1 Introduction

This edited volume seeks to explore new ways of explaining social policy by adopting a mechanism-based approach. The aim of the chapters has been to identify causal mechanisms that can explain the development of one or several social policy programmes. The chapters thus demonstrate how individual case studies can use the tools of a mechanism-based analysis. In this concluding chapter, we aim to show how causal mechanisms can also contribute to cross-case study research. Mechanism-based comparative social policy research aims to achieve an analytical level at which insights from single case studies and comparative analyses can mutually inform and complement each other. This can be achieved through modularisation, that is, by combining several mechanisms to explain a single policy process. Accordingly, individual mechanisms can serve as modules to explain social policy developments in a wide range of cases, but they cannot explain the entire process of introducing a new social policy programme or reforms of existing policies. Mechanisms are thus analytical elements that can explain a particular case when combined with each other. In fact, this is the standard approach in the preceding chapters. In most studies, a historical episode or period is explained through several interacting mechanisms. Some contributions, however, proceed differently. They focus on one or two mechanisms and clarify their specific role in a broader course of development, while not aiming to explain the whole process (see Chaps. 7, 9, and 11). What is more, some contributions also try to break down the complex causal mechanisms into elementary causal mechanisms as elaborated in the Introduction to this edited volume (see Chaps. 8 and 9). These case studies aim for an even more detailed analysis by shedding light on the reasons for individual actors’ actions.

Modularisation facilitates the detection of links between studies on quite different areas of social policy in different countries and at different times. For example, the alarmed middle classes mechanism—which is, like the other mechanisms mentioned here, explained in Sects. 2 and 3—can be found to be just as relevant for pension policy in South Korea (see Chap. 3) as for health policy in Bolivia (see Chap. 12). Likewise, the outcompeting mechanism was identified in the cases of South Korea (see Chap. 3) and Turkey (see Chap. 4), and the transnational cooperation mechanism can capture developments when it comes to unemployment insurance in Turkey (see Chap. 4) and health policy in Albania (see Chap. 9). Moreover, there are proximities or similarities between mechanisms, such as the double benefit mechanism for the cases of South Korea, Vietnam, Sri Lanka, and Malaysia (see Chap. 3) and the mechanism of public resource accumulation for the case of Uganda (see Chap. 6).

Notably, across all contributions, there are some mechanisms that focus on the same aspect when explaining a particular development in social policy, such as transnational influences on national social policies. However, the individual studies highlight how diverse the interaction between national and transnational actors can be. The analyses throughout this volume therefore show that it is not enough to argue that single factors like transnational influences can, to a certain extent, explain social policy developments. Rather, it is important to understand how this factor actually plays out and, accordingly, to explain it through different causal mechanisms.

Finally, it also turns out that the number of mechanisms identified throughout this volume is not so limited that we could derive a simple and general theory of social policy. Rather, we find that the diversity of social policy developments is reflected in a plurality of mechanisms that have been identified as relevant. Therefore, as a next analytical step, it is necessary to categorise the mechanisms and to bring them into a systematic order. This concluding chapter will therefore present a structured compilation of the complex causal mechanisms that were identified by the authors in this edited volume and indicate perspectives for further mechanism-based social policy research.

The complex causal mechanisms compiled in this volume can be distinguished, first, according to whether or not they are mechanisms that are closely linked to a specific social policy sector or to a particular social policy institution, for example social insurance (see also Chap. 3). This distinction into special and more general mechanisms aims at stressing differences in the scope of a mechanism. Can a mechanism occur in all fields of social policy (or even in fields that are not related to social policy), or is it tied to a specific institutional arrangement which, if absent, also indicates that the mechanism cannot be present? This type of clustering is based on how the mechanisms have been defined in the individual chapters of this volume. If they are formulated in such a way that they are tied to the preconditions of a specific social policy field or to a specific type of social policy institution, we classify them as policy-specific mechanisms; if not, we classify them as general mechanisms. Of course, this classification remains to a certain extent provisional, as it is only based on the analyses that were conducted throughout this volume. Follow-up studies can put these particular mechanisms to the test to analyse whether they have a broader scope.

What is even more interesting is how a systematic clustering of the mechanisms mentioned in the individual chapters could be achieved. We distinguish two options. First, the mechanisms can be classified depending on which actors are the reference point of the respective mechanism. This is in line with the actor-centred approach pursued throughout this volume. Second, we can focus on the thematic proximity between the mechanisms and on well-known theoretical approaches within the field of comparative social policy as a way of classifying the mechanisms.

