Premises: Perspectives on the Welfare State

Abstract

This chapter provides a brief overview of influential debates within welfare state research. It then discusses selected functions of the welfare state that contribute to an understanding of the various effects that social policies can have on individuals.

What exactly do we deal with when exploring welfare states? Barr (2001: 4–5) notes several issues that complicate a definition: the different sources of measures generating welfare (employment, insurance, private financial contributions, et cetera), the way in which services are provided, and the issues assigned to the welfare state rather than another policy area. Obviously, the mere existence of some sort of social policy alone does not yet mean that we can speak of a welfare state (Titmuss 1974: 26) and the scope of the term has been discussed for decades now (among earlier contributions on how to conceptualise the welfare state is Briggs 1961). Regardless of the viewpoint, key elements of a welfare state include a strong emphasis governments place on the well-being of citizens (Esping-Andersen 1990) and the acknowledgement of social rights as a crucial part of democracies (Marshall 2000 [1950]; Garland 2014). Furthermore, Titmuss (1976: 14–15) highlights two different points of interest when researching social policies: the institutional organisation of social services and the perspective of those who receive benefits. While the first aspect is closely linked to the historical development of the structures in which a welfare state is embedded, the latter focusses on the needs of those who are protected by it. When studying the individual-level outcomes of welfare policies, both perspectives are equally important.

These different viewpoints give rise to very different definitions of the welfare state, ranging from a narrower emphasis on specific policy programmes to broad conceptions encompassing all mechanisms of social protection against market dynamics and outcomes (as summarised by Otto 2018b: 52). Even though a universal definition does not exist, a useful conception seems to be understanding the welfare state as “a mosaic, with diversity both in its sources and the manner of its delivery” (Barr 2001: 4). This is reflected in the literature, as scholars often focus on specific elements of the wide-ranging institutional net that can be attributed to the welfare state. For the purposes of this book, I will treat the term similarly, not focusing on the welfare state itself, but on its functions and consequences for specific social phenomena—in other words: the reasons why social policies are treated as independent variables in the first place.

In a very general sense, four terms are used in this book: welfare state, social policies, welfare regime, and welfare stateness. The term welfare state addresses the overall institutional set up, in which social protection is offered, organised, and financed. Social policies refer to specific measures, which make up welfare states and their legal basis. There can be social policies in countries that do not classify as welfare states, while the welfare state cannot exist without social policies (Esping-Andersen 1999: 33). Welfare regimes go beyond the welfare state and social policies, as they also incorporate what is sometimes called a ‘welfare residual’ or ‘welfare culture’. Here, not only the welfare provided by the state is included but also the contribution of market and family (Esping-Andersen 1999: 34–35). Lastly, welfare stateness describes how comprehensive a welfare state is. Although this is sometimes used synonymously to the term welfare generosity (Otto 2018a: 765), it is more general than a ranking of more or less welfare provision. Instead, it is frequently used to capture what is meant by the German term Wohlfahrtsstaatlichkeit (e.g. Öktem 2016: 2). This term is actually a very broad expression capturing the general idea that a commitment to social policies is an immanent feature of welfare states. Furthermore, the term emphasises governmental aspects (Schnabel 2017: 212). Even though social policies and the face of welfare states may change, welfare stateness remains a key feature of modern western democracies.

Before turning to the empirical operationalisation of different social policy arrangements as independent variables, this chapter discusses the welfare state from a more general perspective by tracing the evolution of welfare state research. It begins with a brief overview of influential debates in the relevant literature. Those debates are summarised with particular emphasis on how they conceptualise and operationalise the welfare state. The review is kept compact and does not claim to be exhaustive. Instead, I highlight issues that contribute to a better understanding of the research objective of this book, focussing on those debates that have contributed most to the conceptual and empirical approaches we encounter today. This is followed by a discussion of selected functions of welfare states that help to understand the range of impacts that social policies can have on individuals.

2.1 A Short History of Welfare State Research

The welfare state has been part of policy-making for almost one and a half centuries.¹ Otto von Bismarck was the first to introduce comprehensive forms of social insurance in case of sickness, accident, and old age and invalidity between 1881 and 1889 in Germany. In the following decades, similar programmes quickly emerged throughout Europe and many other parts of the world. Even though the temporal sequence seems to suggest it, the assumption that the growing popularity of social insurance schemes was due to imitation of the German case (as proposed by Briggs 1961) is contested. Industrialisation, democratisation, and other decisive developments shaped politics in many parts of the world during the late nineteenth and early twentieth century. This spurred debates about social responsibilities of the state in various countries during a similar time period but with very different rationales (as discussed by Kuhnle & Sander, 2010).

After the end of World War II began what is often termed the ‘Golden Age of the Welfare State’ (e.g. Ferrera 2008). The post-war decades were characterised by fast expansion of welfare policies in the Western World. Benefits now covered entire populations and grew more generous. This development came to a halt in the 1970s when the two oil crises hit the industrialised countries. Further economic, structural, and demographic challenges eventually led the welfare state into crisis from the 1980s onwards, which caused cutbacks and long-lasting transformations in the 1990s and 2000s (the post-war developments of welfare states are discussed in more detail by Kuhnle & Sander 2010; cf. also Bonoli & Natali 2013b). Adding to these challenges, the labour market underwent severe changes in the last decades with growing atypical employment and inequality in the distribution of vulnerability among different population groups (Häusermann et al. 2016: 1047). In opposition to the ‘Golden Age’, such changes led scholars to refer to the contemporary state as a ‘Silver Age’ of austerity (e.g. Taylor-Gooby 2002).

There is a variety of possible approaches to classify sub-phases of this historical trajectory. For the purpose of this book, the most relevant ones are those that left a distinct mark on comparative welfare state research today. Three phases were especially influential in my perception of the literature. (1) The exploration and description of welfare states as a necessary condition of modern democracies shaped the literature until the late 1980s. (2) In the 1980s and 1990s, comparative welfare state research grew in importance as an academic discipline—addressing crisis and future of welfare states on the one hand and exploring patterns and shared paths on the other. (3) Contemporary research focusses especially on welfare state reforms and retrenchment.

These phases of research on the welfare state reflect the actual historical developments in western democracies outlined before but exhibit some delay. Thus, even if the historical perspective nowadays agrees that retrenchment started in the 1980s, research at that time was still focussed on other aspects. Actual and potential crises of the welfare state were explored but the systematic withdrawal of the welfare state was acknowledged more than a decade later (Pierson 1996 is one of the more influential publications in this line of research).

2.1.1 First Phase: Exploration and Description

The inseparable connection between democracy and the welfare state in the post-war years quickly spurred scientific interest in social policies. The focus of theoretical approaches was to explore the nature and importance of social policies (Marshall 1961; Titmuss 1974, 1976; Marshall 2000 [1950]). More empirical approaches aimed at describing specific welfare states in case studies and historical descriptions (e.g. Briggs 1961), while others focussed on conditions for the emergence of specific policies and correlates that drove these developments based on aggregated data (e.g. Cutright 1965; Tompkins 1975).

