Straddling two continents and beyond three worlds? The case of Turkey’s welfare regime

Abstract

This article aims to consider how Turkey has been classified in the welfare regime literature, and on what basis it has been classified. This will then form the basis for exploring whether there appears to be any variation between approaches and methods and/or between the “position” (e.g., location or language) of the authors. Studies of Turkey’s welfare regime exhibit a significant degree of variation in terms of both approaches and conclusions, resulting in little in the way of consensus. Among Turkish-language studies (and some, but not all, Turkish scholars writing in English), there does seem to be a broad consensus that Turkey may be classified as part of the Southern European welfare model, which squares with the modal conclusion of the English-language studies on the topic. However, some “regional” studies conclude that Turkey is part of the Middle East and North Africa (MENA) region, while many of the cluster analyses suggest a wide variety of clusters that are not geographically contiguous.

Introduction

There have been extensive debates on Gøsta Esping-Andersen’s classic text The Three Worlds of Welfare Capitalism, ¹ which is arguably the most cited work in social policy. Much of this debate, however, focuses on his original “three worlds,” which comprises 18 nations. There has been some material adding an extra “world” (e.g., “Southern European”) or a few nations at the margins, but much less focus on moving beyond the original “three worlds” to explore those nations that actually contain the vast majority of the world’s population. While a number of scholars have explored broadening theories of Western welfare states and analyses of Esping-Andersen’s “three worlds” so as to encompass other nations as well, ² it remains unclear if the original concepts and measures that were designed (largely) for Western Europe are appropriate for other worlds.

In this article, we explore these issues by presenting a meta-analysis of the scholarly works that have classified the Turkish welfare regime as part of existing welfare regime clusters. We have chosen to focus on the case of Turkey because it differs in a number of ways from Esping-Andersen’s original 18 nations. First, it is a middle- rather than a high-income country. Second, Turkey’s historical and geographical position within the Middle East has largely shaped its welfare regime, with the population being predominantly Muslim and the Islamic tradition of charity being deeply entangled in the modern Turkish welfare system. ³ At the same time, Turkey is also a member of the Organisation for Economic Co-operation and Development (OECD) and was accepted as an official European Union (EU) candidate in 1999, which was followed by the opening of accession negotiations in 2005. ⁴ Moreover, scholars appear to be uncertain about the Turkish welfare regime, as they have produced a wide and rather confusing variety of different terms and classifications, such as “eclectic social security regime,” ⁵ “successful informal security regime,” ⁶ “Southern European or Mediterranean,” ⁷ “inegalitarian corporatist,” ⁸ “minimalist welfare state,” ⁹ “residual,” ¹⁰ “productive plus” regime, ¹¹ and “egalitarian corporatist system.” ¹²

This article aims to consider on what basis Turkey’s welfare regime has been classified in the existing literature by conducting a structured literature search within the Scopus database. This search will then form the basis for exploring whether there appears to be any variation between approaches and methods and among the “position” of the authors (in terms of, e.g., location or language). For example, are there differences between the conclusions of Turkish and non-Turkish scholars, or between studies written in the English or Turkish languages? We demonstrate that there is an extensive and growing literature on the Turkish welfare regime. The “three worlds” literature continues to expand, a trend that is apparently reflected in the number of studies analyzing the Turkish case. This study will specifically examine the literature on Turkey’s welfare regime in order to help researchers critically evaluate the existing scholarship and locate their prospective work both within and beyond the existing literature. By comparing the Turkish- and English-language literature, as well as large-N, regional, and single-country studies, it will aid in understanding the epistemological taxonomies that underlie welfare regime classification. The main contribution of the study is that it will help students of the Turkish welfare regime observe the dominant tendencies or paradigms that shape the welfare regime classifications of Turkey and assist them in figuring out ways in which they might contribute to the literature, either by criticizing the existing paradigms or by enhancing them.

As will be demonstrated, most scholars have classified Turkey as part of the “Southern European” welfare regime sub-family. This paradigmatic Southern European framework has arguably weakened the possibility of considering Turkey according to any other alternative welfare regime models. As will be shown in the next section, the empirical data from Turkey’s welfare system, including coverage and expenditure on different social programs, indicate that Turkey’s welfare regime seems to have been transformed from a corporatist/Southern European model into a new model that exemplifies characteristics of both liberal and social democratic models, possibly implying an overall new welfare regime. It can thus be argued that Turkey’s welfare regime may be joining a welfare regime family of emerging market economies, in which income-based social assistance programs constitute the central element of an expanding welfare state. Our analysis, as will be seen, indicates that none of the studies that has examined Turkey has considered the possibility that the Turkish welfare regime may have embarked in just such a novel direction.

