The gender dimension of outsiderness in Western Europe: a comparative cross-model analysis

Abstract

Purpose

The article investigates whether and to what extent outsiderness is gendered in Western Europe, both in terms of its spread and degree. It thus explores which male and female post-Fordist social classes are more exposed to the risk of this phenomenon. It also scrutinizes whether such a gendered characterization has varied over time and across clusters of Western European countries.

Design/methodology/approach

Relying on a comparative analysis of the data provided by the European Social Survey (ESS) dataset and comparing two points in time – the early/mid-2000s and the late 2010s – the work provides both a dichotomous and continuous variable of outsiderness, which measure its spread and degree in the female and male workforces of a pooled set of growth models.

Findings

The empirical analysis shows that outsiderness is profoundly gendered in Western Europe and thus a feminized social phenomenon. However, the comparative investigation highlights that outsiderness has been genderized in diverse ways across the four growth models. Different patterns of gendered outsiderness can be identified.

Originality/value

The article provides a comparative and diachronic analysis of outsiderness from a gender lens, putting into a mutual dialogue different literature on labour market, and shows that outsiderness represents a key analytical dimension for assessing gender inequalities.

Citation

Giuliani, G.A. (2023), "The gender dimension of outsiderness in Western Europe: a comparative cross-model analysis", International Journal of Sociology and Social Policy, Vol. 43 No. 13/14, pp. 62-78. https://doi.org/10.1108/IJSSP-12-2022-0317

1. Introduction

The paradigmatic changes occurring in the economic and labour market structures of the Western European countries have given rise to a very informative comparative literature on the new insider/outsider cleavage (e.g. Lindbeck and Snower, 2001; Rueda, 2007; Emmenegger et al., 2012).

The transition toward a post-Fordist economy has been associated with the emergence of rising inequalities and the growth of a new group of workers – the outsiders – who are more likely to face a permanent disadvantage status in the labour market (Saint-Paul, 2002; Rueda, 2007; Emmenegger et al., 2012).

Theoretical and empirical works tend to agree that in the post-Fordist economies, atypical work – i.e. part-time and fixed-term jobs – together with unemployment can be interpreted as conditions of increased social and labour market vulnerability (Schwander and Häusermann, 2013; Palier and Thelen, 2010). In this regard, some studies have particularly focused on the outsiderness risk that post-Fordist occupational classes have to face during their working life (Schwander and Häusermann, 2013; Giuliani and Raspanti, 2022).

In this growing research agenda, gender plays an important role. Scholars have emphasized that outsiderness has a gendered feature (Russell and Barbieri, 2000; Schwander and Häusermann, 2013). Women are indeed commonly portrayed as an outsider category by default (Palier and Thelen, 2010; Bonoli, 2013), since they experience more disadvantaged labour market positions than men, mostly due to the disproportionately shared (if shared at all) burden of caring responsibilities within the households (Leitner, 2003).

The outsiderness literature thus incorporates some theoretical insights provided by the feminist studies on gender inequalities (e.g. Charles and Grusky, 2004; Estévez-Abe, 2006), but there is scarce evidence of engagement with these studies. More specifically, the analysis of the gendered dimension of outsiderness remains deficient in at least three aspects.

First, the current research has not emphasized enough to what extent outsiderness in a given country is gendered. It is still unclear whether the risk of outsiderness is primarily a female issue, thus affecting men only marginally.

Second, categorizing all female workers as outsiders oversimplifies the structure of the social stratification in the advanced economies, not allowing to identify differences across the post-Fordist occupational classes.

Third, a focus on outsiderness in terms of comparative gendered trends is mostly lacking. In institutional terms, the way in which the economic development of the Western European countries has combined with rising inequalities is not univocal (Trigilia, 2022). It follows that different growth models may be detected, which correspond to different configurations of the labour market (Gherardini, 2022) and therefore of male and female stratification. Furthermore, the social classes' exposure to labour market risks is not fixed and can evolve over the years. The gender dimension of outsiderness may thus vary over time and across countries or clusters of countries.

Focusing on a cluster of countries representing the four growth models (Trigilia, 2022) and comparing two points in time – the early/mid-2000s and the late 2010s – this work explores how outsiderness has configured and differentiated in gendered terms in Western European countries. More specifically, the article answers two main research questions:

RQ1.

Is outsiderness gendered in Western Europe? And if so, to what extent?

RQ2.

How has the gender dimension of outsiderness varied across the four growth models and over time?

The article is exploratory in nature: it provides a comparative and longitudinal investigation of the gendered feature of outsiderness and does not aim to elaborate a causal theoretical framework. In this regard, its contributions are theoretical, methodological and empirical.

From a theoretical perspective, the article instigates a mutual dialogue with the literature on outsiderness, the research on post-industrial social stratification and that concerning gender inequalities in the labour market. The study conceptualizes the gendered dimension of outsiderness from a social risk-based approach and looks at outsiderness as a crucial factor for analysing gender inequalities.

From a methodological perspective, it builds on and expands the tools provided by the literature on the insider/outsider cleavage. The article thus provides both a dichotomous and continuous variable of outsiderness, which is able to measure both its spread and degree.

Finally, from an empirical perspective, based on survey data from the European Social Survey (ESS) dataset, it offers information concerning cross-growth models differences and their evolution over time.

The article is structured in the following way. The first section succinctly discusses the conceptual underpinning of the article, that is, the insider–outsider theory. Then, after a concise review on the relations between outsiderness and gender inequalities, a framework for investigating outsiderness from a gender perspective is elaborated. This is followed by the data and method section. In the third empirical section, the configuration of social stratification and outsiderness in the four growth models is analysed. The final part is devoted to the conclusions.

