The social configuration of labour market divides: An analysis of Germany,Belgium and Italy

Abstract

We analyse insecurity-based dividing lines and their social configurations in the German, Belgian and Italian labour markets in 2015, using latent class analysis applied to EU Labour Force Survey data. In contrast to the dual vision of ‘insider-outsider’ approaches, our findings illustrate the existence of five distinctive labour market groups or segments across countries with similar social configurations. We explain this through the social embeddedness of national regulatory systems which generate different degrees of inclusiveness for different groups of workers. This adds to ongoing debates on connecting micro- and macro-levels of analysis, as labour market segmentation as a macro-phenomenon is studied based on its micro-foundations (terms and conditions of employment relationships). We use the interlinkages between national regulatory systems and social categories to explain the findings.

Keywords

Belgium employment Germany insecurity Italy latent class analysis segmentation social divides

Introduction

Although segmentation and inequality are persistent problems in European labour markets, we lack knowledge about where labour market divides are located and which groups of workers are on which side of the divide(s). Sociological studies of inequality (Tilly, 1998) propose focusing on workplaces as the locus where employment conditions are organized, and consequently, dividing lines are generated. Recent employment research calls for connecting macro-level analyses of institutional change and micro-level analyses of the dynamics underpinning these changes as resulting from the social relations at work (Brandl and Lehr, 2018; Pulignano and Doerflinger, 2018a). Studies within the tradition of institutional segmentation theory investigate divides as embedded in country-specific regulatory systems, affecting the form that segmentation takes (Beynon et al., 2002; Osterman, 1994; Rubery, 1978). In accordance, segmentation is framed within the capital–labour rather than the worker–worker relationship, as divides result from the diversity in the extent to which employment conditions – embedded in national regulatory systems – are regulated around decent standards. Although regulatory systems should ensure decent working conditions, inequality could result from inclusive or exclusive models of eligibility (Grimshaw et al., 2018). For instance, countries’ social protection systems may discriminate workers by employment contract or continuity (Grimshaw et al., 2018). Hence, interlinkages between employment and social protection systems may feature varying degrees of inclusiveness. In this vein, inclusive regulatory systems are not sufficient, as regulation promoting encompassing employment practices is equally important (Rubery et al., 2016).

We build on institutional segmentation theory (Osterman, 1994; Rubery, 1978) which acknowledges the interlinkages between employment and social protection systems, tied to labour market regulation and the degree of inclusiveness of employment practices, when examining labour market divides. We define these divides as segmentation lines between secure and insecure workers and consider them as grounded in the terms and conditions of workers’ employment relationships embedded in national regulatory systems. Therefore, we build on existing research claiming that insecurity causes major dividing lines in contemporary labour markets (Eichhorst et al., 2016; Rubery and Piasna, 2016); yet, it is not clear where these dividing lines are located and whether they are identical across countries. Therefore, we aim to uncover the locations of dividing lines and examine whether (and why) they are (not) similar across countries.

We use latent class (LC) analysis based on the EU Labour Force Survey (EU-LFS) to uncover and analyse divides between different groups of workers in Germany, Belgium and Italy in 2015. We expect different kinds of labour market divides in the countries investigated because of specificities in national regulatory systems. Our country choice thus follows the logic of contextualized comparisons (Locke and Thelen, 1995): we compare countries that are generally considered similar, aiming to unveil differences between them. We also explore and illustrate which social categories can be found in the identified segments within and across countries. We call these segmentation lines ‘social divides’. By positioning the study of segmentation within national regulatory systems which intersect with social categories in labour markets, we provide a nuanced analysis of labour market segmentation.

We demonstrate that the regulatory systems in the countries investigated are not politically or ‘socially’ neutral but contested, as they reflect law, agreements, policy and actors’ behaviours (Granovetter, 1992), affecting the extent to which inclusiveness is guaranteed for different groups of workers. In so doing, we add to ongoing debates on connecting micro- and macro-levels of analysis, by incorporating the micro-foundations (terms and conditions of employment relationships) of labour market segmentation as macro-phenomena and using the interlinkages between national regulatory settings and social categories as explanatory variables.

We next present our approach towards ‘social divides’ by explaining its analytical relevance, core assumptions and implications. Having explained aspects of methods, data and operationalization, we then explain the main results of our LC analysis, highlighting the social divides. These findings are explained in the light of national regulatory systems. Finally, our conclusion stresses implications for further research.

An approach to ‘social divides’ in labour markets

Early theories of labour market segmentation developed along institutional and Marxian trajectories. Institutional theories consider labour markets as divided into primary and secondary segments (Doeringer and Piore, 1971), reflecting differences between social categories (so that women or migrants frequently work in secondary segment jobs). In contrast, Marxian theories view workforce divisions as reinforced by capitalist strategies intended to disrupt the formation of class solidarity and possible resistance; the persistence of a casual workforce reflects considerations which are not directly related to the productivity potential of different labour market segments (Rubery, 1978). Despite differences between institutional and Marxian segmentation theories, both emphasize that labour markets should be analysed by studying the characteristics of jobs in the context in which people are employed. Thus, the workplace is the level at which a general analytical framework for understanding the structure and functioning of the labour market develops (Rubery, 2007). In contrast to neoclassical arguments that individual productivity differences explain the existence of secondary segment workers, institutional segmentation theory highlights the role of workplace practices as embedded in national regulatory systems in creating segmentation (Rubery and Piasna, 2016).