The Turkish welfare regime

Esping-Andersen’s welfare regime approach ¹³ appears to have relevance to the Turkish case. Many scholars have argued that the configurations of welfare provision in Turkey comply with the analytical requirements of “being a regime,” ¹⁴ as welfare is generated within a tripartite relationship among state, family, and the market. The key axes of Esping-Andersen’s approach—decommodification, stratification, and defamilialization—are critical to understanding the case of Turkey. Dispossession and rapid rural-to-urban migration have brought about an extensive commodification of labor, and the existing welfare regime has developed institutions for the protection of private and public sector employees for the purpose of decommodification. The welfare system is certainly hierarchical, and it has produced stratification of varying degrees and types in different decades of Turkish welfare regime development. Familialization is also a critical dimension in the Turkish case, as the welfare system can be shown to shape women’s place in the family and in society, with women’s emancipation from family ties being strengthened and weakened by different sets of welfare policies constructed in different political periods. ¹⁵

The roots of the social welfare system in Turkey can be traced back to the endowment (waqf) systems of the Ottoman Empire. Since then, welfare provision has been shaped by both European welfare systems and traditional Islamic values of charity. ¹⁶ As in many other middle-income countries, there were moves toward a modern welfare system after World War II, including the establishment of the Ministry of Labor and Social Security (Çalışma ve Sosyal Güvenlik Bakanlığı) in 1945, but a welfare system containing both formal and informal components remained in place until the 1990s. ¹⁷

Before the 2000s, the Turkish welfare system was based on a corporatist fragmented social provision, in which employees in the state sector, workers, and the self-employed were members of different institutions with different qualities of service and different benefits. Since the turn of the new millennium, a new welfare system has largely eliminated this fragmented structure, creating a social security institution and a general health insurance system to cover all citizens. Within this structure, social welfare programs aimed at the informal poor, most importantly healthcare, have been equalized with those aimed at formal sector employees. In this universal healthcare system, all citizens are covered and the premiums of formal sector employees are paid by the employees, with the premiums of the informal poor being paid by the government and the premiums of the informal non-poor recently being lowered to almost 15 US dollars. The quality of healthcare has significantly improved, which has often been regarded as one of the main pillars of the social policies of the ruling Justice and Development Party (Adalet ve Kalkınma Partisi, AKP). ¹⁸ As a result, Turkish welfare system expenditures as a percentage of the GDP have increased from 3.1 percent in 1980 to 12.5 percent in 2013. ¹⁹

This expansion has been accompanied by a restructuring of welfare provision from fragmented employment-based social security policies to income-based social services and social assistance policies. ²⁰ In 2011, the Ministry of Family and Social Policy (Aile ve Sosyal Politikalar Bakanlığı) was established to administer central government programs and to introduce new social assistance programs. The ministry’s budget increased from 955 million US dollars in 2002—when the relevant affairs were dealt with by a separate agency—to 11.6 billion US dollars in 2014. The ratio of social expenditures to the GNP increased from 0.5 percent to 1.51 percent during the same period. ²¹ Moreover, between 2000 and 2010, the percentage of social assistance spending in total government spending increased by 266 percent. ²² The government has drastically expanded means-tested social assistance, in kind or cash transfers, as well as providing free healthcare programs for the poor, conditional cash transfers, programs for orphans, food stamps, housing, education and disability aid for the poor, and sharply increasing the number of beneficiaries and the share of government budgets allocated. ²³ The coverage of the free healthcare card program for the poor (the Green Card or Yeşil Kart program) increased from 4.2 to 12.7 percent of the population between 2003 and 2009, and in 2012, Green Card holders were included in the new universal healthcare system. ²⁴ As such, the regular in-kind and cash benefits granted by the central government to a poor family adds up to 260 US dollars, while the official minimum wage in Turkey is 370 US dollars. ²⁵

The government has also initiated social housing programs targeting the poor, which provide houses on cheap credit and covered over three million families by 2013. Furthermore, the Ministry of National Education (Millî Eğitim Bakanlığı) distributed all school coursebook materials free of charge to all students in primary and secondary education, amounting to 15 million students. Each year, 600,000 students also take part in the free school transportation program, which provides free lunch at schools as well. The government has put into the constitution affirmative action policies for disabled people, which greatly increased their participation in the labor market, and what is more, since 2002, the coverage of disability benefits has tripled, while the amount of per capita payments has doubled. This is also the case for old age pensions for the poor. Most importantly, if a poor family provides nursing to a disabled family member or elderly person, it receives an amount almost equal to the minimum wage. Similarly, a new program granted poor grandparents a monthly wage of more than 100 US dollars for providing daycare for their grandchildren.