2. Outsiderness in the post-industrial economies: theoretical considerations

The end of the Fordist era in the late 1970s was followed by a wave of structural changes in the labour markets of the advanced economies. In Western Europe, passive labour market policies, rigid employment protection legislation (EPL) and extensive union coverage were accused of distorting the functioning of the markets and thus were held to account for the poor economy performance of these countries (Nickell and Layard, 1999). Increased labour costs in a context of a globalized economy required companies to adopt a flexible social structure of accumulation (SSA), characterized by a combined strategy of de-industrialization, de-unionization and financialization (Rubin, 2014). Accordingly, from the 1980s – and increasingly in the 1990 and 2000s – governments started to de-regulate work contracts in order to boost the job creation potential in the service sector which has come at the detriment of employment security for workers. The flexibilization of the labour markets, however, mainly affected the fringes and not the core of the workforce (Saint-Paul, 2002), thus contributing to the emergence of the outsider–insider cleavage (Rueda, 2007).

The literature on outsiderness does not share a univocal conceptualization and operationalization of this dualism.

In the original formulation of the insider–outsider theory developed in political economy, outsiders and insiders are categorized on the base of their legal job status (Rueda, 2007). Within such a framework, two different approaches can be detected (Rovny and Rovny, 2017). The first one relies on the current labour market status. Rueda (2007) divides insiders and outsiders according to their current employment. Those with secure positions are classified as insiders and those without are outsiders. The second approach proposed an alternative classification based on the outsiderness risk experienced by occupational classes (Schwander and Häusermann, 2013; see also Giuliani and Raspanti, 2022). This literature stresses that the labour market status is a quite unstable category: a person can quickly change types of contract as well as move rapidly from employment to unemployment and vice versa. In the current times, the risk of outsiderness, however, is more structural: some specific categories of workers are more likely than others to experience unemployment or to be in nonstandard jobs during their working life. In this regard, social classes are seen as representing a more stable category for detecting which workers have better/more chances of being outsiders or insiders, since people do not easily and quickly change their occupational class.

Other bodies of research have linked the exclusion in the labour market with a lower entitlement to social rights (e.g. Palier and Thelen, 2010). In several countries – especially in the Continental ones – atypical workers fail to meet the minimum requirement for being eligible for social programs resulting entitled only to means-tested measures. Therefore, outsiders are those workers who can rely only on a second-order social protection based on social assistance rather than on a first-order social protection centred on social insurance (Palier and Thelen, 2010). However, whether labour market and welfare state outsiders perfectly overlap in all the Western European countries remains an empirical question (Davidsson and Naczyk, 2009).

Finally, other studies have conceptualized the insider/outsider divide considering the workers' subjective evaluation of the labour market, that is their perception of re-employability or, more in general, their perceived labour market and job insecurity (e.g. Duman and Kemmerling, 2020; Lowe, 2020).

3. Analysing outsiderness from a gender perspective

When analysing social disadvantages in the post-Fordist economies, gender has been depicted as a key explanatory variable (Leicht, 2008). The transformations of the economic structures of the advanced democracies went in parallel with a social modernization, which has questioned the male breadwinner family model and, consequently, the strict gendered division of paid and unpaid work within the household (Esping-Andersen, 2009). The massive entrance of women into the labour market since the 1970s has triggered a – still ongoing – revolution, which has facilitated the transition toward a dual-earner family model (Leitner, 2003).

However, the comparative literature in the field agrees that substantial inequalities have characterized women's entrance into the labour market (e.g. Hakim, 2006).

First, the female employment rate varies substantially across European countries, even in the post-Fordist age. While these rates are very high in the Scandinavian and Anglo-Saxon countries, women in the Southern countries – especially in Italy and Greece – continue to show low participation within the labour market. In the continental cluster, female employment has substantially increased over time, but it continues to be markedly lower than that for males.

Second, female employment has been mostly concentrated in specific occupational sectors – first of all, the service sector – where women emerge as disproportionally represented compared to men, while other economic fields – for example, the technical and manufacturing sectors – remain substantially masculinized. In other words, the increase of female labour force participation has been accompanied by a high level of horizontal segregation and the strengthening of the pink-collar occupational ghettos (Charles and Grusky, 2004).

Third, there is a structural penalization affecting women – compared to men – in terms of the types of job they have usually been offered – kinds of contract, work intensity and quality of the job (Estévez-Abe, 2006). These gender inequalities have been usually associated with the persistence of unequal, gendered family responsibilities (Saraceno, 2022). Despite an acceleration of the modernization process in terms of gender culture in several European countries (Inglearth, 1990, 2018; Giuliani, 2022) and an increase in women's educational attainments (Blossfeld et al., 2015), women still have to cope with the double burden of caring and work. Such a burden is likely to negatively influence women's career planning and the employers' decisions concerning hiring and promotion (Blossfeld et al., 2015).

The literature on outsiderness does indirectly take into account these gender inequalities that characterize post-industrial economies. Women have been mostly depicted as an outsider group, thus more exposed to the risk of being unemployed or atypical workers – first of all, part-time (e.g. Emmenegger, 2010; Schwander and Häusermann, 2013; Bonoli, 2013; Beramendi et al., 2015). This research field has thus indirectly admitted that outsiderness has a gendered dimension. However, such an aspect has been treated, theoretically and methodologically, only as a second-order issue in the literature.

To properly uncover to what extent this social phenomenon is gendered, an ad hoc theoretical framework needs to be elaborated. Following the most recent literature (e.g. Rovny and Rovny, 2017; Schwander, 2019; Natili and Negri, 2022), we consider the “classic” political economy conceptualization of the insider/outsider cleavage – based on the legal status of the job – the most appropriate starting point for this exploratory analysis. Within this framework, the choice is to follow the risk-based approach to the study of outsiderness (Schwander and Häusermann, 2013). The categorization of the outsiders/insiders is thus based on the risk exposure displayed by the post-industrial social classes (Oesch, 2006). An occupational group is labelled as outsider when it shows a higher risk of being unemployed and/or employed in an atypical job compared to the whole workforce mean.

