Who is in and who is out? A risk-based conceptualization of insiders and outsiders

Abstract

With the post-industrialization and flexibilization of European labour markets, research on social and economic correlates of labour market vulnerability and weak labour market attachment is growing. Part of this literature conceptualizes these correlates in terms of dualization and insider–outsider divides in an attempt to explore their political implications: this article is written in order to contribute to this strand of research. In this article, we propose a conceptualization and measurement of labour market insiders and outsiders, based on their respective risk of being atypically employed or unemployed. We propose both a dichotomous measure of insiders/outsiders and a continuous measure of the degree of an individual’s ‘outsiderness’. We argue that such risk-based measures are particularly suited for research on the policy preferences and political implications of insider–outsider divides. On the basis of EU-SILC and national household panel data, we provide a map of dualization across different countries and welfare regimes. We then explore the correlates of labour market vulnerability – that is, outsiderness – by relating it to indicators of income and upward job mobility, as well as labour market policy preferences. The results consistently confirm an impact of labour market vulnerability, indicating a potential for a politicization of the insider/outsider conflict.

Keywords

atypical employment insider/outsider theory labour market policy political preferences unemployment

Introduction

Since the 1970s, labour markets and family structures have undergone tremendous changes. With these transformations, employment patterns have changed. The standard post-war model of stable full time employment – often lifelong for the same company – does not correspond to the employment biographies of most individuals anymore. There are two reasons for this, the first being unemployment, especially long-term unemployment, which has again become common in most Western societies (Organisation for Economic Co-operation and Development, 2006). Second, an ever-growing share of the workforce can be found in so-called atypical forms of employment. Part-time and temporary work has become widespread and accounts for most of the job creation in the EU since the 1990s (Organisation for Economic Co-operation and Development, 2006, 2010; Plougmann, 2003). Since welfare states were created in the industrial era to cover the ‘average production worker’, this deviation from previously standard employment patterns may result in specific risks of poverty and welfare losses, especially if welfare entitlement is closely linked to employment, as it is in the social insurance welfare states of Continental Europe (Palier, 2010).

As a consequence of this transformation of labour markets and its welfare implications, the links between weak or flexible labour market attachment, social and economic (dis-)advantages and political opinions have become important topics on the research agendas in labour market sociology (Burgoon and Dekker, 2010; Polavieja, 2005; Ranci, 2010; Tomlinson and Walker, 2012) and political science, mostly in the context of the rapidly growing literature on ‘dualization’, and the divides between labour market insiders and outsiders (see for example Emmenegger, 2009, 2010; Emmenegger et al., 2012; King and Rueda, 2008; Lindvall and Rueda, 2012; Rueda, 2005, 2006, 2007; as well as Davidsson and Naczyk, 2009, for an overview). We consider this literature as an attempt to systematize and aggregate different forms of labour market vulnerability and flexibility in a way that allows for an exploration of their political implications, both at the individual level of preferences and political behaviour, and at the macro-level of policy reforms. However, few contributions have actually invested theoretically and empirically in discussing and validating the conceptualization and measurement of insiders and outsiders. Rather, most of them have referred to the conceptualization and measurement previously used in micro-economics (for example, Lindbeck and Snower, 1998, 2001; Saint-Paul, 1998), which is based on an individual’s current labour market status (fully employed vs atypically or unemployed). We would like to argue, however, that political preferences and political behaviour are not only shaped by individuals’ current situations but also by their expectations and perceptions concerning their (future) labour market risks. Such expectations depend strongly on the socio-structural environment of an individual, that is, on the occurrence of labour market vulnerability in one’s specific social group. Individuals in specific social and occupational groups are more or less likely to experience different forms of atypical employment (often alternating with spells of unemployment) throughout their employment biography. Many women, for example, work full time while young, withdraw from the labour market on account of childrearing and re-enter the labour market years later to take a part-time job. Hence, their employment trajectory clearly differs from the standard model of full time insider employment, but a snapshot of their employment status at a specific point in time would not reveal this. Since their entire (prospective or retrospective) employment trajectory is likely to affect their political behaviour and views, we may want to evaluate their risk of being (and thinking like) an outsider on the basis of the social and occupational group to which they belong, rather than on the basis of their short-term labour market situation.

Therefore, we would like to propose and explore a conceptualization of insiders and outsiders based on their risk of being atypically employed or unemployed. We do this as both a dichotomous and a continuous variable.

The article is structured as following. We first theorize our measures and compare them with other definitions of insiders/outsiders. In a second step, we operationalize the risk of atypical employment and unemployment, and we propose a ‘map of dualization’ that presents insiders and outsiders across regimes and countries, based on EU-SILC data from 2007 and additional household panel survey data for countries not included in EU-SILC. In a third step, we relate our measures to the key indicators of labour market advantage and disadvantage, income and upward job mobility prospects. In the last part of the article, we show that our measures predict differences in insider/outsider-preferences for active and passive labour market policies as hypothesized in the literature.

Theory

Post-industrial foundations of dualized labour markets

Over the past 30 years, the economies of advanced Western democracies have transitioned to a post-industrial social and economic structure. In the industrial era, the industry and the growing public sector were able to provide stable, full time and well-insured jobs for virtually the entire male workforce. In the post-industrial era, however, unemployment rates and especially long-term unemployment rates have increased (Organisation for Economic Co-operation and Development, 2006), and most of job growth is due to atypical employment forms such as part-time employment and temporary or fixed term contracted work (Plougmann, 2003). As a consequence, fewer labour market participants work in the stable standard employment relations that were so typical of the booming post-war decades. For instance, the number of workers on temporary contracts across the European Union has been growing by 15–20 percent annually since the 1980s – a figure that represents about ten times the overall rate of employment growth (Standing, 1993: 433; see also Esping-Andersen, 1999a). Similarly, part-time employment accounts for close to 80 percent of the net job creation in the EU since the mid-1990s (Plougmann, 2003). Of course, part of this flexibilization can be seen as a response to increased claims for more flexible employment conditions, for example, working part-time to combine family and paid work. However, even voluntary atypical employment leads to lower social right than continuous standard employment in the European social welfeare states. Furthermore, research shows that fixed-term contracts tend to imply economic disadvantages and cannot be considered to be reliable ‘bridging’ jobs into permanent employment (Booth et al., 2000, 2002; Gash, 2008; Kalleberg et al., 2000). Furthermore, even voluntary atypical work leads to lower social rights in the European social insurance welfare states. Hence, atypical work and unemployment can generally be interpreted as conditions of increased social and labour market vulnerability. This growing segmentation of the labour market in secure jobs and more ‘vulnerable’, unstable jobs is known as ‘dualization of the labour market’ (Palier and Thelen, 2010; Rueda, 2005, 2007; Saint-Paul, 2002). It is a trend that affects all advanced post-industrial economies, but which differs in its extent and social stratification regarding who is affected. Atypical employment – denoting here all employment relations that deviate from standard and permanent employment – is, for example, clearly gendered in many countries. Especially for women in continental Europe, atypical employment is generally the norm rather than the exception (Esping-Andersen, 1999b, 2009). Similarly, atypical employment is more widespread among younger labour market entrants in a range of Continental and Southern Europe European countries (for example, Chauvel, 2009) than among the elderly workforce.