With its current structure, the Turkish welfare system could be considered part of an emerging welfare regime family together with other emerging market economies. Piven and Minnite argue that this emerging family signifies a bifurcation in the global welfare system: while advanced capitalist countries have undergone significant welfare retrenchments that reduced the capacity to decommodify labor and level out inequalities, emerging market economies have expanded their welfare states, mostly on the basis of novel types of social assistance programs that have significant decommodifying effects. ²⁶ Similarly, the International Labour Organization (ILO) argues that a bifurcation is occurring between an eroding “European social model” and a rising “inclusive growth through social protection” model in emerging markets. ²⁷ Over the last three decades, many emerging market economies have rapidly expanded their social assistance programs and services for the poor by increasing program coverage and benefits and by relaxing the terms and conditions of eligibility, while in most developed countries social assistance-type schemes continue to have a more residual role. ²⁸

Data and Analysis

This study is based on a structured search for all studies that have attempted to classify the Turkish welfare regime. The parameters for the search aim to discover those sources that focus on social welfare (rather than economic classifications) in Turkey and that involve placing the country within some sort of typology, classification, or grouping. Two searches were carried out, using the same search terms in the English and Turkish languages. The search for this article was carried out on the Scopus database in April 2016, using the following search terms:

Turk$ AND ‘social welfare’ AND regime$ OR class OR typ$

The English-language search produced 58 results. The main criterion for inclusion was that the article had to classify Turkey within a regime, model, or cluster. Snowballing was carried out, with the final number of studies being 28. The Turkish-language search initially produced 18 results, including 10 articles and 7 books. Application of the inclusion criteria produced 6 studies.

Studies written in English

Studies written in English can be classified into three broad approaches to classifying Turkey: large- or medium-N global studies; “regional studies,” although there is limited consensus on the “region” in question (see below); and single-country case studies.

Large-N studies

The articles in this category are generally “large-N”, statistically classifying a large number of countries on the basis of data from sources such as the OECD via techniques such as cluster analysis. We will demonstrate that there is no consensus among “large-N” studies on the classification of the Turkish welfare regime. Gough regarded Turkey as part of the Southern European or Mediterranean welfare regime cluster. ²⁹ Eardley et al. identified at least seven “social assistance regimes,” placing Turkey in the “rudimentary assistance” regime, together with the countries of southern Europe. ³⁰ Gough et al.’s ³¹ analysis of social assistance in OECD countries slightly modified Eardley et al. to find eight “social assistance regimes”; however, Turkey was again considered as part of the “rudimentary assistance” regime, along with Italy, Spain, Portugal, and Greece. Gough ³² further noted that the eight “social assistance regimes” identified by Gough et al. derived from their own judgements, though now he used cluster analysis on the same data so as to test their earlier judgements. Due to inadequate data, Iceland, Italy, the Netherlands, and Turkey were excluded from the analysis, leaving 20 countries. The best solution produced seven clusters: the “minimal extent, exclusive, very low benefits” cluster included “Greece, Portugal (Turkey?),” with the closest 1997 grouping being Greece, Italy, Portugal, Spain, and Turkey. In other words, it appeared that the classification of Turkey (which was excluded due to inadequate data) was a best guess. Gough claimed that his results challenged Esping-Andersen’s threefold division of welfare regimes, as there were divisions within the Mediterranean sub-regime as well, suggesting that this fourth regime was not fully cohesive. He also changed his view, overturning his earlier assertion ³³ that there was a distinctive Southern European assistance regime.

Abu Sharkh and Gough ³⁴ carried out cluster analysis on 65 nations. They found four substantial cluster groups in 1990 and eight clusters in 2000. In both years, Turkey was in Cluster B (“Successful Informal Security Regime”), which contained 21 countries in 1990 (one-third of the total, representing several world regions, together with Iran and Tajikistan) and a more complex cluster of 16 countries in 2000 (together with Iran). They pointed out that the countries in this group were primarily, but not necessarily, low-middle income with high growth rates, but were relatively undemocratic and unequal. Moreover, it was a disparate group in cultural and historical terms.