Following the first option of clustering via actors, we can structure the general mechanisms according to different groups of actors that have been found to be decisive for social policy development, which are (a) transnational actors, (b) policymakers and administrators, (c) political parties, (d) voters, and (e) interest groups and social movements. However, some chapters also identify mechanisms that we cannot directly assign to a single actor or type of actor, such as the mechanism of anti-communist backlash (see Chap. 8). Therefore, an actor-centred classification of mechanisms does not cover the full set of mechanisms that were identified throughout this volume. Moreover, a classification that is only actor-centred is not linked to the established approaches in social policy research and therefore not able to show how the identified mechanisms relate to existing theoretical approaches.

Given this background, as a second criterion of clustering, different mechanisms can be linked to established approaches in comparative social policy research. Here, a distinction between (a) the socio-economic theory, (b) the power resource approach, (c) the parties matter approach, approaches focusing on (d) political institutions, (e) on globalisation and policy diffusion, and (f) on policy heritage and path dependency as core elements of explanation can serve as inspiration (e.g. Obinger and Schmidt 2019). Such a clustering has two analytical advantages: First, while applications of these theories often tend to leave processes un-investigated, our mechanisms help to illuminate the black box of what is happening between x and y. For example, instead of arguing that parties matter for expanding social benefits, causal mechanisms are able to explain what the causally relevant steps in such a process are, that is, precisely how parties and their interaction with other relevant actors lead to the expansion of social benefits. Second, it also becomes clear from such a clustering that the mechanisms in this volume are by no means meant to replace existing theories and approaches in comparative social policy. Rather, they offer building blocks for explaining social policies with clear links to existing approaches. Importantly, sometimes these mechanisms also allude to causal relationships between the elements in a theoretical framework that tend to be overlooked in empirical applications, but that clearly belong to the core of the respective theoretical approaches. The risk of clustering solely on the basis of established theoretical approaches is to reproduce these approaches in the language of mechanisms. Yet the modularised approach offers many opportunities to identify causal relationships that also indicate gaps in or between these established approaches.

In the following, we therefore work tentatively with a combination of both forms of clustering, that is, actor-centred and theory-centred clustering. Our approach is facilitated by the fact that many of the established approaches also place specific actors at the centre of their explanatory approach. Still, some caveats to our combined actor-centred and theory-based clustering should be mentioned: First, not all theoretical approaches outlined above can serve as a reference point for our classification of mechanisms. We exclude the socio-economic theory (e.g. Obinger 2019) due to its different methodological background: It focuses on macro-level causal effects and not on the causal links between independent and dependent variables, which is at the core of mechanism-based analysis. Moreover, the socio-economic theory does not theorise on actors and is therefore not compatible with our actor-centred approach. Second, with regard to the actor-centred clustering, it is well known that social policy developments can rarely be explained by the actions of a single actor (Heclo 2010). Although the clustering via actors thus highlights single actors, most mechanisms focus on the interaction of different actors—for example, transnational actor mechanisms trace the interaction between transnational actors and national actors. Moreover, our modular approach can capture how different actors can be involved in a particular process, as studies can combine, for example, transnational actor mechanisms and interest group mechanisms. Third, some chapters in this volume also refer to mechanisms that have been discussed in the social policy literature, but without the emergence of a separate theoretical approach. In these cases, the criterion of actor-centred clustering can function as a way of systematically integrating such mechanisms as well. Taken together, this type of clustering thus serves the main purpose of this book, namely to complement, expand, deepen, and possibly also correct existing approaches to social policy (see Chap. 1).

In the following, we collate the mechanisms that were identified throughout this volume and complement them with additional mechanisms that can be theorised from the literature, to arrive at a comprehensive list of causal mechanisms that can contribute to explanations of social policy developments, while not claiming that this list is exhaustive. We begin by presenting the general mechanisms and focus afterwards on the policy-specific mechanisms.

2 General Mechanisms

2.1 Transnational Actor Mechanisms

A vast body of literature acknowledges the important role that developments beyond the national level play for social policy-making. Scholars focusing on globalisation have studied how this very process shapes the level of national social benefits (Starke and Tosun 2019). Scholars focusing on policy diffusion have studied to what extent social policies that are adopted in one country have been influenced by social policies from other countries or International Organisations (IOs), distinguishing coercion, competition, emulation, and learning as types of diffusion (Maggetti and Gilardi 2016; Obinger et al. 2013). However, this literature has often been criticised for not acknowledging processes and the active role that different actors play within them (Kuhlmann et al. 2020). A more systematic search for causal mechanisms can contribute to a refinement and modification of the causal relations that shape diffusion processes.