Two conceptualisations of the welfare state received particular attention. In the first, the welfare state is analysed in terms of effort (e.g. Tompkins 1975; Wilensky 1975). In order to describe and compare social policy arrangements, scholars thus focussed especially on social expenditure, which was perceived as the best indicator for the amount of work a welfare state puts into the well-being of its citizens. One of the key assumptions and empirical finding was that industrialisation leading to economic development would be a universal motor for the growth of social expenditure and population-coverage of social insurance schemes. Such interdependency between economic growth and the welfare state was supported by the Keynesian assumption that increasing public welfare would soften economic fluctuations (Quadagno 1987: 110). Since these processes affected all western capitalist democracies in the post-war decades, a certain degree of convergence between industrialising nations was expected. However, such hypotheses about the convergence of social policies did not prove to be true (as is discussed in more detail by Skocpol & Amenta 1986: 133–134). Instead, a variety of welfare states emerged following different principles and shifting the focus of scholarship on the matter from economic forces to political institutions (cf. Myles & Quadagno 2002) inspiring growing commitment to the power research perspective and new institutionalism (Pierson 1996).²

Such different theoretical viewpoints have manifested in discussions about the empirical conceptualisations of welfare stateness. From an empirical point of view, restricting the explanation for variations in welfare commitment to singular features (like economic development), has met criticism, since other political and social factors were viewed as equally important (e.g. Castles & McKinlay 1979; Esping-Andersen 1990: 19). In addition, social expenditure was criticised as being an inadequate and oversimplified proxy for the multidimensional nature of welfare state institutions (Korpi 1989: 310).

A second conceptualisation of the welfare state highlights social rights. This approach is based on Marshall’s essay ‘Citizenship and Social Class’ (2000 [1950]). He describes social rights as key features of citizenship, parallel to civil and political rights and conceptualises the welfare state in terms of benefits it offers to citizens based on those rights. In this context, Marshall also highlights the link between benefits and social inequality. He identifies four factors influencing the degree of equality in modern welfare states of his time:

[…] whether the benefit is offered to all or to a limited class; whether it takes the form of money payment or service rendered; whether the minimum is high or low; and how the money to pay for the benefits is raised (Marshall 2000 [1950]: 39).

This differentiation introduces two important aspects, which influence the discussion about the nature and measurement of social policies until today. First, Marshall distinguishes entitlement criteria, mode of benefits (income replacement vs. services), generosity of benefits, and financing of benefits. This suggests a much more differentiated way of potentially capturing differences between welfare states compared to simply focussing on welfare effort. Second, by adding the link to equality, Marshall emphasises the relationship between welfare provision and social inequality based on social class. Thus, different modes of income replacement and other social services potentially shape inequalities in different ways. This points to a function of welfare states that goes well beyond providing social security, as it frames it as a system of social stratification (this will be discussed in more detail later).

Empirically, the social rights perspective was first measured through social expenditure as well. However, even controlling for recipient population, this indicator was never able to adequately model the different internal logics of redistribution in different countries (Stephens 2010: 515). Thus, expenditure could not capture Marshall’s conception of “what is the core of the modern welfare state—the extent and quality of the social rights that constitute social citizenship” (Korpi 1989: 310). Perhaps the most influential contribution to a more elaborate operationalisation has been the Social Citizenship Indicator Program (SCIP), which Walter Korpi started at Stockholm University in the early 1980s. It provides detailed data on an array of different indicators of social rights in 18 OECD countries from 1930 until 2005 (Korpi & Palme 2008). This data constitutes the basis for a variety of publications, which have had a very strong impact on welfare state research until today (e.g. Korpi 1989; Esping-Andersen 1990). Other datasets and operationalisations followed this tradition (e.g. Scruggs 2006).

In addition to such attempts to pinpoint and operationalise key aspects of the welfare state (effort and social rights), there were also early contributions dealing with the exploration of different types of social policy arrangements in a cross-cultural comparative perspective. Two ideas were especially influential: (1) the demarcation of different extreme poles of contribution and financial organisation and (2) the classification of multi-faceted ideal-typical social policy arrangements.

The first tradition most prominently distinguishes between what is called the Bismarckian and the Beveridgean social system.³ Named after Otto von Bismarck and William Beveridge as key figures in the establishment of very typical ways of organising and financing social policies, these terms are quite established in the literature until today. The key feature of Bismarckian welfare states is that they are based on social insurances. Tied strongly to employment, they are financed through contributions by employers and employees and benefits relate to previous income. Contrary to that, the principle of the Beveridge system is universal and independent of employment. Benefits consist of flat rates and—even though this may not be in line with Beveridge’s initial ideas—are tax-financed and either universal or means-tested (Bonoli 1997: 357).

In the second tradition, different types are identified on the basis of additional aspects, like the distinction between market, family and state. An early example for an approach of that kind is given by Titmuss (1974: 30–32). He distinguishes three different models of social policy-making. In the Residual Welfare Model of Social Policy, private markets and the family constitute the main providers of welfare. The state only intervenes if those main providers fail to meet individual needs and it does so only temporarily. In contrast, in the Industrial Achievement-Performance Model of Social Policy needs are compensated based on the individual performance and productivity at work. Lastly, the Institutional Redistributive Model of Social Policy grants universal services, which are independent of the market and needs testing. While these types incorporate some elements of the distinction between Bismarckian and Beveridgean policies, they go beyond this dichotomy by adding aspects (like the role of the family) and offering combinations of Bismarckian and Beveridgean elements. Such attempts to find similarities and shared paths in comparative analyses of the welfare state, which—as Titmuss puts it—“help us see some order in all the disorder and confusion of facts, systems and choices concerning certain areas of our economic and social life” (1974: 30) grew especially popular in the 1990s. Thus, they will be discussed in more detail in the next section of this chapter.

2.1.2 Second Phase: Similarities and Classifications

The results of economic, demographic and other societal crises during the 1970s and 1980s were reflected in the literature of the period. They inspired more academic interest in the subject and led to a considerable growth of comparative welfare state research as a discipline within the social sciences (Skocpol & Amenta 1986; Olsson 1987). At the same time, this increased the need for clearer conceptual work on how to explain and measure variations of social policies between countries and periods (Lockhart 1984). As Skocpol and Amenta conclude in a review article (1986), aggregated quantitative analyses reached a limit and new insights had to come—in their view—from more in-depth comparative-historical analyses of specific cases or trajectories. Another influential impetus came from the aforementioned approaches to operationalise the welfare state as a multidimensional phenomenon by focussing on a variety of indicators instead of relying solely on social expenditure (e.g. Korpi 1989).

Indeed, comparative welfare state reseach in the 1990s focussed on more complex conceptualisation of welfare states and on historical analyses. In addition to case studies, this research intention was implemented in attempts to classify archetypical systems of welfare provision. Drawing on the above-mentioned ideas of differentiating distinct aspects of social policy arrangements (by for instance adding the social rights perspective), this period contributed a vast amount of proposals to capture different (ideal-) types of welfare states. While the focus of research on social policies had previously been on effort—the so-called “how much dimension” (Bonoli 1997: 352)—these attempts acknowledged that welfare states are multi-facetted and require more complex measurements than simply using expenditure-based indicators (Esping-Andersen 1990). However, in contrast to measuring the different facets through composite indices or a variety of single indicators, many of the studies aimed at proposing a typology of welfare states.

