The four global worlds of welfare capitalism: Institutional,neoliberal,populist and residual welfare state regimes

Abstract

What welfare state regimes are observed when the analysis is extended globally, empirically and theoretically? We introduce a novel perspective into the ‘welfare state regimes analyses’ – a perspective that brings developed and developing countries together and, as such, broadens the geographical, empirical and theoretical scope of the ‘welfare modelling business’. The expanding welfare regimes literature has suffered from several drawbacks: (i) it is radically slanted towards organisation for economic co-operation and development (OECD) countries, (ii) the literature on non-OECD countries does not use genuine welfare policy variables and (iii) social assistance and healthcare programmes are not utilized as components of welfare state effort and generosity. To overcome these limitations, we employ advanced data reduction methods, exploit an original dataset (https://glow.ku.edu.tr/) that we assembled from several international and domestic sources covering 52 emerging markets and OECD countries and present a welfare state regime structure as of the mid-2010s. Our analysis is based on genuine welfare policy variables that are theorized to capture welfare generosity and welfare efforts across five major policy domains: old-age pensions, sickness cash benefits, unemployment insurance, social assistance and healthcare. The sample of OECD countries and emerging market economies form four distinct welfare state regime clusters: institutional, neoliberal, populist and residual. We unveil the composition and performance of welfare state components in each welfare state regime family and develop politics-based working hypotheses about the formation of these regimes. Institutional welfare state regimes perform high in social security, healthcare and social assistance, while populist regimes perform moderately in social assistance and healthcare and moderate-to-high in social security. The neoliberal regime performs moderately in social assistance and healthcare, and it performs low in social security, and the residual regime performs low in all components. We then hypothesize that the relative political strengths of formal and informal working classes are key factors that shaped these welfare state regime typologies.

Keywords

welfare state regimes welfare state social policy social assistance populism social movements working class emerging markets cluster analysis model based clustering

Introduction

How do welfare state regime families appear if we analyze them from a global perspective, by using genuine welfare policy variables and by considering social assistance and healthcare as central pillars of welfare provision in addition to traditional social security? How can we then explain the formation of new global welfare state regimes? In this article, we shed some light on these questions by taking a first step towards developing a global welfare state regime typology and putting forward a set of working hypotheses pertaining to the historical dynamics that have shaped the formation of global welfare state regimes.

The scholarship on welfare states has created a ‘race to discover new regime types’ however, it has been radically slanted towards Western or organisation for economic co-operation and development (OECD) countries in case selection (Ebbinghaus, 2012; Powell and Barrientos, 2015; Walker and Wong, 2013). Although a handful of studies have expanded their scope by examining welfare provision in non-OECD countries, scholars have not compared these countries with their Western counterparts, disregarding the possibility that geographically or culturally distant countries might converge into common welfare regimes (Powell and Barrientos, 2015). Furthermore, studies focusing on the non-West have predominantly used developmental outcome and contextual variables, such as the Gini Index, Gross Domestic Product (GDP) measures or poverty indicators, as proxies for welfare variables leading to a mismatch between conceptualization and operationalization of welfare provision (Yörük et al., 2019b). Lastly, the existing literature has ignored healthcare and social assistance policies although these policies are integral parts of welfare provision both in the (OECD) and non-OECD countries (see, Bambra, 2005; Saraceno, 2002).¹

We aim to overcome these limitations by comparing OECD and non-OECD countries, by focusing on welfare policy variables, or the supply side, rather than outcome variables, and by including health and social assistance policies to supplement the traditional policy domains studied previously. We show that when these shortcomings are addressed, there are four worlds of welfare state regimes: institutional, neoliberal, populist and residual welfare state regimes.

Our analysis relies on a novel dataset of welfare generosity and welfare effort designed specifically for this study, the Global Welfare Dataset (https://glow.ku.edu.tr/), which is an outcome of the ERC- funded Emerging Welfare Project (emw.ku.edu.tr). As the first of its kind, our dataset allows us to conduct a detailed analysis using genuine welfare policy variables – for example, replacement rates, benefit durations, employee–employer contribution ratio and so on – that measure the welfare effort and welfare generosity. We deploy factor and cluster analyzes to reach a robust welfare state typology. In doing so, we take a major step forward to align conceptualization and operationalization of welfare state families across the OECD and non-OECD countries.

We also contribute to the literature by providing some working hypotheses pertaining to political dynamics generating different welfare state regime types. The origins of The Three Worlds of Welfare Capitalism literature, that is, the work of Esping-Andersen (1990) makes a ground-breaking contribution as it provides a descriptive analysis of welfare regimes and then brings a political economic explanation for the divergence among welfare regimes. We follow in the footsteps of this approach and present possible mechanisms based on political factors. Parallel to Esping-Andersen (1990), we focus on the role of social movements and political mobilization of the working class to account for the emergence of four global welfare state regimes. In concrete terms, we hypothesize that the level and intensity of formal worker class mobilization (as measured by the number of strikes) and of poor people’s mobilization (as measured by the number of riots) are determinants of (i) how different countries develop and weight different welfare state components and (ii) how this results in the emergence of institutional, populist, neoliberal and residual welfare state regimes.

The way we explain the development of welfare state regime types informs our choice of labels to designate the welfare state regimes. The term ‘populist’ refers to the two historical waves of populism and accompanying social movement dynamics that have created the populist welfare state regime. The term ‘neoliberal’ is coined because the neoliberal welfare state regime typology has mostly taken shape during the neoliberal era following a previous wave of poor people’s activism. We borrow the terms ‘institutional’ and ‘residual’ from the welfare state literature (see, Furniss and Tilton, 1977; Titmuss et al., 1974; Wilensky and Lebaux, 1958) to differentiate between developed and underdeveloped welfare regimes, while taking into account the dynamics of poor people’s movements and labour unrest.