In his chapter on the adoption of social protection policies in Africa, Devereux (Chap. 7) introduces the mechanism of policy pollination and traces the process in which single actors from international development agencies travel to different countries to promote their preferred policies, thereby applying different strategies. Focusing more on the social construction of social policies as they have, for example, also been discussed in the literature on policy emulation, Sirén in his chapter on healthcare reform processes in Bolivia (Chap. 12) traces how transnational epistemic communities shaped an understanding of health as a citizenship right through a mechanism of expert theorisation.

However, it would be misleading to neglect the role that national actors play in such processes. IOs make use of different strategies to convince national actors of their preferred policies, ranging from loans or technical assistance to expert exchange through workshops or publications (Orenstein 2008). National actors might then acknowledge the promoted policy as a suitable option. Yet, they are not passive actors that are just doing what IOs ask them to. Rather, they have considerable leeway in shaping and implementing IO suggestions (Kuhlmann and ten Brink 2021; Leisering et al. 2017). The mechanism of transnational cooperation captures this interaction, focusing on how IOs and national actors build alliances to push through a social policy reform. For example, Druga (Chap. 9) traces this mechanism in the case of social policy-making in post-communist Albania, showing how the World Bank and the Albanian government worked together on the introduction of social health insurance. What is interesting here is that the World Bank, which did not favour the Albanian decision for an insurance scheme, nevertheless continued to cooperate with the national government, hoping to be able to influence the final law. Druga’s study is therefore a remarkable example of how national actors play a crucial role in transnational cooperation, keeping a clear eye on national considerations. Öktem (Chap. 4) also refers to the mechanism of transnational cooperation with regard to the role of the European Union in shaping the development of unemployment insurance in Turkey.

Another way in which the interaction between transnational and national actors can also clearly follow the ideas and preferences of national actors is stressed in the evasion mechanism, which captures how national actors highlight the compatibility of IO suggestions and national policy proposals, while actually pursuing their own agendas. This has, for example, been observed with regard to pension policy in some Eastern, South-Eastern, and Southern Asian countries (Chap. 3). Finally, the seeking solutions abroad mechanism (Chap. 8) captures how national policymakers actively look for policies developed in other countries or possibly also IOs, which again stresses the proactive role that the national level can play in transnational interactions.

2.2 National Policymaker and Administrator Mechanisms

This group of mechanisms focuses on actors within the national political arena. We conceptualise policymakers as actors within the political system who have authority to make policy decisions. Alongside mechanisms focusing on policymakers, we consider mechanisms that highlight the role of actors within the administration, who have often been found to play a crucial role not only when it comes to policy formulation, but also in the agenda-setting phase (Klenk 2019). To begin with, existing analyses on the role of administrators might be translated into a bureaucratic inspiration mechanism: Decision-making processes are strongly driven by individual actors within the administration. Individual officials often have decades of experience in a policy field. Because of their central positions in ministries, they can pursue long-term goals in social policy and enjoy professional recognition within their organisation and among the political elites (Hassenteufel et al. 2010). Thus, they might push for introducing or expanding a particular social policy inside their organisation.

Ten Brink, Müller, and Liu (Chap. 2) provide an example of how mechanisms that stress the role of policymakers and administrators interact and can, when combined, explain the introduction of social insurance schemes for urban areas in China. The policy experimentation mechanism explains how administrative units within the state or semi-state administrations set out to find new solutions to perceived social and economic developments, thereby making use of the internal differentiation of the state apparatus and its federal multi-level structure. In the case of China, instead of implementing one policy solution, policymakers tested different variants of social insurance schemes at the regional level, which were inspired by domestic and international experiences. More specifically, ten Brink, Müller, and Liu distinguish a strategic policy experimentation mechanism and a more neutral policy experimentation mechanism. While in the former policy experimentation is a means to push particular policies in a context of political conflict, in the latter policy experimentation serves to provide information on how different social policies work. Moreover, ten Brink, Müller, and Liu identify an elite cooperation mechanism, which explains how decisions within the governing party-state elite were made, either building on a broad consensus between the involved actors or reflecting a compromise. Finally, the top-leader intervention mechanism explains how hierarchical decisions are made by single powerful actors within the administration.