Much recent literature on segmentation stems from the dualization debate rooted in political science (Emmenegger et al., 2012; Rueda, 2014). Dualization scholars understand differences between labour market insiders (those employed on secure contracts) and outsiders (the unemployed or those in atypical employment). This debate tends to adopt a binary vision of the labour market, differentiating between secure and vulnerable labour market positions (Palier and Thelen, 2010). Although some dualization research recognizes the limitations of the binary division (Häusermann and Schwander, 2013), it is accepted as revealing broad socio-structural patterns of divides and preferences. Yet, from a sociological perspective, there is a major weakness to this approach: it tends to treat insiders and outsiders as homogeneous social categories, neglecting within-group variation, despite the fact that employment research has highlighted more fragmented configurations (Blossfeld and Mayer, 1988; Yoon and Chung, 2016). Relatedly, the variable contract type, which is often used to distinguish insiders and outsiders, may not be the best measure because of changes across and within contractual groups in recent years. Employment literature highlights the changing nature of standard and non-standard employment, as well as heterogeneity within these categories (Lautsch, 2002; Stone and Arthurs, 2013).

We adopt the approach of studies arguing that a plurality of segmentation patterns exists rather than dualization between two groups (Blossfeld and Mayer, 1988; Fernández-Macias, 2012). We assume that these fragmented configurations differ across countries given the specificities of national regulatory systems. Therefore, we propose an analytical approach centred on ‘social divides’, with the aim of advancing knowledge in three ways.

First, we investigate patterns of segmentation empirically instead of studying assumed divides between (two) pre-defined labour market groups. Earlier studies, such as Biegert (2014), often used pre-defined groups (insiders and outsiders) and examined the nature and/or extent of differences between them, for instance, regarding income or employment stability. In so doing, there is a risk of overlooking the main dividing lines within and across labour markets, because there is no empirical assessment of the causes of segmentation. Moreover, within-group variation remains hidden when studying pre-defined groups. Yoon and Chung (2016) provide an exception, as their study empirically assesses segmentation patterns in Britain, identifying a third segment between insiders and outsiders.

Second, we advance knowledge on cross-national segmentation patterns by comparing Germany, Belgium and Italy. This is because such patterns are likely to vary across countries, reflecting differences in socio-political, economic and cultural institutions as well as regulatory systems. Yet knowledge of cross-national segmentation patterns is scarce, which could be related to the difficulty of comparing institutions across countries (Locke and Thelen, 1995). Numerous studies highlight possible cross-national differences and the need to explore them (Davidsson and Naczyk, 2009); however, there are only a few studies which do so qualitatively (Palier and Thelen, 2010) and quantitatively (Biegert, 2014).

Third, we aim to uncover the social configuration of the divides identified, since labour markets are characterized by an ‘unequal distribution of possibilities of access of some groups to certain material and symbolic conditions, or to “social goods”’ (Kreckel, 1980: 528). This is in line with social reproduction theory, which stresses the heterogeneity of the workforce, not only in terms of employment contracts or skills but also regarding the position workers occupy in the social structure of the labour market (Fagan and O’Reilly, 1998). Thus, identifying the social groups with different degrees of access to ‘social goods’ across and within countries enables analysis of the ‘social divides’.

Workplaces are central contexts in which segmentation is created, as working conditions are largely regulated at this level, which is embedded in national regulatory systems (Beynon et al., 2002). It is at this level that the micro-foundations for labour market segmentation as macro-level phenomenon are laid (Pulignano and Doerflinger, 2018a). We therefore focus on the conditions underpinning a worker’s employment relationship when examining social divides, because they determine a worker’s degree of (in)security. In accordance with manifold contributions (Eichhorst et al., 2016; Rubery and Piasna, 2016), segmentation today can be viewed in particular in terms of divides between workers with secure and with insecure jobs.

The employment relationship and the insecurity-based divide

Debates on the progressive shift from standard to non-standard employment relationships (Stone and Arthurs, 2013) highlight that different degrees of (in)security can cause segmentation. However, the standard employment relationship may not be a good proxy for the degree of (in)security, as various studies highlight that workers with standard contracts may still suffer from insecurity (Doerflinger, 2016; Lautsch, 2002; Lewchuk, 2017; Weil, 2014). Building upon such studies, we argue that instead of looking solely at the form of the employment relationship (contract type), we need to study its terms and conditions to assess divides between secure and insecure jobs. In line with Gallie et al. (2017), we use a broad definition of insecurity which is not only linked to the risk of job loss. Instead, we claim that the concept encompasses the terms and conditions of employment relationships at the workplace. This has three advantages. First, it allows for a nuanced analysis of insecurity by acknowledging its multidimensional and relational nature, as insecurity relates to different characteristics of the employment relationship. Second, it entails capturing variations other than contract type within contractual groups which affect the degree of (in)security. Third, this variation can be linked to job-related and personal characteristics, which permits us identify the ‘social’ divides.

Rodgers (1989) distinguishes between four dimensions of insecurity: temporal, organizational, economic and social. We argue that these are inherently related to the terms and conditions of the employment relationship. Temporal insecurity refers to the continuity of employment: insecure jobs involve short-time horizons and the risk of job loss. We use two indicators to measure this dimension: the type of employment contract (fixed-term as against open-ended or ‘permanent’ and the number of weekly working hours (full-time or part-time). Although some part-time jobs could provide relative stability, others, particularly in private services, may not only be involuntary but also unstable as the availability of work may fluctuate in accordance with demand. Thus, such jobs may not be ‘continuously available’, which may make them temporally insecure. To capture the possible interlinkages between working hours and sectors, sector (based on NACE codes) is used as an ‘external’ variable in the analysis.