Following this approach, outsiderness is defined as a dichotomic variable (being or not being an outsider). However, outsiderness can be also conceptualized in continuous terms by considering two dimensions: the spread and the degree. The former refers to the share of workers included in the occupational group categorized as the outsider. The higher the share, the greater the spread of outsiderness for that class. The degree concerns the intensity of outsiderness. Some class groups can be labelled as outsiders, but the intensity of their disadvantage can be minimal, while for other groups it can be marked.

Both the outsiderness spread and degree are tied with social stratification and its re-configuration at the national or cross-national level. For instance, low-skilled social groups are expected to show a higher degree of outsiderness compared to high- or medium-skilled ones. At the same time, in some advanced economies, the size of some occupational classes categorized as outsiders can be greater than in others, thus resulting in a higher spread. However, social stratification is strongly gendered per se: some occupational groups are historically male-oriented (or masculinized) while others are typically female-oriented (or feminized). It follows that post-Fordist social groups may be disaggregated according to gender. For example, it is possible to conceptually split the blue-collar worker class in two distinct occupational groups: male blue-collar workers and female blue-collar workers. Keeping male and female classes separate allows us to better explore the gender dimension of outsiderness both at the micro- and the macro-level.

At the micro-level, it is possible to detect which female social classes are outsiders and to measure their spread and degree. It is also possible to explore whether the same female classes are categorized as outsiders in different national and cross-national contexts or whether differences may be identified.

At the macro-level, the average risk displayed by those female and male post-industrial social groups labelled as outsiders provides an overall value of female and male outsiderness in terms of both spread and degree.

The gap between female and male outsiderness provides an accurate indication concerning the gendered nature of outsiderness, the possible cross-countries/cross-models similarities and differences, as well as the changes or stability over time.

4. Case selection, data and methods

To explore the gendered dimension of outsiderness in Western Europe, we focused on a pooled sets of what have been defined the four growth models (Trigilia, 2022): the Anglo-Saxon countries of the non-inclusive growth (NIG: Ireland and the United Kingdom), the Scandinavian group of the egalitarian inclusive growth (EIG: Denmark, Finland, Iceland, Norway and Sweden), the Continental cluster of the dualistic inclusive growth (DIG: Austria, Belgium, France, Germany, the Netherlands and Switzerland) and the Southern constellation of the non-inclusive low growth (NILG: Italy, Portugal and Spain). These four growth models display significant macro-economic differences (Gherardini, 2022), especially when considering how economic growth has combined with social inequalities, which have led to different configurations of their labour market structures (Rizza, 2022) and, consequently, of social stratification (Gherardini and Giuliani, 2022). In other words, growth models could offer sufficient variation to identify diverse patterns of gendered outsiderness [1].

Post-industrial classes have been detected through Oesch's class schema, which has been able to integrate women better into stratification analysis compared to the traditional class schemata. The occupational scheme – which combines two different analytical dimensions: the extent of marketable skills (vertical dimension) and the type of work being done (horizontal dimension) – provides a more in-depth differentiation between sectors characterized by an over-representation of the female labour force. The eight social classes have been disaggregated by gender for a total of 16 occupational groups (Table 1) [2].

Our measurement of outsiderness is based on the conceptualization and methodology proposed by Schwander and Häusermann (2013), integrated with some changes. Outsiders are defined as those individuals experiencing a higher probability of (a) being unemployed (job-seeking and/or inactive) and (b) being employed part-time [3] and/or (c) with fixed-term contracts. We also count as atypical workers those employees with other types of non-standard contracts, such as temporary help agencies and contingent work. The probability of experiencing unemployment/atypical work depends on the frequency within the relevant occupational category of an individual. For all the 16 social groups, we compared the group-specific unemployment and atypical employment rates with the average rate in the model workforce. For the dichotomic variable, we labelled as outsiders those classes showing a rate of outsiderness which is significantly (p < 0.05) higher than the model mean. As for the continuous variable, the outsiderness degree was built at the micro-level by subtracting the average workforce rate from the group-specific rate.

At the macro-level, we created six measures for each welfare model to detect the outsiderness spread and the degree in the total workforce, female workforce and male workforce.

For the outsiderness spread, the measures are calculated as follows:

WS=ΣOi

where O_i indicates the share of the outsider social group (i) in the total workforce (with “i” ranging from 1 to 16).

FS=ΣFK

where F_K indicates the share of the female outsider social group (k) in the female workforce (with “k” ranging from 1 to 8).

MS=ΣMj

where M_K indicates the share of the male outsider social group (j) in the male workforce (with “j” ranging from 1 to 8).

The outsiderness spread ranges from 0 (no spread) to 100 (very high spread – all the social classes are outsiders).

The gender spread gap is calculated as follows:

GSG=FS−MS

For the outsiderness degree, the measures are calculated as follows:

WD=1/WS Σ(Doi*Oi)

where Do_i indicates the outsiderness degree for the outsider social group (i), (with “i” ranging from 1 to 16).

FD=1/FS Σ(Dfk*FK)

where Df_k indicates the outsiderness degree for the female outsider social group (k) (with “k” ranging from 1 to 8).

MD=1/MS Σ(Dmj*Mj)

where Dm_j indicates the outsiderness degree for the male outsider social group (j) (with “j’ ranging from 1 to 8).

The gender degree gap is calculated as follows:

GDG=FD−MD

Positive values of the outsiderness spread and degree indicate that outsiderness is gendered.

We applied this operationalization to the ESS dataset [4]. We aggregated Rounds 1–3 for detecting outsiderness in the early/mid 2000s, while pooling Rounds 8–9 for investigating outsiderness in the late 2010s. In the middle of these two temporal points, the economic and financial recession of 2008 occurred. The crisis exacerbated the outsider/insider divide. However, reactions to counterbalance the negative effects varied, since countries were equipped with diverse sets of welfare and labour market policy institutions (Trigilia, 2022). In the late 2010s, employment rates were back at pre-crisis levels in most European Union countries, though atypical job rates continued to increase – a trend that started well before the crisis of 2008. Focusing on the late years of the 2010s allows us to take into account the role played by the crisis but also to avoid an overestimation of its more critical short-term effects on outsiderness. The crisis is therefore considered a catalyst of the process of economic and structural change of the labour market that emerged from the end of the Fordist era onwards.