Conceptualizing labour market vulnerability: defining insiders and outsiders at the micro-level

One may ask whether it makes sense to aggregate different forms of atypical employment and unemployment into specific groups, especially two groups of insiders and outsiders only. From a labour market sociology perspective, the answer is probably negative, but from a political science perspective, the aim is to identify broad socio-structural patterns of preferences and divides, which may be thought of as latent conflicts that may or may not be politicized and mobilized in terms of dualization. Furthermore, existing research shows that despite the heterogeneity of the groups of insiders and outsiders, this distinction is more than a mere academic notation, as it has political implications in terms of individual political preferences (for preferences on job protection, see Rueda (2005, 2007) and Emmenegger (2009); on party preferences see Lindvall and Rueda (2012); on preferences for social policy see Burgoon and Dekker (2010); see also Häusermann and Schwander (2011) and Häusermann and Walter (2010)).¹

We would like to contribute to this literature by proposing a new conceptualization of insiders and outsiders and two new measures of it, one dichotomous (similar to the existing one) and one continuous, which allows a more fine-grained measurement of individual labour market vulnerability. In most of the existing literature, insiders and outsiders are distinguished on the basis of their employment status at a particular point in time (that is, the point when a particular survey is conducted). All respondents who are in stable employment are coded as insiders, while all ‘unemployed, involuntary fixed-term employed and involuntary part-time employed’ are coded as outsiders (Rueda, 2007: 14–15; see also Emmenegger, 2009; Lindbeck and Snower, 1998, 2001; Saint-Paul, 1998, 2002). The validity of conceptualizations obviously depends on the specific research question one investigates. Hence, if one is interested in labour market processes (for example, wage negotiations), the conceptualization on the basis of employment status may indeed meet its analytical purpose. However, if we are interested in politics, that is, policy preferences and mobilization, we may need a conceptualization that classifies insiders and outsiders on the basis of less ephemeral social and economic characteristics, which have an impact on the opportunities and constraints of individuals over a longer time span. This means that individuals might develop political preferences depending on their expectations about labour market risks, expectations that are strongly linked to the labour market prospects of their social group or ‘milieu’.²

A definition of insiders and outsiders based on risk comes with a number of advantages and disadvantages compared with the more widespread measure based on labour market status. The disadvantage is that we attribute characteristics to an individual that are derived from its specific social group, that is, we might attribute to an individual a labour market risk that is never going to become manifest. This implies a number of empirical problems, especially for the dichotomous measure, which we address below when discussing the operationalization. On the other hand, we see three possible interests in a more risk-based measure: first, it is less vulnerable to the problem of volatility, that is, the fact that labour market status may be too unstable to affect an individual’s political preferences (see Emmenegger, 2009 for a similar argument). Indeed, if people repeatedly move back and forth between standard and non-standard employment, that is, if post-industrial societies are fluid and mobile, a categorization of insiders and outsiders on the basis of their current labour market status may lead to problems of misclassification. Therefore, we argue in favour of a conceptualization of labour market risk that is based on a more stable category, namely occupational classes. People may change from unemployment to employment within a few months, and they may even change jobs within the same time span, but they do not change their occupational class (that is, the ‘type’ of job they are in) quickly (Goldthorpe et al., 1987; Mayer, 2000). Of course, even occupational categories are not the perfect empirical basis for evaluating long-term employment trajectories. Ideally, we would rely on data tracing employment biographies over their working life. Such data, however, are not available on a comparative basis. We therefore rely on occupational categories as a proxy for employment biographies. They measure permanent, structural disadvantages more reliably than a snapshot of the labour market status. Think of women in Continental Europe who may be employed full time at a young age, but who will experience periods of career interruption or atypical employment later on, a fact of which they are generally well aware, meaning that the anticipation of future atypical employment will shape their attitudes and preferences today. They do have a vulnerable labour market biography, irrespective of particular spells of full time employment. In sum, our argument is that people form identities and preferences not on the basis of a momentary labour market status, but with regard to their general, expected employment biography. We will argue below that post-industrial class theory holds the adequate conceptual tools to approximate these employment biographies.

A second advantage of a risk-based measure is that a conceptualization based on current labour market status suggests the idea of two relatively homogeneous groups of insiders and outsiders. However, outsiders are a heterogeneous category. We find groups of people with ‘typically atypical’ work biographies both among high- and low-skilled, in different economic sectors, age groups, and so on. Consider these examples of typical outsiders: a woman working part-time in retail, a graphic designer working freelance on fixed-term projects, a recent university graduate who is being repeatedly employed on the basis of 1-year contracts, or an unskilled unemployed worker. All of them are typical outsiders (in particular countries), but they are different in many aspects regarding their social risks and economic opportunities, which may be relevant depending on the research question. In a similar vein, Esping-Andersen (1999a), Kitschelt and Rehm (2006) and Häusermann (2010) show that the ‘B-team’ of post-industrial societies is very heterogeneous, as the category of outsiders contains very different social groups. What these groups share – and what separates them from insiders – is a high risk of experiencing ‘atypical’ employment during the course of their lives. A measure based on occupational profiles allows for differentiating between the heterogeneous group of outsiders and insiders in theoretically and empirically meaningful ways depending on particular research interests.

Nevertheless, and this third advantage is linked to our previous point, one may ask whether it makes sense to conceptualize insiders and outsiders in two groups at all, since these two groups will necessarily have strong within-group heterogeneity. Such a dichotomy only makes sense theoretically if the two groups share a certain degree of social closure, which may structure their political preferences. As a consequence, insiders and outsiders may be mobilized by political actors. Our approach to this is on one side empirical: if we find significant differences between insiders and outsiders in terms of labour market characteristics and political preferences despite the heterogeneity of the two groups, it means that the distinction of insiders and outsiders makes sense.

However, we consider the dichotomous measure generally as a weakness, because it entails a loss of information on different degrees of labour market vulnerability that is analytically problematic. Given that we measure outsiders based on risk rather than on status, we are able to develop a continuous measure of the extent of labour market vulnerability, which we may also call a degree of outsiderness.

Measuring the risk of atypical employment and unemployment

Following the above arguments, we define as labour market outsiders those individuals who incur a particularly high probability of being in atypical employment and/or unemployment. The question is how we can measure this risk. We propose to categorize individuals based on the characteristics of their occupational reference group rather than on mere individual-level characteristics. The probability of experiencing unemployment or atypical employment obviously depends on the frequency – or rate of occurrence – within the relevant occupational category of an individual. We argue that class, gender and age form the relevant categories that relate the individual to a social group sharing similar risks regarding atypical employment. Classes are socio-structural groups characterized by a particular situation in the production process (that is, in the labour market), which shapes their resources, latent interests and preferences.³ Class schemes are based on occupational profiles (Erikson and Goldthorpe, 1993; Oesch, 2006; Wright, 1997), because people in similar professions tend to share permanent, structural commonalities, meaning that classes are characterized by a certain degree of ‘social closure’. Post-industrial societies are still structured in different, relatively stable groups or classes, which share similar employment conditions. Class is therefore a meaningful starting point for the identification of group-specific risks of unemployment and atypical employment. We empirically rely on the class schema by Oesch (2006), which is explicitly developed to reflect post-industrial societies in two regards: (1) it takes into account a heterogeneous middle class (Kriesi, 1998); and (2) it distinguishes between different types of low-skilled employees who can no longer be reasonably subsumed under a single category of (blue-collar or manual) workers (Oesch, 2006: 98ff.).

We follow Kitschelt and Rehm (2005), who argue that the Oesch class schema can be regrouped into five ‘class groups’. The location of the five class groups in the occupational structure is shown in Table 1. Capital accumulators are high-skilled managers, self-employed and experts. Socio-cultural professionals are high-skilled professionals in interpersonal professions, most of them in the public and private service sector. Lower-skilled workers are differentiated into three groups: blue-collar workers are unskilled and skilled workers mostly in industry; low service functionaries are unskilled and skilled employees in interpersonal services; and mixed service functionaries are routine and skilled workers in jobs with mostly organizational work logic. Table 1 presents the location of the five classes in the class schema.

Table 1.

The post-industrial class schema.

