Getting more unequal: Rising labor market inequalities among low-skilled men in West Germany
Introduction
Labor market inequalities have risen substantially over recent decades in most advanced economies (Dolton et al., 2009, OECD, 2008). A practically universal feature of this trend is that labor market outcomes of low-skilled people have deteriorated compared to those of high-skilled people (e.g., Blackburn et al., 1989, Gesthuizen et al., 2011, Maxwell, 2008). Yet evidence is accumulating that inequalities have also risen within relatively homogeneous socio-demographic and educational groups (e.g., Barth and Lucifora, 2006, Devroye and Freeman, 2001). As early as 1993, Juhn, Murphy, and Pierce (1993) highlighted the contribution of rising earnings inequality within educational groups to overall inequality growth in the United States (cf. also Lemieux, 2006; for Western Europe: Barth & Lucifora, 2006; for Germany: Dustmann, Ludsteck, & Schönberg 2009). Few studies, however, have taken an in-depth look at changing within-group inequalities. In this paper, we therefore seek to advance our understanding of inequality trends by zooming in on one specific group of workers in one large industrialized country: low-skilled men in West Germany (the former Federal Republic of Germany including West-Berlin).3
Low-skilled men in West Germany are an interesting case. Not only has overall wage inequality in Germany risen substantially during recent decades, previous research also documents exceptionally large increases in inequality among low-skilled men (Dustmann et al., 2009, Erlinghagen, 2006, Giesecke and Heisig, 2010).4 These findings are underscored by a comparative study of twelve Western European countries for the period from the 1980s to 2003: Only in the cases of Germany and Finland was wage dispersion within the low-skilled group higher than among middle-educated workers (Barth & Lucifora, 2006: 30).
Our analysis of low-skilled men complements previous studies of within-group inequalities, which have mostly focused on the higher educated. Many researchers predict that educational expansion boosted inequalities among the higher educated, as workers with lower overall “ability” were added to this group (“skills erosion”, cf. Barth & Lucifora, 2006) and/or because the supply of higher-educated workers rose faster than demand (“over-education”, cf. Lindley & McIntosh, 2010). Low-skilled workers, by contrast, are often expected to have become more homogeneous due to the outflow of individuals with higher ability (Barth and Lucifora, 2006, Leuven et al., 2004, Solga, 2002).5 Some researchers draw the seemingly obvious conclusion that labor market inequalities among the less educated have remained stable or even decreased over time (e.g., Barth & Lucifora, 2006: 6).
It is important to clarify at the outset whom we treat as “low-skilled” in the West German context. We define this group as including all men who hold neither a vocational certificate nor a university degree and who are not currently pursuing either of those degrees. In other words, we treat men as low-skilled if they have dropped out of secondary education or only completed a general secondary (as opposed to vocational secondary, post-secondary, or tertiary) program. According to our definition, men with an Abitur – the degree conferred by the highest tier of Germany's tracked system of secondary education – belong to the low-skilled if they do not hold a vocational or university certificate.6 This is appropriate because the German labor market emphasizes occupation-specific skills acquired through vocational training or specialized higher education. Even though men with only an Abitur have completed upper secondary education, they will often lack the occupation-specific skills (and credentials) that are so important for success on the German labor market— and it is in this sense that they can be considered as “low-skilled”.7
Using data from the German Socio-Economic Panel we will examine inequality trends with respect to two outcomes: the risk of long-term non-employment and earnings conditional on employment. Fig. 1 gives a first impression of changes in these outcomes between the mid-1980s and mid-2000s. The lower graph displays the percentage of low-skilled men who were “chronically non-employed,” that is, who did not earn any labor income during a given three-year period. The upper graph depicts the distribution of three-year average annual earnings (in 2008 Euros) among those with positive earnings.8
Fig. 1 provides clear evidence of growing labor market inequalities. The share of 20–54-year-old low-skilled men who spent a whole three-year period outside of paid employment is subject to cyclical fluctuations, but clearly shows an upward trend. Over the observation period, it rose from less than 6 percent in the mid-1980s to 13 percent in the mid-2000s (lower graph in Fig. 1). Among those with positive earnings, the shape of the earnings distribution also changed quite dramatically (upper graph in Fig. 1). In the first three-year period, 1984–86, the distribution of low-skilled men's earnings was clearly unimodal and not too far away from a normal distribution, with the majority earning an annual average between €15,000 and €40,000. By 2005, the distribution had become considerably flatter and acquired a heavy lower tail, with many men making only modest annual incomes below €20,000. The middle and upper parts of the distribution likewise became more dispersed from the mid-1980s to the mid-2000s. Detailed inspection of results for the intermittent periods (lighter lines in upper graph) reveals that this trend is secular rather than cyclical.