5. Changes in social stratification across the four growth models

Table 2ab illustrates the structures of female and male stratification and their change over time. The data illustrate that women are primarily concentrated in three classes: clerks, socio-cultural (semi-) professionals (SCPs) and low-service functionaries (LSFs). On the contrary, compared to men, they are under-represented among the blue-collar workers (BCs) and technical (semi-) professionals (Tech) social classes but also among traditional bourgeoisie (TB) and small business owners (SBOs) – though to a lower degree. In other words, the clerk, SCP and LSF classes are strongly feminized in all four growth models: approximately, 60% of the female workforce belongs to one of these occupational groups, while the value is below 30% when considering the male workforce. Additionally, among these three classes, the LSF group absorbs the highest share of female workers (around 30%). Data thus stress that women are mostly employed in jobs belonging to the interpersonal service. More interestingly, a high percentage of them work in those occupations which require only vocational or low skills (clerks and LSFs).

Although this phenomenon seems structural, a declining trend can be identified. Over time and in all four growth models, the percentage of workers belonging to these three classes has scaled down within the female workforce. However, there are interesting cross-model differences when considering SCPs and LSFs.

In the NIG/Anglo-Saxon and EIG/Scandinavian countries, we note an upgrading process taking place within the feminized occupations. Female employment has indeed increased in the high-skilled, high-educated service sector (the female SCPs) while decreasing in the low-skilled, low-educated one, i.e. the female Clerks and LSFs though this latter sector continues to employ many female workers. Such an upgrading process has been followed by a de-gendering trend: while women continue to be over-represented in these occupational classes compared to men, a slow, gender convergence process can be observed, especially when considering the low service sectors. In the EIG countries, the share of male LSFs in the male workforce has indeed increased (+2.1), while in the NIG model, they absorb a relatively high quota of the total male workforce (16%) – the highest among the four models. Such a (slow) de-gendering trend may be explained by considering two factors. First, in both the Anglo-Saxon and Scandinavian clusters, the group of male BCs has substantially downsized (−4.4 and −6 pp) as a consequence of the process of de-industrialization and tertiarization that in these countries started earlier than in the Bismarckian countries (Pierson, 2001). Low-skilled male workers have thus fewer chances than in the past to find an occupation in the manufacturing sector and are pressured to enter the feminized low-skilled service sector. Second, it could be the case that male migrants have mostly entered this sector. Given their expected more significant difficulties in finding jobs compared to native men – especially in the high-skilled sector – they could be more inclined to enter feminized occupations.

In the DIG/Continental model, these trends have not occurred yet. With regard to upgrading, the share of SCPs and LSFs remains by and large stable. Furthermore, these occupations remained strictly gendered, especially when considering LSFs. Considering the male workforce, the percentage of men belonging to this social class has remained constant (around 10%) and lower than in the NIG and EIG clusters. Such a scenario can be partially explained considering that, despite a decline (−5 pp), the share of BCs within the male workforce remains high (31%), by a comparative standard. In other words, this class can still absorb a significant quota of low-skilled male workers who are under less pressure to enter feminized occupations.

Finally, in the NILG/Southern cluster, a downgrading process seems to occur: over time, the new jobs for female workers have been primarily created in the low-skilled service sector (the female LSFs), while fewer women are employed in the high-skilled one (the female SCPs). Interestingly, the increase in the share of female LSFs was followed by a de-gendering trend. Men have started to enter the low-skilled sector (the share of male LSFs has increased by 1.7% points), despite a shrinking, but still large, BC class. This can be explained if we consider that in most Southern countries, the labour market has taken a low road (Kazepov and Ranci, 2017), with the bulk of new jobs being low-skilled, while job opportunities in the high-skilled sectors tend to be scarce. Furthermore, compared to the other growth models, the number of low-educated workers is high, especially among men, and cannot be absorbed by the manufacturing sector. In other words, in the NILG, de-gendering occurs in a general context of lack of social capital among broad sectors of society.

6. Exploring the gender dimension of outsiderness in Western Europe

Table 3ab shows the results of the outsiderness analysis in the four growth models. Before considering the gender dimension, it is helpful to detect who actually are the outsider social classes and how outsiderness has been configured among the total workforce of each model.

Female BCs and LSFs are two outsider groups in all four models, both in the early/mid-2000s and the late 2010s. Their outsiderness degree is, on average, the highest among all the outsiders though it has declined over time, especially in the EIG countries. Similarly, even male LSFs are categorized as outsiders, though in the DIG and NILG models, their positions worsened in the late 2010s.

Considering these results, outsiderness seems to be concentrated mainly among those workers with general/vocational or low skills, while the high-skilled ones would appear to be better protected. However, such a conclusion is only partially true. In the Continental and Southern countries, even the high-skilled female SCPs (and in the former, even male SCPs) have an above-average risk of being unemployed/atypical workers. In other words, even high-skilled workers in the post-Fordist age can experience labour market vulnerability. This insight is also confirmed by the outsiderness conditions of the female TBs and SBOs in the Anglo-Saxon and – partially – Scandinavian models in the late 2010s (which is primarily due to the high part-time rate displayed by these two classes). Finally, it is interesting to note that male BCs – commonly described by the literature as insiders (e.g. Beramendi et al., 2015) – are an outsider group in the NILG model (but also in the NIG model in the early/mid-2000s). In other words, from the 2000s onwards, in the Mediterranean countries such a social group has already lost its “sheltered” status inherited from the Fordist age. In contrast, they remain relatively insulated from the outsiderness risk in the EIG and DIG models.

Table 3b provides information regarding the outsiderness spread and degree among the total workforce in the four growth models.