Independent work logic	Technical work logic	Organizational work logic	Interpersonal work logic
Large employers, liberal professionals and petty bourgeoisie with employees (e.g. entrepreneurs, lawyers) Capital accumulators (CA)	Technical experts (e.g. executive engineers) Capital accumulators (CA)	Higher-grade and associate managers (e.g. financial and managing executives) Capital accumulators (CA)	Socio-cultural (semi)-professionals (e.g. teachers, health professionals) Socio-cultural professionals (SCP)	Professional/ managerial
	Technicians (e.g. engineers) Mixed service functionaries (MSF)			Associate professional/managerial
Petty bourgeoisie without employees (small shopkeepers)Mixed service functionaries (MSF)	Skilled crafts and routine operatives (e.g. machine operators, labourers in construction)Blue-collar workers (BC)	Skilled and routine office workers (e.g. office clerks)Mixed service functionaries (MSF)	Skilled and unskilled service (e.g. salespersons, waiters)Low service functionaries (LSF)	Generally/vocationally skilledUnskilled

Note: Based on Oesch (2006) and Kitschelt and Rehm (2005); adapted from Häusermann (2010). For the classification of occupations (ISCO-2d codes) see Appendix Table 1.

We use these five classes as a starting point. We also know that post-industrial labour market advantages are structured not only by class but also by gender and age. Much of the relevant literature points out that the insider–outsider divide is clearly gendered (Emmenegger, 2010; Esping-Andersen 1999b: 308; 2009; Häusermann and Schwander, 2012; Kitschelt and Rehm, 2006; Taylor-Gooby, 1991), and that research on dualization must be linked to research on gender segregated labour market (Davisson and Naczyk, 2009:5).Other studies point to the fact that post-industrial labour markets also tend to hold different occupational prospects for younger and older workers, confronting younger workers with more volatile and instable labour markets, while older workers enjoy more job protection (Esping-Andersen, 1999a; Kitschelt and Rehm, 2006; Chauvel, 2009). (Chauvel, 2009; Esping-Andersen, 1999a; Kitschelt and Rehm, 2006).⁴ We therefore disaggregate the post-industrial classes further according to gender and age (except for capital accumulators, which are clearly the most privileged group in the labour market and therefore considered insiders by definition). The combination of four classes, two sexes and two age groups (below/above 40 years old) leaves us with 17 occupational groups, which are the basis of our measurement of unemployment/atypical employment risk. We limit ourselves to only two age groups and draw the line at 40 years old for practical reasons and because most European countries have still a considerable number of young adults in education at the age of 30 years (Couppié and Mansuy, 2003). Considering that acquiring a firm position in the labour market requires another couple of years, a substantial share of people in their 30s must still be counted as labour market entrants.

Once the 17 groups are established, we compare the group-specific rates of unemployment and atypical employment (combined) with the average rate in the workforce. Atypical employment includes involuntary part-time employment, fixed-term employment and helping family members.⁵ These group- and workforce-specific rates can be calculated both for pooled sets of welfare regimes and for countries individually. Previous work on the insider–outsider divide across welfare regimes has shown that variation in the composition of insiders and outsiders exists both between and within welfare regimes (Häusermann and Schwander, 2012). As this is an explorative article suggesting various operationalizations, we do both: the welfare regime-specific operationalization implies that we pool our occupational groups across all countries belonging to a regime and compare rates of atypical employment to the regime average, while the country-specific operationalization implies that we do the same for each country individually.

For the dichotomous measure of insiders and outsiders, we select all groups that have a rate of atypical employment and unemployment that is significantly (p<0.05) higher than the workforce average and we code all individuals in these groups as outsiders. For the continuous measure of outsiderness, we subtract the workforce average rate from the group-specific rate and use the difference as value of labour market vulnerability or ‘degree of outsiderness’ that we then attribute to all individuals in this specific group.

In this article, we apply this operationalization to EU-SILC data from 2007 (complemented by three national household panel surveys for countries missing in the EU-SILC⁶). The level of detail of household panel data and the number of respondents (3500–8250 respondents for each country in the EU-SILC) is unrivalled by other comparative surveys. It thus allows a precise measurement across countries even for those groups that are naturally small (such as old female blue-collar workers). The high number of respondents is also crucial, since we rely on labour market conditions that may affect small portions of the electorate only (unemployment, temporary employment, and so on).

Table 2 presents descriptive information on the distribution of individuals in the different categories and their group-specific deviation (in percentage points) from the workforce average rate of unemployment and atypical employment (which corresponds to their ‘degree of outsiderness’). We have highlighted all groups whose group-specific average significantly exceeds the workforce average (‘outsiders’). We show only the regime-specific operationalization in this table for reasons of space. Table 3 is based on Table 2, showing the share of outsiders among different groups of the workforce, as well as the mean of labour market vulnerability – that is, outsiderness – of these groups.

Table 2.

Map of dualization: difference between the group-specific rate of atypical employment/unemployment and the rate among the entire workforce.

	Liberal regime			Nordic regime			Continental regime			Southern regime		Atypical work/Unempl.
	(Australia, Canada, Britain, Ireland, USA)	N	Atypical work/Unempl.	(Denmark, Finland, Norway, Sweden)	N	Atypical work/Unempl.	(Austria, Belgium, France, Germany, Netherlands, Switzerland)	N	Atypical work/Unempl.	(Greek, Italy, Portugal, Spain)	N	Atypical work/Unempl.
Outsiders	LSF young women^a	5019	36.5	LSF young women^a	2267	30.1	LSF young women^a	3177	30.9	LSF young women^a	3789	26.4
Insiders	LSF young men^a	3370	19.3	LSF young men^a	962	13.7	LSF young men	1439	−5.2	LSF young men^a	2014	3.7
	LSF old women^a	4035	18.8	LSF old women^a	3039	16.3	LSF old women^a	3974	33.9	LSF old women^a	3939	22.6
	LSF old men	2389	−8.0	LSF old men	847	−4.3	LSF old men	1568	−14.0	LSF old men	1867	−17.1
	SCP young women^a	5342	6.9	SCP young women^a	2390	9.4	SCP young women^a	4418	17.5	SCP young women^a	2842	8.1
	SCP young men	2287	−5.1	SCP young men	1285	−5.0	SCP young men	2367	−14.0	SCP young men	1606	−8.4
	SCP old women	5291	0.0	SCP old women^a	4128	2.1	SCP old women^a	5384	17.5	SCP old women	2898	−13.6
	SCP old men	2856	−14.0	SCP old men	2395	−9.8	SCP old men	3940	−21.1	SCP old men	2533	−27.3
	BC young women^a	1074	12.2	BC young women^a	557	9.8	BC young women^a	869	16.9	BC young women^a	1648	20.5
	BC young men	7282	−2.4	BC young men	3291	−7.6	BC young men	4950	−12.6	BC young men	6446	−2.1
	BC old women	1466	4.6	BC old women^a	1005	2.9	BC old women	1222	22.2	BC old women	2602	33.2
	BC old men	7676	−9.9	BC old men	4637	−9.6	BC old men	5720	−17.9	BC old men	7124	−8.4
	MSF young women^a	4491	10.1	MSF young women^a	770	15.2	MSF young women^a	2792	15.8	MSF young women^a	2144	10.2
	MSF young men	2644	1.5	MSF young men	741	−6.3	MSF young men	2187	−17.7	MSF young men	1646	−7.8
	MSF old women	4782	−0.5	MSF old women	1546	6.5	MSF old women^a	3323	20.3	MSF old women	1588	−4.1
	MSF old men	2335	−15.1	MSF old men	1058	−10.7	MSF old men	2665	−23.8	MSF old men	1750	−25.6
	CA	16,903	−18.6	CA	5693	−13.6	CA	8764	−21.1	CA	5004	−23.2
	Total	77,164		Total	36,319		Total	57,256		Total	50,636
	Entire workforce		29.8	Entire workforce		20.3	Entire workforce		38.6	Entire workforce		36.3
	Minimum		−18.6	Minimum		−13.6	Minimum		−23.8	Minimum		−27.3
	Maximum		36.5	Maximum		30.1	Maximum		33.9	Maximum		33.2

Note: Values are the difference between the group-specific rate of atypical employment/unemployment and the rate among the entire workforce (outsiderness). ^aGroups with significantly higher rates than the workforce average (outsiders); based on EU-SILC 2007; data for Canada, USA and Australia are based on country-specific national household panels from 2007.