Thyen and Schlichte (Chap. 6) explore the role of policymakers and administrators in a colonial context, focusing on Tunisia and Uganda. The imperial staffing mechanism explains how colonial governments create pensions for specific staff such as civil servants and soldiers to ensure the loyalty of groups that they perceive as important. Moreover, the authors identify an appropriation mechanism in the process of decolonisation, referring to existing colonial structures in the field of social policies that governments now relied on for their own aims. To maintain or increase legitimacy, governments expanded colonial pension schemes for public officials to greater and greater parts of the population and introduced social insurance schemes for additional groups.

Political and economic expectations within the broader population are also highly relevant for some mechanisms. A mechanism that we can theorise but that was not identified in our chapters is the anticipatory reforming mechanism, which centres on how governments anticipate possible future problems. For example, if a socio-economic situation is experienced as upheaval or if fundamental changes are taking place and these could also shape political conflicts in the future, parties can advocate for or push the introduction or expansion of social policies, even if there is no pressure from other political actors and especially other parties (Rimlinger 1971). Heinrich, Isabekova, and Pleines (Chap. 5) explain the introduction of mandatory health insurance schemes in the post-Soviet region by what they define as a resistance avoidance mechanism. The mechanism explains how policymakers introduced social insurance although it was not their preferred policy solution. Yet, the adoption can be traced back to social policy expectations within the population and no clear opposition to the reform. In their chapter on the migrantisation of long-term care in Germany, Safuta, Noack, Gottschall, and Rothgang (Chap. 10) show how policymakers within the political arena promote the recruitment of care workers from other countries to prevent a lack of care workers, which they describe as state-supported migrantisation. Finally, in his chapter on healthcare reform in Bolivia (Chap. 12), Sirén identifies what he names a social movement–state interaction mechanism, stressing that actors within the administration can also be comprised of people with a social movement background: Activists from social movements conquer positions in the state bureaucracy and continue to cooperate—now as public officials—with these movements, thus taking up certain demands.

2.3 Political Party Mechanisms

Most scholars in comparative social policy would probably agree that political parties play an important role for the development of social policy, not least given that national governments—who formulate and implement social policies—are predominantly constituted through political parties (Zohlnhöfer 2019). Classical studies focused, for example, on how left- and right-wing parties differ in their social policies, with left-wing parties being more in favour of redistribution than right-wing parties, which could then be ascribed to the different constituencies of both parties (Häusermann et al. 2013). The question if parties matter for social policy development has often been studied in quantitative analyses. In these studies, the policy process is often highly simplified. Newer research on partisan politics takes a closer look at the constituencies of different parties, acknowledging that links between voters and parties have changed (Häusermann et al. 2013; Gingrich and Häusermann 2015). Yet also this line of research remains very much focused on voters and their preferences and therefore to a certain extent remains centred on partisan politics as transmission belts. In contrast to these accounts, some authors have stressed that the agency of political parties—expressed through both party members and party elites—should not be neglected (Wenzelburger and Zohlnhöfer 2021).

Given this general background, two chapters in this volume identify an outcompeting mechanism, in which the development of social policy, most notably its introduction and expansion, is explained as a result of party competition. Political parties try to win elections or compete with one another by making promises on social policies, which could be traced for the case of pension policy in South Korea in the countries’ democratisation process (Chap. 3) and for the case of unemployment insurance in Turkey (Chap. 4). While the outcompeting mechanism is focused on democratic settings, we can add the gaining acceptance spiral mechanism as another political party mechanism. Although the mechanism focuses on autocratic regimes, these can also be constituted by political parties who introduce or expand social policies to increase their legitimacy within the population. In this book, examples of such processes could be identified for the introduction or expansion of pension policies in South Korea before democratisation, in Vietnam during the economic liberalisation process, and plausibly also for Malaysia in the late colonial period (Chap. 3).

Importantly, in this understanding political parties are not conceptualised as actors that aggregate voter preferences, although the mechanisms that are highlighted here focus on how political parties respond to such actual or perceived voter preferences. Rather, in our mechanism-based approach we see political parties as corporate actors who make decisions. Still, we can identify voter-oriented mechanisms as a subgroup of our political party mechanisms. In his chapter on political responses of conditional income transfer recipients, for example, Barrientos (Chap. 13) identifies a support for redistribution mechanism, capturing that people who receive conditional income transfers vote for the incumbent who is perceived as a politician committed to values like poverty reduction and social justice.