Organizational insecurity involves the degree of control a worker has over the work process. The less control a worker has over working conditions, wages and the pace of work, the more insecure a job tends to be. Such control is implicitly linked to hierarchies in companies as well as to the degree of standardization of jobs. Higher level positions (less standardized) tend to involve greater control, whereas lower level positions (more standardized) often provide lower control. We use two variables to capture the organizational insecurity dimension. First, supervisory responsibility is used as a proxy for control when measuring labour market segmentation. Second, we use occupation based on ISCO codes as external variable, as there is an implicit link between occupation and levels of control. For instance, we assume control levels (and the degree of standardization) to be rather low in such groups as ‘elementary occupations’ or ‘plant and machine operators’, whereas ‘managers’ and ‘professionals’ tend to have more control over their working conditions, wages and pace of work.

Economic insecurity relates to wages. Low-wage jobs can be associated with insecurity, as they may generate difficulties to make ends meets and in-work poverty. The employment relationship defines the wage that an employer pays to the employee in exchange for the work performed, but this wage level may of course be influenced by national and collectively agreed minima. We include wage deciles to cover the economic insecurity dimension. We also include holding a second job as external variable, because this can be associated with difficulties to make ends meet.

Finally, social insecurity generally relates to workers’ extent of protection against contingencies like discrimination or unfair dismissal, as well as protection by social security benefits. Such protections normally derive from national laws or collective agreements. Therefore, it is not simply the terms and conditions of the employment relationship that entitle a worker to certain protections, but rather the fact of having an employment contract. Yet depending on the national regulatory system, entitlements may differ in accordance with types of contract. Considering the social insecurity dimension in our quantitative analysis would imply studying a country’s entire population to find variation. However, we solely focus on employed people, who are generally entitled to protections based on their employment contracts, although variations in the level of entitlements may exist. Furthermore, we primarily focus on the micro-level (social divides between employees) and not the macro-level (country-level social protections).

Data and methods

To explore social divides in the German, Belgian and Italian labour markets, we use data from EU-LFS, an annual cross-sectional survey curated by Eurostat. Data are collected in the 28 EU Member States, two candidate countries and three countries of the European Free Trade Association. Our analyses focus on the employed population, because we consider insecurity as grounded in the terms and conditions of employment. It would not make sense to include the unemployed because of the absence of an employment relationship; the self-employed are excluded for the same reason.

To investigate insecurity-based labour market divides, we use LC analysis (Hagenaars and McCutcheon, 2002; McCutcheon, 1987): a model-based clustering method that creates groups of observations that are similar on a number of characteristics. In other words, LC analysis allows us to identify labour market segments based on a set of variables linked to insecurity (see Table 1). Having identified the segments, we subsequently investigate their population using various socio-demographic and job-related variables (see Table 1). Generally, the use of LC analysis is based on the theoretical assumption that observed indicator variables are statistically associated because of an unobserved common factor rather than being causally related (McCutcheon, 2002). LC analysis thus offers scope to investigate patterns of labour market segmentation that are not directly observable.

Table 1.

Operationalization of insecurity.

Item	Values	Insecurity
Measurement model
Working hours	Full-time/part-time	Temporal
Contract type	Permanent/temporary	Temporal
Supervisory responsibilities	Yes/no	Organizational
Pay from main job	Low (lowest four deciles)/medium (middle three deciles)/high (highest four deciles)	Economic
External variables
Gender	Male/female
Age (years)	17–32/33–50/51–70
Nationality	Country national/EU national/Non-EU-national
Education level	Low/medium/high
Occupation according to ISCO codes	Armed forces/managers/professionals/technicians/clerical/service and sales/skilled agricultural/craft and trades/plant, machine operators and assemblers/elementary	Organizational
Sector according to NACE codes	Agriculture and mining/care services/industry and manufacturing/other services/private services/public administration
Firm size	Small (1–10)/medium (11–49)/large (50 or more)
Second job	Yes/no	Economic

We use a bias-adjusted three-step approach (Bakk et al., 2013; Vermunt, 2010): first, we created an LC model based on the variables linked to insecurity, then we assigned subjects into the extracted labour market segments and in the last step, we analysed associations between labour market segments and various sociodemographic and job-related variables (external variables). Data were weighted by the yearly weighting factor supplied by EU-LFS. More information on the quantitative methodology, the selected multi-group LC model, as well as model fit criteria can be found in the Online Appendix (Supplemental material), Tables 1 and 2 and in Lukac et al. (2019).

Table 2.

Profiles of the labour market segments (%).

	Standard 1	Standard 2	Managers	Part-time	Temps
Group size	12	31	27	26	4
Indicators
Contract type
Permanent	96	91	97	82	1
Working hours
Full-time	97	79	97	31	100
Supervisory
Yes	0	22	61	9	2
Wages
Low	2	2	1	99	81
Middle	59	85	16	1	18
High	39	13	82	0	1
External variables
Gender
Male	69	45	78	26	68
Age (years)
17–32	14	35	14	33	98
33–50	53	35	46	37	2
51–70	34	30	40	30	0
Education level
Low	46	9	2	30	57
Medium	49	71	40	59	35
High	5	20	58	11	7
Nationality
Country national	89	91	96	82	88
EU-national	3	5	3	7	5
Non-EU-national	8	4	1	11	7
Sector
Agriculture/mining	2	1	1	2	7
Care services	6	16	8	15	12
Industry, manufacturing	42	27	35	14	36
Other services	3	4	2	13	2
Private services	29	38	33	47	34
Public administration	19	15	22	9	9
Firm size
50 or more	32	49	71	24	47
11–49	41	33	23	33	36
1–10	27	18	6	43	17
Second job
Yes	0	5	4	4	1

Latent profile probabilities are presented; numbers are rounded.