With regard to the outsiderness spread, we can see a decreasing trend in the Anglo-Saxon Scandinavian models (−13.5 and −7.9): in the late 2010s, fewer workers were exposed to the outsiderness risk compared to the early/mid-2010s. The EIG model has always shown a lower outsiderness spread compared to the other welfare model: in the late 2010s, only 23.4% of the total workforce belonged to occupational classes that were more likely to be outsiders. On the contrary, in the DIG and NILG countries, the spread has substantially increased over time (+7.4 and + 12.3, respectively), and the values are high, especially in the Mediterranean cluster.

Regarding the outsiderness degree, values have scaled down in all four models. In the late 2010s, the NIG countries displayed the highest outsiderness degree (17.0), while in the DIG and NILG, the intensity of the risk is quite similar (14.4 and 16.2). However, the drastic drop in the Mediterranean cluster's value (−7.9) is partially due to a general increase in the workforce's unemployment/atypical work mean rate (shifting from 31.9 to 34.9%). Finally, the EIG model shows the lowest outsiderness degree.

Having illustrated how outsiderness configures in the four growth models, it is now possible to analyse its gender dimension. Overall, the empirical data shown in Table 3ab point out that outsiderness in Western Europe is strongly gendered. The gender spread and degree gaps both in the early/mid-2000s and late 2010s reveal that female occupational classes are disproportionally exposed to the risk of being unemployed and/or atypical workers compared to male social classes. Outsiderness is thus a feminized social phenomenon, i.e. a disadvantage mainly affecting women. That being said, cross-model differences can be identified.

6.1 The NIG/Anglo-Saxon model

The NIG cluster was the model where the gender spread gap rose the most (+35.8). Such an increase is only partially explained by a higher circulation of the outsiderness risk among the female workforce. The change seems mostly due to a significant improvement within the male labour force and more specifically of male BCs. In other words, at the beginning of the century, outsiderness was more equally distributed across the male and female workforces in the context of high social inequalities. The male BCs egress from their outsider status in the late 2010s and consequently led to a decrease in the male outsiderness spread, which is now concentrated among male LSFs. The high female spread depends on the fact that two of the three larger female occupational classes – clerks and LSFs – are outsiders. Even though these two classes are downsizing, the reduction is not sufficient to alter the outsiderness spread. Furthermore, while female TB and SBOs have increased their share in female social stratification, in the late 2010s, they were more exposed to the outsiderness risk compared to the workforce mean. New female jobs have (partially) emerged in two more disadvantaged sectors.

The increase in female outsiderness spread has been followed by a decrease in the female outsiderness degree (−7.9) and a consequent fall in the degree gender gap. While more women are likely to be outsiders, such a risk has become less severe, and its intensity is now closer to that experienced by male workers. Nevertheless, the female outsiderness risk remains higher than in the EIG and DIG groups.

To summarize, the de-gendering and upgrading trends in social stratification were not sufficiently strong to substantially alter the gendered dimension of outsiderness in the NIG model. The increase in the high-skilled female SCPs – who remain insulated from the outsiderness risk – was counterbalanced by an increase of two new outsider groups – female TB and SBOs. Male outsiderness regards only a single low-skilled group, the LSFs, which – though relatively large by a comparative standard – still represent only a minor fraction of the male workforce.

6.2 The DIG/Continental model

In terms of outsiderness spread, the DIG model emerges as the most gendered. Indeed, the spread gap is markedly high (65.4 in the early/mid-2000s and 51.3 in the late 2010s) – higher even than that shown in the NIG model – and its fall (−14.1) is primarily due to a worsening of the male workers' conditions rather than to an improvement in the female workforce. As a matter of fact, male outsiderness spread has scaled up by 11.4 points while the female counterpart has decreased by only 2.7 points, and in the late 2010s, almost 70% of the female workforce was exposed to the risk of unemployment and atypical work. The growth of the outsiderness spread among the male labour force can be explained by the deterioration of the LSFs' status: while at the beginning of the century, they were insulated by the labour market risks, they ended up as an outsider category at the end of the 2010s. On the other hand, the outstanding level of female outsiderness spread is due mainly to the fact that the three main feminized occupational classes – LSFs, MSFs and SCPs – are outsider groups. Finally, it is worth noting the status of the female SCPs: though high-skilled, they are vulnerable, suggesting that skills do not protect them from “gender penalization”.

The marked gendered characterization of the outsiderness in the DIG cluster also emerges when looking at the gender degree gap, which has increased over time. Both the female and male degrees have declined at the disaggregated level, but the fall has been more marked among the male labour force (−5.3 points) and less significant in the female occupational class.

Furthermore, the drop does not dramatically change the unequal condition between male and female workers: for the former, the outsiderness degree is superficial (below 4), while for the latter, outsiderness is severe (above 15).

The spread and degree values that have been observed in the continental countries mirror the lack of a real de-gendering trend within the social stratification and the stagnation of the upgrading process within the female labour force. Furthermore, the data suggest that a possible strengthening of female worker skills in the service sector might not be sufficient to reverse gender inequalities since high-skilled female SCPs remain an outsider group. The increase in the female employment rate in the DIG countries since the end of the 1990s has been achieved by promoting the one-and-a-half-earner family model – rather than the dual-earner one (Esping-Andersen, 2009) – and by developing a strongly deregulated service sector. This has relegated women to part-time and less secure jobs, thus gendering outsiderness.

