Family Background and Youth Labour Market Outcomes Across Europe: Did the Crisis Exacerbate Inequalities?

There is a robust consensus in the literature about the existence of a positive correlation between unemployment or worklessness of children and their parents (Corak et al. 2004; Bratberg et al 2008; Oreopulos et al. 2008; Ekhaugen 2009; Macmillan 2010, 2013; Gregg et al. 2012; Mader et al. 2014). Almost all of these studies, focus on a single country and on the occupational condition of fathers, showing that children of non-working fathers are less likely to be employed themselves. Some of them are interested in testing whether the latter is really a causal effect, while others aim at disentangling the different transmission mechanisms (transmission of preferences, financial constraints for human capital investments, wellbeing and mental health, child’s attitude towards unemployment). The aim of our paper is to document the differences across European countries in the effect of parental worklessness on their children labour outcomes, before and after the economic crisis. The contribution of this paper is twofold. First, to the best of our knowledge, this is the first comparative study at the European level about the effect of parental background on youth unemployment risk. A cross-country comparison is important not only to assess the degree of this specific type of inequality, but also to understand the link between the latter and different labour market institutions and welfare regimes. Second, we consider the occupational condition of both parents, and not only the father's one. This allows us to study the role of the gendered division of household labour (male-breadwinner vs dual-earner models) in explaining youth unemployment risk and, especially, daughters' participation in employment.

This study is based on EU-SILC data which encompass extensive comparable cross-sectional and longitudinal microdata both at the household and individual level for all EU-27 Member States. We exploit the 2005 and 2011 waves because they provide several information on the parental educational and occupational background through two ad hoc modules of intergenerational transmission of poverty (2005) and disadvantages (2011). All variables relate to the period when the individual was around 14 years old and are collected only for individuals aged between 24 and 60 at the time of the interview. For all young individuals aged 25-34 we estimate the probability to be “employed”, "unemployed", "in education" or "inactive" (using a multinomial logit model). We control for a set of individual characteristics (age, education, parenthood status, citizenship), the presence of parents (living with both parents, with only one parent or without parents) and the household occupational structure. We distinguish between two-parent households with both working parents (work-rich households), two-parent households with only one working parent, one-parent households with working parent and one/two-parent households with no working parents (work-poor households). Estimations are carried out separately for four groups of countries: Nordic (Denmark, Finland, Norway and Sweden), Continental (Austria, Belgium, France, Germany, Switzerland and the Netherlands), Anglo-Saxon (Ireland and the United Kingdom) and Mediterranean (Cyprus, Greece, Italy, Malta, Spain and Portugal) countries. The four groups of countries are representative of the great heterogeneity of European labour market institutions and welfare regimes.

Descriptive statistics point out that in Nordic countries youth unemployment shares decreased between 2005 and 2011 for youngsters belonging to one or two parent households with working mother and in two parent work-poor households, while the share of young unemployed increased significantly for children of non working lone mothers. In the Anglo-Saxon countries unemployment shares increased for all types of families but especially for work-poor households. Also in Mediterranean countries, unemployment shares increased for all types of families but especially for children of non-working lone mothers and for those living in a two-parent household where only the mother works. In Continental countries, unemployment shares decreased for all types of families,  except for two-parent households with only one working parent.

A different picture emerges from the econometric analysis when we control for individual characteristics and parental background (at the age of 14). Preliminary results show that the effect of the family occupational structure is quite different before and during the crisis, as well as across countries. Coeteris paribus, before the crisis in Nordic and Continental countries the unemployment risk was really heterogeneous across family occupational structures. The odd-ratios of unemployment were particularly high in Nordic countries for children of lone-mothers, independently on the maternal employment status, and for those living in work-poor households. Also in Continental countries, the probability of being unemployed for individuals belonging to work-poor households was well above the average. Differently, the probability of being unemployed was almost comparable across household types in both Anglo-Saxon and Mediterranean countries. The economic crisis reduced inequalities in the unemployment risk associated with the parental occupational structure in Nordic countries, it left them almost unaffected in Continental countries, while it increased them both in Anglo-Saxon and Mediterranean countries. More precisely, during the global economic downturn, in Nordic counties the unemployment probability has decreased significantly for young people with a working lone mother, while in Anglo-Saxon countries it increased remarkably for those with a non-working lone mother. In Mediterranean countries, the crisis almost doubled the unemployment probability of children of lone mothers, whatever the mother's employment status. The crisis also induced a sort of polarization among young individuals living with both parents at the age of fourteen. Indeed, the odd-ratios of unemployment rose more for youngsters whose parents were both working or both workless, compared to those living in households were only the father or the mother was working.

By examining the way in which the economic crisis affected the intergenerational transmission of unemployment in the various countries, we highlight the different vulnerable groups that need policy attention in the various groups of countries. In addition, the evidence of an intergenerational persistence of unemployment suggests that policies should pay attention to both youth and parental unemployment. In particular, the results about polarization deserves great attention because it may be related to the dual role of the family of origin that can act either as  a "trap" or a "protection" for young people against the risk of being unemployed.