The items used for operationalizing insecurity in line with Rodgers (1989) are illustrated in Table 1. Insecurity-related variables are used to build up the measurement model, and sociodemographic and job-related variables are utilized as external variables to investigate the populations of the segments identified in a subsequent step. It is important to underline that this subsequent analysis does not affect the formation of segments.

Uncovering social divides in the German, Belgian and Italian labour markets

We identify five distinct labour market segments, demonstrating that there is substantively more fragmentation and complexity than assumed by binary approaches. Tables 2 and 3 illustrate the profiles of these segments. First, almost everybody in the group which we label ‘standard workers 1’ has a permanent, full-time contract, nobody has supervisory responsibilities and wage levels tend to be medium or high. The second group, ‘standard workers 2’, is similar to the former, but there are differences in the share of workers on temporary contracts, working part-time and having supervisory responsibilities. These two groups of standard workers face low to medium levels of temporal and economic insecurity and medium to high levels of organizational insecurity. Third, the group members labelled as ‘managers’ have permanent, full-time contracts, most have supervisory responsibilities and earn high wages. Hence, the ‘managerial workforce’ faces low levels of temporal, organizational and economic insecurity. We label the fourth group ‘part-time workers’, since most – though not all – workers who fit the overall pattern of this group are employed on a part-time basis. Most have permanent contracts, very few have supervisory responsibilities and wages tend to be low for the entire group. Finally, in the ‘temporary employees’ group, almost all work full-time, hardly any have supervisory responsibilities and the great majority earn low wages. Part-time workers and temps generally face high levels of temporal, economic and organizational insecurity.

Table 3.

Occupational profiles (%).

	Standard 1	Standard 2	Managers	Part-time	Temps
Armed forces	29	2	62	2	4
Managers	0	7	89	4	0
Professionals	8	26	55	9	1
Technicians and associate professionals	11	36	35	15	3
Clerical support workers	21	38	12	27	3
Service and sales workers	11	28	6	51	4
Skilled agricultural and forestry	11	37	7	30	15
Craft and trades	29	36	12	16	7
Plant and machine assemblers	37	33	10	17	3
Elementary occupations	13	20	1	62	4

Latent profile probabilities are presented; numbers are rounded.

There is a clear pattern regarding the social categories belonging to these segments. The two groups of standard workers are relatively balanced in their social configuration. In contrast, managers are mostly older men with high education levels, whereas part-time workers and temps tend to be young with low to medium education levels. The majority of part-time workers are female, and temps are predominantly male. Furthermore, the probability of being a non-national is the highest in these two segments. Therefore, the divides identified reflect categorical divisions between different social groups, which the labour market as a social institution seems to reproduce.

In terms of occupations (see Table 3), the patterns identified are hardly surprising. Groups featuring relatively standardized work (elementary occupations, sales and service workers) with low degrees of control are particularly common among part-time workers and temps. In contrast, those in non-standardized occupations (professionals and managers) with high levels of control belong to the managerial workforce. Occupations featuring medium levels of standardization (plant, machine operators and assemblers, craft and trade workers, technicians and associate professions, clerical support workers) are likely to be part of the standard workforce. With regard to the other external variables (sector, firm size, second job), no clear patterns emerge.

Figure A3 in the Online Appendix displays group sizes in the countries investigated. Belgium has the least fragmented labour market, with three broad groups, standard workers 1 (47 percent), part-time workers (27 percent) and managers (24 percent). Standard workers 1 are not present, and the share of temps is tiny (less than 1 percent). In Italy, there are four broad groups, part-time workers (29 percent), standard workers 1 (29 percent), managers (23 percent) and standard workers 2 (16 percent). In contrast, temps only cover 3 percent of the workforce. In Germany, there are three broad labour market groups, standard workers 2 (39 percent), managers (29 percent) and part-time workers (23 percent), as well as temps (5 percent) and standard workers 1 (4 percent).

Explaining social divides: the role of national regulatory systems

According to institutional segmentation theory, the characteristics of national regulatory systems provide explanations for the uncovered social divides. Calculating R² values (see Table 4) sheds light on national variation, as these values indicate how well the extracted LCs (labour market segments) predict the indicators (variables linked to insecurity) in the model. The closer the R²-value to 1, the stronger the connection between the indicator and the latent measurement of segmentation. In the pooled model comprising all countries, wages have the highest explanatory power for insecurity-based divides, followed by working hours, contract type and supervisory responsibilities. Although the combination of indicators shapes insecurity, our analysis demonstrates that economic insecurity is the most important dimension, followed by temporal and organizational insecurity. Table 4 also highlights national specificities: while wages and working hours are important everywhere, contract type matters more in Germany compared to Belgium, and supervisory responsibilities are more in Italy than in Germany and Belgium. Thus, social divides are not caused equally across countries; there are differences in accordance with regulatory systems, which intersect with social categories.

Table 4.

R² values (rounded).