6.3 The NILG/Southern model

The outsiderness spread in the NILG welfare model in the late 2010s was relatively and equally distributed across the male and female occupational classes. Consequently, the gender outsiderness gap records the lowest value among the four growth models (only 7 pp). However, this does not mean the spread is low: outsiderness is extensive and affects 46.7% of the male workforce and 53.9% of females. Furthermore, unlike the other models, there is an increasing trend: a wider share of female and male workers were affected by labour market disadvantage in the late 2010s compared to the early/mid-2000s. We can thus observe a sort of gender paradox: high and growing class inequalities go hand-in-hand with a less gendered characterization of outsiderness. This does not imply that women are insulated from the labour market risk: the lower gender inequalities mainly depend on the fact that a disproportional share of the male workforce is at-risk of outsiderness compared to the other three growth models. The men's disadvantaged position is explained when considering that in NILG countries, male BCs are an outsider group and that the growing LSF group worsened its positions in the late 2010s. These two classes represent altogether almost 50% of the male labour force. In general, in the NILG countries, outsiders belong to those occupational categories where female and male workers are more concentrated, regardless of skills. Furthermore, although the female outsiderness spread is lower compared to that recorded in the NIG and DIG countries, it is important to remember that in the Southern countries – and especially in Italy – female employment rates have been historically low, while the percentage of housewives is still relatively high. These countries have indeed been historically characterized by implicit familialism – i.e. there are neither publicly provided alternatives to nor financial support for family care – which has reinforced the male-breadwinner family model (Leitner, 2003). In other words, a vast share of the female population remains excluded from the labour force tout court: the female workforce is much narrower than that of the other growth models, and approximately, half of it is exposed to the outsiderness risk.

Regarding outsiderness degree, there is a declining trend when considering the female and male classes (−9.7 and −6.8, respectively). However, the gap remains high (8 points). The more equal distribution of outsiderness has been followed by a gendered severity of disadvantage: women experience a higher intensity of unemployment and atypical work risks than men – for whom vulnerability (12.2 in the late 2010s) is, however, higher than that displayed by the male workers in the other three models.

The downgrading and de-gendering trends observed in the social stratification of the NILG countries have led to a gender paradox for which – in a context of high social inequalities – outsiderness affects female and male workers quite indiscriminately. However, a substantial rate of women remains excluded from the workforce, and a high percentage of those included are exposed to a vulnerability risk. Furthermore, the disadvantage they perceive is much more severe than that of men.

6.4 The EIG/Scandinavian model

The EIG model shows a clearer gender-equality trend. With regard to the female outsiderness spread, the value was high in the early/mid-2000s (52), but it substantially decreased in the late 2010s (−18.3), while that of males grew (+2.0). Consequently, the gender gap decreased and is substantially lower than that recorded in the NIG and DIG groups. The fact that the figure is higher than in the NILG countries should not be misinterpreted, since the two contexts are very different. In the Scandinavian model, inequalities are much less spread within the workforce, and the female employment rate is historically high due to the institutionalization of the dual-earner family model (Leitner, 2003). Furthermore, here the drop in the gender spread gap is mainly explained by the improvement of the female workforce conditions and not merely by a worsening of the male workers' status. More importantly, in the Nordic countries, the occupational classes with the most female representation are not necessarily the most disadvantaged: in the late 2010s, only female LSFs were an outsider category, while the female clerks and the high-skilled, growing SCP group were insulated from the risk of unemployment and atypical work.

Also when considering the outsiderness degree, an improvement in gender inequality can be seen. The intensity of outsiderness has scaled down for women (−5.3), while it has remained stable for men. Even in this case, the gender degree gap has decreased primarily due to a real improvement in the female condition in the workforce. Women remain more disadvantaged than men (12.6 vs 5.2 in the late 2010s) but record the lowest scores of outsiderness degree among the four models.

The de-gendering and upgrading processes affecting the social stratification of the EIG countries have positively impacted the gendered dimension of outsiderness since they have managed to improve female workers' conditions rather than deteriorate those of males. The EIG labour market does still present pink occupational ghettos, but in the late 2010s, only a relative minority – though significant – group of female workers experienced disadvantage – the LSFs. The fact that female SCPs do not show a significant degree of vulnerability suggests that investment in skills does have an economic and social return, contrary to what happens in the DIG and NILG growth models.

7. Conclusions

This article has investigated whether and to what extent outsiderness is gendered in Western Europe, both in terms of its spread and degree. It has also explored whether such a gendered characterization has varied over time and across the four growth model. The empirical analysis shows that outsiderness is profoundly gendered in all the clusters of countries examined. Female occupational classes are disproportionally exposed to the risk of being unemployed and/or atypical workers compared to male social classes. Outsiderness is thus a feminized social phenomenon, i.e. a vulnerable condition mainly affecting women. However, the comparative investigation has also highlighted that the ways in which outsiderness has genderized over time have been diverse across the four growth models. In other terms, different varieties of gendered outsiderness can be identified. In the NIG model, de-gendering and upgrading trends in social stratification may be observed, but they were not sufficiently strong to alter the gendered dimension of outsiderness. In the DIG countries, we can refer to a sort of frozen landscape, since outsiderness has remained strongly gendered, both in terms of spread and degree, due to the lack of a real de-gendering trend in the labour force and the stagnation of the upgrading process within the female workforce. In the NILG model, the downgrading and de-gendering trends in the social stratification have led to a gender paradox for which higher social inequalities are combined with lower gender inequalities. Nevertheless, a high rate of women is a priori excluded from the workforce and those in the labour market are still highly exposed to a severe risk of outsiderness. Finally, the EIG countries show a more gender-equal trend since the de-gendering and upgrading processes in the social stratification have managed to improve female workers' conditions rather than simply deteriorate those of males.

The empirical analysis triggers some theoretical considerations.

First, a multi-dimensional approach to the study of outsiderness – thus focusing on both its spread and its degree – leads to a more fine-grained analysis of a multifaceted phenomenon, and this, in turn, enables scholars to grasp better the different impacts among the male and female occupational classes.

Second, a close connection with the social stratification literature – combined with a diachronic perspective – has helped detect those trends that may affect outsiderness over time and across countries. De-gendering and upgrading are two critical factors, albeit not always sufficient to alter the gendered dimension of outsiderness. Skill and education do not automatically lead to social returns in terms of better protection from labour market risks, as demonstrated in the DIG and NILG case studies. Similarly, gender convergence is ineffective if new female outsider groups emerge, as in the NIG model.

Third, the analysis of the gendered dimension of outsiderness should also consider the configuration and change of male outsiderness. The deterioration of the male workforce status can artificially diminish the gender gap, generating the false perception of an improvement in gender equality. However, such a process hides an increase in widespread social inequalities that do not alter the women's disparities in the labour markets, as stressed by the NILG case.