LSF, low service functionaries; SCP, socio-cultural professionals; BC, blue-collar workers; MSF, mixed service functionaries; CA, capital accumulators (see Table 1); Young, <40 years old; Old, >40 years old.

Table 3.

Labour market vulnerability in different groups of the workforce.

	Liberal regime		Nordic regime		Continental regime		Southern regime
	% of outsiders	Mean of outsiderness	% of outsiders	Mean of outsiderness	% of outsiders	Mean of outsiderness	% of outsiders	Mean of outsiderness
Among total workforce	34.5	1.4	32.6	−1.8	47.0	1.5	38.0	0.4
Among women	59.5	8.4	62.0	7.6	94.3	21.4	73.9	14.3
Among young	54.8	6.6	51.6	3.1	48.0	2.7	53.8	5.2
Among low-skilled individuals	40.1	3.2	37.1	−0.6	50.6	3.8	39.0	1.9
Among high-skilled individuals	31.0	−1.6	24.3	−1.9	43.0	−2.7	34.8	−4.5

Note: ‘% of outsiders’ means the percentage of individuals who belong to a social group with a rate of unemployment and atypical employment that is significantly higher than the workforce average. ‘Mean’ denotes the average deviation (in percentage points) of the group-specific rates from the average workforce rates (see Table 2).

Numbers are based on the regime-specific operationalizations and based on EU-SILC 2007; data for Australia, Canada and USA are based on country-specific national household panels.

Two main insights result from Tables 2 and 3. First, even though there are variations in the composition of insiders and outsiders across regimes, female and young labour market participants experience atypical employment and unemployment more strongly than men and elderly employees in all regimes. In all four regimes, young female low service functionaries are more strongly affected by these forms of labour market vulnerability than any other group. Their rate of atypical employment and unemployment exceeds the regime-specific average rate by 25.2 to 34.2 percentage points. By contrast, high-skilled older men and capital accumulators experience the lowest risk of atypical employment and unemployment. The clear gender bias is particularly strong in the Continental regime: 94.3 percent of women are outsiders when using the dichotomous measure and their mean value of outsiderness lies by more than 20 percentage points above the mean of the workforce. This is more than in the other three regimes, where women’s outsiderness exceeds the average by 7.6 to 14.3 percentage points. Young labour market participants are particularly strongly affected in the Liberal and Southern regimes, where more than half of the young belong to social groups disproportionally affected by atypical employment and unemployment. In all regimes, the mean of outsiderness among the young is higher than the mean of the entire workforce.

The second insight resulting from Tables 2 and 3 is that both insiders and outsiders are heterogeneous in terms of skills and education level. The high-skilled are not shielded from atypical employment. Between 24.3 (in the Nordic regime) and 43 (in the Continental regime) percent of the high-skilled belong to groups experiencing rates of unemployment and atypical employment that lie significantly above the workforce average. Many high-skilled women, for example, work in part-time jobs and many graduates find their way into the labour market through fixed-term contracts only. Even though it may make sense – depending on the research question – to analyse low- and high-skilled labour market vulnerability separately, it is important to notice that atypical work in particular has spread widely into the higher skilled classes, with all the social and political correlates in terms of poor social security coverage and weak political mobilization that this implies (for a discussion of the issue of highly skilled outsiders, see also Davidsson and Nacyk (2009) and Polavieja (2005)).

However, it is true that ‘high-skilled outsiderness’ is particularly a matter of involuntary part-time employment and fixed-term contracts, rather than being driven by unemployment. To illustrate this more clearly, and to differentiate these forms of labour market vulnerability, we calculated the rates of unemployment and the rates of involuntary part-time and temporary work separately. The resulting tables are shown in Appendix Table 2 and Appendix Table 3. Not unexpectedly, the risk of unemployment turns out to be more strongly skill-biased than the risk of atypical employment: if we were to define outsiders solely on the basis of unemployment, they would almost exclusively be found among low service functionaries and blue-collar workers. Nevertheless, age and gender remain relevant in structuring the risk for unemployment. In all regimes, blue-collar workers, as well as female and young male low service functionaries have particularly high rates of unemployment. In contrast to unemployment, atypical employment is generally more gender-and less skill-related, as both high and low-skilled women massively tend to work in involuntary part-time or in temporary work.

Due to space restrictions we cannot present a table with the specific classification of insiders and outsiders for each country in this article.⁷ In general, the regime-specific and the country-specific operationalizations provide very similar results (r=0.86 for the dichotomous measure, r=0.94 for the continuous measure when correlating them at the individual level), but intra-regime variance is stronger in some cases than in others. The Liberal regime is quite homogeneous. The pattern in Australia, Ireland, the US and Britain corresponds to the pattern of the pooled analysis where skills and gender are important in structuring the risk for atypical employment and unemployment. In Canada, the risk for being an outsider is somewhat less gendered than generally in the Liberal regime and more structured by age (contrary to the pooled analysis, young blue-collars and socio-cultural professionals are disproportionally affected by atypical employment and unemployment). In the Nordic regime, we find that in Sweden and Denmark, there are generally fewer workers affected by a disproportionate risk for atypical employment⁸ and unemployment than in Norway and Finland. In Finland, the risk of being an outsider is stronger age-biased than in the pooled analysis, but in all countries, capital accumulators clearly face the lowest rate of outsiderness.

In the Continental regime, the country-specific patterns of outsiders and outsiderness are very consistent with the pooled regime-analysis. In all Continental countries, gender is an extremely strong predictor of atypical work and unemployment. Only France deviates: here, skill level and age are equally important in structuring labour market vulnerability. As a result, and contrary to the other countries, French young male low service functionaries are counted as outsiders, whereas older female socio-cultural professionals are not. Overall, the French pattern has more similarities with the Southern regime than with the other Continental countries.

Finally, the countries of the Southern regime form the most homogeneous pattern. Young women and elderly low-skilled women are counted as outsiders in all countries. Age plays an important role too: in Spain and Italy, low-skilled young men also experience slightly higher labour market vulnerability than the national average (this refers to blue-collar workers in Spain and low service functionaries in Italy). Overall, it is noteworthy that the range of outsiderness between the minimum value of −27.3 (old male socio-cultural professionals) and the maximum value of 33.2 (old female blue-collar workers) is highest in the Southern regime, indicating strong inequality in terms of labour market vulnerability.

Assessing the validity of the risk-based conceptualization of insiders and outsiders

To further assess our risk-based conceptualization of insiders and outsiders, we discuss both the criterion and construct validity of our conceptualization (see Evans, 1992; Oesch, 2006: 94ff.). The criterion validity assesses whether a concept measures what it is indented to measure. It depends on measuring outcomes or characteristics that are supposed to be directly linked to the concept one wants to validate (Evans, 1992: 212). For the risk-based conceptualization of insiders and outsiders this implies that we assess whether the conceptualization is indeed related to actual differences in labour market (dis)advantages. Construct validity, on the other hand, is assessed by testing if a concept predicts other, more distant variables in a theoretically meaningful way (Evans, 1992: 212). The literature on dualization postulates different policy preferences of insiders and outsiders due to their different positions in the labour market (Emmenegger, 2009; Rueda, 2005). Consequently, we will test whether the risk-based conceptualization is able to predict differences in such labour market policies preferences.

Analysis of labour market advantages

In this section, we first assess the criterion validity by discussing two key indicators of labour market advantages: work income and upward mobility (see also Oesch, 2006). We expect outsiders to fare worse on both indicators of labour market advantages. Datawise, we rely on the ISSP Work Orientation III survey (2005), which includes 15 countries. We use the four insider/outsider measures developed in this article: the dichotomous and continuous measures based on the regime-specific rates of atypical employment – outsider/outsiderness (regime) – as well as both measures based on country-specific rates of atypical employment – outsider/outsiderness (country). To control for country-specific differences, we include country dummies in all models and use clustered standard errors. For both dependent variables we specify four models: the first model shows the estimate for the regime-specific, dichotomous measure. Model 2 refers to the country-specific, dichotomous measure. Model 3 uses the regime-specific, continuous measure, and model 4 shows the estimate for the country-specific, continuous measure. All models in Table 4 control for age, gender, education, union membership, church attendance, if an individual lives in a couple household and (for upward mobility) income, following the literature in this field (Burgoon and Dekker, 2010; Emmenegger, 2009). Details regarding the operationalization are documented in Appendix Table 4.