Our goal is to better understand and shed more light on potential drivers of these striking trends. Following standard approaches in the inequality literature, we distinguish between two general sources of distributional changes among low-skilled men: compositional changes with respect to earnings- and employment-related characteristics and changes in the effects of these characteristics. We distinguish two broad sets of relevant characteristics: worker characteristics (e.g., health or migration background) and job characteristics (e.g., sector or firm size).
The paper is structured as follows. We first highlight a few essential details of the German institutional context (Section 2). Section 3 discusses potential explanations for the increase in inequality. We then describe our empirical approach (Section 4) and present our empirical results in Section 5. Section 6 concludes with a discussion of our main findings and some questions for future research.
Section snippets
Institutional context
Germany is often characterized as a prototypical coordinated market economy (Hall & Soskice, 2001) with a conservative welfare state (Esping-Andersen, 1990). Strict employment protection is argued to spawn marked insider–outsider divides (e.g., OECD, 2004): Whereas labor market insiders enjoy stable employment patterns and earnings streams, individuals who want to enter or re-enter the labor market face increased risks of being excluded from secure, well-paying jobs, with insider–outsider
A survey of possible explanations
How can we account for the striking changes in labor market inequalities displayed in Fig. 1? Our point of departure is a standard decomposition framework where the distribution of labor market outcomes (e.g., chronic non-employment or earnings) is determined by (a) the distribution of worker and job characteristics and (b) the effects of these characteristics on the considered outcome. Within this framework, changes in the distribution of labor market outcomes can arise because of
Data and methods
Our data come from the German Socio-Economic Panel (GSOEP; Wagner, Frick, & Schupp 2007), a panel study based on annual interviews with a probability sample of German households. The GSOEP began in 1984 and has been expanded and refreshed repeatedly. We include all samples except the high-income sample drawn in 2002.11
Compositional changes
To what extent can the increase in labor market inequalities among low-skilled men be attributed to changes in the distribution of worker and job characteristics? Table 3 reports our findings for the share of chronically non-employed men. As noted above and shown in the first row of Table 3, the prevalence of chronic non-employment more than doubled from the mid-1980s to the mid-2000s.
Rows 2–6 of Table 3 display the results of the compositional adjustments. We present five scenarios that
Conclusion
The goal of the paper was to better understand the growth of labor market inequalities among low-skilled West German men. We showed that the risk of long-term exclusion from work rose markedly between the mid-1980s and mid-2000s and that earnings dispersion grew dramatically among those who were employed. Our results further suggest that these changes are mostly secular rather than cyclical. In order to better understand these trends, we related them to both compositional changes among
References (40)
- et al.
Skills, tasks and technologies: implications for employment and earnings
- et al.
Decomposition methods in economics
Employment protection reforms, employment and the incidence of temporary jobs in Europe: 1996–2001
Labour Economics
(2010)- et al.
Do new labour activation policies work? A descriptive analysis of the German Hartz reforms. WZB Discussion Paper SP I 2011-211
(2011) - et al.
Unemployment benefit and wages: the impact of the labor market reform in Germany on (reservation) wages. Ifo Working Paper 101
(2011) - et al.
Wage dispersion, markets and institutions: The effects of the boom in education on the wage structure. IZA Discussion Paper 2181
(2006) - et al.
The declining economic position of less-skilled American males. NBER Working Paper 3186
(1989) - et al.
The perverse effects of partial labour market reform: fixed-term contracts in France
Economic Journal
(2002) - et al.
Youth on globalised labour markets: Rising uncertainty and its effects on early employment and family lives in Europe
(2011) - et al.
Does inequality in skills explain inequality in earnings across advanced countries? NBER working paper 8140
(2001)