This article makes a critical contribution to the rapidly growing literature on welfare state regimes. We employ the same methodological and theoretical approaches used to understand European welfare state regimes in our effort to classify and explain global welfare state regimes. In that way, we extend the geographical scope of welfare regime analysis while retaining the scientific rigour of the previous literature and by overcoming the ‘variable selection problem’ of the global ‘welfare modeling business’ (Yörük et al., 2019b). We build our analysis on the main conclusions of the Western welfare regime literature (that is, the existence of three worlds of welfare), but illustrate the existence of additional welfare state regimes when the analysis is extended into a global scale. Therefore, what seems to scholars of European welfare as three worlds turns out to be a single welfare regime when analyzed from a broader perspective.

This article is organized as follows. The following section presents the methods used in this study. The next two sections report the results of the cluster analysis and offer a discussion on the composition of constitutive welfare state components, respectively. The following section develops our working hypotheses about the underlying political factors leading to the formation of these global welfare regimes and the final section concludes.

Methods

Research on welfare state regimes indicates that they are multidimensional and require a variety of indicators to capture different aspects (Bonoli, 1997). To account for this multidimensionality, we develop a two-step process.² We first conduct a principal components factor analysis on a set of welfare indicators, and the results suggest that there are two main dimensions of welfare provision identified by the indicators in Table 1. We refer to these two dimensions as welfare generosity and welfare effort, the construction of which is consistent with the tendency in the welfare state regimes literature (Bambra, 2006; Nelson et al., 2020; Scruggs et al., 2014). In the second step, we use the two-dimensional indicators in the cluster analysis to classify welfare regimes.

Table 1.

Components of welfare state regimes: Generosity and welfare state effort.

Welfare generosity		State effort
Old age pensions	Standard benefit replacement rate Minimum benefit replacement rate Qualifying age Expected duration Employee–employer funding ratio	Pension expenditure, % GDP Coverage rate, % + population
Sickness cash benefits	Standard benefit replacement rate Waiting days Qualifying period Benefit duration Employee–employer funding ration	Sickness expenditure, % GDP Coverage rate, % insured
Unemployment insurance	Standard benefit replacement rate Waiting days Qualifying period Benefit duration Employee–employer funding ratio	Unemployment expenditure, % GDP Coverage rate, % insured
Social assistance	Average per capita transfer Social pension replacement rate Social pension as % GDP per capita Social pension expected duration	Total social assistance expenditure, % GDP Social pension expenditure, % GDP Social assistance coverage, % first quintile Social pension coverage, % eligible population
Healthcare	Health expenditure per capita	Public health expenditure, % GDP

The welfare generosity indicator captures the various programmatic rules and components of social programmes that make benefits more or less generous across cases (Scruggs 2014; Nelson et al., 2020). In other words, it shows the ‘how’ dimension of social provision (Bonoli, 1997: 352). We construct an additive index, using the variables listed in Table 1, to measure welfare generosity for each policy area.

The second dimension is welfare effort. Our factor analysis shows that expenditure and coverage rates tend to load on the same factor across policy areas. Hence, we construct an additive index for each programme to capture the ‘how much’ dimension of welfare provision (see, Bonoli, 1997: 352). This index represents the extent to which states invest in social provision. Although the use of expenditure data has often been bedevilled (see, Clasen and Siegel, 2007 for a broad discussion), they capture an essential element of welfare states, namely, their size and the spending priorities of different regimes (Bonoli, 1997; Castles, 2009; Obinger and Wagschal, 2010). After all, a system that spends 1 percent of GDP with high levels of coverage and replacement rates should not be the same in substantive terms as a system that has equally high levels and spends 15 percent of GDP. Furthermore, since we do not use expenditure data exclusively to make a judgement about the characteristics of a welfare state regime, our analysis captures the essential components of social welfare provision that are deemed crucial (see Esping-Andersen, 1990).

In the second step, we deploy cluster analysis to identify the welfare state families. Cluster analysis is an exploratory method that allows for systematic classification of cases in distinct groups (Abu Sharkh and Gough, 2010; Gough, 2001; Gough and Wood, 2004; Kangas, 1994; Powell and Barrientos, 2004; Pribble, 2013). In this study, we employ hierarchical, k-means and model-based clustering methods (MBC).

The mechanics of these three data clustering techniques differ and using them together allows us to achieve a robust welfare state regime typology. Hierarchical cluster analysis minimizes the distance between cases of the same group while simultaneously maximizing the distance between groups (Tan et al., 2013; Tryfos, 1998). In this way, clusters emerge from the data, facilitating the identification of welfare state regimes. The k-means cluster procedure is often conducted alongside other clustering methods (Gough, 2001; Cramer, 2003), but requires the researcher to set the number of clusters in an ad hoc manner. This algorithm works iteratively to assign each data point to one of the groups based on the features that are provided. The data points are clustered based on feature similarity (Mirkin, 2016). MBC, on the other hand, assumes that data are generated by a mixture of probability distributions in which each component represents a different cluster (Fraley and Raftery, 2002). MBC selects the model which optimizes the fit between the data and models using Bayesian information criteria (BIC) (Dasgupta and Raftery, 1998; Fraley and Raftery, 1998; Biernacki et al., 2000; Keribin, 2000)³

Results

Each of the clustering techniques listed above yields similar results, and we use them in combination to check the robustness of our findings. Obtaining consistent results from different clustering techniques permits us to conclude that there are four welfare state regimes across OECD and emerging market economies: institutional, neoliberal, populist and residual.