One of the most influential contributions to the literature on classifications of welfare systems is Gøsta Esping-Andersen’s Three Worlds of Welfare Capitalism (TWWC) (1990). Based on a historical discussion and a very comprehensive database, he clusters 18 OECD countries into three welfare regimes. Being a former member of the team working on the Social Citizenship Indicator Program (Korpi & Palme 2008), he includes Marshall’s (2000 [1950]) ideas regarding social rights of citizenship. In order to capture social rights and their connection to social inequality, he focusses on the degree of decommodification and social stratification in the examined countries. He defines decommodification as the extent to which individual dependency on labour market participation is reduced through social policies (Esping-Andersen 1990: 21) and the ability to uphold a “socially acceptable standard of living independent of market participation” (Esping-Andersen 1990: 37). Social stratification captures how far the organisation, distribution and financing of social policy reduces, reproduces or increases inequality. This again draws on Marshalls concept, in which—using Esping-Andersen’s words—the “status as a citizen will compete with, or even replace, one’s class position” (Esping-Andersen 1990: 21). These different dimensions are measured with the help of an array of different indicators, including a decommodification score⁴ and three stratification indices.⁵ In addition, Esping-Andersen includes the role of state, market and family in the provision of welfare (although the latter is more prominently discussed in his later publications).

His theoretical and empirical analysis yields three types of welfare regimes, which show some resemblance to the one proposed Titmuss (1974). The first is the Social Democratic (or Socialist) regime, which is characterised by generous benefits, a high degree of decommodification and a prominent role of the state as the main provider of welfare. A strong focus on universalism and equality as guiding principles means that this type of welfare state reduces inequality. Many Nordic countries resemble this ideal. In sharp contrast, the second regime emphasises the market. In this Liberal regime, decommodification is low, benefits and services are provided based on means-tested targeting following the objective of poverty-relief. Due to strict eligibility criteria and less generous benefits, inequalities are potentially even increased in this type of welfare regime. Anglo-Saxon countries show resemblance with this type of social security system. Lastly, the Conservative (or Corporatist) welfare state is based on social insurance tied to labour market participation. It follows the principle of status preservation and emphasises the role of the family. Germany serves as an example for a country closely resembling this type.

What followed was an era one might term typology enthusiasm, largely marked by the introduction of various typologies of welfare regimes in the 1990s and early 2000s, inspired by Esping-Andersen’s seminal study. His intuitive threefold classification quickly inspired what Abrahamson (1999) memorably dubbed ‘welfare modelling business’. The Three Worlds turned into a true classic over the last decades (Emmenegger et al. 2015), although the approach has also drawn criticism addressing both conceptual (Orloff 1993; Room 2000; Kasza 2002) and methodological (Bambra 2006; Scruggs & Allan 2006) issues. Adjustments led to even more classifications, introducing criteria like defamilialisation (Esping-Andersen 1999), additional types such as a Southern European type (Leibfried 1992; Ferrera 1996) or the extension of a typology to countries which were not previously classified or classified ambiguously. As a result, we are confronted with an enormous body of literature offering a great number of distinct typologies and various summarising contributions (e.g. Arts & Gelissen 2002; Bambra 2007; Ferragina & Seeleib-Kaiser 2011; van Kersbergen & Vis 2015). In light of the importance of welfare state typologies for research until today and their frequent use as an explanatory variable, a more elaborate discussion of problematic issues is necessary.

The theoretical foundation of most typologies focusses on broad similarities in the historical genesis of social security systems which are depicted as path-dependent—in line with Esping-Andersen (1990) and earlier historical institutionalist approaches. In this context, the terms ideal type and real type or typology are often distinguished (e.g. Kvist 1999; Aspalter 2011; Rice 2013)—ideal type usually meaning an archetypical way of organising and financing welfare provisions that serves as a guideline for interpreting actual policy-making. In contrast, real types refer to clusters of countries that are based on similar patterns determined purely empirically based on indicators of actual policy-making. When considering how such types—regardless of being ideal or real—are translated into empirical classifications, a great variety concerning both indicators and methods can be found. These differences will be discussed in more detail in the next chapter. For now, suffice it to summarise that they concern a multitude of relevant aspects: the selection of indicators on which a typology (of real cases) is based, the methods applied to determine clusters of countries, the sample of countries that is analysed and the conceptual premises chosen to interpret results. In light of the severity of these differences in the theoretical conceptualisation and empirical operationalisation, it does not surprise that even though almost all scholars since the early 1990s base their work on Esping-Anderson’s Three Worlds, the number, title, and composition of the regimes differ significantly. Even if we accept that comparable regimes are named differently,⁶ barely any country has been attributed to the same type in all studies.

Regardless of the critical debated surrounding welfare regime typologies (Aspalter 2011; van Kersbergen & Vis 2015), taking a meta-perspective on the literature reveals that there are also some common elements in the various different typological approaches. In an attempt to systematise the literature, Ferragina and Seeleib-Kaiser (2011) assess the overlap between Esping-Andersen’s TWWC typology and 22 other prominent typologies published between 1990 and 2009. They find that although there are often more than the initial three types, most countries are attributed to either the Liberal, the Social-Democratic or the Conservative regime in more than 50 percent of the examined studies.⁷ The most frequent addition to the threefold classification is a Southern (or Mediterranean) type. The resulting fourfold classification seems to be a somewhat robust finding—at least until a decade ago. However, Central and Eastern European countries remain entirely ignored. Reviews including more recent typologies and post-socialist countries are rare and often focus on the country selection instead of the classification (Kim 2015). In general, there are not many new typologies introduced after 2010. The few attempts that were proposed, used very different perspectives and indicators—such as aggregated individual welfare attitudes and values (Vrooman 2013) and aggregated data on policy outcomes (Ferragina & Seeleib-Kaiser 2015). Perhaps, the TWWC and related regime approaches have indeed lost some of their lustre.

This may be one of the reasons, why enthusiasm subsided recently. Instead of offering new classifications, the existing theoretical and empirical typologies are more intensively examined for their usefulness as analytical tools (Reinprecht et al. 2018). Still, even those scholars contesting the prevalent conceptual strategies agree that typologies remain a popular analytical tool and that Esping-Andersen’s premises should be transformed but not rejected entirely (Rice 2013). Others argue in favour of concentrating on distinct areas of the welfare state, which means for instance separating healthcare systems and educational systems from overall typologies of welfare stateness, as they follow different objectives and are less related to decommodification and other key elements at the heart of ideal-typical welfare states (Ferragina & Seeleib-Kaiser 2011: 587). Following this line of arguments, these systems inspired their own distinct typologies (cf. Beckfield et al. 2013).

Only few authors take a more radical view, even suggesting that research which draws on Esping-Andersen’s original typology starts to show signs of Kuhnian normal science since follow-up research always tests and contests previous findings within the existing paradigm (van Kersbergen & Vis 2015). Indeed, studies without any reference to Esping-Andersen are rare and usually address a very different perspective like Künzel’s (2012) analysis of sub-national variations in welfare provisions. However, the sources of variation discussed and the resulting lack of comparability are quite consequential when typologies are used as independent variables. Thus, they will be quite important in the following chapters.

Overall, the old classifications distinguishing a Liberal, Social-Democratic, Southern and Conservative type are still seen as useful tools, albeit the fact that some of the foremost archetypical countries (such as Germany and Sweden) move continuously further away from the corresponding ideal types (Reinprecht et al. 2018).