Gough and Abu Sharkh ³⁵ explored how the composition of public revenues in terms of sources (e.g., taxation, social insurance contributions, mineral rents, and foreign aid) is associated with different welfare regimes and social policy outcomes. Using cluster analysis, they identified four clusters of countries where one form of revenue is dominant, and a fifth where this was not the case. The cluster with substantial tax revenues was, unexpectedly, limited to just two world regions: the Middle East and North Africa (MENA), and southern and eastern Africa. Every MENA country for which they had data, excepting only Yemen and Iran, was in this group. Cluster B, however, contained considerable variation, with Turkey (20 percent of GDP), Chile (17 percent), and Korea (16 percent) recording tax revenues higher than many in cluster A. They concluded that the Gough-Wood label “informal security regime” did not do full justice to such countries.

Rudra ³⁶ argued that not all least developed countries (LDCs) have “productive welfare states” that direct welfare efforts primarily toward encouraging wage labor. Using a combination of policy, spending, and outcome (i.e., literacy, mortality, and immunization rates) variables, she carried out a cluster analysis of 32 nations, with a three-group solution being the most distinct solution. Clusters 1 and 3 appeared to favor the productive and protective components of welfare, respectively. Turkey was in Cluster 3, together with 12 countries—including Egypt, Iran, Tunisia and Morocco—while Cyprus, Greece, Israel, and Kuwait were in Cluster 1.

Hudson and Kühner ³⁷ examined the protective and productive dimensions of welfare for 55 high- and higher-middle-income countries by means of fuzzy set ideal type analysis (FSITA). They used data on income protection, employment protection, training investment, education investment, and the percentage of non-wage workers in total employment. Turkey was placed in the “Productive Plus” group, which also included Denmark, Estonia, Norway, Poland, and Mexico. They concluded that welfare state types did not simply interface with regional blocs in terms of the balance between welfare’s productive and protective dimensions.

Ağartan ³⁸ regarded Turkey as an “inegalitarian corporatist system” that guaranteed employment-related benefits for most formal sector workers but excluded wide sections of the population. In a section entitled “Theorizing Turkish Welfare Regime,” she argued that, of Seeking’s ³⁹ three types, the Turkish welfare system seems to best be described by the “inegalitarian corporatist” type: a formal social insurance system that provides social protection to privileged groups and excludes rural populations and the urban poor. ⁴⁰ Turning to Gough et al., ⁴¹ she noted that, while Turkey is not included in the analysis except for a brief reference, with its above-average welfare outcomes in all areas—i.e., life expectancy, literacy, and freedom from poverty—and its below-average levels of public expenditure, it seems to belong to either their first or second category. She continued that, after 1960, the Turkish welfare system attained a dual character that, to use Gough et al.’s terminology, ⁴² combined elements of a welfare state regime and an informal security regime. She further noted that, while the Turkish welfare regime had been undergoing major transformation since the mid-1990s, it was nevertheless not possible to argue that the Turkish welfare system had evolved towards a European-type welfare state. Rather, a dual welfare structure emerged, consisting of formal social security schemes for industrial workers and civil servants and an informal welfare system to provide some degree of protection to those left out of the formal system.

Aysan ⁴³ analyzed the Turkish pension system and pension reforms in relation to welfare regime discussions. He stated that, among the G20 and OECD countries, Turkey, with its contradictory socioeconomic features, emerged as an unusual case. Contrary to claims that the Turkish welfare state and social assistance system were residual, ⁴⁴ he argued that the Turkish state had a significant role in welfare distribution, particularly regarding the elderly. Examining eleven pension and labor market variables for the period of 2009–2011 for 19 OECD nations, he performed a cluster analysis that revealed three broad pension regimes: Southern European, Continental European, and social-liberal. The Southern European regime included Greece, Spain, and Turkey (as well as Denmark, though Italy joined the Continental European group while Portugal merged with the social-liberal group). He concluded that Turkey can be considered under the Southern European welfare regime, which is a distinct welfare regime with mixed features of citizenship-based Beveridgean universal national health services and occupation-based Bismarckian income transfers.

As shown in Table 1, all these studies are very diverse in terms of the number of countries they consider, as well as in terms of their underlying concepts and variables. Most importantly, there is little consensus as to the classification of Turkey, which is placed under a number of different headings and classed alongside a diverse mixture of other countries.