Hierarchical cluster analysis (using Ward’s linkage method) indicates that the 52 countries analyzed in this study form four clusters, as shown in Figure 1. MBC analysis also brings about precisely the same set of clusters, except for India, which moves from the populist cluster to neoliberal cluster. Lastly, in our k-means analysis, we impose k=4 since hierarchical and model-based cluster analyzes suggest four clusters. The results are very consistent. Only three countries are placed in different clusters across techniques. Malaysia and Indonesia are grouped within the residual cluster, and Canada is grouped within the neoliberal cluster, while India goes back to the populist cluster.

Following the very logic that characterized The Three Worlds (Esping-Andersen, 1990), we argue that welfare state regimes in both developed countries and emerging markets should be conceptualized and explained with respect to two factors: (i) the composition and performance of constitutive welfare state components, and (ii) the underlying political economy, including developmental strategies and political exigencies. The next section presents the variation in the characteristics of four welfare state regimes, and the section that follows develops hypotheses about the possible political-economic mechanisms that have brought these welfare state regimes into existence.

Populist, institutional, residual and neoliberal welfare state regimes

In what follows, we describe the characteristics of the four welfare state regimes, populist, institutional, residual and neoliberal, by analyzing the relative positioning of these regimes with respect to their performance in welfare state effort and generosity across major welfare state components.⁴ The dendrogram results (Figure 1) indicate that our sample is bifurcated first into two main groups, one composed of populist and institutional regimes and one composed of residual and neoliberal regimes. Figure 2 and the raw generosity and state effort scores, presented in Supplemental Appendix and Supplemental Table, illustrate that the first two clusters are more generous in welfare provision and their welfare effort is higher. Institutional regimes, however, score higher than the populist regimes in social security, social assistance and healthcare. Table 2 shows the membership in each of the four clusters.

Table 2.

Four global welfare regimes.

Institutional	Neoliberal	Populist	Residual
Australia	Indonesia²	Argentina	Bangladesh
Austria	Israel	Brazil	Mexico
Belgium	South Korea	Bulgaria	Pakistan
Canada³	Malaysia²	Czech Republic	Peru
Denmark	United States	Chile	Philippines
Finland		China
France		Colombia
Germany		Estonia
Iceland		Greece
Ireland		Hungary
Italy		Greece
Luxembourg		Hungary
New Zealand		India¹
Netherlands		Japan
Norway		Latvia
Spain		Poland
Sweden		Portugal
Switzerland		Romania
United Kingdom		Russia
		Slovenia
		Slovak Republic
		South Africa
		Thailand
		Turkey
		Ukraine

Source: Authors’ compilation

MBC: model-based clustering methods.

¹India: Populist Welfare State Regime (WSR) in hierarchical and K-means while neoliberal WSR in MBC.

²Indonesia: Neoliberal WSR in hierarchical and MBC while residual WSR in K-means.

³Canada: Institutional in hierarchical and MBC while neoliberal in K-means

One of our main contributions is to introduce and describe the populist welfare regime as a distinct new type of welfare regime, and so we focus our attention on this particular regime type more than on the others. Other welfare state regime types, institutional, residual and neoliberal, have been discussed in the existing literature to varying degrees (see, Powell et al., 2020b for a review). The populist welfare state regime consists mostly of emerging market economies from Asia, Latin America, the Post-Communist Region and Southern Europe, in addition to Japan. The countries in the populist welfare state regime are characterized by moderate levels of generosity and state effort across all policy areas. The populist cluster has just above average scores in terms of social security provisions (old-age pension, sickness and unemployment) and just below average scores in terms of social assistance and healthcare (Figure 2). It is important to note here that the famous Mediterranean or Southern European cluster, first pointed out by Ferrera (1996), appears to split between the populist and institutional regimes. In other words, Mediterranean and post-Communist clusters in Europe become sub-parts of the populist cluster when we shift the focus from Europe to a broader range of cases. It is intriguing to see, for example, that Greece and Portugal are more similar to Brazil and China, than to the Netherlands or France.

As for Japan, (Powell et al., 2020: 2) asked an important question amid an atmosphere of scholarly confusion concerning the regime typology of the country: ‘Can Japan best be regarded as one of the three worlds, a fourth world, or a hybrid or unique case? More fundamentally, can the Three Worlds approach, essentially based on Europe, capture a very different nation?’ This question echoes the warning of Ferragina and Seeleib-Kaiser (2011: 11) who argued that ‘as Japan’s political and societal system is rooted in the tradition of neither liberalism nor Catholicism nor social democracy, it seems inappropriate to characterize the Japanese welfare state as belonging to either of the regimes’. We believe that this difficulty stems from the tendency to group Japan with advanced Western countries, while our analysis suggests that the country clusters better with emerging market countries and Southern Europeans.

Institutionally, the populist welfare state regime is an outcome of two patterns that characterize the history of welfare state development in this heterogeneous set of countries: (i) the post-war development of fragmented and corporatist social security systems that benefited the privileged segments of formal sector workers and civil servants at the expense of the informal urban and rural poor (Cook, 2010; Eder, 2010; Haggard and Kaufman, 2008; Huber and Bogliaccini, 2012; Walker and Wong, 2005) and (ii) the rapid development of social assistance and healthcare programmes during the neoliberal period that target the previously excluded urban and rural poor (Buğra and Keyder, 2006; Cook, 2010; Gao, 2017; Seekings and Nattrass, 2015; Thachil, 2014; Zucco, 2013). As a result of these historical patterns, the populist cluster is narrowing the previous wide gap with its European counterparts in welfare effort and generosity. Hence, the populist cluster is a result of an overall welfare state expansion that spans the period since the Second World War, first through social security development and then through social assistance and healthcare system development.