2.4 Interest Group Mechanisms

Interest groups are another key actor in the comparative social policy literature. Since the conflict between labour and capital has been found to be crucial for explaining the introduction and expansion of social policies, trade unions (Korpi 1983) and employers’ associations (Hall and Soskice 2001) are particularly important. Similar to other types of mechanisms introduced in this book, the mechanisms here are also based on the interaction between two groups of actors. On the one hand, these are interest groups or social movements, NGOs, or other civil society actors. On the other hand, these are political elites, governing parties, or other actors within the executive. For example, living and working conditions might be experienced as problematic, or demands for more social rights might arise due to fairness considerations. Consequently, demands for introducing or reforming social security systems increase in some parts of society. Interest groups, social movements, NGOs, or other associations take up and strengthen these demands (Mesa-Lago 1978; Korpi 1983). A mobilisation involving many parts of society emerges, which is expressed in demonstrations, strikes, and other forms of protest.

A key assumption of Korpi’s power resource theory (1983) is that labour unions, who represent the interests of the working class, mobilise their constituencies to push their demands on social policies. Sirén (Chap. 12), in his study on healthcare reform in Bolivia, relies on this literature to identify a mechanism of class-based mobilisation. While the traditional power resource literature focuses especially on trade unions in the industrial sector, for the Bolivian case Sirén argues that the mobilisation of the indigenous majority through peasants’ unions and neighbourhood organisations was the driver of social policy reforms. Thyen and Schlichte (Chap. 6) illuminate the role of interest groups from a different angle. With regard to social insurance in the case of Tunisia, they identify a mechanism of labour incorporation. It focuses on trade unions who had supported their country’s aspirations for independence. Shortly after national independence, the working-class groups mobilised by these unions were then rewarded with a comprehensive social insurance programme. While the previous mechanisms rather focus on the role of labour, the business-led reform mechanism that Öktem identifies in his analysis of unemployment insurance in Turkey stresses how business actors can also convince policymakers to adopt their policy proposals (Chap. 4). In this particular case, pressure from business actors contributed to a more prominent role for active labour market policies within the system.

Finally, a mechanism that we can theorise but that was not identified in our case studies is the mechanism of cross-class mobilisation. In market economies, the world of labour is predominantly or at least partly determined by wage-dependent formal employment. This usually results in the formation of employer and employee organisations. If more stable, informal, or institutionalised forms of cooperation between employers and the working class emerge due to a certain rational compatibility of interests, this can also lead to a welfare state coalition that jointly supports the introduction or the expansion of social policies (Hall and Soskice 2001). Employee organisations advocate for social security for their core clientele. Employer organisations might favour such policies because they protect workers, for example in cases of sickness, and therefore in the long run contribute to productivity, or because of normative convictions that value “decent” working conditions. In this coalition mechanism, however, the course of social policy expansion will be limited as the interests of the organised core groups (large companies, sectors with a high degree of unionisation) will be in the foreground (Palier and Thelen 2010; Yang 2017).

2.5 Political Heritage Mechanisms

As already previously highlighted, there are several mechanisms that do not focus on a specific actor. Some of them can be classified as mechanisms that are in line with what is often called the political legacy approach. Especially in historical institutionalism, which pays special attention to the temporal dimensions of politics (Thelen and Mahoney 2015), the concept of path dependence plays a prominent role. In a broad understanding, it simply implies that within a temporal sequence, previous sequences are relevant (Pierson 2000). Many studies take the notion of path dependence as their theoretical point of departure, stressing that social policies are shaped by previous developments. However, it is especially in the historical institutionalist tradition that the role of actors and their interactions within historical processes is detailed (e.g. Streeck and Thelen 2005). Against this background, historical institutionalist approaches are highly compatible with actor-centred mechanism-based approaches (see also Chap. 1).