	Variable	R² pooled	R² Germany	R² Belgium	R² Italy
Indicators	Wages	0.64	0.67	0.68	0.59
	Working hours	0.39	0.40	0.34	0.40
	Contract type	0.31	0.37	0.11	0.26
	Supervisory responsibilities	0.27	0.24	0.18	0.36

We use Goodman–Kruskal tau-b coefficients as R² statistic (Southwood, 1974), indicating how well an indicator is explained by the latent variable, which is similar to conventional interpretations of the variance explained in analysis of variance and to item communalities in factor analysis (Vermunt and Magidson, 2005). The values are provided in Latent GOLD 5.1 output.

The extent to which working hours matter depends on the regulation of full-time and part-time work and the gap between them. A wide range of studies has investigated the part-time work penalty, highlighting gaps in wages and social security contributions, access to training and promotion and the difficulty of moving back to full working hours (Jepsen et al., 2005; Russo and Hassink, 2005; Sándor, 2011). While the share of part-time work is similar across the countries (22.2 percent in Germany, 18.5 percent in Italy, 16.4 percent in Belgium according to 2017 OECD data), its regulation differs, with implications for the level of insecurity accompanying this form of work.

In Germany, where working hours matter the most, normal part-time work and marginal part-time work need to be distinguished. While normal part-time workers enjoy the same entitlements as full-time employees relative to the lower amount of working hours, marginal part-time work (‘mini-jobs’) involves lower social security contributions and benefits. This has two effects: First, it causes higher (social) insecurity for the workers, particularly in the long run, as contributing to the pension fund is voluntary. Second, lower contributions make hiring marginal part-time workers cheaper for firms, which have an incentive to create such (insecure) jobs (Stuart et al., 1998). In contrast, marginal part-time work does not exist in Belgium and Italy, where equal provisions relative to working hours are formalized by law, and inferior treatment of part-time workers is generally forbidden.

Working hours are intertwined with social categories, as a large proportion of part-time workers in the investigated countries are female. The part-time penalty is interlinked with the fact that women frequently work in private services and in smaller firms (which is also confirmed by our data), entailing weaker collective regulation in many countries (Rubery and Fagan, 1995). While legal regulation (such as statutory minimum wages) sets universal standards, collective regulation at sectoral and company levels is likely to create differences across industries and establishments, which may cause particular disadvantages for women because of their greater presence in weakly regulated sectors and small establishments. This coincides with wage differences between men and women, as wage levels in private services tend to be lower than in traditional ‘male’ manufacturing sectors. Furthermore, women may be more vulnerable to lower wages because traditional role models consider them as secondary household earners, which is reflected in pay levels (Bendl et al., 2012). The average unadjusted gender pay gap in Europe amounted to 16.2 percent in 2016; however, it was 21.5 percent in Germany, but only 6.1 percent in Belgium and 5.3 percent in Italy (Eurostat, 2018), indicating huge gender wage differentials in Germany.

This gap also reflects specificities of their national bargaining systems. In particular, decentralization and declining coverage rates in Germany (only 56 percent in 2016, according to OECD data (OECD, 2018)) have accentuated inequality, particularly between traditional manufacturing sectors and services (Pulignano and Doerflinger, 2018b). In contrast, the bargaining systems in Belgium and Italy remain more centralized, with coverage rates of 96 percent (2016) and 80 percent (2015), respectively, though recent changes in Italy (such as Renzi’s Jobs Act of 2014) foster decentralization (Pinelli et al., 2017). Wage dispersion in Italy is higher than in Belgium, which could become more pronounced since recent legal changes encourage low-paid employment at the margins of the labour market. However, there is more pay equality across sectors in Belgium and Italy than in Germany, where collective regulation may largely reflect the gender structure of the labour market, setting lower standards for women (Rubery and Fagan, 1995) and causing them higher levels of temporal and economic insecurity.

Furthermore, our analysis reveals differences with regard to contract type across countries. Contract type has very limited explanatory power in Belgium because regulation is equally strong for permanent and temporary contracts. Equal pay and working conditions are stipulated by law, and if there are differences regarding voluntary fringe benefits, strong local unions often bargain supplementary agreements to close the gap between permanent and temporary employees (Pulignano and Doerflinger, 2013). In contrast, the regulations for permanent and temporary contracts differ substantially in Germany, leading to gaps in wages and working conditions, particularly between permanent and temporary agency workers (who are mostly employed on fixed-term contracts). Equal provisions do not necessarily apply because of an opening clause which gives priority to collective agreements. As the existing agreements for the agency work sector set lower standards than in most other sectors, inequalities are generated between permanent and agency workers (Pulignano and Doerflinger, 2013). Italy is situated in between Germany and Belgium regarding the importance of contract type. While equal provisions used to apply in Italy, recent legal changes (the. 2012 Fornero Act and the Renzi Jobs Act) have reduced dismissal protection for permanent workers while increasing protection for temporary workers. At the same time, temporary employment has been facilitated, for example, by lifting the obligation to provide a valid reason for using it (Cubellotti, 2015). Since we analyse data from 2015, the effects of recent reforms may not yet be reflected in our analysis.