In the populist welfare state regime, a new world of welfare with unique programmes, services, budgets and extensive coverage is emerging, which does not fit into any of the existing welfare regime classifications.⁵ While the pioneering West had up until recently set the example for the rest of the world in terms of welfare provision, now, emerging markets have taken the lead by developing new social assistance programmes, which neither follow in the footsteps of the older poverty relief programmes of the West nor imitate them. In essence, they are inventing new programmes, most prominently conditional cash transfer programmes (Piven and Minnite, 2016). This invention is a root characteristic of the contemporary welfare state regime structure when the analysis is extended to emerging markets, and it implies that this welfare development is not only a quantitative expansion but also corresponds to a radical qualitative shift in the history of the welfare state.

The institutional welfare state regime consists of advanced OECD countries. This cluster consists of the countries with the highest scores across all the policy areas in terms of welfare generosity and welfare effort, and particularly strong in healthcare, social assistance and unemployment benefits. The institutional welfare state regime combines a variety of cases established as families of welfare in previous studies, including a mix of liberal, corporatist and social democratic regimes. Figure 1 reveals that several sub-clusters in the institutional welfare state regime family are in line with existing studies, which increases the face validity of our findings: first, two liberal sub-clusters, one group comprising Australia, New Zealand and Luxembourg, and another comprising Ireland and the UK; second, a corporatist cluster, including Austria, Belgium, Finland, France, Germany and Spain; third, a social democratic cluster composed of Denmark, the Netherlands, Norway and Sweden. Therefore, when observed from a narrower perspective of an OECD sample, countries in the institutional welfare state regime form distinct clusters, but when the analysis is extended to a more global sample, these countries form a single larger regime type. In other words, the three worlds of Esping-Andersen appear to be sub-clusters of the institutional regime. Our findings, hence, do not challenge the central finding of the Three Worlds literature, but posits that it matters which lens scholars use: from a large distance, liberal, corporatist and social democratic regimes look similar, as part of the institutional welfare state regime. What we challenge is the Eurocentrism of the literature as well as lack of the methodological rigour in those studies that analyze the non-West.

The residual welfare state regime consists of the less developed emerging markets of Bangladesh, Mexico, Pakistan, Peru and the Philippines. Countries in this regime are characterized by a complete lack of one or more of the social security programmes, and thus tend to have lower overall scores than the rest of the cases in populist and institutional regimes. Specifically, these countries are well below the mean in four of the five policy areas, as shown in Figure 2. There are no unemployment benefits in any of the countries in this regime. Old-age pensions and healthcare provision are also limited compared to populist and institutional regimes. Only in sickness benefits are these cases above average in welfare generosity and welfare effort.

Similar to the institutional regime, we designate this cluster as the residual regime, relying on the residual–institutional contrast within the welfare state literature (Furniss and Tilton, 1977; Titmuss et al., 1974; Wilensky and Lebaux, 1958). As seen in Figure 2, although countries in the residual welfare state regime do not have social security programmes as extensive as other emerging markets in the populist regime, they have commensurable performance in social assistance. Most importantly, Prospera in Mexico, Pantawid Pamilya Pilipino in the Philippines, BSIP in Pakistan, Juntos in Peru and several programmes in Bangladesh cover 13–25 percent of the population in these highly populated countries (Durr-e-Nayab and Farooq, 2014; Perova and Vakis, 2012; World Bank, 2015). Yet, although social assistance programmes have high coverage, the content of these programmes shows a lower level of generosity compared to other regimes.⁶

The final cluster, the neoliberal welfare state regime, consists of a relatively dissimilar set of cases including Indonesia, Israel, Korea, Malaysia and the United States. Similar to the residual regime, countries in this cluster are well below the mean of the sample in four of the five policy areas. The group has relatively better scores in healthcare, social assistance and old-age pensions, but they do not have statutory sickness cash benefits, and they have relatively meagre unemployment benefits. The main differences between residual and neoliberal welfare state regimes are that the residual regime lacks unemployment while the neoliberal lacks sickness, and the residual regime performs much worse in social assistance and healthcare. Therefore, we claim that the residual welfare state regime is an underperforming welfare state regime in all areas, while the neoliberal regime provides a more limited set of benefits.

The infamous American exceptionalism in welfare studies has put the United States in this group. Despite the burgeoning revisionist literature on the American welfare state (Prasad, 2016), studies show the multiple channels of welfare provision in the United States, through tax expenditures and ‘private welfare’ (Hacker, 2002; Howard, 1999) and at the state rather than the national level (Scruggs and Hayes, 2017). In the present analysis, we focus only on national public programmes, allowing us to show that with respect to these policies, the case differs significantly from other developed country cases. The US performs better than developing country cases and is on par with OECD cases in terms of health and social assistance effort and generosity, but has lower scores in terms of old-age pensions. A significant difference between the United States and other developed country cases is the former’s lack of legislated national sickness cash benefits (Social Security Administration , 2016), which results in a significant decline in the USA scores in our framework.

Possible mechanisms of global welfare state regime development

Having described the content of welfare state regimes, we now turn to provide some educated hypotheses about how these regimes may have come into existence and clarify why we designate our clusters with these specific terms. We follow in the footsteps of Esping-Andersen (1990), whose analysis provided, first, a classification of welfare regime typologies and, second, an explanation for emergence of welfare typologies from a political economic and political sociological perspective.