In our volume, the old system departure mechanism as identified by Malinar (Chap. 8) can be named as an example of a mechanism that is focused on the political heritage. Policy actors perceive the old, long-standing system as so bad, delegitimised, and ineffective that they will find any political solution better than continuing with the present system. Statements such as “Things cannot go on like this” serve as evidence for this mechanism, which demands a turning away from the past, but does not provide a new policy solution. As Malinar shows for the case of the Croatian healthcare system, the combination of this mechanism with two other mechanisms, the actor-oriented doctors enter politics mechanism, and another transnational mechanism, the seeking solutions abroad mechanism, can be combined into the mechanism of anti-communist backlash, which now contains an idea about the policy solution to be adopted. This solution is no longer just a matter of creating something new. It must not resemble the old system—in this case the communist system—, quite on the contrary, it should be as far away from the previous system as possible. Political elites will therefore avoid any similarity between new policies and the old system because there is too much resistance to the old system, mostly for emotional reasons. Rejecting old policies enables all political actors to profile a solution that has to meet one key requirement, namely, not to have any similarity with the old regime. This opens considerable scope for action and creates room for fundamental institutional ruptures and the overcoming of path dependencies.

While not represented in this volume with a case study, a positive reference to historical heritage is also possible. A situation of regime or system change can also lead to the rejection of certain new ideas that will destroy what was worth preserving in the previous regime. Thus, despite the dictatorial and authoritarian constitution of a regime, in the society there can be a prevailing perception that certain areas or outcomes, such as pensions, were good and should be maintained because they provided social security. Consequently, national actors might oppose privatisation programmes in the wake of the system transformation from socialism to democratic market economies. The new political elites are not able to implement certain radical alternatives but must follow previous paths despite the fundamental transformation of state and society. The honouring one’s legacy mechanism explains how an old institutional system is protected and defended against political attempts to break with this tradition.

3 Policy-Specific Mechanisms

The mechanisms listed in the preceding section can be relevant in all fields of social policy and for different institutional arrangements of the welfare state; in fact, some might even be traced in cases that are not connected to social policy at all. In addition, the following mechanisms are linked to specific institutional arrangements of the welfare state or to specific policy sectors within the broad spectrum of social policy. One approach to clustering this set of mechanisms is to proceed according to these institutional characteristics and to distinguish, for example, mechanisms that can only become relevant either in Bismarckian or Beveridge systems, or that can only occur in health policy, but not in labour market and old-age security policy. However, it is also possible to apply the type of clustering that we used for the general political mechanisms, that is, a combined actor-centred and theory-based clustering.

3.1 Medical Profession Mechanisms

Social policy does not only depend on actors who advocate for or introduce policies. It also depends on occupational groups who deliver social policies. The expansion of social services leads to the growth and increased importance of specific professional groups in the respective policy field (Pierson 1994). Notably, throughout our edited volume, the medical profession turned out to be a particularly important group, that has been found to often have a special say when it comes to health decisions (Hassenteufel and Genieys 2021). However, precisely how the medical profession influences social policy decisions can be very different.

In situations of regime or system change, the upheaval in the political elite can be so fundamental that new actors enter the political stage. The doctors enter politics mechanism captures how the medical profession becomes a key political actor and can thus influence the health system in line with its interests (see Chap. 8). If, on the other hand, the political elites and the government are dominated by other political forces, the physicians’ position can also be captured in what Sirén terms the professional autonomy mechanism in the case of Bolivian healthcare reform (Chap. 12). Here, too, the medical profession is an important actor, but rather when it comes to defending the medical profession’s privileges.

Focusing on Chile and Uruguay in the early twentieth century, González de Reufels and Huhle (Chap. 11) show how governments and the medical profession worked together in a kind of modernisation alliance to legitimise each other. By supporting the medical profession, the state becomes more modern in terms of public healthcare institutions and well-equipped hospitals, while the medical profession receives political support for the further expansion of the healthcare system. The legitimisation mechanism enables the further professionalisation of the medical profession and the acceptance of physicians as political actors. This legitimisation is supported when the medical profession organises on a transnational level and can thus also expand its legitimacy from one country to the entire region. A second mechanism that González de Reufels and Huhle focus on in this regard is the competitive cooperation mechanism. The transregional cooperation of the medical profession also drives competition, as nations compare themselves with each other and compete for the position of regional leader. At the same time, doctors and states in the region perceive a strong competition with other regions—in their case, with Europe. Taken together, this supports a form of competition that does not amount to a zero-sum game, but to an overall higher level of modernity in health policy.