2.1.3 Third Phase: New Risks and Retrenchment

Contemporary research on the welfare state is characterised by a strong emphasis on change. Such change manifests in two ways. First, economic and social structures are changing, leading to new challenges for welfare states, which in consequence have to adapt. As Gilbert puts it:

It is not demographic factors or tax ceilings, globalization, or the normative changes shaped by knowledge and experience with social policies or the rising faith in the market economy that by themselves account for the fundamental change in the character of social protection; rather, it is the combination of these forces. (Gilbert 2004: 42)

Second, adapting to such new challenges mainly manifests in two ways: cutbacks—in other words retrenchment of the welfare state (Pierson 1996)—and transformation such as reallocation of responsibility to the individual and other responses to new risks (Bonoli & Natali 2013a). Overall, there are two main perspectives on a similar subject: one focusses on external factors driving change and one on the way welfare states adapt internally.

Even though there are still attempts to classify differences between welfare states in typologies (cf. Ferragina et al. 2015), the perspective on the welfare state has shifted. While different trajectories of social policies have been viewed as path-dependent and thus somewhat stable during most of the twentieth century, a key focus of recent welfare state research is on the politics of the new welfare state as Bonoli and Natali (2013a) title an edited volume on the matter.

The first phenomenon receiving much attention is retrenchment of the welfare state (Pierson 1996, 1998). Retrenchment addresses cutbacks in welfare benefits and services. As Pierson argues, such cutbacks are unpopular and thus sometimes disguised within structural shifts including more means-testing, growing individual responsibility to provide one’s own security through private insurance schemes and changes in eligibility rules (Pierson 1996: 157). Indeed, the majority of research agrees that cutbacks have taken place in most advanced welfare states, which do not primarily show up in a reduction of expenditure but in other areas such as replacement levels (for a review cf. Starke 2006).

However, cutbacks and changes in how the welfare state is financed and how resources are redistributed represent only one side of the politics of the new welfare state. There is also growing complexity in the risks and needs welfare states are confronted with. Male breadwinners not being able to generate income was the main risk in post war welfare states. This changed considerably. Social and economic transformations led to a wide array of new risks and needs which have to be met and partly appear regardless of employment (Bonoli 2005, 2007). As Zutavern and Kohli (2010: 175–176) summarise, new needs and risks include changes in the labour market, such as tertiarisation, feminisation and flexibilisation, as well as changes in life courses and life forms, including longevity, fertility and family stability. Overall, such changes shifted the focus of social policies from income replacement to the promotion of labour market participation (Bonoli & Natali 2013b: 5–6) and individual autonomy and responsibility (König 2017). This goes along with growing attention devoted to new risk groups. The compatibility of family and employment (especially for women) and the dualisation of labour markets, producing ‘insiders’ who continue to enjoy traditional social policy provision and ‘outsiders’ who face new insecurities (Bonoli & Natali 2013b: 8), are only two issues receiving growing attention in this field. In addition to changes in the ‘old’ agendas of social protection, the ‘new’ demands thus introduce efforts in other areas. This includes activating policies increasing labour market participation, which gain importance relative to passive policies focussed on income replacement. The changing objective of welfare policies embedded in this trend is sometimes referred to as moving from a welfare state towards an enabling state (cf. Gilbert 2004: 44).

Within approaches highlighting such activating policies, social investment is a policy agenda, which received growing interest during the last decade. From a political point of view, it is seen as a promising way of grasping answers to the new challenges of welfare states (Kuitto 2016: 442). Broadly speaking, social investment refers to social policies, which target the above-mentioned new challenges and risks. Instead of focussing on income replacement, social investment incorporates policy measures, which enable individuals to take responsibility for their own welfare. Examples for such efforts being investments in human capital and education (cf. Andersson 2018: 109) as well as childcare provision (León 2017). A main agenda of social investment is to cater to the new rationales of social security, which promote individual over collective responsibility (cf. Ellison & Fenger 2013). As Vaalavuo (2013: 516) points out, the focus shifted from a redistribution of income to a redistribution of opportunities. However, it is important to emphasise, that this does not mean, that the former necessarily decreases or counteracts the latter. Instead, preventing risks (investment strategy) and compensating risks if they materialise (protection strategy) overlap and potentially aid each other (Vandenbroucke & Vleminckx 2011: 451).

A last field, which is strongly related to new risks and risk groups, can be found in research on the relationship between family and welfare state. These debates emerged out of criticism of old approaches in comparative welfare state research, which underemphasised the role of the family (cf. Orloff 1993). A key concept in this line of research is defamilialisation (and familialisation as its counterpart). Analogous to decommodification, this term captures the extent of individuals’ (in-)dependence from the family (for an overview of the debate see Lohmann & Zagel 2016). Defined more specifically, defamilialisation refers to

the degree of women’s autonomy from the family (e.g. the spouse) in achieving financial resources and the degree to which women’s unpaid work in the family, particularly unpaid caregiving, is substituted by paid labour from outside the family by means of public, market or third-sector services (Keck & Saraceno 2012: 454).

Summarising the developments above, we can identify two dominant topics in recent literature on welfare states in Europe. One addresses retrenchment of the welfare state and asks whether austerity emerges as a stable characteristic. The other focusses on new strategies in social policy-making (such as social investment) and explores the potential of such new emphasis in social protection.

In line with the growing complexity of welfare states and the risks they respond to, research on the welfare state grew more complex as well. One of the biggest differences compared to previous strands of research on the welfare state is the detachment from historical path-dependence. Welfare states do not seem to expand further in their previous trajectories. A significant part of the literature identifies retrenchment as a stable and rather universal development in western capitalist democracies (Bonoli & Natali 2013b), which challenges the old conception of distinct welfare cultures or regimes.⁸ The same holds true for a growing relevance of new risks and strategies to deal with them. Thus, the changing nature of social policy arrangements and objectives raises methodological questions as old operationalisations (like the identification of distinct welfare regimes) are challenged. In particular, the literature on welfare state change and retrenchment has not only introduced new conceptual perspectives, but also triggered a methodological debate about the selection of suitable measurements of social policy-making—the so-called dependent variable problem (e.g. Clasen & Siegel 2007). While researching change, scholars observed the lack of a common understanding of what the object of research—the welfare state as dependent variable—entails and how it should be measured (Green-Pedersen 2004). As a result, there is an ongoing discussion based on the repeated observation that different conceptual and operational strategies lead to different results (cf. Kühner 2007; Bolukbasi & Öktem 2018). Similar discussions have emerged in the literature on social investment. While this strand of literature initially relied dominantly on expenditure data, the demand for more nuanced operationalisations was raised more recently (cf. Andersson 2018).

For the purpose of this book, these strands of literature are relevant, as they provide and test indicators, which are taken up as independent variables. Within the retrenchment debate, especially policies related to old risks are highlighted. In particular, three perspectives on welfare states and social policy provision exist.⁹ The first focusses on expenditure (e.g. welfare effort), the second on social rights, and the third on benefit receipt. While the first two operationalisations represent well-established perspectives on social policy arrangements, the focus on benefit receipt is comparatively rare (van Oorschot 2013: 225; Otto 2018a). Regarding the operationalisation of effort, reducing the operationalisation to measures of expenditure is still criticised continuously (following the arguments discussed in the previous sections of this chapter). Focussing on social rights, including benefit entitlement and generosity, is thus a popular alternative.¹⁰

Figure 2.1 illustrates the relationship between the three components according to van Oorschot (2013: 230). Here, social rights define the access to benefit receipt in specific benefit areas, while social expenditure represent the cost associated with benefit provision and take-up. Using van Oorschots terminology, social rights thus represent policy outputs, while benefit receipt signals social outcomes and expenditure cost outcomes.

Figure 2.1

(Slightly modified version of schema by van Oorschot 2013: 230)

Social rights, benefit receipt and social spending.