Table 4.

Determinants of labour market advantages: income and promotion chances.

	Income				Chances for promotion in current job
	M1	M2	M3	M4	M1	M2	M3	M4
Outsider (Regime)	−0.277**				−0.372***
	(0.093)				(0.079)
Outsider (Country)		−0.242**				−0.184*
		(0.088)				(0.095)
Outsiderness (Regime)			−0.013***				−0.016***
			(0.004)				(0.004)
Outsiderness (Country)				−0.011**				−0.011***
				(0.005)				(0.004)
Female	−0.226**	−0.247***	−0.169*	−0.152*	0.015	−0.091	0.050	−0.017
	(0.087)	(0.080)	(0.091)	(0.074)	(0.059)	(0.078)	(0.077)	(0.092)
Age	0.004**	0.005***	−0.001	−0.000	−0.035***	−0.032***	−0.034***	−0.032***
	(0.001)	(0.002)	(0.002)	(0.002)	(0.003)	(0.003)	(0.003)	(0.003)
Education	0.292***	0.298***	0.279***	0.230**	0.262***	0.281***	0.255***	0.273***
	(0.090)	(0.092)	(0.083)	(0.086)	(0.060)	(0.067)	(0.061)	(0.055)
Union membership	−0.120	−0.113	−0.027	0.124	−0.050	−0.042	−0.067	−0.070
	(0.119)	(0.123)	(0.122)	(0.075)	(0.075)	(0.073)	(0.077)	(0.076)
Church attendance	−0.016**	−0.016**	−0.019**	−0.012	0.024*	0.024*	0.026*	0.027*
	(0.007)	(0.007)	(0.008)	(0.007)	(0.014)	(0.014)	(0.015)	(0.015)
Living in a couple household	0.163*** (0.048)	0.158*** (0.047)	0.141*** (0.043)	0.126** (0.048)	0.042(0.061)	0.034(0.060)	0.053(0.060)	0.054(0.059)
Income					0.260***	0.266***	0.232***	0.240***
					(0.015)	(0.014)	(0.012)	(0.013)
Country fixed effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Pseudo R²	0.740	0.740	0.742	0.782	0.035	0.033	0.035	0.034
N	8577	8577	10010	9149	6983	6983	6505	6505

Notes: Values in parentheses are standard errors; OLS regression for income, ordered logistic regression for promotion chances, regressions with clustered standard errors and country dummies; country dummies not shown due to space restriction.

Pseudo R² is the McKley and Zavoina R².

Significant at the 0.1 level, **significant at the 0.05 level, ***significant at the 0.01 level.

Data source: ISSP Work Orientations III (2005). For details on operationalization, see Appendix Table 4.

The effect of outsiderness on income is clear-cut and as expected: being an outsider has a consistent, strong and negative effect on income regardless of the measure we use. The negative effect of being an outsider (net of all control variables such as gender, age, education, and so on) lowers the respondent’s income by 277 Euros when we use the regime-specific, dichotomous operationalization, and by 242 Euros using the country-specific, dichotomous operationalization. Regarding the effects of the continuous measures, we calculate substantive effects by comparing the estimated income of individuals with the maximum outsiderness value to individuals with the minimum outsiderness value while holding all other variables constant: this difference is 850 Euros for the regime-specific operationalization and 880 Euros for the country-specific operationalization.

The control variables show no surprising effects: gender and church attendance are negatively related to income, while age, high education and living in a couple household have a positive association with income. Union membership does not affect income significantly.

The disadvantaged position of outsiders in the labour market also shows in the subjective assessment of upward mobility. Outsiders are consistently and significantly less likely than insiders to agree to the statement that chances for advancement in their current job are high. In order to discuss the effects substantively, we compare the predicted probabilities of agreeing to that statement for insiders and outsiders. We calculate predicted probabilities for an individual with average income and age, holding the other variables at zero, that is, a 42-year-old, not-religious, low-skilled male outsider, who lives alone, is not a union member, earns 1878 Euros and lives in France. His probability of saying that he definitely expects to be promoted is 9.1 percentage points lower compared with being an insider (regime-based, dichotomous operationalization). The difference is 4.3 percentage points when we use the country-specific, dichotomous measure. To estimate the substantive effects of the continuous measure, we compare the probabilities of the same individual, attributing to him the highest and lowest value of outsiderness (26.4 and −16.5 in France): the net effect of outsiderness amounts to 15.6 percentage points. If we use the regime-specific measure, an individual with the highest value (31.9) has a likelihood of 15.7 percent to say that he expects definitively to be promoted, while the likelihood for the same individual with the lowest labour market vulnerability (–18.5) is 33.6 percent. This is a difference of 17.9 percentage points. Hence, being an outsider and labour market vulnerability more generally have substantial net effects on promotion prospects, which are independent from other determinants such as age, gender and education.

Turning to the control variables, we see that high-skilled employees and high-income earners are more optimistic about their career advancement prospects while elder employees assess their career chances less optimistically. The other control variables do not display significant effects, with the exception of church attendance being slightly positively related with the subjective assessment of upward mobility.⁹

We conclude from this section that the risk-based measure of outsiders is valid, insofar as we find significant differences in labour market advantages between insiders and outsiders, even for the dichotomous measures, despite the heterogeneity of these two groups. Literature on social closure of post-industrial classes (Oesch, 2006) and reproduction of poverty spells over generations (Tomlinson and Walker, 2012) shows that these differences are persistent over time. This longitudinal analysis, however, lies beyond the scope of this article.

Analysis of preferences for labour market policies

We now move to the discussion of the construct validity. The literature on dualization suggests that insiders and outsiders hold different preferences regarding labour market policies. We replicate these analyses of the existing insider/outsider literature using our measures. For this part of the analysis, we rely on ISSP Role of Government IV (2006) survey data, which includes the same 15 countries as before plus the Netherlands. To predict labour market policy preferences of insiders and outsiders we use ordered logit regressions. Our main focus in this analysis is on the direction and consistency of effects, rather than the magnitude and substantive differences, which are notoriously small in all micro-level analyses (see Emmenegger, 2009; Rueda, 2005).

The existing literature (most clearly so Rueda, 2005) has evidenced insider–outsider preference divides with regard to active labour market policies. As an indicator of preferences for active labour market policies (ALMP), we use a question asking respondents whether they agree that the government is responsible for providing a job for everyone who wants one. We operationalize preferences for passive labour market policies (PLMP) with a question asking if the respondent agrees that the government should spend more on unemployment benefits.¹⁰

Table 5 shows the estimates of preferences for active and passive labour market policies. For each dependent variable, models 1–4 test the influence of labour market vulnerability with regard to the four different risk-based measures of outsider and outsiderness. The models include the same set of control variables as before.

Table 5.

Determinants of labour market preferences.