3.2 Social Service Provider Mechanisms

The literature on comparative social policy has also stressed the important role that social service providers can play for social policy (e.g. Noordegraaf and Steijn 2013). This becomes also relevant in some chapters of this book. Sirén in his chapter (Chap. 12) traces a provider resistance mechanism, focusing on the interests of the people who are employed in the Bolivian health insurance funds. On a more general level, providers are public, private for-profit, or private non-profit institutions, including not only insurance funds but also hospitals or pharmaceutical manufacturers. They can also perceive themselves as key actors in the health system and try to resist certain reforms that they consider unfavourable. The stakeholder pressure mechanism that Safuta, Noack, Gottschall, and Rothgang identify in their chapter (Chap. 10) points in a similar direction. Here, it is private care providers who advocate a certain approach to care and migration policy to be able to maintain their business model. What is characteristic for both cases is a rational pursuit of interests, but the actor constellations are different in the health system than in long-term care, which can result in different strategies and political alliances.

3.3 Status Group/Social Insurance Mechanisms

Another group of mechanisms that we can identify refers to the characteristics of social insurance, including its focus on employees instead of the entire population, on financing through social contributions instead of taxes, on relative autonomy from the state instead of administrative subordination, and on a performance-based understanding of social justice (e.g. Klenk et al. 2012). One mechanism that belongs to this group is the alarmed middle classes mechanism, which captures how fears among groups already incorporated into the social insurance system about either needing to pay more or receiving fewer benefits leads to an only modest expansion of social policies to new groups. Kuhlmann and Nullmeier (Chap. 3) identify this mechanism in the pension reform process in South Korea. Sirén (Chap. 12) identifies the same mechanism when it comes to formal sector workers who are integrated into the health insurance funds in Bolivia.

Another mechanism, which was not identified in a case study, can be found among those groups that have not yet been included in a social security system. If a social insurance scheme was initially introduced for only a limited group of employees and if this social security programme is successful, whether this group is privileged will soon become a contested issue. Demands arise that other groups, namely those with higher and lower incomes than the group that is already included in the social insurance system, should also benefit from these services or even receive a better form of social security. Other groups (e.g. white-collar workers, farmers, self-employed, fishermen) develop comparative preferences, which can be explained by the status group competition mechanism. These groups also want to be included in the social insurance scheme. This competition between socio-economic status groups can lead to an extension of social insurance coverage if there are no massive disadvantages (e.g. financial burdens) for the groups that are already included.

3.4 Public Budget/Public Actor Mechanisms

We can also identify a group of mechanisms that focuses on the financing of social security systems and the management of budgets in the various and often highly fragmented welfare state institutions. A striking example here is the double benefit mechanism, as Kuhlmann and Nullmeier (Chap. 3) show for the case of pension policy in South Korea, Vietnam, Sri Lanka, and Malaysia, in which governments make sure that there is always enough money in the pension fund because this money is used for other purposes, most importantly to finance the public debt or to boost the country’s economic development (see also Holliday 2000; Koreh 2017). By identifying a mechanism of public resource accumulation, Thyen and Schlichte (Chap. 6) refer to a very similar process in Uganda. In his chapter on Turkey, Öktem shows how the double benefit mechanism can also explain processes in the field of labour market policy, as money from unemployment insurance was also used by national governments to finance budget deficits (Chap. 4). As the case of Vietnam shows, policymakers also use money from the pension system to fulfil other social policy-related tasks, such as facilitating early retirement to ease a tight job market (Chap. 3), which can be explained by a variant of the double benefit mechanism. Moreover, the Vietnamese case also serves as an example of the crisis management by going further mechanism, in which policymakers expand a social insurance system to new groups so that the system remains financially sustainable in times of economic upheaval.

The fight for state funding mechanism can be considered as a counterpart to this mechanism. It refers to a conflict about the financing of social security systems within the public sector, as the chapter by Heinrich, Isabekova, and Pleines shows (Chap. 5). Decentralisation, fragmentation, or multi-level interdependence are the initial conditions for such financial conflicts between different public actors. If a social security system is significantly underfunded, for example due to fixed social contribution rates at the federal level, substantial financial support is required, which can only be obtained directly from public budgets at the state level, regional level, or municipal level for the decentralised institutions. It then depends entirely on these authorities whether state subsidies are provided or not.

Heinrich, Isabekova, and Pleines (Chap. 5) also identify a mechanism in which an institution is established in the social security system that acts as a driving force for reforms and financial stabilisation. They describe this as the reform supporter mechanism. It is precisely when social groups, parties, and associations have no interest in shaping the extremely complex social policy (in their case: health policy) that the creation of a separate central public institution can appear as a decisive factor in an otherwise shunned and uncontested field. The public institution then becomes the motor for reform and a self-sustaining state-based reform dynamic emerges.

