence rate of unemployment for all respondents living in the four countries was then computed. Obviously, the labour market statuses of each interviewee were observed for as many times as the duration of his/her presence in the ECHP sample. This means that the data set used in the analysis is a two-levelled one. The occasions of observation of individuals’ labour market position represent the unit of first level, while individuals represent those of second level. A random intercept Poisson regression model for longitudinal data was specified for each individual country in order to measure the net effects – in that country – of ESeC class position on individuals’ annual unemployment incidence rate, controlling for year of observation, and a set of individuals’ time varying (age, level of schooling, civil status, and health conditions) and time constant (gender) characteristics. Obviously the model controls also for unobserved heterogeneity between individuals. The parameters of the four national models expressing the net effects of ESeC classes on the annual unemployment incidence rates of interviewees support the construct validity of ESeC scheme. Actually, in Austria, Denmark, Italy and UK risks of experiencing unemployment are strongly influenced by the social class to which individuals belong. Moreover the effects of each individual social class on the unemployment incidence rates are reasonably similar in all the four countries studied. These results can also be considered as a falsification of the individualisation thesis of unemployment risks. Even after controlling for several individual characteristics and individual unobserved heterogeneity, social class deeply affects the unemployment risks of the individual living in the four EU countries studied. As a consequence, it can be maintained that in these countries social classes continue to represent a crucial factor of structuring unemployment risks and related economic inequalities, whatever their respective welfare arrangements, their labour market regulations, their economic structure and conjuncture.