The majority of temps are young people. Temporary contracts are often a way to enter the labour market, and regulatory provisions exist in all the countries investigated. In Belgium, ‘labour market entry’ is one of the six reasons allowing companies to use temporary agency work which usually involves fixed-term contracts for the workers (Håkansson et al., 2017). Similarly, German labour law allows labour market entrants to be given temporary contracts without the normally required valid reason for this. Apprentices are hired on temporary contracts, too. Since Germany features the highest proportion of temps in our countries – a group mostly made up by young men in manufacturing – we assume that this group contains many apprentices. Similarly, temporary apprenticeship contracts for young people also exist in Italy, as well as another form of employment contract (contratto di lavoro intermittente) for workers aged 24 years or younger. Generally, our data support the well-documented idea (Passaretta and Wolbers, 2019) that temporary contracts function as stepping stones to other forms of employment. This is because more than 98 percent of temps belong to the youngest age cohort; older cohorts are hardly present among this sub-group, which is a strong indication for a trajectory starting with temporary contracts which are subsequently upgraded over time. The fact that many temps are young workers or apprentices also explains their rather low wages. Overall, our analyses demonstrate that the importance of the contract type variable is interlinked with the regulatory gap between permanent and temporary contracts. Where this gap exists (as in Germany), the variable contributes to shaping social divides. Our results also emphasize the intersection between contract type and age, as young workers in particular tend to be employed on temporary contracts.

Organizational insecurity contributes the least to shaping social divides in Germany, Belgium and Italy. Although higher levels of control over working conditions, wages and the pace of work may come along with higher security, the economic and temporal dimensions of insecurity may outweigh this.

Overall, our results show the interconnections between national regulatory settings and social categories shaping social divides in the three labour markets. Our analysis highlights the social embeddedness of regulatory systems; they are shaped and reshaped by actors, and this influences their degree of inclusiveness, which in turn creates the observed insecurity-based labour market divisions. Thus, actors generate the micro-foundations of social (macro-) phenomena (Pulignano and Doerflinger, 2018a). In sum, our analysis yields two major results. First, it demonstrates that segmentation research can benefit from connecting micro- and macro-levels. Specifically, we have used micro-level variables (terms and conditions of the employment relationship) to uncover dividing lines in the German, Belgian and Italian labour markets and have studied them as grounded in national regulatory settings. To make sense of the divides uncovered, we examined in particular the interlinkages between national regulatory settings and social categories, which provide explanations for the observed social divides. Second, we have shed light on the existence of complex socioeconomic configurations in the labour markets investigated. These tend to be more fragmented than generally assumed; therefore, instead of approaching the topic of segmentation as solely linked to insiders and outsiders, our research suggests that it is more accurate to talk about different degrees of ‘insiderness’ and ‘outsiderness’. These variations result from interactions between the terms of conditions of a worker’s employment relationship, his or her social characteristics and regulatory settings. As regulatory settings differ, our analysis has revealed variation across three labour markets which are often considered as similar (Hall and Soskice, 2001). Therefore, our findings are in line with recent research advocating a shift away from broad categories such as liberal and coordinated market economies and exploring dynamics and the possibly resulting variations within categories or country clusters (Thelen, 2014).

Conclusion

This article has analysed social divides in the German, Belgian and Italian labour markets. Our results highlight the existence of five segments across countries, with Belgium’s configuration being less fragmented compared to Italy and Germany. While the two identified groups of standard workers are relatively balanced in terms of age, gender and education, managers are predominantly highly educated, older men. In contrast, part-time workers tend to be (young) women and temps young men, with low to medium education levels. Therefore, the social divides identified reflect categorical divisions in society.

Our analyses generate three essential implications for further research. First, there is a certain level of convergence, as similar labour market groups could be identified across countries despite differences in national regulatory systems. However, these differences are crucial for explaining the divides uncovered, which are not functionally equivalent across countries. For example, our results demonstrate that contract type is less important than wages and working hours in shaping insecurity-based divides. Previous research has often associated security with permanent contracts, but our results demonstrate that other indicators linked to the terms and conditions of the employment relationship may more strongly contribute to generating (in)security. Moreover, the importance of contract type differs across the three countries, reflecting specificities of regulatory systems. Future research should therefore not treat indicators as functionally equivalent across countries and rather follow the idea of contextualized comparisons (Locke and Thelen, 1995) when selecting cases and analysing research results. Relatedly, one should be cautious when comparatively studying the same pre-defined groups across countries: for instance, the atypical workforce in Germany has a very different labour market position compared to its Belgian counterpart.

Second, the social configurations identified reflect categorical inequalities, particularly along the lines of gender, age, education level and ethnicity. These categories intersect, creating additional labour market disadvantages. Our results suggest that such intersections through social categories may contribute to revealing the processes through which inequalities are perpetuated in society. As with existing intersectionality research (McBride et al., 2015), our results confirm that a combined understanding of (for example) gendered and age-based work processes could lead to more comprehensive explanations of the production of social disadvantages in contemporary labour markets.

Third, our analyses stress the existence of insecurity-based labour market divides, which are intertwined with regulatory systems and social categories. Segmentation along the dimensions of temporal, economic and organizational insecurity particularly disadvantages women, migrants, young and poorly educated workers and is deeply rooted in sectoral and national regulatory systems. Therefore, our study highlights that drivers of segmentation can (co)exist at various (regulatory) levels including national, sectoral and individual. This stresses the need to develop novel theoretical and methodological approaches enabling us to investigate social divides at these different levels, and particularly to study their interaction.