3.5 User Mechanisms

Finally, we focus on the clients, users, or beneficiaries in social security systems who are affected by social policy at the implementation level. First of all, the role of informal practices comes to the fore here, which has been described as an important characteristic in social policy-making (see also Niedzwiecki 2018). In their analysis of mandatory health insurance in the post-Soviet region, Heinrich, Isabekova, and Pleines identify an informalisation mechanism (Chap. 5), which captures a special form of user behaviour that occurs in combination with doctors’ economic interests when the health system is underfunded. As regular funding is insufficient, medical personnel and patients agree on informal payments, leading to unregistered service provision. As these informal payments provide patients with additional services and medical personnel with additional income, a stable alliance emerges that unfolds an intermediate form of public and private service provision, since services are mostly provided within the framework of the public institutions. Another mechanism that can be classified as a user mechanism is identified by Safuta, Noack, Gottschall, and Rothgang in their analysis of long-term care in Germany (Chap. 10). They trace a turn to the market mechanism, which explains how existing structures within the policy sector, such as unregulated cash benefits to beneficiaries that were established with the introduction of long-term care insurance, encourage relatives of elderly people who are in need of care to consider market solutions. In both mechanisms, a calculatory orientation is very pronounced.

4 Future Perspectives for a Mechanism-Based Approach in Social Policy Research

The clustering of mechanisms presented in this concluding chapter is an initial attempt to compile a list of complex causal mechanisms that can further stimulate research on social policy developments, especially in a comparative perspective. Yet, this can only be the first step of a more encompassing research agenda on mechanism-based analysis in social policy. In further refining this approach to social policy, we suggest three future directions, which can mutually reinforce each other.

First, by clustering the mechanisms identified in this volume, we have emphasised existing links to established theories of comparative social policy research. However, mechanism-based analysis should engage more profoundly with these approaches. Until now, two debates have been at the fore: Either established theories are understood as holistic approaches, and the design and orientation of the case studies that apply these theories is also entirely determined by this approach. Or these theories are interpreted as implying relevant explanatory factors that are then built into statistical models, for example in macro-quantitative research, to be able to assess their explanatory power. A mechanism-based approach as introduced in this volume now offers a different option, as case studies that were based on single theoretical approaches can be subjected to a re-analysis. This means that existing studies can be decomposed to find out which causal mechanisms are included in these studies, although a mechanism-based approach was not explicitly pronounced in these studies (see also Starke 2021). Instead of assuming one unified theory, such an approach presupposes that several mechanisms are referred to within one theoretical approach. Established approaches in social policy research can thus serve as the reference point for this first direction of future research on causal mechanisms in comparative social policy.

Second, instead of analysing political developments over a longer period of time, research can focus on individual mechanisms. Already in the studies contained in this volume, authors were able to identify the same mechanisms in very different countries and policy fields, as well as at very different times. Such cases can thus provide an opportunity for a more detailed analysis of a particular mechanism: Is it really the same mechanism that is at play in these cases? Or do we see variations that rather lead to identifying two or more different mechanisms? These questions should be answered in follow-up analyses of already existing case studies or by bringing in new case studies based on the expectation that a particular mechanism might be relevant there. Additionally, we encourage more in-depth analysis of single complex causal mechanisms by identifying the elementary mechanisms that underlie them (see also Chaps. 8 and 9). Such an approach can rely on both theoretical work and empirical analysis and decompose the identified complex causal mechanism step by step. The reference point for such an approach is the elementary causal mechanisms that have been identified in other research traditions (see Chap. 1).

Third, this technique of decomposition can also be applied to the analysis of individual decision-making processes in social policy. This presupposes that a study focuses on a limited period of time and can analyse the processes that are taking place in great detail. Such studies require a systematic application of process tracing and the methodology of qualitative social research to identify causal pathways and mechanisms (see, e.g. Beach and Pedersen 2019; Nullmeier 2021). However, not all guidelines for process tracing use the actor-centred and modular conception of mechanisms as suggested throughout this volume. Along with the precise exploration of individual processes, the tools and concepts for such micro-analyses still need to be refined.

Taken together, these three future directions of mechanism-based research can pave the way for case study-centred social policy research to develop a set of instruments that also open up broader comparative perspectives. This can also prevent comparative social policy research from producing findings that barely relate to each other or using case studies to confirm or reject entire theoretical traditions. Modularity at the level of complex causal mechanisms enables case studies on different countries, fields of social policy, and historical periods to speak more profoundly to each other.