	Active labour market policies				Passive labour market policies

	Government should provide a job for everyone				Government should spend more on unemployment benefits
	M1	M2	M3	M4	M1	M2	M3	M4
Outsider (Country)	0.014				0.077
	(0.073)				(0.087)
Outsider (Regime)		0.208***				0.213**
		(0.073)				(0.107)
Outsiderness (Country)			0.004*				0.006*
			(0.002)				(0.003)
Outsiderness (Regime)				0.009***				0.010***
				(0.002)				(0.003)
Female	0.152**	0.006	0.071	−0.021	0.027	−0.088	−0.052	−0.123
	(0.060)	(0.076)	(0.070)	(0.058)	(0.101)	(0.097)	(0.110)	(0.105)
Age	−0.003	−0.002	−0.002	−0.000	0.022***	0.022***	0.023***	0.024***
	(0.003)	(0.003)	(0.003)	(0.003)	(0.003)	(0.003)	(0.003)	(0.003)
Income	−0.258***	−0.252***	−0.251***	−0.242***	−0.214***	−0.209***	−0.205***	−0.198***
	(0.031)	(0.031)	(0.032)	(0.030)	(0.037)	(0.038)	(0.038)	(0.038)
Education	−0.508***	−0.494***	−0.490***	−0.470***	−0.191*	−0.179*	−0.170*	−0.153
	(0.065)	(0.064)	(0.066)	(0.064)	(0.099)	(0.098)	(0.100)	(0.104)
Union membership	0.280***	0.278***	0.280***	0.277***	0.367***	0.366***	0.368***	0.365***
	(0.070)	(0.069)	(0.070)	(0.070)	(0.107)	(0.108)	(0.107)	(0.107)
Church attendance	0.013	0.012	0.012	0.012	−0.019	−0.019	−0.019	−0.019
	(0.012)	(0.012)	(0.012)	(0.012)	(0.012)	(0.012)	(0.012)	(0.012)
Living in couple household	−0.114***	−0.113***	−0.113***	−0.110***	−0.234***	−0.234***	−0.233***	−0.232***
	(0.033)	(0.032)	(0.033)	(0.033)	(0.059)	(0.059)	(0.059)	(0.059)
Country fixed effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Pseudo R²	0.059	0.059	0.059	0.060	0.077	0.077	0.077	0.078
N	8401	8401	8401	8401	8453	8453	8453	8453

Notes: Values in parentheses are standard errors, ordered logistic regression with clustered standard errors and country dummies; country effects not shown due to space restriction.

Pseudo R² is the McKley and Zavoina R².

Significant at the 0.1 level, **significant at the 0.05 level, ***significant at the 0.01 level.

Data source: ISSP RoG IV 2006. For details on operationalization, see Appendix Table 4.

We start by discussing the estimates for active labour market policy preferences. Models 2–4 show that outsiders clearly and significantly want more active job creation than insiders – which is in line with the expectations in the literature. The effect of the country-specific outsider dummy variable points to the right direction, but does not reach significance. This might be due to the fact that the comparison with country-specific averages (the overall variance being lower than in the regime-comparison) leads to a dichotomous measure with too heterogeneous groups, an observation that adds value to the continuous measure. In terms of substantive effects, we calculate probabilities for the same individual as specified before:¹¹ being an outsider increases his probability of agreeing that the government is (definitely or probably) responsible for providing a job for everyone by about 6 percentage points when we use the regime-specific, dichotomous operationalization. Using the country-specific continuous measure, going from the highest to the lowest value of outsiderness has a net effect of about 15 percentage points (from 78.5 to 63.4 percent) on the likelihood that a respondent agrees that the government should provide a job for everyone. The same effect is even 18.4 percentage points when using the regime-specific continuous measure.¹² Turning to the control variables, income, education and whether a person lives in a couple household influence preferences for active labour market policies negatively, while union members are more likely to be in favour of active labour market policies than non-union members. The frequency of church attendance exerts no influence on the preferences for active labour market policies and gender has no consistent effect either.

For passive labour market policy preferences we find very similar results. As expected, outsiders have stronger preferences for passive labour market policies, that is, they tend to agree more often to the statement that the government should spend more on unemployment benefits. Again, the coefficient for the country-specific dummy variable does not reach significance, but the effect goes in the expected direction. In terms of substantive effects, the net difference between insiders and outsiders is 4.2 percentage points for the regime-specific, dichotomous measure.¹³ Regarding the continuous measure of outsiderness, we find that net of all other determinants, an individual with the highest value of labour market vulnerability is about 13.1 (regime-specific measure) or 14.4 (country-specific measure) percentage points more likely to favour more generous unemployment benefits than an individual with the lowest score of labour market vulnerability. The effects of the control variables are similar to the effects we found for active labour market policy preferences. Income, education and if a person lives in a couple household have a significant negative influence on passive labour market policy preferences, while union members and elder employees are more in favour of passive labour market policies than non-members and younger individuals. Religiosity, again, exerts no influence on the preferences.¹⁴

Overall, we find evidence for the construct validity of our risk-based measures of outsiders and outsiderness with regard to preferences for active and passive labour market policies: outsiders are consistently more likely to agree that the government should provide a job for everyone, and that the government should spend more on unemployment benefits. Overall, we argue that the empirical analysis supports the new measure that we have theoretically argued and empirically developed in this article. Indeed, preferences for active labour market policy are the core measure on which insiders and outsiders are expected to differ. For this variable, we clearly find evidence for the risk-based measures. When it comes to passive labour market policies, expectations in the literature are somewhat less clear, since income-related passive transfers benefit insiders too, given that insiders are likely to have much better contribution records. However, insiders are also – by definition and by measurement – less likely to become unemployed at all, so that we still expect outsiders to be more favourable towards generous unemployment benefits, a hypothesis that is confirmed by the data.

Conclusion

This article proposes a new conceptualization and measurement of labour market insiders and outsiders, and of labour market vulnerability more generally. We have argued that if we want to explain political preferences of insiders and outsiders, it may be useful to operationalize insiders and outsiders on the basis of stable social and economic characteristics, which are likely to shape their (current and future) life chances and constraints. We suggest a conceptualization of insiders and outsiders that is based on the risk of individuals to find themselves in atypical employment or unemployment, this risk being measured by the specific rates of unemployment/atypical employment of the social group to which they belong. On this basis, we propose both a dichotomous operationalization of insiders/outsiders and a continuous measure of ‘outsiderness’, or more generally labour market vulnerability. The dichotomous measure codes individuals as outsiders if they belong to a social group whose rate of unemployment and atypical employment is significantly higher than the workforce average, all other individuals being coded as insiders. The continuous measure departs from the dichotomous view of insiders and outsiders that prevails in the dualization literature, but allows a more fine-grained measurement of labour market vulnerability by attributing each individual the difference between his/her group-specific rate of unemployment/atypical employment and the average rate in the workforce as a value of outsiderness. With these measures, we developed a map of dualization across four welfare regimes and individual countries, which shows that, overall, low-skilled service sector employees, women and young labour market participants tend to be most strongly affected by labour market vulnerability, with a few regime- and country-specific differences.

We also assessed the criterion and construct validity of our conceptualization: outsiderness (as measured in terms of the risk of unemployment/atypical employment) clearly contributes to a disadvantage in terms of income and job mobility, net of other factors such as gender, age or education. We also showed that our measures predict active and passive labour market policy preferences of insiders and outsiders as expected in the literature of dualization and insider–outsider divides. Based on these findings, we argue that the definition of outsiders and outsiderness as developed in this article is empirically and theoretically relevant for the research on dualization, and that it might be of interest to other studies in this area. However, we are also aware that our measures are more complex and require more fine-grained data (especially in terms of occupational classes) than the standard operationalization, which simply codes outsiders based on their current labour market status. Hence, one may ask whether it is worthwhile choosing the more complex path. We would argue that it is, given its added theoretical value with regard to specific research questions (notably if we are interested in analysing not the immediate labour market disadvantages of outsiders, but the political consequences of dualization) and its added degree of differentiation regarding the continuous measure of outsiderness and labour market vulnerability. In addition, the map of dualization we presented in Table 2 can be used in subsequent analyses very easily, and country-specific values are available from the authors.