Table 1 Classification and Analysis of Studies on Turkish Welfare Regime (English-Language Studies)

Regional Studies

The regional approach is based on the assumption that geographical location is the key to classification. In other words, clusters of nations are not based on characteristics produced by techniques such as cluster analysis, but rather on geographical location. “Regions,” therefore, are made up of neighboring countries. However, as we shall see, there appears to be little consensus about which region Turkey should be placed in, as a wide variety of terms with different constituent nations, have been used, sometimes even by the same author: e.g., the Arab region; ⁴⁵ the Middle East; ⁴⁶ the Middle East and North Africa (MENA); ⁴⁷ “South Eastern Europe”; ⁴⁸ Southern Europe; ⁴⁹ an extended family of Mediterranean nations; ⁵⁰ or a hybrid Southern Europe/MENA model, reflecting Turkey’s place straddling the two continents. ⁵¹

Henry and Springborg ⁵² classified Turkey as one of the 22 countries within the MENA region, which is defined by Islam and the tradition of external intervention. Manning ⁵³ claimed that, in terms of welfare state classification, Turkey has less in common with Esping-Andersen’s core types than with the so-called “Latin Rim,” where a strong state and a family-centered tradition prevailed. Deacon and Stubbs ⁵⁴ classified Turkey as “inegalitarian corporatist,” but noted that it is an outlier within South East Europe, as it was the only state that lacked a commitment to universal health and education.

Grütjen ⁵⁵ argued that Turkey fit the key characteristics of the ideal type of the Southern European Model, as cross-national comparison revealed only a few deviations from this model, while the predominance of the state and family in social provision was striking. For this, he cited the World Bank’s claim ⁵⁶ that it is “difficult to overstate the importance of marriage, family and extended family ties” in the Turkish context. He used the key characteristics of the Southern European Model of welfare to classify the Turkish welfare regime: comparing data from Turkey and the four representatives of the Southern European Model (Spain, Italy, Portugal, and Greece), as well as representatives (the United Kingdom, Sweden, and Germany) of the three welfare regimes outlined by Esping-Andersen, he emphasized the similarities between Turkey and Southern Europe. His argument was that, in Turkey much as in Southern Europe, the social security system is polarized and protects an occupational core, the level of state penetration in the social realm is extremely low, and a safety net in the form of a social assistance scheme is absent. The most significant resemblance, though, is the importance of the family as a main institution of welfare. In other respects, however, Turkey did not match the ideal type of a Southern European Model.

Jawad ⁵⁷ pointed to the main themes of recent texts concerning Middle Eastern Welfare States: the influence of international intervention; a “statist” political-economic approach to social policy; how state social policy is residual in character, with public policy focused on the “economic” rather than the “social”; clientelism, the politicization of welfare, and the instrumental use of social policy by the state to gain power and political legitimacy; and the influence of Islam on social policy. In the end, however, she concluded that this was only a partial analysis of social policy in the region, presenting a revised history of social policy based on the welfare state as a particular cultural settlement.

Gal ⁵⁸ claimed the existence of an “extended family of Mediterranean welfare states” consisting of eight nations (Cyprus, Greece, Israel, Italy, Malta, Spain, Portugal, and Turkey), which share a number of interesting and crucial commonalities while differing in other ways. For example, this family diverged from the Southern European Model in that it did not have a Bismarckian social protection legacy. He noted that, even though Turkey was clearly an outlier with regard to living standards and social expenditure, observers both within the country and abroad included it in a discussion of the Mediterranean family of nations. ⁵⁹

Aybars and Tsarouhas ⁶⁰ considered Turkey to be part of the geographic Middle East, sharing historical, political, and cultural traits with other states in the region. Nevertheless, it represented an exceptional case, especially with respect to Europeanization processes in a variety of policies, including social policy. Looking at major trends in social security and labor market arrangements, they argued that Turkey represented a “hybrid” model, one that embodied the characteristics of both Middle Eastern and South European welfare models, yet remained an outlier to both in certain respects.

Gencler and Colak ⁶¹ claimed that Turkey matched the characteristics of the so-called Southern European Model of welfare, ⁶² along with Spain, Greece, Italy, and Portugal. Their cross-national comparison revealed only a few deviations from this model and noted that “the similarities are remarkable.” They went on to state that the most significant common trait of the welfare regimes in Turkey and the rest of Southern Europe was the importance of the family as a main institution of welfare. The existing welfare system in Turkey, they stated, can be considered a minimal and indirect, informal security regime.