The methodological debates emanating from the dependent variable problem have taken up arguments and perspectives—such as the differentiation between social rights and welfare effort—from earlier welfare state research. Nevertheless, they also offer important additional insights due to a very strong focus on the operationalisation and selection of very specific indicators and their treatment in empirical analyses. This leads to detailed discussions about the measurement of particular aspects of social policy-making and specific indicators. An example for this is the very detailed discussion about the measurement of net replacement rates (Scruggs 2013; Wenzelburger et al. 2013).

While the dependent variable problem refers mainly to perspectives related to old risks, the measurement of policies related to new risks and risk groups, is partly detached from these approaches. Especially the field addressing family policies and gendered outcomes works with their own sets of operationalisations (Keck & Saraceno 2012; Lohmann & Zagel 2016). There are several reasons, why research on family policies follows a somewhat different path than research on the old risks. First, this strand of literature emerged later and—as noted above—out of criticism of the oversight of family care responsibilities in prominent approaches in welfare state research. Second, enabling families to combine parenthood and employment is a comparatively new agenda in social policy-making and emerged partly diagonally to old paths within welfare states. This manifests among other things in contradictory policies within a welfare state—as discussed by Lohmann & Zagel (2016: 48–49) using the example of German family policy, which encourage female care responsibility through financial benefits and promote female labour market participation at the same time. Still, while indicators of (de-)familialisation are usually provided separately from indicators of the old welfare state functions, they show many similarities. Like in the case of old risks (esp. unemployment, sickness, and old age), indicators of eligibility, access and generosity exist (Keck & Saraceno 2012). Similarly to the already described methodological approaches, these indicators are used as single measures, as well as combined in indices (Lohmann & Zagel 2016).

In addition to a focus on the particular field of family policies, contributions on other responses to new risks such as social investment and active labour market policies also spawned methodological debates. This led to various types of measurement, including typologies of active labour market policies (Bonoli 2012) and spending on fields tied to investment, which emphasise activating components (Kuitto 2016: 447–448). However, as noted above, the discussion of how to operationalise social investment—especially going beyond expenditure—is ongoing (Kuitto 2016; Andersson 2018).

Summing up the main debates in more recent comparative welfare state research, the literature on welfare state change and retrenchment has led to new perspectives and empirical challenges. It has inspired a detachment from analyses of historical trajectories and led to a stronger focus on operationalising relevant features of welfare state change. Several different perspectives on the welfare state have been highlighted in more detail and provoked new debates about how to measure them. Thus, while they inspired many substantial contributions on the matter, recent debates in research on measuring welfare states are still far from agreeing on a best-possible solution to measuring the welfare state (cf. Otto 2018b). There is a great number of contributions addressing methodological issues and proposing specific ways of operationalising different welfare state arrangements. Nevertheless, if there is a trend in comparative welfare state research at the moment, it seems to be one towards complexity.

2.1.4 Summary: Influential Debates

This section briefly traced influential debates in research on the welfare state since the 1950s. Throughout the last 80 years, a variety of different perspectives can be identified (a summary of key premises, debates and conceptualisations is provided in Table 2.1). Each of the perspectives offers distinct proposals for the empirical operationalisation of the welfare state or specific dimensions of social policies. Since all of them are taken up by research that includes the welfare state as an explanatory variable, the origins and premises of the different methodological approaches are important for the following evaluation of their applicability as macro-level independent variables.

Table 2.1 Summary of key premises, debates and conceptualisations

Overall, we can distinguish two different premises when it comes to the conceptualisation and empirical measurement of social policies. The first aims at grasping the welfare state as a whole, while the second focusses on specific dimensions of welfare provision. While the first conceptualises the welfare state as a multidimensional phenomenon (in a typology or composite index), the second highlights specific social polices either as proxies for overall welfare orientation or as singled out issues relevant for specific perspectives on the welfare state. The different focal points, debates and conceptualisations summarised in Table 2.1 are not restricted to one period even though they emerged roughly in the described order. Until today, we find conceptualisations highlighting welfare effort, as well as typological approaches following the TWWC.

2.2 Functions of Welfare States

The debates outlined above have shaped the theoretical and empirical conceptualisation of welfare states to this day. While these debates help to understand what welfare states are, how they have evolved, and why there are differences and similarities, this still does not reveal how different welfare states affect societies and individuals within them. Exploring these processes is quite important to the research objective of this book, as the tasks and goals of welfare states and differences in how they are approached shape individual outcomes. As such, they are the reason why features of the welfare state are treated as independent variables in the first place. The literature offers various approaches to characterising welfare states’ objectives and functions. While functionalist theories focus on the relationship between economy and social policies, institutionalist and political approaches highlight the importance of political and administrative institutions and actors (for a more comprehensive summary of these theoretical perspectives cf. Myles & Quadagno 2002). All of these approaches offer important perspectives. However, when asking how the welfare state interacts with individuals the last two are especially insightful. Here, the institutionalist perspective offers insights about processes within the welfare state, while the political approach helps contextualise political systems and channels of policy feedback.

In this section, I will briefly summarise a selection of important ways in which social policies influence individuals and highlight those that I believe are very relevant to contemporary research treating the welfare state as an explanatory concept in the social sciences.¹¹ The term function is not meant here in the sense of a quasi-automatic reaction to policies, but as a mode of action. The discussion remains at a general level, focussing on how different welfare states are linked to individual outcomes before these links are gradually fleshed out in the course of this book.¹²

Overall, five key functions of the welfare state receive particular attention. This includes the functions of (1) providing security, (2) (re-)distributing resources, (3) shaping social stratification, (4) enabling and incentivising, and (5) socialising individuals. They incorporate processes that shape individual outcomes directly or indirectly and thereby provide the groundwork for the conceptualisation of the welfare state as an explanatory factor in following parts of this book (especially in chapters 4 and 5).

2.2.1 Security: the Welfare State as a Safety Net

The manner and efficiency of welfare states’ responses to needs—such as a need for income or a job—and risks—such as the risk of poverty or unemployment—tell us much about their nature. Needs are met foremost by protecting against risks (Zutavern & Kohli 2010: 169). In this sense, meeting needs by avoiding risks or moderating their outcomes is the key function of the welfare state—one may call it a ‘minimum’ function (Eger & Breznau 2017: 441). The main risk to be avoided is poverty. In its most basic form, this function is provided by social assistance schemes. Such—targeted or universal—basic transfer payments aim at avoiding the direst manifestations of need, such as poverty (Garland 2014: 342).

As more contemporary literature points out, many changes in social policies may be attributed to severe changes in needs and risks—as was outlined in the previous part of this chapter. Furthermore, it is argued that addressing risks has become more important than responding to needs, as the latter are less profound in contemporary societies (Kemshall 2002). In addition, what actually counts as legitimate need is subject to social and political deliberation (Sachweh 2016). There is however evidence, that new risks are not covered as well as old ones by most welfare states. Following contributions on the dualisation of labour markets (e.g. Emmenegger et al. 2012), changing risks have created new risk groups (e.g. like single motherhood or youth unemployment), which slip through the net when it comes to the old way of social protection (Brady et al. 2017: 771). When asking how and why welfare states react to risks and needs from a contemporary perspective, it is thus important to bear in mind the old as well as the new risks.