More generally, we think that this article bears insights for the literature on dualization and insider–outsider divides irrespective of the precise measurement we use. The distinction between a core workforce of insiders, which is fully integrated in the labour market, and a more marginal and vulnerable part of the workforce – the outsiders – is empirically validated, and we were able to identify a pattern of socio-structural groups of insiders and outsider who – despite the heterogeneity of these two groups – face distinct labour market disadvantages and hold distinct preferences. We also think that our continuous measure of outsiderness allows linking the more political science-oriented dualization literature with the more sociological literature on labour market vulnerability. Finally, we were able to show cross-country and cross-regime differences in the degree and structure of labour markets vulnerabilities and dualization. This raises many questions for future research: where and to what extent will an insider–outsider divide be mobilized politically? Do the political preferences of insiders and outsiders differ only regarding labour market policies, or do they hold different preferences regarding the welfare state in general? Moreover, to what extent does labour market vulnerability affect vote choices and electoral preferences for political parties? These are questions that current research on dualization and insider–outsider divides needs to address, in order to spell out the social and political consequences of changing labour markets.

Footnotes

Appendix

Appendix Table 4.

Table of operationalization.

Variable	Operationalization
Income	ISSP WO III 2005; monthly mean income, individuals are attributed the mean value of their income group (mostly deciles) in 1000 Euros.
Promotion chances in current job	ISSP WO III 2005; opportunities for advancement are high; recoded V31; 1=strongly disagree, 2=disagree, 3=neither/nor, 4=agree, 5=strongly agree
Preferences for active labour market policies	ISSP 2006 RoG IV; government is responsible for providing a job for everyone who wants one; recoded V25; 1=definitely should not be, 2=probably should not be, 3=probably should be, 4=definitely should be
Preferences for passive labour market policies	ISSP 2006 RoG IV; government should spend money on unemployment benefits; recoded V23; 1=spend much less, 2=spend less, 3=spend the same as now, 4=spend more, 5=spend much more
Outsider (regime and country)	Dummy variable, based on a comparison of group-specific rates of atypical employment/unemployment and the regime (country)-specific average rate. EU-SILC 2007 For the USA: American Time Use Survey (ATUS) For Canada: Survey of Labour and Income Dynamics (SLID) For Australia: The Household, Income and Labour Dynamics in Australia (HILDA)
Outsiderness (regime and country)	Continuous variable, difference between group-specific rates of atypical employment/unemployment and the regime (country)-specific average rate. EU-SILC 2007 For the USA: American Time Use Survey (ATUS) For Canada: Survey of Labour and Income Dynamics (SLID) For Australia: The Household, Income and Labour Dynamics in Australia (HILDA)
Regimes	Liberal countries: Australia, Canada, Ireland, Great Britain, United States Nordic countries: Denmark, Finland, Norway, Sweden Continental countries: Austria, Belgium, France, Germany, Netherlands Switzerland Southern countries: Italy, Greece, Portugal, Spain (Data for Austria, Belgium, Italy and Greece lacking in Tables 3 and 4, data for the Netherlands lacking in Table 3)
Classes	ISCO-2d codes, recoded into CA, MSF, BC, SCP, LSF; see Appendix Table 1 EU-SILC 2007: pl050 HILDA (AU): gjbm682 SLID (CA): nocj2e6, nocg, manag1 ATUS (USA): peio1ocd, prdtocc1
Unemployment	EU-SILC 2007; dummy variable measuring unemployment, recoded from EU-SILC: pl030 HILDA (AU): gesdtl SLID (CA): altstat28 ATUS (USA): pemlr
Involuntary part-time	EU-SILC 2007; dummy variable measuring involuntary part-time work, recoded from pl030 (self-classification of respondents) and pl120 (reason for part-time work) HILDA (AU); gesdtl (self-classification of respondents) and gjbptrea (reason for part-time work) SLID (CA); scsum28 (self-classification of respondents) and reawpt1 (reason for part-time work) ATUS (USA); prwkstat (self-classification of respondents) and pehrwant (reason for part-time work)
Fix-term contract	EU-SILC 2007; dummy variable measuring fix-term contract work, recoded from pl140 HILDA (AU); gjbmcnt SLID (CA): prnjb1 ATUS (USA); CPS-Supplement (2005) ‘Contingent and Alternative Employment Arrangements’
Atypical work/Unemployment	EU-SILC 2007, HILDA (AU), SLID (CA), ATUS (USA) Dummy variable measuring atypical employment (involuntary part-time, fix-term work, helping family member) and unemployment among all other employment status;
Church attendance	ISSP RoG IV 2006, ISSP WO III 2005; recoded from ATTEND (how often do you go to church), 8=several times a week, 7=once a week, 6=2 or 3 times a month, 5=once a month, 4=several times a year, 3=once a year, 2=less frequently, 1=never
Living in a couple household	ISSP RoG IV 2006, ISSP WO III 2005; dummy variable measuring if respondent lives in a couple household (MARITAL and COHAB), 1=living in a stable couple (married or not), 0=divorced, widowed, single, separated
Education	ISSP RoG IV 2006, ISSP WO III 2005; dummy variable based on highest completed degree (DEGREE), 1=completed higher secondary education, 0=below higher secondary education

Acknowledgements

Earlier versions of this article have been presented at the Annual Meeting 2009 of the American Political Science Association in Toronto, CA and at the 17th International Conference of Europeanists 2010 in Montreal, Canada. We would like to thank David Rueda, Margarita Estévez-Abe and John D. Stephens and the participants of these conferences for helpful comments.

Notes

References

Booth

A.L.

Francesconi

Frank

(2000) ‘Temporary Jobs: Who Gets Them, What Are They Worth and Do They Lead Anywhere’, ISER-Working Paper 205. Colchester: Institute for Social and Economic Research, University of Essex.

Booth

A.L.

Francesconi

Frank

(2002) ‘Temporary Jobs: Stepping Stones or Dead Ends?’, Economic Journal 112: 189–213.

Burgoon

Dekker

(2010) ‘Flexible Employment, Economic Insecurity and Social Policy Preferences in Europe’, Journal of European Social Policy 20: 126–41.

Chauvel

(2009) ‘Comparing Welfare Regime Changes: Living Standards and the Unequal Life Chances of Different Birth Cohorts’, in Jones

I.R.

Higgs

Ekerdt

D.J.

(eds) Consumption and Generational Change: The Rise of Consumer Lifestyles. New Brunswick: Transaction Publishers.

Couppié

Mansuy

(2003) ‘Young People and New Entrants in European Labour Markets: The Timing of Gradual Integration’, in Mueller

Gangl

(eds) Transitions from Education to Work in Euro The Integration of Youth into EU Labour Markets. Oxford: Oxford University Press.

Davidsson

Naczyk

(2009) ‘The Ins and Outs of Dualisation: A Literature Review’, RECWOWE Working Paper 2/2009. Edinburgh: Reconciling Work and Welfare in Europe (RECWOWE), University of Edinburgh.

Emmenegger

(2009) ‘Barriers to Entry, Insider/Outsider Politics and the Political Determinants of Job Security Regulations’, Journal of European Social Policy 19: 131–46.

Emmenegger

(2010) ‘Gendering Insiders and Outsiders, Labour Market Status and Preferences for Job Security’, Economics, Management, and Financial Markets 5: 88–128.

Emmenegger

Careja

(2012) ‘From Dilemma to Dualisation: Social and Migration Policies in the “Reluctant Countries of Immigration”’, in Emmenegger

Häusermann

Palier

Seeleib-Kaiser

(eds) The Age of Dualisation: The Changing Face of Inequality in Europe. New York: Oxford University Press.

10.

Emmenegger

Häusermann

Palier

Seeleib-Kaiser

(2012) The Age of Dualisation: The Changing Face of Inequality in Europe. New York: Oxford University Press.

11.

Erikson

Goldthorpe

J.H.

(1993) The Constant Flux. Oxford and New York: Oxford University Press.

12.

Esping-Andersen

(1990) The Three Worlds of Welfare Capitalism. Princeton, NJ: Princeton University Press.

13.

Esping-Andersen

(1999a) Social Foundations of Postindustrial Economies. Oxford and New York: Oxford University Press.

14.