Future research could expand the current study in three complementary ways. First, the comparative profiling of the female outsider classes could be continued, focusing on whether they show a penalization also in terms of social rights (e.g. access to a complementary pension, unemployment benefits, childcare, etc.). Second, a household perspective can be added, thus assessing whether and to what extent resources and risk pooling among partners may offset the outsiderness risk. Third, studies can search for those explanatory packages of factors that may account for cross-model or cross-country variation of gendered outsiderness. These analyses could maintain a model-oriented perspective or adopt a national-oriented lens to identify possible intra-model differences and new cluster configurations.

Table 1

Oesch's 8 class schema disaggregated by gender

Traditional bourgeoisie	Technical (semi-) professionals	Associate managers	Socio-cultural (semi-) professionals
1. (Male TB)	5. (Male Tech)	9. (Male ASs)	13. (Male SCPs)
2. (Female TB)	6. (Female Tech)	10. (Female ASs)	14. (Female SCPs)
Small business owners	Blue-collar workers	Clerks	Low-service functionaries
3. (Male SBOs)	7. (Male BCs)	11. (Male Clerks)	16. (Male LSFs)
4. (Female SBOs)	8. (Female BCs)	12. (Female Clerks)	17. (Female LSFs)

Source: Authors' own work based on Oesch (2006) and Schwander and Häusermann (2013)

Table 2

Social stratification, early/mid-2000s – late 2010s

	NIG model			EIG model			DIG model			NILG model
	T1	T2	Change	T1	T2	Change	T1	T2	Change	T1	T2	Change
a) Female social stratification (% of total female workforce)
Female TB	1.5	2.9	1.4	1.3	1.2	−0.1	2.2	1.9	−0.3	2.1	3.1	1.0
Female SBOs	5.2	8.4	3.2	5.7	6.4	0.7	6.3	8.1	1.8	14.2	13.4	−0.8
Female Tech	2.8	4.0	1.2	4.2	5.9	1.7	3.0	4.4	1.4	2.0	2.8	0.8
Female BCs	5.6	6.5	0.9	6.6	3.4	−3.2	7.4	7.5	0.1	12.1	9.7	−2.4
Female AMs	15.7	14.1	−1.6	15.9	19.5	3.6	16.6	16.3	−0.3	10.1	12.7	2.6
Female Clerks	20.3	16.3	−4.0	12.6	10.3	−2.3	17.8	16.1	−1.7	14.4	14.0	−0.4
Female SCPs	17.2	18.9	1.7	20.9	24.3	3.4	19.5	19.3	−0.2	15.4	13.7	−1.7
Female LSFs	31.6	29.0	−2.6	32.8	29.0	−3.8	27.3	26.4	−0.9	29.7	30.5	0.8
TOT	100	100		100	100		100	100		100	100
(N)	2,295	1,687		4,375	3,446		5,459	6,072		2,072	2,749
Female Clerks + SCPs + LSFs	69.1	64.2	−5.1	66.3	63.6	−2.7	64.6	61.8	−2.8	59.5	58.2	−1.3
FemaleTB + SBOs + Tech + BCs	15.1	21.8	6.7	17.8	16.9	−0.9	18.9	21.9	3	30.4	29	−1.4
Note(s): T1 = early/mid-2000s; T2 = late 2010s
Source(s): Authors' own work based on ESS data, round 1–3 (for the early/mid-2000s); round 8–9 for the Late 2010s. Data have been weighted (anweight)
b) Male social stratification (% of total male workforce)
Male TB	4.9	4.4	−0.5	3.6	3.5	−0.1	3.4	3.1	−0.4	4.6	4.0	−0.6
Male SBOs	17.2	18.2	1.0	11.7	10.8	−0.9	10.2	11.5	1.3	21.2	19.2	−2.0
Male Tech	8.5	11.5	3.0	12.7	14.7	2.0	10.8	12.3	1.5	6.3	8.6	2.3
Male BCs	24.0	19.7	−4.3	31.8	25.8	−6.0	36.0	31.0	−5.0	34.6	32.5	−2.1
Male AMs	19.0	18.8	−0.2	16.2	16.4	0.2	15.5	16.5	1.0	9.3	10.3	1.0
Male Clerks	4.1	5.2	1.1	3.9	4.8	0.9	7.0	7.7	0.7	6.1	6.8	0.7
Male SCPs	6.0	6.2	0.2	7.7	9.5	1.8	6.6	7.0	0.4	5.5	4.4	−1.1
Male LSFs	16.2	16.0	−0.2	12.4	14.4	2.0	10.5	11.0	0.5	12.5	14.2	1.7
TOT	100	100		100	100		100	100		100	100	100
(N)	2,898	1,939		4,808	3,945		7,226	6,871		2,951	3,292
Male Clerks + SCPs + LSFs	26.3	27.4	1.1	24.0	28.7	4.7	24.1	25.7	1.6	24.1	25.4	1.3
Male TB + SBOs + Tech + BCs	54.6	53.7	−1.0	59.8	54.8	−5.0	60.4	57.9	−2.5	66.7	64.3	−2.4

Note(s): T1 = early/mid-2000s; T2 = late 2010s

Source(s): Authors' own work based on ESS data, round 1–3 (for the early/mid-2000s); round 8–9 for the late 2010s. Data have been weighted (anweight)