Dedeoglu ⁶³ discussed residual welfare states in the MENA region before turning to feminist typologies of welfare states. She pointed out Esping-Andersen’s focus on decommodification, which omitted the central role of informal institutions in the provision of welfare in low-income countries. She drew on Moghadam, ⁶⁴ who argued that the “developmentalism” of MENA states was undermined by its “neopatriarchal” approach to women, gender, and the family, yet Turkey was somehow different from the other Middle East countries in that, in the mid-1920s, it had placed limitations on the Islamic influence on the organization of the state and headed toward a secular state structure. Women’s status in the Turkish welfare system showed a dualistic character of “veiled Europeanisation,” one lying between the Middle East and European countries. While there had been a move from the male breadwinner family to a model of universal breadwinner, ⁶⁵ there yet remained a “decorative approach to gender issues.” She concluded that the Turkish welfare regime is a synthesis of neoliberal and Islamic notions of social solidarity.

Karshenas and Moghadam ⁶⁶ and Karshenas et al. ⁶⁷ regarded Turkey as part of MENA, a highly gendered “authoritarian corporatist social welfare regime,” with female disadvantage continuing to be a hallmark of the MENA region well into the 1990s.

Ultimately, the regional approach is of limited value in classification, as there is little consensus on the nature of the “region.” As we have seen, writers have placed Turkey within a wide variety of regions—Arab, Middle Eastern, MENA, South Eastern Europe, Southern Europe, the extended family of Mediterranean nations, and hybrid Mediterranean/MENA—which undermines the possibility of reaching scholarly consensus.

Single-Country Case Studies

The final approach of the studies written in English examines Turkey as a single-country case study. The most significant tendency here is the classification of Turkey as part of the Southern European welfare regime cluster. Buğra and Keyder co-authored the most cited texts on Turkish social policy, in which they argued that Turkey is part of the Southern European family. In “New Poverty and the Changing Welfare Regime of Turkey,” ⁶⁸ they considered Turkey part of the Southern European welfare regime, pointing to a high level of self-employed and unpaid family workers as compared to the low formal employment; informal housing, patronage and clientelistic networks, and traditional and family-based support and solidarity mechanisms; a highly fragmented and hierarchical pension and health insurance system with extreme inequalities based on job status; and a limited role for social assistance for the poor. Buğra and Keyder ⁶⁹ and Yakut-Cakar ⁷⁰ pointed out how the welfare regime of Turkey resembled the corporatist social security systems of Southern Europe. Kılıç ⁷¹ also noted that Turkey’s welfare state policies could especially be likened to those of the so-called Southern European welfare regimes. ⁷² Şentürk ⁷³ argued that in Southern European welfare regimes, which includes Turkey, networks such as family, community, neighborhood, hometown, and political affiliation play a role in sustaining an individual’s life during unemployment, sickness, or old age, in addition to regular and formal state interventions and market relations.

Eder ⁷⁴ stated that Turkey’s welfare state has long been limited and inegalitarian. Strong family ties, coupled with indirect and informal channels of welfare, have compensated for the welfare vacuum. At first glance, Turkey’s welfare reform—which emerged from the 2000–2001 crisis—appeared a classic case of moving toward a minimalist, “neoliberal” welfare regime, but there was also evidence of the expansion of state power, with the cumulative effect of a mutated “institutional welfare-mix.” The characterization of the welfare regime change in Turkey as a convergence toward a neoliberal paradigm via the retreat of the state, or else as a completely divergent case of new welfare étatism too unique to be compared to other countries, could be highly problematic.

Morvaridi ⁷⁵ pointed to “neoliberal welfare governance,” where a new form of philanthropy emerged on the back of the neoliberal economic globalization that has encouraged the concentration of wealth in the hands of a few “super-rich” individuals. He also discussed faith-based philanthropy and Islamic philanthropic organizations, pointing to “a neoliberal model” of the Turkish state and philanthropy, where patronage and clientelism play an important role in welfare distribution.

Akan ⁷⁶ argued that the primary informal social protection mechanism in the Turkish system was the extended family. The existing variables are far from composing a regime of any kind, and therefore it seems more sensible to discuss a social policy structure that has become a “ragtag collection” as a result of the haphazard reform initiatives attempted in response to short-term budgetary and electoral concerns. Buğra and Candas ⁷⁷ regarded Turkey as an “eclectic social state” formation best described as a dual citizenship model with a Bismarckian formal social security system that also incorporated informality and clientelism.