Esping-Andersen

(1999b) ‘Politics without Class? Post-industrial Cleavages in Europe and America’, in Kitschelt

Lange

Marks

Stephens

J.D.

(eds) Continuity and Change in Contemporary Capitalism (pp. 293–316). Cambridge: Cambridge University Press.

15.

Esping-Andersen

(2009) The Incomplete Revolution: Adapting to Women’s New Roles. Cambridge and Malden: Polity Press.

16.

Evans

(1992) ‘Testing the Validity of the Goldthorpe Class Scheme’, European Sociological Review 8: 211–32.

17.

Gash

(2008) ‘Bridge or Trap? Temporary Workers’ Transitions to Unemployment and to the Standard Employment Contract’, European Sociological Review 25: 651–68.

18.

Goldthorpe

J.H.

Llewellyn

Payne

(1987) Social Mobility and Class Structure in Modern Britain. Oxford: Clarendon Press.

19.

Häusermann

(2010) The Politics of Welfare State Reform in Continental Europe, Modernization in Hard Times. Cambridge and New York: Cambridge University Press.

20.

Häusermann

Schwander

(2011) ‘Who are the Outsiders and What Do They Want? Explaining Welfare Preferences in Dualized Societies’, Les Cahiers européens de Sciences Po, 1. Paris: Centre d’études européennes at Sciences Po.

21.

Häusermann

Schwander

(2012) ‘Varieties of Dualisation? Labour Market Segmentation and Insider–Outsider Divides across Regimes’, in Emmenegger

Häusermann

Palier

Seeleib-Kaiser

(eds) The Age of Dualisation: The Changing Face of Inequality in Europe. New York: Oxford University Press.

22.

Häusermann

Walter

(2010) ‘Restructuring Swiss Welfare Politics, Post-industrial Labour Markets, Globalisation and Attitudes Towards Social Policies’, in Hug

Kriesi

(eds) Value Change in Switzerland (pp. 143–68). Lanham: Lexington Press.

23.

Iversen

Soskice

(2001) ‘An Asset Theory of Social Policy Preferences’, American Political Science Review 95: 875–93.

24.

Kalleberg

Rasell

Cassirer

Reskin

Hudson

Webster

D.L.

Appelbaum

Spalter-Roth

R.M.

(1997) Nonstandard Work, Substandard Jobs: Flexible Work Arrangements in the U.S. Washington, DC: Economic Policy Institute.

25.

Kalleberg

Reskin

Hudson

(2000) ‘Nonstandard Employment Relations: Part-time, Temporary and Contract Work’, Annual Review of Sociology 26: 341–65.

26.

King

Rueda

(2008) ‘Cheap Labour, The New Politics of “Bread and Roses” in Industrial Democracies’, Perspectives on Politics 6: 279–97.

27.

Kitschelt

Rehm

(2005) ‘Work, Family and Politics: Foundations of Electoral Partisan Alignments in Postindustrial Democracies’, paper presented at the Annual Meeting of the American Political Science Association.

28.

Kitschelt

Rehm

(2006) ‘New Social Risk and Political Preferences’, in Armingeon

Bonoli

(eds) The Politics of Post-Industrial Welfare States: Adapting Post-war Policies to New Social Risks. London: Routledge.

29.

Kriesi

(1998) ‘The Transformation of Cleavage Politics: The 1997 Stein Rokkan Lecture’, European Journal of Political Research 33: 165–88.

30.

Lindbeck

Snower

D. J.

(1998) ‘Cooperation, Harassment, and Involuntary Unemployment: an Insider-Outsider Approach’, The American Economic Review 78: 167–188.

31.

Lindbeck

Snower

D.J.

(2001) ‘Insiders versus Outsiders’, Journal of Economic Perspectives 15: 165–88.

32.

Lindvall

Rueda

(2012) ‘Insider–Outsider Politics, Party Strategies and Political Behaviour in Sweden’, in Emmenegger

Häusermann

Palier

Seeleib-Kaiser

(eds) The Age of Dualisation: The Changing Face of Inequality in Europe. New York: Oxford University Press.

33.

Lister

(2004) ‘The Third Way’s Social Investment State’, in Lewis

Surender

(eds) Welfare State Change: Towards a Third Way? (pp. 157–81). Oxford and New York: Oxford University Press.

34.

Mayer

K.U.

(2000) ‘Promises Fulfilled? A Review of 20 Years of Life Course Research’, European Journal of Sociology 41: 259–82.

35.

Oesch

(2006) Redrawing the Class Map: Stratification and Institutions in Germany, Britain, Sweden and Switzerland. London: Palgrave Macmillan.

36.

Organisation for Economic Co-operation and Development (2006) OECD Employment Outlook. Paris: OECD.

37.

Organisation for Economic Co-operation and Development (2008) Growing Unequal? Income Distribution and Poverty in OECD Countries. Paris: OECD.

38.

Organisation for Economic Co-operation and Development (2010) OECD Employment Outlook. Paris: OECD.

39.

Palier

(2010) A Long Good-Bye to Bismarck? The Politics of Welfare Reform in Continental Europe. Amsterdam: Amsterdam University Press.

40.

Palier

Thelen

K. A.

(2010) ‘Institutionalizing Dualism: Complementarities and Change in France and Germany’, Politics & Societies 38(1):119–48.

41.

Plougmann

(2003) ‘Internationalisation and the Labour Market in the European Union’, in Andersen

G.J.

Jensen

P.H.

(eds) Changing Labour Markets: Welfare Policies and Citizenship (pp. 15–38). Bristol: The Policy Press.

42.

Polavieja

J.G.

(2005) ‘Flexibility or Polarization? Temporary Employment and Job Tasks in Spain’, Socio-economic Review 3: 233–58.

43.

Ranci

(2010) Social Vulnerability in Europe: The New Configuration of Social Risks. Basingstoke: Palgrave Macmillan.

44.

Rehm

(2009a) ‘Risk and Redistribution: An Individual-Level Analysis’, Comparative Political Studies 42: 855–81.

45.

Rehm

(2009b) ‘Risk Inequality and the Polarized American Electorate’, British Journal of Political Science 41: 363–87.

46.

Rehm

Hacker

J.S.

Schlesinger

(2012) ‘Insecure Alliances: Risk, Inequality, and Support for the Welfare State’, American Political Science Review 106: 386–406.

47.

Rueda

(2005) ‘Insider–Outsider Politics in Industrialized Democracies: The Challenge to Social Democratic Parties’, American Political Science Review 99: 61–74.

48.

Rueda

(2006) ‘Social Democracy and Active Labour-Market Policies, Insiders, Outsiders and the Politics of Employment Protection’, British Journal of Political Science 36: 385–406.

49.

Rueda

(2007) Social Democracy Inside Out: Partisanship and Labour Market Policy in Advanced Industrial Democracies. Oxford and New York: Oxford University Press.

50.

Saint-Paul

(1998) ‘A Framework for Analyzing the Political Support for Active Labour Market Policy’, Journal of Public Economics 67: 151–65.

51.

Saint-Paul

(2002) ‘The Political Economy of Employment Protection’, Journal of Political Economy 110: 672–704.

52.

Standing

(1993) ‘Labor Regulation in an Era of Fragmented Flexibility’, in Buechtemann

C. F.

(ed) Employment Security and Labor Market Behavior. Ithaca, NY: ILR Press. pp. 425–41.

53.

Taylor-Gooby

(1991) ‘Welfare State Regimes and Welfare Citizenship’, Journal of European Social Policy 1: 93–105.

54.

Tomlinson

Walker

(2012) ‘Labour Market Disadvantage and the Experience of Recurrent Poverty’, in Emmenegger

Häusermann

Palier

Seeleib-Kaiser

(eds) The Age of Dualisation: The Changing Face of Inequality in Europe. New York: Oxford University Press.

55.

Wright

E.O.

(1997) Class Counts. Cambridge and New York: Cambridge University Press.