Figure 2.2

Links between the welfare state and risks and needs

In general, risks are distributed unequally amongst individuals, based on class, age, and gender (Esping-Andersen 1999: 32). The way in which the welfare state responds to them, thus actively shapes individual vulnerability as expressed by their socio-economic position and—as a result—patterns of social stratification on an aggregate level. The latter will be addressed in more detail in a proceeding section of this chapter (cf. Section 2.2.3). Here, the focus is on the aim of lowering individual risk. Pinpointing the role and function of the welfare state underlying its relationship to risks and needs can be best summarised as providing security. This objective of meeting risks and needs ties the welfare state to the individual in different ways, which are illustrated in Figure 2.2.

• 1. Lowering risks directly: there is a direct effect of social policies on individual risks. In this sense, welfare states can shape how many individuals are actually at risk. Such a link is represented by two arrows in Figure 2.2 as risks can be independent as well as dependent variables. For example, if unemployment is perceived as a factor increasing the risk of poverty and the need for new employment or income replacement, policies reducing the prevalence of unemployment (a) and policies responding to poverty (b) are equally relevant.

• 2. Lowering risks indirectly: an indirect effect of social policies means that the link between welfare state and an outcome (c3) is at least partly mediated by another variable. Taking up the above-mentioned example, the welfare state might reduce poverty by lowering unemployment (c1), which is a predictor of poverty (c2).

• 3. Moderating the impact of risks: there are also perspectives, in which social policies serve as moderators (d). If persons are at risk because they enter unemployment, the welfare state shapes how this risk manifests (it moderates the outcome of unemployment as a determinant). Compared to the indirect relationship described before, the welfare state interacts with another variable, instead of operating via a mediator. Evidently, combinations of the two in the sense of moderated mediations are conceivable.

Its specific response to risks determines the nature of a welfare state. Risks and needs can be handled with varying efficiency and generosity, and lead to more or to less equality in society. Furthermore, how responsibility is divided amongst state, family and market reveals the underlying welfare regimes. In the terminology promoted by Esping-Andersen (1990, 1999), the stronger the role of the welfare state, the less responsibility is placed on the market (decommodification) and the family (defamilialisation).

Responding to risks and needs entails another noteworthy element, one might call a sub-function, which is the avoidance or at least reduction of uncertainty (Barr 2001: esp. chapter 2; Crouch & Keune 2013; Garland 2014). It is argued that this might even be one of the reasons social policies were implemented in the first place. As Sigerist (1943: 375–376) points out, uncertainties due to economic cycles and resulting constant risk of unemployment were among the “major grievances” workers in Germany (but this can be extended to other countries as well) had to face. The insurance schemes introduced by Bismarck were in parts a reaction to that—albeit not out of charity but in order to weaken socialist movements. By providing a safety net and increasing certainty regarding how risks and needs are met if ever occurring, welfare states potentially influence individual outcomes—especially those tied to cost-benefit considerations (Iversen & Soskice 2001).

As laid out in this part of the chapter, risks are at the heart of welfare state responsibilities. The function of lowering risks and moderating the outcome for individuals at risk is thus important to determine a welfare state’s responsiveness in a comparative perspective. Evidently, varying responsiveness can be an important explanation when examining why different social policy arrangements have different effects on individual outcomes. Therefore, the different possible links between social policies and risks displayed in Figure 2.2 already suggest specific hypotheses, which will be discussed in detail later in this book.

2.2.2 Redistribution: Robin Hood, Piggy Bank and More

Another key function of the welfare state is the redistribution of resources. Such redistribution happens in different modes. On the one hand, it entails shifting income from stronger participants of the labour market to weaker ones. On the other, it relocates resources over the span of life and economic phases from times of strength to times of need. These two redistributive functions are sometimes referred to as the Robin Hood function and the Piggy Bank function (Barr 2001). Regardless of the mode of redistribution adopted by a specific welfare state, the instrument of shifting income from one group or time to another can in the most basic sense be described as the tool through which welfare states fulfil their most elementary obligation: social protection and thus the security function introduced before. While this shows that there is potential overlap between different functions, distinguishing the different components is important in order to understand the different facets in which welfare states operate.

Redistribution, which aims at relocating income from rich to poor individuals or families, falls under the Robin Hood type of resource allocation (Kvist et al. 2013: 322). The objective here is to reduce inequality between social groups by working towards vertical equity (Barr 1993: 10). This is found especially in targeted welfare state models, where means-tested benefits are provided to those in need. Such modes of equalising individuals seem to correspond to popular notions of fairness. However, evidence shows that there seems to be a so-called paradox of redistribution. This refers to the finding that the Robin Hood approach to redistribution appears to be less effective than a universal approach, which is funded by all and distributes to all. This paradox was most prominently discussed by Korpi and Palme, who conclude:

the more we target benefits at the poor only and the more concerned we are with creating equality via equal public transfers to all, the less likely we are to reduce poverty and inequality (Korpi & Palme 1998: 681–682).

This seemingly counterintuitive finding is explained by several arguments. Broadly summarised, universalism is expected to increase public support for redistributive policies—especially in the middle class—and thus strengthen political efforts implementing such policies, which increases the overall budget (Jacques & Noël 2018: 72). Recent research on the matter finds some changes in the link between logic of redistribution, public support and poverty. While the link between universalism and reduction of poverty proves to be stable, the relationship between universalism and preference for redistribution—as a bridge hypothesis—produces contradicting results (e.g. Brady & Bostic 2015). However, this seems to depend on the chosen operationalisation of universalism—considering the aim of this book, this is a small but important finding. Evidence suggests that capturing the institutional design of redistribution¹³ instead of its outcomes¹⁴ supports the link between universalism and policy support (Jacques & Noël 2018: 82).

Concluding, the Robin Hood logic of redistributing from one part of the society to another can follow different objectives. At one extreme, it can mean the targeted reallocation of resources from richer parts of the population to poorer ones. At the other extreme, we find universal redistribution which takes “from all to give to all” (Jacques & Noël 2018: 71). As the short summary in this chapter shows, such different faces of redistributive logic potentially influence individuals in a variety of ways. Among other things, they exhibit different efficiency in lowering risks, shape political support and redistributive preferences.

A second type of redistribution disperses resources throughout the life course. The welfare state is referred to here as a piggy bank (Barr 2001). Social risks are unevenly distributed throughout an individual’s life, and obvious situations of high vulnerability occur in childhood and old age. However, as labour market participation is no longer just a matter of the male breadwinner, new risks such as motherhood gain importance, which arise at specific stages in the life course (Esping-Andersen 1999: 42). Redistribution, then, is not just an exchange of resources between individuals, but also within a life course from times of economic prosperity to times of need. Insurances against risks which may occur at later stages in life, result in consumption smoothing (e.g. Barr 2001: 5) by allowing individuals to decrease uncertainties and achieve more economic stability over their life course.

This perspective can also be found in areas outside of financial redistribution. For example, the social investment approach shows resemblance with the described logic. Like other policy measures, social investment follows a generational path in which—at least in an ideal-typical way—individuals switch from being recipients of investment (in childhood and youth) to being contributors (while participating in the labour market) and finally back to being recipients in old age (Andersson 2018: 110). In this sense, investments in human capital and other individual traits in younger years can be cashed in during times of need later in life. From the point of view of the individual, redistribution over the life course not only impacts risks (especially those arriving in later life), it also influences the perception of uncertainty and predictability of life.