Our hypotheses are mostly focused around political–economic dynamics and working-class mobilizations, combining facets of the power resources theory of Esping-Andersen (1990); Korpi (1983), and the political containment theory of Piven and Cloward (1971). The power resources approach analyzes welfare states expansion as a function of the political power of the working class and the parliamentary dominance of social democratic parties, especially in Northern Europe (Esping-Andersen, 1990; Korpi, 1983). Therefore, social welfare programmes are regarded as political achievements of the democratic class struggle of the working classes. Piven and Cloward’s path-breaking work, Regulating the Poor, claimed that to maintain order and legitimacy, the modern state responded to poor people’s insurgencies by expanding social assistance programmes. In times of social turmoil, social assistance programmes expanded as a means of establishing control over the disorderly (Piven and Cloward, 1971). Thus, this theory suggests that what mainly drives social assistance policies is not social need but social disorder.

We hypothesize that welfare state regime compositions, that is, welfare generosity and welfare state efforts of these regimes, are associated with the strength (that is, bargaining power and the threat) of formal and informal working classes in the developmentalist and neoliberal periods. As such, government responses to contain and mobilize the power of different segments of the working classes have resulted in different welfare state regime typologies. Comprehensive support for this hypothesis would require multivariate analysis, controlling for economic development, state capacity and changing demographics, but such a full analysis is beyond the scope of this article. Given the space limitations, we provide some empirical evidence in the form of descriptive statistics to support our ‘educated hypotheses’ for each regime type using data from the Cross-National Time-Series (CNTS) dataset.⁷ Figure 3 represents the change in riots (defined as ‘any violent demonstration or clash of more than 100 citizens involving the use of physical force’ (Banks and Wilson, 2021), and Figure 4 represents labour strikes, for each of the welfare state regimes between 1960 and 2016. We take riots to represent the political activism of the poor informal populations, that is, the main independent variable of the Piven–Cloward hypothesis (Zarate-Tenorio, 2014), and labour strikes to represent the power of formal working classes, that is, the power resources theory.

The populist welfare state regime cluster is based on the semi-corporatist fragmented social security-based welfare systems established in the post-war period, on the top of which extensive social assistance and healthcare policies were built during the neoliberal era. We call this regime ‘populist’, because of two historical waves of populism that facilitated its rise and transformation: the first is the traditional populism of the second and third quarters of the 20th century that came to be associated with the political machines of leaders in many developing countries, such as Vargas in Brazil, Peron in Argentina, Ecevit in Turkey, Allende in Chile, Salazar in Portugal and Indira Gandhi in India (Grigera, 2017; Jaffrelot and Tillin, 2017; Keyder, 1987; Ramos da Silva Lopes, 2006). This populism, as a developmentalist style, expanded the welfare state by extending social security benefits to the privileged minority of formal sector workers in private and public sectors in a dual effort to contain working-class radicalism (as seen in Figure 4 between the 1960s and the 1990s) and to mobilize massive popular support from the working class (Conniff, 2012; Wolfe, 2010). This effect of populist regimes was paralleled by the protectionist labour regimes of the Soviet Bloc and China that provided benefits to rural and urban workers (Imbrogno, 1991; Li, 2012).

The second populist wave materialized in the 2000s with an explicit emphasis on people versus elites and has been led by leaders such as Lula in Brazil, Kirschner in Argentina, Erdoğan in Turkey, Thaksin in Thailand, Orban in Hungary, Modi in India, Putin in Russia, Xi Jinping in China, as well as Tsipras in Greece. This time, the poor, or informal working classes, have risen as the primary source of political threat and support for these leaders, as shown in Figure 3, particularly for the 2010s, who extended substantial social assistance and healthcare policies to the poor as a way to contain their activism and to mobilize their political support (Aytaç and Öniş, 2014; De la Torre, 2017; Kotwas and Kubik, 2019; Levitsky and Roberts, 2011; Mudde and Kaltwasser, 2019; Remmer, 2012). At the same time, the associational and structural bargaining power of the formal working classes has remained intact, especially in the early 1990s and 2010s (see, Figure 4), and the working class managed to resist the retrenchment of existing social security benefits to a significant extent (Silver, 2003). As a result, these benefits remain above the inter-cluster average. The steep increase in riots in the populist regime also appears to be a key factor for the contemporary expansion of social assistance policies (Figure 3) (Yörük, 2012; Yörük et al., 2019a).

As for the institutional regime, there is an extensive literature that explains the formation of welfare state regime families in OECD countries, and our explanation relies on those strands of theories, that is, the power resources theory, which focus on various forms and scales of democratic class struggle as depicted in Figure 4 for formal working-class power (Esping-Andersen, 1990; Korpi, 1983). Although the underlying dynamics of welfare regime formations have evolved significantly since the publication of The Three Worlds of Welfare Capitalism (1990), there are a long list of studies that confirm that the basic Three Worlds structure and its founding principles in Western countries still largely hold (Bambra, 2006; Esping-Andersen, 1999; Korpi and Palme, 2003; Powell and Barrientos, 2004; Vrooman, 2012; also see, Powell et al. (2020b) for a recent review of the welfare regimes literature).

In the residual regime, we hypothesize that social security systems (old age pension, unemployment and sickness benefits) are not substantially developed because the post-war developmentalist period in these countries did not materialize in a strong working-class movement with enduring policy effects, and the working-class mobilization was not strong enough during the neoliberal period to successfully resist the retrenchment of previously established social security systems (see, Figure 4). Although there is some level of poor people’s movements that radicalize on different ethnic and racial grounds (Figure 3), these countries also lack enough financial resources to sustain social assistance and healthcare programmes that are as strong as those in the populist welfare state regime type, which decelerated the expansion of social assistance programmes (Yörük et al., 2019a).