Kılıç ⁷⁸ discussed the current social policy reform process in Turkey from a gender perspective. The changes undertaken within the scope of the reforms might signal a move from the ideal of the “male-breadwinner family” toward the model of the “universal breadwinner,” yet this rough shift from a “gender-differentiated” toward a “gender-neutral” citizenship did not seem promising for the achievement of equal citizenship for women. Elveren ⁷⁹ stated that the welfare regime in Turkey, as a developing country, could be referred to as an “indirect and minimalist welfare regime.” ⁸⁰

While overall, the Southern European welfare regime thus seems to be the modal conclusion, nevertheless different writers stress different aspects of Turkey’s welfare characteristics, and it remains difficult to discover areas of agreement and classification.

Studies written in Turkish

In contrast to the great diversity of methods and conclusions presented in studies written in the English language, the studies in Turkish lacked original empirical analysis, and they concluded that Turkey was part of the Southern regime.

Rakıcı and Yılmaz ⁸¹ argued that Turkey shared the following characteristics common to other members of the Southern regime: fragmented and hierarchical structure due to a large informal economy; premium-based healthcare insurance; social assistance and services decentralized, family-centred, and loosely connected with social insurance system. Taşçı ⁸² focuses on how the elderly were perceived by society at large. He stresses the polarizing and fragmented structure of the regime, which provides “hyper-protection” for formal sector employees an underprotection for others; the exposure of public welfare institutions to partisan and clientelist intervention; the lack of innovation in social assistance programs; and the centrality of religious institutions and the family in welfare provision. He pays particular attention to the tradition of sadaka (Islamic charity) in Turkey, which strengthens the Southern European typology of the country. As in other Southern European countries, welfare provision and the distribution of social risks are based upon the solidarity networks of family, with the state getting involved mainly in a non-systematic manner. Kalaycı ⁸³ states that the Turkish welfare system is characterized by the centrality of the family, a fragmented and polarized structure in welfare institutions, and a minimalist and neoliberal nature shown especially by an escalation in privatization. Kol ⁸⁴ focuses on healthcare, pointing out how healthcare care is based on premium payments, rather than on a universalistic principle. She also stresses fragmented and hierarchical social rights; the high degree of discretion, clientelism, and corruption in social policy programs; and the centrality of religion and family in welfare provision. The level of social assistance benefits is low and assistance is granted according to household poverty, thereby strengthening the centrality of family. Dericiler ⁸⁵ notes the parallels between Turkey and the Southern European welfare regime cluster: employment dominated by self-employment and petty bourgeoisie or unpaid family labor; wide informality and remittances; social rights unevenly distributed; and a limited state role in welfare provision. Family and religious institutions and local governments play a greater role than the central government in sustaining individuals and coping with risk situations. Toprak ⁸⁶ highlights the contrast between the inclusion and high degree of protection for formal state employees, with social assistance and social service provision for the rest of population, which is highly fragmented, family-centered, and disconnected from the social security system.

Table 2 Classification and Analysis of Studies on Turkish Welfare Regime (Turkish-Language Studies)

Discussion

Summing up the evidence of all the very diverse studies is problematic. It can be seen that the studies in Turkish all tend to speak with one voice: through discussion of previous work, all seven studies suggest that Turkey is part of the Southern Model. The English-language studies, on the other hand, arrive at a variety of conclusions. Of the 28 studies in English, the modal conclusion from ten studies is “Southern European.” In addition, there are some close variations on this, such as the “Latin Rim”, the “rudimentary assistance” regime (composed of Southern European nations), and the “extended family of Mediterranean welfare states.” However, while all the Turkish- and many of the English-language studies conclude that Turkey is part of the Southern Model, Esping-Andersen ⁸⁷ claimed that this was not a distinct grouping, but at most a sub-grouping of the corporatist welfare regime; Manow, ⁸⁸ meanwhile, regarded the continental-conservative regime as “ugly” because it is an “undecided regime” that is essentially a residual but diverse category. The next category is that of the “residual welfare state of the MENA region,” with three studies. A few studies point to various types of a hybrid or dualistic character. And finally, some of the cluster analyses suggest rather diverse geographical groupings, such as Denmark, Estonia, Norway, Poland, Mexico, and Turkey.

It is also not clear if we should equally weight studies via a simple form of “vote counting,” or whether some studies should be weighted more than others. For example, empirical studies could be weighted more than literature reviews or an “eyeballing” of data. Similarly, more stress might perhaps be placed on more highly cited studies, ⁸⁹ though it is clear that these studies are not highly cited due to how they classify the Turkish welfare regime. More weight could also be placed on Turkish-language studies, or English-language studies by Turkish scholars, since the authors are more familiar with the nation, but these tend to be less cited.