Table 3

Outsiderness in the four growth model, early/mid-2000s – late 2010s

a) Outsider occupational classes
	NIG model		EIG model		DIG model		NILG model
	T1	T2	T1	T2	T1	T2	T1	T2
Male TB	−20.5	−18.3	−15.8	−11.4	−23.3	−15.9	−27.9	−29.1
Female TB	−10.2	23.6a	−9.8	2.7a	−13.4	−17.5	−4.3	−23.5
Male SBOs	−31.7	−21.9	−18.5	−9	−26.2	−21.2	−23.6	−25.9
Female SBOs	0.6	2.5a	−7.1	−0.4	−6.9	−9.7	−19.5	−16.6
Male Tech	−7.4	−19.1	−12.2	−8.3	−16.7	−18.6	−8.7	−15.2
Female Tech	−7.5	−4.7	−4	−8.1	−2.1	3.9	−8.4	−9.7
Male BCs	11.9a	−1.5	3.9	1.8	2.3	−5.3	19.0a	12.9a
Female BCs	21.0a	21.0a	17.3a	10.8a	22.9a	16.1a	26.7a	22.3a
Male AMs	−10.3	−7.4	−12.9	−13.5	−21.3	−19.4	−14.9	−19.3
Female AMs	−10.3	−14.4	−10	−8.1	−3	−1.5	0.5	−13.2
Male Clerks	0.3	−8.4	−3.2	3.2	−8.9	−3.4	−13	−12.8
Female Clerks	22.1a	9.9a	9.8a	3.7	10.1a	7.5a	5.6	−1.6
Male SCPs	13.0a	−9.7	−5.6	2.9	9.1a	4.2a	7.6	−2.5
Female SCPs	1.7	−1.7	1.7	2	17.0a	16.0a	6.8	5.6a
Male LSFs	17.4a	13.9a	7.3a	7.4a	−1.8	9.1a	4.7	10.6a
Female LSFs	29.0a	25.7a	21.1a	13.2a	25.4a	22.5a	31.4a	26.3a
Total Entire Workforce	41.5	39.6	28.6	21.3	36.6	32.4	31.9	34.9
(N)	3,426	3,589	5,369	7,370	7,866	12,934	3,465	6,003
Note(s): Values are the difference between the group-specific rate of atypical employment/unemployment and the rate among the entire workforce (outsiderness). a: groups with significantly higher rates than the workforce average (outsiders)
T1 = early/mid-2000s; T2 = late 2010s
Source(s): Authors' own work based on ESS data, round 1–3 (for the early/mid-2000s); round 8 and 9 for the late 2010s. Data have been weighted (an weight)

b) Outsiderness spread and degree
	NIG model			EIG model			DIG model			NILG model
	T1	T2	Change	T1	T2	Change	T1	T2	Change	T1	T2	Change
Spread workforce	51.3	37.8	−13.5	31.3	23.4	−7.9	34.7	42.1	7.4	37.6	49.9	12.3
Spread female	57.5	63.1	5.6	52.0	33.7	−18.3	72.0	69.3	−2.7	41.8	53.9	12.1
Spread male	46.3	16.0	−30.3	12.4	14.4	2.0	6.6	18.0	11.4	34.6	46.7	12.1
Gender spread gap	11.3	47.1	35.8	39.6	19.3	−20.3	65.4	51.3	−14.1	7.2	7.2	0.0
Degree workforce	19.8	17.0	−2.8	15.7	10.9	−4.8	18.0	14.4	−3.6	24.1	16.2	−7.9
Degree female	25.8	17.9	−7.9	17.9	12.6	−5.3	19.1	16.5	−2.6	30.0	20.3	−9.7
Degree male	14.0	13.9	−0.1	7.3	7.4	0.1	9.1	3.8	−5.3	19.0	12.2	−6.8
Gender degree gap	11.8	4.0	−7.8	10.6	5.2	−5.4	10.0	12.72	2.7	11.0	8.1	−2.9

Note(s): T1 = early/mid-2000s; T2 = late 2010s

Source(s): Authors' own work based on ESS data, round 1–3 (for the early/mid-2000s); round 8 and 9 for the late 2010s. Data have been weighted (an weight)

Table A1

Table of operationalization

Variable	Operationalization (European social survey and aggregated datatsets)
Post-industrial social classes (Oesch's schemata)	isco08/isco88, emplrel, emplno, agea, gndr Stata script available at https://people.unil.ch/danieloesch/scripts/
Growth models	NIG countries: Ireland, Great Britain
	EIG countries: Denmark, Finland, Iceland, Norway, Sweden
	DIG countries: Austria, Belgium, France, Germany, Netherlands, Switzerland
	NILG: Italy, Portugal, Spain
Unemployment	Dummy variable measuring unemployment, recoded from mnactic
Atypical work	Dummy variable measuring combination of fix-term contract work, part-time work, and “no contract”, recoded from wrkctra and wkhct
Outsiderness	Dummy variable measuring unemployment and atypical work

Notes

1. The choice to consider clusters as the unit of analysis rather than single countries is widespread in the literature (e.g. Schwander and Häusermann, 2013; Schwander, 2019). Schwander and Häusermann (2013, p. 257) have highlighted that the cluster-specific and the country-specific operationalizations provide very similar results. However, we are aware that intra-cluster differences may occur and that within a specific country, regional differences may be relevant (Giuliani and Raspanti, 2022). From a methodological perspective, aggregating countries allows us to have a number of respondents sufficiently ample to measure the level of outsiderness for all the 16 social groups.

2. Schwander and Häusermann (2013) disaggregated social classes also by age (old vs young). In this exploratory study, we preferred to keep separate the gender and generational dimension of outsiderness; therefore, age was not considered.

3. We did not distinguish between voluntary and involuntary part-time since the ESS does not provide these data. Empirical research has highlighted that also voluntary part-time leads to lower social rights compared to standard employment.

4. See the Appendix for more information.

Appendix The gender dimension of outsiderness in Western Europe

Outsiderness operationalization has been applied to the ESS datasets.

We aggregated Rounds 1–3 for detecting outsiderness in the early/mid-2000s, while pooling Rounds 8–9 for investigating outsiderness in the late 2010s. Round 3 was added for the Liberal and Southern regimes since the number of respondents in the aggregated dataset for each social class was not appropriately broad (N ≥ 30, see Agresti and Finlay, 2008, p. 94).

The number of respondents in the aggregated datasets is sufficiently wide to measure the outsiderness at the welfare regime level.