We designate the final cluster the neoliberal regime because the composition of welfare components reflects the central ideology of the liberal and neoliberal welfare logic: cut-offs in social security benefits and the existence of social assistance that targets those who fail in the market (Esping-Andersen, 1990). We hypothesize that while the lack of strong working-class movements during the post-war period has hindered the development of strong social security policies (see, Figure 4), the existence of political radicalization of poor people in the 1960s has triggered social assistance policies as political containment strategies (Figure 3), echoing the central argument of Piven and Cloward (1971). As we stated earlier, these educated hypotheses are intended for indicating potential mechanisms that triggered welfare state regimes and for encouraging further research that could test such arguments with comprehensive quantitative and qualitative comparative analyzes, which would need to control for other confounding factors.

Overall, our working hypotheses for the mechanisms of welfare regime development can be summarized as follows:

High levels of poor people’s mobilization during the neoliberal era, which followed high levels of labour movements during the developmentalist period, we contend, have pushed the governments in the Global South to generate a populist welfare state regime.

High levels of poor people’s activism during the post-war period without the presence of a strong labour movement are associated with the formation of a neoliberal welfare state regime.

Low levels of labour militancy during both post-war and neoliberal period, accompanied by intermediate levels of poor people’s activism, are associated with the formation of the residual welfare state regime.

The presence of high and intermediate levels of labour militancy during post-war and neoliberal period, respectively, accompanied by the absence of a strong poor people’s movement, is associated with the formation of the institutional welfare state regime

Conclusion

This article contributes to the welfare state regimes literature by illustrating that there are four worlds of welfare state regimes when the analysis is extended from developed countries, an expanded set of cases, including emerging markets. To address the challenges of the welfare state regimes literature, we introduce a novel dataset that contains genuine welfare policy variables, which represent the most crucial welfare generosity components such as pensions, unemployment schemes, sickness benefits, social assistance and healthcare. We conduct three different methods of cluster analysis and show that 52 countries in our sample, which consist of OECD members and emerging market economies, constitute four welfare state regimes: populist, institutional, residual and neoliberal.

We observe that institutional regime cases are the most generous welfare states, but they differ in the extent to which health and social assistance play a role in the overall regime. The populist regime is characterized by generous social insurance policies and relatively less generous health and social assistance programmes. Furthermore, in the last two clusters, we observe a lack of one of the principal components of welfare analyzed in this study–unemployment insurance in the residual regime and sickness benefits in the neoliberal regime, while the rest of the programmes are performing very poorly.

We then hypothesize that the relative strength of formal and informal working classes is a key factor that expanded and shaped welfare state regime typologies, as is observed by the mid-2010s. We call for future studies that will examine the historical developments of these regimes and present the mechanisms behind them with full empirical support. Only then we can reach a broader welfare regime theory that is likely to ‘emerge from global comparative research’ (Powell and Barrientos, 2015: 244).

Supplemental Material

sj-pdf-1-esp-10.1177_09589287211050520 – Supplemental Material for The four global worlds of welfare capitalism: Institutional, neoliberal, populist and residual welfare state regimes

Supplemental Material, sj-pdf-1-esp-10.1177_09589287211050520 for The four global worlds of welfare capitalism: Institutional, neoliberal, populist and residual welfare state regimes by Erdem Yörük, İbrahim Öker and Gabriela Ramalho Tafoya in Journal of European Social Policy

Footnotes

Acknowledgments

The authors would like to thank the team members of the Emerging Welfare ERC project (), Kerem Yıldırım, Burcu Yakut-Çakar, Lyle Scruggs, Armando Barrientos, the editor of the Journal of European Social Policy, and two anonymous reviewers for their comments on earlier drafts of this article. We also benefitted immensely from feedback given during presentations at American Sociological Association Annual Meeting (2019), International Sociological Association, RC19 Meeting (2019) the International Conference on Public Policy (2019), Midwest Political Science Association Annual Conference (2018), and East Asian Social Policy Conference (2019).

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study is supported by H2020 European Research Council (714868).

ORCID iD

Erdem Yörük

Supplemental material

Supplemental material for this article is available online.

Notes

References

Abu Sharkh

Gough

(2010) Global welfare regimes. Global Social Policy 10(1): 27–58.

Acemoglu

Naidu

Restrepo

, et al. (2019) Democracy does cause growth. Journal of Political Economy 127(1): 47–100.

Aytaç

Öniş

(2014) Varieties of populism in a changing global context: The divergent paths of Erdoğan and Kirchnerismo. Comparative Politics 47(1): 41–59.

Bambra

(2005) Cash versus services: ‘worlds of welfare' and the decommodification of cash benefits and health care services. Journal of Social Policy 34(2). Cambridge, UK: Cambridge University Press: 195–213.

Bambra

(2006) Research note: Decommodification and the worlds of welfare revisited. Journal of European Social Policy 16(1): 73–80.

Banks

Wilson

(2021) Cross-National Time-Series Data Archive. Databanks International. Jerusalem, Israel. https://www.cntsdata.com/.

Biernacki

Celeux

Govaert

(2000) Assessing a mixture model for clustering with the integrated completed likelihood. IEEE Transactions on Pattern Analysis and Machine Intelligence 22: 719–725.

Böhmelt

Clayton

(2018) Auxiliary force structure: paramilitary forces and progovernment militias. Comparative Political Studies 51(2): 197–237.

Bonoli

(1997) Classifying welfare states: A two-dimension approach. Journal of Social Policy 26(3): 351–372.

10.

Buğra

Keyder

(2006) The Turkish welfare regime in transformation. Journal of European Social Policy 16(3): 211–228.

11.

Castles

(2009) What welfare states do: A disaggregated expenditure approach. Journal of Social Policy 38(1): 45–62.

12.