Some of these diverse conclusions can be explained by the different approaches adopted. Unsurprisingly, for example, regional approaches tend to find regional clusters (although different regions, such as Southern European or MENA, are put forward), while cluster analyses tend to find clusters that are not geographically contiguous, such as “successful informal security regimes.” However, the studies may also be associated with different criteria and variables. It is noticeable that few of the studies draw on the Esping-Andersen criteria of decommodification, stratification, welfare mix, or defamilialization. In some cases, studies justify their approach, often modifying Esping-Andersen, but in other cases variables (sometimes expenditure variables, which were rejected by Esping-Andersen) are produced with little justification. Some of the studies appear rather descriptive, and require better concepts and measures. For example, Eder ⁹⁰ focuses on the welfare mix, but with no direct reference to Esping-Andersen. ⁹¹ Some of the feminist studies do not discuss defamilialization. ⁹² Manow ⁹³ points to the neglected role of religion in Esping-Andersen’s three worlds. Similarly, it has been argued that there is a need to reconstitute future social policy in the region by taking better account of how religious values and ideals influence social welfare. ⁹⁴

Moreover, some of the earlier conclusions reached may no longer hold. First, Gough ⁹⁵ overturned his earlier view ⁹⁶ that there is a distinctive Southern European assistance regime, although the earlier conclusion is still cited by other writers. Second, a number of writers have argued that the Turkish welfare regime has undergone a major transformation since the mid-1990s. ⁹⁷ However, it is not certain exactly what this “regime” is, ⁹⁸ and while Özkan ⁹⁹ discusses welfare regime change, it is not entirely clear where the regime is moving from or to. Although no scholar classified Turkey as a “liberal welfare regime”, several of them noted neoliberal influences, which may suggest the possibility of “regime change” in the future. For example, in the 1980s, in parallel to the overall neoliberal paradigm in the world, the regime started to change toward a market-oriented structure. ¹⁰⁰ Akan ¹⁰¹ discusses the sociopolitical context of Turkish social policy in neoliberal times. Gencler and Colak ¹⁰² state that the social security system of Turkey has been restructured, particularly since the early 1990s, in line with the neoliberal paradigm. Morvaridi ¹⁰³ points to “neoliberal welfare governance.” A number of writers point to the neoliberal influence of external agencies such as the World Bank and IMF. ¹⁰⁴ However, Eder ¹⁰⁵ claims that the picture is more complex than neoliberalism, while Özbudun ¹⁰⁶ points to the growing trend towards conservatism, noting especially the AKP’s shift towards conservatism and authoritarianism.

Conclusion

In this article, we have conducted a meta-analysis of studies that classify Turkey within existing welfare regime families. We have brought together all the studies that have this objective in mind and drawn conclusions about the main trends in the literature based on the meta-analysis. It appears that studies classifying the Turkish welfare regime exhibit a significant degree of variation in terms of both their approaches and their conclusions, which results in little in the way of consensus and a notable degree of difficulty in differentiating between a large and confusing range of alternative terms. There are important differences between the views of different groups of scholars. Some of these differences may reflect familiarity with the country; that is, a deep knowledge of Turkey as compared to an observation made within statistical modeling. However, there are also different views on classification: while the majority view is that Turkey is part of the “Southern Model,” other accounts suggest that it is part of a wide variety of other “regions.” Whereas the studies in Turkish tend to agree that Turkey is part of the Southern Model, the English-language studies arrive at a variety of conclusions, though their overall modal conclusion is for the “Southern European” model. This, however, goes against the view of Esping-Andersen, ¹⁰⁷ who denies that such a model exists. In his view, this is not a distinct cluster but, at most, a sub-grouping of the corporatist welfare regime. Moreover, other “regional” studies conclude that Turkey is part of the MENA region, while many of the cluster analyses suggest a wide variety of clusters that are not geographically contiguous. In short, despite the extensive literature on the “worlds” of welfare capitalism, this study has shown that application to one country is problematic in that there is little consensus regarding approach, method, concepts, data, and conclusions. Finally, the studies reviewed are necessarily dated, having been written at a range of dates in the past. They all tend to focus on static classifications made at one particular point in time. However, as noted above, it can be argued that the Turkish welfare regime is rapidly changing. In particular, it may be transforming into part of a welfare regime family of emerging market economies, and it is unclear if the tools of the Esping-Andersen approach are suitable for a country such as Turkey. This may suggest a potentially fruitful research agenda for future scholars of the welfare regimes in Turkey.