Supplemental Material

doerflinger_et_al_online_supp – Supplemental material for The social configuration of labour market divides: An analysis of Germany, Belgium and Italy

Supplemental material, doerflinger_et_al_online_supp for The social configuration of labour market divides: An analysis of Germany, Belgium and Italy by Nadja Doerflinger, Valeria Pulignano and Martin Lukac in European Journal of Industrial Relations

Footnotes

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship and/or publication of this article: This research was developed in the framework of two interlinked research projects: ‘Hoe ontstaan sociale scheidslijnen in hedendaagse Europese arbeidsmarkten?’ (How are social divides produced in contemporary European labour markets?), financed by the Flemish Fonds Wetenschappelijk Onderzoek (FWO), project number G071716N and a KU Leuven C1 grant (number C14/16/015).

Supplemental material

Supplemental material for this article is available online.

Author biographies

Nadja Doerflinger is a Postdoctoral Researcher in the Research Group ‘Employment Relations and Labour Markets’ at the Centre for Sociological Research (CeSO), KU Leuven, Belgium.

Valeria Pulignano is a Professor in Sociology of Labour and Industrial Relations and Head of the Research Group ‘Employment Relations and Labour Markets’ at the Centre for Sociological Research (CeSO), KU Leuven, Belgium.

Martin Lukac is a PhD student at the Centre for Sociological Research at KU Leuven.

References

Bakk

Tekle

Vermunt

(2013) Estimating the association between latent class membership and external variables using bias-adjusted three-step approaches. Sociological Methodology 43(1): 272–311.

Bendl

Hanappi-Egger

Hofmann

(2012) Diversität und Diversitätsmanagement. Vienna: facultas.

Beynon

Grimshaw

Rubery

Ward

(2002) Managing Employment Change: The New Realities of Work. Oxford: Oxford University Press.

Biegert

(2014) On the outside looking in? Transitions out of non-employment in the United Kingdom and Germany. Journal of European Social Policy 24(1): 3–18.

Blossfeld

Mayer

(1988) Labor market segmentation in the Federal Republic of Germany: An empirical study of segmentation theories from a life course perspective. European Sociological Review 4(2): 123–140.

Brandl

Lehr

(2018) A problem of embeddedness: A micro-theoretical model of trade union action and interaction in Europe. Employee Relations 40(3): 500–514.

Cubellotti

(2015) Italy: Changes to fixed-term contracts. Available at: https://www.eurofound.europa.eu/publications/article/2015/italy-changes-to-fixed-term-contracts

Davidsson

Naczyk

(2009) The ins and outs of dualisation: A literature review. Working Papers on the Reconciliation of Work and Welfare in Europe, REC-WP 02/2009. Edinburgh: RECWOWE Publication, Dissemination and Dialogue Centre.

Doerflinger

(2016). Different worlds of work? A study on labour market regulatory institutions and contingent work in Belgium and Germany. Unpublished PhD Thesis, University of Leuven, Leuven.

10.

Doeringer

Piore

(1971) Internal Labor Markets and Manpower Analysis. Lexington, KY: D. C. Heath and Company.

11.

Eichhorst

Marx

Wehner

(2016) Labor market reforms in Europe: Towards more flexicure labour markets? IZA Discussion Paper no. 9863. Bonn: IZA.

12.

Emmenegger

Häusermann

Palier

Seeleib-Kaiser

(2012) The Age of Dualization: The Changing Face of Inequality in Deindustrializing Societies. Oxford: Oxford University Press.

13.

Eurostat (2018) Gender pay gap statistics. Available at: http://ec.europa.eu/eurostat/statistics-explained/index.php/Gender_pay_gap_statistics

14.

Fagan

O’Reilly

(1998) Conceptualising part-time work: The value of an integrated comparative perspective. In: O’Reilly

Fagan

(eds) Part-Time Prospects: An International Comparison of Part-Time Work in Europe, North America and the Pacific Rim. London: Routledge, pp. 1–32.

15.

Fernández-Macias

(2012) Job polarization in Europe? Changes in the employment structure and job quality, 1995–2007. Work and Occupations 39(2): 157–182.

16.

Gallie

Felstead

Green

Inanc

(2017) The hidden face of job insecurity. Work, Employment and Society 31(1): 36–53.

17.

Granovetter

(1992) Institutions as social constructions: A framework for analysis. Acta Sociologica 35(1): 3–11.

18.

Grimshaw

Cartwright

Keizer

Rubery

(2018) Market exposure and the labour process: The contradictory dynamics in managing subcontracted services work. Work, Employment and Society 33(1): 76–95.

19.

Hagenaars

McCutcheon

(2002) Applied Latent Class Analysis. Cambridge: Cambridge University Press.

20.

Håkansson

Pulignano

Isidorsson

Doerflinger

(2017) Explaining job insecurity for temporary agency workers: A comparison between Sweden and Belgium. Economic and Industrial Democracy. Epub ahead of print 6 June. DOI:10.1177/0143831X17707824.

21.

Hall

Soskice

(eds) (2001) Varieties of Capitalism: The Institutional Foundations of Comparative Advantage. Oxford: Oxford University Press.

22.

Häusermann

Schwander

(2013) Who is in and who is out? Proposing a risk-based operationalisation of insiders and outsiders. Journal of European Social Policy 23(3): 248–269.

23.

Jepsen

O’Dorchai

Plasman

Rycx

(2005) The wage penalty induced by part-time work: The case of Belgium. Brussels Economic Review 48(1–2): 73–94.

24.

Kreckel

(1980) Unequal opportunity structure and labour market segmentation. Sociology 14(4): 525–550.

25.

Lautsch

(2002) Uncovering and explaining variance in the features and outcomes of contingent work. ILR Review 56(1): 23–43. Available at: https://doi.org/10.1177/001979390205600102

26.