Clasen

Siegel

(2007) Investigating Welfare State Change. Cheltenham, UK: Edward Elgar.

13.

Conniff

(2012) Populism in Latin America. Tuscaloosa, AL: The University of Alabama Press.

14.

Cook

(2010) Eastern Europe and Russia. In: Castles

Leibfried

Lewis

, et al. (eds) The Oxford Handbook of the Welfare State. Oxford, UK: Oxford University Press.

15.

Cramer

(2003) Advanced Quantitative Data Analysis. Maidenhead, UK: Open University Press.

16.

Dasgupta

Raftery

(1998) Detecting features in spatial point processes with clutter via model-based clustering. Journal of the American Statistical Association 93(441): 294–302.

17.

De la Torre

(2017) Populism in Latin America. In: Kaltwasser

Taggart

Espejo

, et al (eds) The Oxford Handbook of Populism. Oxford, UK: Oxford University Press, 195.

18.

Durr-e-Nayab Farooq

(2014) Effectiveness of cash transfer programmes for household welfare in Pakistan: The case of the Benazir income support programme. The Pakistan Development Review 53(2): 145–174.

19.

Ebbinghaus

(2012) Comparing Welfare State Regimes: Are Typologies an Ideal or Realistic Strategy? Edinburgh: ESPAnet Conference.

20.

Eder

(2010) Retreating state? Political economy of welfare regime change in Turkey. Middle East Law and Governance 2(2): 152–184.

21.

Esping-Andersen

(1999) Social Foundations of Postindustrial Economies. Oxford, UK: OUP.

22.

Esping-Andersen

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

23.

Ferragina

Seeleib-Kaiser

(2011) Thematic review: Welfare regime debate: Past, present, futures? Policy & Politics 39(4): 583–611.

24.

Ferrera

(1996) The 'Southern model' of welfare in social Europe. Journal of European Social Policy 6(1): 17–37.

25.

Fraley

Raftery

(1998) How many clusters? Which clustering method? Answers via model-based cluster analysis. The Computer Journal 41(8): 578–588.

26.

Fraley

Raftery

(2002) Model-based clustering, discriminant analysis, and density estimation. Journal of the American Statistical Association 97(458): 611–631.

27.

Furniss

Tilton

(1977) The Case for the Welfare State: From Social Security to Social Equality. Bloomington, IN: Indiana University Press.

28.

Gao

(2017) Welfare, Work, and Poverty: Social Assistance in China. Oxford, UK: Oxford University Press.

29.

Gough

(2001) Social assistance regimes: A cluster analysis. Journal of European Social Policy 11(2): 165–170.

30.

Gough

Wood

(2004) Insecurity and Welfare Regimes in Asia, Africa and Latin America: Social Policy in Development Contexts. Cambridge University Press.

31.

Grigera

(2017) Populism in Latin America: Old and new populisms in Argentina and Brazil. International Political Science Review 38(4): 441–455.

32.

Hacker

(2002) The Divided Welfare State: The Battle over Public and Private Social Benefits in the United States. New York, NY: Cambridge University Press.

33.

Haggard

Kaufman

(2008) Development, Democracy, and Welfare States: Latin America, East Asia, and Eastern Europe. Princeton, NJ: Princeton University Press.

34.

Howard

(1999) The American welfare state, or states? Political Research Quarterly 52(2): 421–442.

35.

Huber

Bogliaccini

(2012) Latin America. In: Castles

Leibfried

Lewis

, et al. (eds) The Oxford Handbook of the Welfare State. Oxford, UK: Oxford University Press, pp. 644–655.

36.

Imbrogno

(1991) State and welfare: Comparative social policy in the USA and USSR. Journal of International and Comparative Social Welfare 7(1–2): 15–30.

37.

Jaffrelot

Tillin

(2017) Populism in India. In: Kaltwasser

Taggart

Espejo

, et al. (eds) The Oxford Handbook of Populism. Oxford, UK: Oxford University Press.

38.

Kangas

(1994) The Politics of social security: On regressions, qualitative comparisons, and cluster analysis. In: Janoski

(ed) The Comparative Political Economy of the Welfare State. Cambridge, UK: Cambridge University Press, 346–364.

39.

Keribin

(2000) Consistent estimation of the order of mixture models. Sankhyā: The Indian Journal of Statistics, Series A 62(1): 49–66.

40.

Keyder

(1987) State and Class in Turkey: A Study in Capitalist Development. Brooklyn, NY: Verso.

41.

Korpi

Palme

(2003) New politics and class politics in the context of austerity and globalization: Welfare state regress in 18 countries, 1975–95. The American Political Science Review 97(3): 425–446.

42.

Korpi

(1983) The Democratic Class Struggle. London and Boston, MA: Routledge.

43.

Kotwas

Kubik

(2019) Symbolic thickening of public culture and the rise of right-wing populism in Poland. East European Politics and Societies: And Cultures 33(2): 435–471.

44.

Levitsky

Roberts

(2011) The Resurgence of the Latin American Left. Baltimore, MD: Johns Hopkins University Press.

45.

(2012) Welfare state changes in China since 1949. In: Greve

(ed) Routledge Handbook of the Welfare State. London, UK: Routledge, 222–231.

46.

Mirkin

(2016) Clustering for Data Mining: A Data Recovery Approach. Boca Raton, FL: CRC, Chapman and Hall.

47.

Mudde

Kaltwasser

(2019) Populism: A Very Short Introduction. Oxford, UK: Oxford University Press.

48.

Nelson

Fredriksson

Korpi

, et al. (2020) The social policy indicators (SPIN) database. International Journal of Social Welfare 29(3): 285–289.

49.