2.2.3 Social Stratification: the Welfare State and Social Inequality

As pointed out at the beginning of this chapter, the link between the welfare state, social class and social inequality was already established by Marshall (2000 [1950]). It was later prominently taken up again by Esping-Andersen, who sums up pointedly: “The welfare state may provide services and income security, but it is also, and has always been, a system of social stratification” (Esping-Andersen 1990: 55). In this sense, the manner in which a welfare state is organised and financed, the kind of redistribution it is based on and the generosity of the provided benefits cause patterns of inequality in societies. In light of the present research question, it is not just important to establish that the welfare state has a stratifying function and that different regimes differ in how they stratify. The main interest should be why such a link exists in the first place, as the answer to this question might later reveal how to conceptualise the welfare state.

At a broad level, the three ideal-typical welfare regimes introduced by Esping-Andersen (1990) help to identify explanations for different stratification outcomes. Countries, which incorporate many features of the Social-Democratic ideal equalise the most. Because of universal programmes and high benefit generosity, they achieve lower income inequality and higher equality of opportunity. In contrast, countries approaching the Liberal ideal are characterised by social policies, which may even increase inequality. Targeted means-tested and ungenerous benefits, whose receipt is associated with stigmatisation potentially weaken the situation of those at the bottom of the social hierarchy even more and widen the gap between contributors and recipients. Lastly, welfare states approaching the Conservative spectrum tend to reproduce existing inequalities since benefits and services are closely linked to previous employment status and earnings. Furthermore, a strong focus on the male-breadwinner may even increase inequality as it marginalises women (Esping-Andersen 1990: 55–78; Sachweh & Olafsdottir 2012: 152–153; Esping-Andersen 2015).

Even though the primary aim of welfare states is not to equalise class structures, it can be argued that efforts towards more equality and equal opportunities have become increasingly important in political debates and agendas in the last decades of the twentieth century (Esping-Andersen 2015: 125). Thus, reducing social inequality can be seen as a more recent objective and function of welfare states. Lastly, we have to take into account that the stratifying role of welfare states not only manifests in objective indicators of social inequality. Since the latter is a highly controversial topic, it impacts attitudes about inequality and preferences for specific stratification patterns as well (Sachweh & Olafsdottir 2012).

Concretising such general explanations, a direct link between social policies and social stratification can manifest (for instance) in the generosity of income replacement in the sense that high income replacement can reduce the gap between income groups, while a minimal replacement can even widen existing gaps. Different patterns of stratification, manifesting in social inequality and mobility, are thus a side effect of the most basic function of lowering risks and meeting needs. They reveal how strong the impact of different principles behind welfare states can be on class structures, class struggles and the salience of class itself (e.g. Esping-Andersen 1990: 55). Therefore, the welfare state’s role in shaping social stratification is still considered an independent function that goes beyond responsiveness and provision of security. The welfare state impacts social stratification directly by levelling out (or widening) income gaps.

2.2.4 Activation: an Enabling and Incentivising Institution

Another function of welfare states is to enable as well as incentivise specific behaviour. Overall, such activation is a relatively new agenda within welfare states, which has gained importance since the 1990s. A key trend, which has increased the promotion of activating policies, lies in the shift towards more individualisation of welfare provision. As discussed in the first part of this chapter, a variety of structural, demographic, and normative changes led to transformations of welfare stateness. Among the consequences of such processes is the rising importance of individual responsibility for one’s own social security. Thus, behaviour preventing risks needs to be incentivised. This especially refers to labour market participation and privatisation of insurance schemes. In this context, the term enabling state gained relevance. Initially used to describe the Anglo-Saxon approach to welfare provision, this agenda is gaining momentum throughout Europe (Gilbert 2004: 42).

Activation is embedded foremost in the strategies of social investment that were discussed in a previous section of this chapter (cf. 2.1.3). Regardless of the policy field, the objective of such activating policies is always similar: to prevent risk by investing either in individuals or in the contextual structures surrounding them. In many cases, such investment is linked to the labour market. Since new risks often affect those, who were traditionally not part of the labour force, it is increasingly necessary to activate those particular groups. This applies especially to mothers (Kowalewska 2017; Dotti Sani & Scherer 2018), but also to other groups such as chronically ill or disabled individuals (Holland et al. 2011).

Within the field of social investment, the labour market plays an important role because the support of employment is seen as one way of securing individuals against risks such as poverty. Thus, active labour market policies (ALMP) include measures such as training, job-creation and more, which aim at increasing participation in the labour market and ending spells of unemployment by incentivising and enabling individuals to re-enter the labour market (Bonoli 2012). ALMPs are discussed critically in the literature and reviews confirm only partial effectiveness (Crépon & van den Berg 2016). While training and investments in human capital have an impact—at least in the long run—other measures such as public sector employment programs appear to be ineffective (Card et al. 2018), produce more costs than benefits or help one group at the expense of another (Crépon & van den Berg 2016). When it comes to increasing the chances of employment, ALMPs are more effective for women and long-term unemployed individuals and the impact of training manifests a few years after completing the programme (Card et al. 2018). Taking into account the reduction of risks, the relationship between investment in labour-market programs, reduction of unemployment and reduction of poverty varies considerably depending on wage structures (Cronert & Palme 2017).

Overall, while the importance of social investment within the labour market and beyond grows in policy-making, its actual effectiveness is not yet clear. Differences in the design and effectiveness of social investment between countries are the result of a number of issues. Here, not just the policies themselves differ, but since they are linked to more traditional social policy measures of the ‘social protection’ agenda (Ellison & Fenger 2013: 612), the combination of compensating policies (such as income replacement) and activating policies seems to determine their successfulness (Kuitto 2016: 445). In general, a growing emphasis on activating policies is in line with more general sociological contributions on social change and especially the individualisation of risk and responsibilities (Beck 1986; Beck & Beck-Gernsheim 2002), which speaks in favour of a continuous expansion. Furthermore, considering that the Europe 2020 strategy for smart, sustainable and inclusive growth includes social investment as a key element (European Commission 2013), convergence of social investment policies within the European Union is possible. However, there is evidence, that the continuous introduction of new policies in line with the social investment perspectives came to a halt with the Euro crisis (Ronchi 2018).

Regardless of its successfulness, activation is an important function of welfare states as it represents an increasingly relevant goal in policy-making and a distinct way of approaching social security. It represents a possible strategy to respond to new risks on the one hand and lower dependency on income replacement and other state-provided benefits on the other.

In sum, this active role of the welfare state is somewhat new, but it still goes along with the more classical function of providing income replacement and other benefits. Variations between countries can thus be attributed to how they combine an activating agenda with more conventional measures. For instance, what gained popularity under the term flexicurity describes the combination of extensive active labour market policies and generous benefits (Bonoli 2013: 15). However, old and new approaches to welfare provision can also undermine each other. Using the example of family policies, incentives in one field (such as female employment) can be combined with deactivating measures (such as monetary incentives for staying at home) in another. Other variation stems from the character of activating policies. As Kowalewska (2017: 4) summarises, a broad differentiation of activating agendas entails identifying the extent to which either employment or investments in human capital are emphasised. Moreover, the activating agenda of the welfare state is not restricted to classical labour market policies, but can also manifest in other areas such as family policies (Bonoli 2013: 26). For example, defamilialisation represents a key component of enabling parents in general and mothers in particular to participate in the labour market.

