Is the Spanish Recession Increasing Inequality? Male Migrant-native Differences in Educational Returns Against Unemployment

Abstract

In this paper, we explore the impact that the current economic crisis is having on the extent to which education protects migrants and natives from unemployment in Spain. To do so, we use random constant models on a pooled sample of respondents to the Labour Force Survey (LFS) from 2003 to 2012. Specifically, we analyse the market value of the educational credentials held by African, Latin American and East European migrants compared with Spanish males aged 16–50. Our conclusions indicate that inequality in the returns provided by education to migrants and natives has been growing fast in recent years. In other words, migrants have been less able to grant themselves an employment through their education than natives since 2008 when the Spanish economy started to contract and that this process has also been speeding up since. We also identify important differences within the migrant category: while the potential to prevent unemployment among better-educated migrants from Eastern Europe decreased during the economic recession, it is Africans and Latin Americans who are worse off.

Keywords:

• Migration
• Unemployment
• Returns to Education
• Recession
• Spain

Acknowledgement

We are grateful for the comments given by two anonymous referees and Luis Garrido Medina.

Notes

[1] These are our estimations from the LFS and the Local Registers (Padrón Municipal Continuo). In 2012, 36% of the foreign-born population residing in Spain were born in a Latin American country, 41% from Europe, 16% from Africa, mainly from Morocco, and only 6% from Asia, mostly from China.

[2] Despite the larger impact of unemployment among less-educated workers regardless of their migrant status, the educational profile of newcomers remained pretty stable since 2006 (Fernández-Huertas 2014).

[3] Among the OECD member states, the impact of the current economic crisis concentrates on poorer educated individuals. The unemployment gap between well-educated young people and early dropouts from school widened during the crisis. On average, 5% of those with university degrees were unemployed, compared to 13% of those without upper secondary education. In the period 2008–2011, the unemployment rate for the second group increased by 4%, compared to 1.5% among those with tertiary education (Van Damme 2013).

[4] Some scholars continue to believe in the existence of period rather than cohort effects. These authors blame Borjas for not considering the changing context in the host society where human capital is increasingly under-priced (Chiswick 1986; Duleep and Regets 1992; La Londe and Topel 1992; Yuengert 1994).

[5] The primary segment is characterised by stable jobs, high wages, good prospects for promotion and strong protection against economic downturns. The secondary segment, however, is more precarious, poorly paid and more vulnerable to change. Migrants are more likely to be confined to the secondary segment of the labour market because (i) social capital plays an important role in the migration and incorporation process since it grants a different level of access to relevant resources across ethnic origins (Portes and Rumbaut 1996; Portes 1998; Espinosa and Massey 1999; Sanders, Nee, and Sernau 2002; González-Enríquez 2009); (ii) discrimination on the part of employers also appears to be a driver (Gill 1989; Moore 2010; Kaas and Manger 2012).

[6] Note our hypotheses refer to a general context of increasing returns to education (regardless of migrant status), which adapts better to the characteristics of the current economic crisis. Unemployment rates among lower-educated men and women have been higher than among those who have attained a tertiary education. The economic crisis only widened these gaps (OECD 2013, 75).

[7] Bulgarians and Romanians are vastly overrepresented in this category, where they represent 14,6 and 64.1%, respectively. A distinction between migrants of different origins is statistically unstable. Setting the meaningful difference between EU and non-EU nationals (since the former group enjoys circumstances more conducive to settling in member states), implies no changes in our conclusions.

[8] Asians were eliminated from the analyses since the subsample corresponding to this group was insufficient.

[9] We restrict our analysis to the men population between the ages of 16 and 50. This is due to several reasons. From a substantive point of view, the most dramatic change in the labour market status of migrants prior and after the crisis happened among men workers. Besides, the computational load of our models (which are estimated on the basis of very large sample sizes) required some sort of sample restriction. The analyses we present here have also been estimated for the sample of women workers and the results are available upon request.

[10] Using the regional yearly level of unemployment implies no changes in the results.

[11] The quantification of the importance of the clustering is calculated using the variance of the constant error term and the individual one according to the following formula:

[12] The mixed model seems to be the most appropriate technique for modeling this process since the LR test versus the standard LPM is statistically significant.

[13] From the decomposition of the variance, we know that most of the variation is at the individual level [Var(εi)] but that there is also some variation over time. This variation mostly affects the constant rather than the slope of education. This should be interpreted as a confirmation that most of what determines the changing probability over time of being employed is related to unobserved processes that are not captured by the changing impact of education during the crisis.

Additional information

Funding

This work was supported by the Spanish Ministry of Economy and Competitiveness [grant number CSO 2010-21004].