Obinger

Wagschal

(2010) Social expenditure and revenues. In: Castles

Leibfried

Lewis

, et al. (eds) The Oxford Handbook of the Welfare State. London, UK: Oxford University Press.

50.

Perova

Vakis

(2012) 5 years in Juntos: New evidence on the program’s short and long-term impacts. Economia 35(69): 53–82.

51.

Piven

Cloward

(1971) Regulating the Poor: The Functions of Public Welfare. New York, NY: Pantheon Books.

52.

Piven

Minnite

(2016) ‘Poor People’s Politics’. In: Brady

Burton

(eds) The Oxford Handbook of the Social Science of Poverty. London, UK: Oxford University Press.

53.

Powell

Barrientos

(2004) Welfare regimes and the welfare mix. European Journal of Political Research 43(1): 83–105.

54.

Powell

Barrientos

(2015) Introduction: Twenty-five years of the welfare modelling business. In: Ruby

(ed) Social Policy and Society. Cambridge, UK: Cambridge University Press, Vol. 14, 241–245.

55.

Powell

Kim

(2020a) The puzzle of Japans welfare capitalism: a review of the welfare regimes approach. Journal of International and Comparative Social Policy 36(1): 92–110.

56.

Powell

Yörük

Bargu

(2020b) Thirty years of the three worlds of welfare capitalism : A review of reviews. Social Policy and Administration 54(1): 60–87.

57.

Prasad

(2016) American exceptionalism and the welfare state: The revisionist literature. Annual Review of Political Science 19(1): 187–203.

58.

Pribble

(2013) Welfare and Party Politics in Latin America. Cambridge, UK: Cambridge University Press.

59.

Ramos da Silva Lopes

(2006) Welfare Arrangements, Safety Nets and Familial Support for the Elderly in Portugal. London, UK: London School of Economics and Political Science.

60.

Remmer

(2012) The rise of leftist-populist governance in Latin America. Comparative Political Studies 45(8): 947–972.

61.

Saraceno

(2002) Social Assistance Dynamics in Europe: National and Local Poverty Regime. Bristol, UK: Policy Press.

62.

Scruggs

(2014) Social Welfare Generosity Scores in CWED 2: A Methodological Genealogy. Working Paper 01. University of Connecticut. Storrs, Connecticut.

63.

Scruggs

Jahn

Kuitto

(2014) Comparative Welfare Entitlements Dataset 2. Storrs, CT: University of Connecticut. Version 2014 3.

64.

Scruggs

Hayes

(2017) The influence of inequality on welfare generosity. Politics & Society 45(1): 35–66.

65.

Seekings

Nattrass

(2015) Income support through the welfare state. In: Seekings

Nattrass

(eds) Policy, Politics and Poverty in South Africa: Developmental Pathways to Poverty Reduction. London, UK: Palgrave Macmillan, 133–161.

66.

Silver

(2003) Forces of Labor: Workers’ Movements and Globalization since 1870: Cambridge Studies in Comparative Politics. Cambridge, UK: Cambridge University Press.

67.

Social Security Administration (SSA) (2016) Social Security Programs throughout the World: The Americas, 2015. Washington, DC: SSA.

68.

Tan

P-N

Steinbach

Kumar

(2013) Introduction to Data Mining. 2nd edition. Boston, MA: Addison-Wesley.

69.

Thachil

(2014) Elite parties and poor voters: theory and evidence from India. American Political Science Review 108(2): 454–477.

70.

Titmuss

Abel-Smith

Titmuss

(1974) Social Policy: An Introduction. London, UK: George Allen and Unwin.

71.

Tryfos

(1998) Methods for Business Analysis and Forecasting: Text and Cases. New York, NY: Wiley.

72.

Vrooman

(2012). Regimes and cultures of social security: Comparing institutional models through nonlinear PCA. International Journal of Comparative Sociology 53(4–6): SAGE Publications Ltd, 444–477.

73.

Walker

Wong

(2005) Introduction: East Asian Welfare Regimes. In: Walker

Wong

(eds) East Asian Welfare Regimes in Transition: From Confucianism to Globalisation. Bristol, UK: University Press Scholarship online (Policy Press)Policy Press, 3–20.

74.

Walker

Wong

(2013) The ethnocentric construction of the welfare state. In: Kennett

(ed) A Handbook of Comparative Social Policy. 2nd Edition. Cheltenham, UK: Elgar, 98–114.

75.

Wilensky

Lebaux

(1958) Industrial Society and Social Welfare. New York, NY: Russell Sage Foundation.

76.

Wolfe

(2010) Populism and developmentalism. In: Holloway

(ed) A Companion to Latin American History. Hoboken, NJ: John Wiley & Sons, 347–364.

77.

World Bank (2015) The State of Social Safety Nets 2015. Report no 97882. Washington, DC: The World Bank.

78.

Yörük

(2012) Welfare provision as political containment. Politics & Society 40(4): 517–547.

79.

Yörük

Öker

Şarlak

(2019a) Indigenous unrest and the contentious politics of social assistance in Mexico. World Development 123, 104618.

80.

Yörük

Öker

Yıldırım

, et al. (2019b) The variable selection problem in the three worlds of welfare literature. Social Indicators Research 144(2): 625–646.

81.

Zarate Tenorio

(2014) Social spending responses to organized labor and mass protests in Latin America, 1970–2007. Comparative Political Studies 47(14): 1945–1972.

82.

Zucco

(2013) When Payouts Pay Off: Conditional Cash Transfers and Voting Behavior in Brazil 2002–10. American Journal of Political Science 57(4): 810–822.

Supplementary Material

Please find the following supplemental material available below.

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

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

0.00 MB

0.57 MB