Lastly, while the activation of individuals is often an explicitly stated objective, which clearly manifests in policies, welfare states can also incentivise behaviours unintendedly. Again, the case of family policies can serve as an example, where generous benefits for stay at home care responsibilities can potentially deactivate women with regard to their labour market participation. This can be an unintended consequence of attempts to incentivise other behaviour such as child-bearing. Beyond this example, such unintended deactivation would be in line with neo-Marxist theory, which assumes that truly extensive social benefits (in any policy field) would disincentivise labour market participation and thus counteract the principles of capitalist market economy (Quadagno 1987: 114–115).

2.2.5 Socialisation: the Welfare State, Endogenous Norms and Values

The last objective discussed in this chapter, is to socialise individuals. While responding to risks, shaping inequality and redistributing resources are often politically steered objectives, socialisation is not a specifically targeted goal. Instead, the welfare state acts in an invisible and often unintended manner in this case.

There is a philosophical side to this issue. In a sense, welfare states can be understood as manifestations of certain ethical ideals. They represent principles of egalitarianism and they affect individual autonomy and responsibility (White 2010). If the nature of a welfare state is thus not just perceived as a system of security and inequality but also as a cultural authority, it can affect individuals much more than just through socioeconomic channels. The kind of equity promoted in different modes of redistribution on the one hand, and different ways of shaping social inequality on the other, dictate solidarity and justice principles. In this sense, welfare states can be seen as institutionalised ideas about social justice (Sachweh 2016). However, not only justice principles can be linked to social policies. As Gangl and Ziefle (2015: 511) summarise, they also legitimise and institutionalise aspects in other domains such as gender role beliefs, which in turn manifest in labour market participation. The welfare state is especially seen as a socialising institution when it comes to shaping outcomes associated with endogenous norms (Lindbeck 1995; Bisin & Verdier 2004).

Overall, key characteristics of the welfare state and the main principles underlying redistributive efforts (such as universalism or targeting) represent distinct manifestations of solidarity and justice principles (Arts & Gelissen 2001), gender equality (Bonoli & Natali 2013b: 3) and more. Living or growing up in a welfare state thus means being embedded in a cultural context, which represents and legitimises specific ideals (Pfau-Effinger 2005). This can influence individual perceptions of such ideals and lead to an increased probability of adapting underlying principles—such as egalitarian views. It may in turn even produce a feedback effect (e.g. through political participation) and increase egalitarian social policies (Lindbeck 1995: 488). In this line of argument, living in a welfare state during the formative years can—for instance—shape attitudes towards related issues such as redistribution (Neundorf & Soroka 2017). It also influences role-models related to gender and promotes and facilitates how such models manifest in reality—for example when it comes to division of housework (e.g. Fahlén 2016) or norms about the balance between labour market participation and care responsibilities for women (Barbieri & Bozzon 2016: 103). Especially in the field of gender inequality, social policies and particularly family policies are closely related to gender ideologies (Grunow et al. 2018).

Figure 2.3

(Modified version of theoretical model proposed by Arts & Gelissen (2001: 288))

The welfare state, norms and principles.

Arts and Gelissen (2001) propose a theoretical model capturing how welfare state regimes are linked to the formation of normative frames. Their model can be generalised in some respects in order to be applicable to areas beyond solidarity and justice principles (such as principles of gender equality). Figure 2.3 provides a display of such a generalised version of their approach and sketches the processes described by Arts and Gelissen. In short, the welfare state triggers cognitive processes, in which individuals (1) learn the dominant norms, (2) form habits and (3) frame situations. This is of course not a deterministic process and welfare states do not strictly impose their endogenous norms and reference frames. However, they possess the potential to coin individual perspectives in relevant areas. The three cognitive processes can thus manifest on the individual level as knowledge of rules and principles and their habituation as well as in adapting to normative frames with varying intensity, leading to variations in the distribution of such notions on the aggregate level. To elaborate using the above-mentioned examples, the socialising function of welfare states can manifest as varying degrees of adopting principles of gender equality or universalism underlying the nature of different welfare states.

In sum, the function of socialisation is tied to what is sometimes referred to as a ‘welfare culture’. Here, social policies and ideologies are closely interrelated (Pfau-Effinger 2005). They present a context in which individuals are embedded and thus shape what is perceived as legitimate and dominant principles within policy arrangements.

2.2.6 Summary: Functions and Variations of Welfare States

The brief summary of the broad literature surrounding each of the functions highlighted in this chapter, gives various reasons why welfare stateness is an important independent variable. It also reveals a variety of potential dependent variables beyond the most obvious ones. This includes manifestations of inequality (and especially poverty), attitudes, behaviour and many other phenomena, which appear to be closely tied to the main functions of the welfare state. Furthermore, it shows that some of the functions are related. The role of the welfare state in addressing risks and needs for instance is strongly related to the redistributive function, as redistribution is one of the most important tools in financing social security. In this sense, redistribution is not only a function of welfare states, but also describes their modus operandi.

Furthermore, some functions represent explicit policy objectives (such as risk reduction), while others are characterised by rather subtle or unintended consequences. For example, the stratifying function has only recently become a more clearly defined objective in some European countries. As was argued, reducing inequality and increasing equal opportunities are no longer just a by-product of different social policy arrangements, but can be an explicit political goal. Activation can be such an explicit policy goal as well—for instance in the case of active labour market policies or social investment—it can however also be an unintended consequence, triggering individual behaviour unintentionally. The briefly discussed disincentivising effects are mostly due to such unintentional functions as well—in many instances one could even call them dysfunctions. Lastly, the socialising function represents another example for a process that is in most cases unintended. Albeit following different objectives and leading to different outcomes, the functions are related to one another. Figure 2.4 graphically summarises this.

Figure 2.4

Summary of introduced functions of welfare states

2.3 From Dependent to Independent Variable

As a key element of modern democracies, the welfare state is an important object of research in the social sciences—especially in cross-cultural comparative research. Due to differences emanating from historical trajectories and extents to which the economic, societal and demographic changes since the second half of the twentieth century affected countries, the study of the welfare state is characterised by considerable complexity. Different conceptual and empirical perspectives on the welfare state discussed in the first part of this chapter reveal a variety of approaches and they outline distinct historical and theoretical conceptualisations and empirical operationalisations. However, the summary in this chapter also shows a considerable amount of heterogeneity in the approaches and the literature is far from reaching a consensus when it comes to conceptualising the welfare state. Furthermore, while the discussion of literature on the welfare state as a dependent variable may reveal popular positions and debates, we still know little about how scholars include the welfare state as an independent variable—theoretically as well as empirically.

Reasons for treating the welfare state as an independent variable were discussed in the second part of this chapter. While a great share of the literature focusses on how to grasp different social policy arrangements—conceptually and empirically—, this discussion shows that the welfare state also has important functions. These functions reveal why welfare states in general and social policies in particular are not only important dependent variables but also important independent ones. Depending on how welfare states fulfil the different tasks (either intended or unintended), a variety of other phenomena is impacted. Even though I focussed mainly on the description of functions so far, this already hints at important outcomes. Social security shapes individual risks and reduces uncertainty, redistribution shapes policy preferences, socialisation shapes attitudes, incentives trigger behaviour, and social stratification—from a sociologist’s perspective—potentially shapes almost everything. To explore the welfare state or features of it as an explanatory factor—not only when examining outcomes on the societal but especially those on the individual level—is more than evident. This leads us to essential questions, which are addressed in the following chapters of this book. How can the welfare state be operationalised if treated as an independent variable? Does this require different perspectives on the welfare state than the ones present in comparative welfare state analysis? And if so, what might such different perspectives look like?