Lewchuk

(2017) Precarious jobs: Where are they, and how do they affect well-being? Economic and Labour Relations Review 28(3): 402–419.

27.

Locke

Thelen

(1995) Apples and oranges revisited: Contextualized comparisons and the study of comparative labor politics. Politics & Society 23(3): 337–367.

28.

Lukac

Doerflinger

Pulignano

(2019) Developing a cross-national comparative framework for studying labour market segmentation: Measurement equivalence with latent class analysis. Social Indicators Research 145(1): 233–255.

29.

McBride

Hebson

Holgate

(2015) Intersectionality: Are we taking enough notice in the field of work and employment relations. Work, Employment and Society 29(2): 331–341.

30.

McCutcheon

(1987) Latent Class Analysis. London: SAGE.

31.

McCutcheon

(2002) Basic concepts and procedures in single- and multi-group latent class analysis. In: Hagenaars

McCutcheon

(eds) Applied Latent Class Analysis. Cambridge: Cambridge University Press, pp. 56–88.

32.

OECD (2018) Part-time employment statistics. Available at: https://data.oecd.org/emp/part-time-employment-rate.htm

33.

Osterman

(1994) How common is workplace transformation and who adopts it? ILR Review 47(2): 173–188.

34.

Palier

Thelen

(2010) Institutionalizing dualism: Complementarities and change in France and Germany. Politics & Society 38(1): 119–148.

35.

Passaretta

Wolbers

(2019) Temporary employment at labour market entry in Europe: Labour market dualism, transitions to secure employment and upward mobility. Economic and Industrial Democracy 33(1): 76–95.

36.

Pinelli

Torre

Pace

Cassio

Arpaia

(2017) The recent reform of the labour market in Italy: A review. European Economy Discussion Paper no. 72. Luxembourg: Publications Office of the European Union.

37.

Pulignano

Doerflinger

(2013) A head with two tales: Trade unions’ influence on temporary agency work in Belgian and German workplaces. International Journal of Human Resource Management 24(22): 4149–4165.

38.

Pulignano

Doerflinger

(2018a) Expanding social actor-based explanations in labour market dualisation research: A combined macro-micro and micro-macro approach. Employee Relations 40(1): 75–88.

39.

Pulignano

Doerflinger

(2018b) Labour markets, solidarity and precarious work: Comparing local unions’ responses to management flexibility strategies in the German and Belgian metalworking and chemical industries. In: Doellgast

Lillie

Pulignano

(eds) Reconstructing Solidarity: Labour Unions, Precarious Work, and the Politics of Institutional Change in Europe. Oxford: Oxford University Press, pp. 104–123.

40.

Rodgers

(1989) Precarious work in Western Europe: The state of the debate. In: Rodgers

Rodgers

(eds) Precarious Jobs in Labour Market Regulation. Geneva: ILO, pp. 1–16.

41.

Rubery

(1978) Structured labour markets, worker organisation and low pay. Cambridge Journal of Economics 2(1): 17–36.

42.

Rubery

(2007) Developing segmentation theory: A thirty years perspective. Economies et Societés 41(6): 941–964.

43.

Rubery

Fagan

(1995) Comparative industrial relations research: Towards reversing the gender bias. British Journal of Industrial Relations 33(2): 209–236.

44.

Rubery

Piasna

(2016) Labour market segmentation and the EU reform agenda: Developing alternatives to the mainstream. ETUI Working Paper no. 2016.10. Brussels: European Trade Union Institute.

45.

Rubery

Keizer

Grimshaw

(2016) Flexibility bites back: The multiple and hidden costs of flexible employment policies. Human Resource Management Journal 26(3): 235–251.

46.

Rueda

(2014) Dualization, crisis and the welfare state. Socio-Economic Review 12(2): 381–407.

47.

Russo

Hassink

(2005) The part-time wage penalty: A career perspective. IZA Discussion Paper no. 1468. Bonn: IZA.

48.

Sándor

(2011) Part-Time Work in Europe. Dublin: Eurofound.

49.

Southwood

(1974) Goodman and Kruskal’s Tau-b as correlation ratio: Some implications. Sociological Methods & Research 3(1): 82–110.

50.

Stone

Arthurs

(eds) (2013) Rethinking Workplace Regulation: Beyond the Standard Contract of Employment. New York: Russell Sage Foundation.

51.

Stuart

Fagan

Rubery

(1998) Where and why is part-time work growing in Europe? In: O’Reilly

Fagan

(eds) Part-Time Prospects: An International Comparison of Part-Time Work in Europe, North America and the Pacific Rim. London: Routledge, pp. 35–56.

52.

Thelen

(2014) Varieties of Liberalization and the New Politics of Social Solidarity. Cambridge: Cambridge University Press.

53.

Tilly

(1998) Durable Inequality. Berkeley, CA: University of California Press.

54.

Vermunt

Magidson

(2005) Latent GOLD 4.0 User’s Guide. Belmont, Massachusetts: Statistical Innovations Inc.

55.

Vermunt

(2010) Latent class modeling with covariates: Two improved three-step approaches. Political Analysis 18(4): 450–469.

56.

Weil

(2014) The Fissured Workplace: Why Work Became So Bad for So Many and What Can Be Done to Improve It. Cambridge, MA: Harvard University Press.

57.

Yoon

Chung

(2016) New forms of dualization? Labour market segmentation patterns in the UK from the late 1990s until the post crisis in the late 2000s. Social Indicators Research 128(2): 609–631.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.58 MB