The Political Economy of Crisis Recovery

The aftermath of the global financial crisis marked another stress test for welfare states and varieties of capitalism. More than ever before, governments were forced to consider substantial reforms to welfare provision and enact flexibility-enhancing measures in order to improve financial solvency and economic performance. The crash, however, was not only a regionally uneven process in its origins but also led to makeshift or uneven policy responses. As a result, the socio-economic effects of the downturn and political reactions to it varied considerably among countries. Nevertheless, there have been some common trends in outcome measures. These have served to blur the dividing lines between different welfare states and production systems, so vividly captured in the mainstream political economy literature.


Introduction
One of the most enduring areas of research in political economy has revolved around the clustering of developed countries into distinct political and economic systems. Commonly referred to as the 'worlds of welfare' and 'varieties of capitalism' approaches, these categorizations are based on the assumption that countries can be defined by the types and combinations of policies, institutions and ideologies they employ. The worlds of welfare approach focuses mainly on differences in the structures of welfare states by examining their extent of decommodification, social stratification and the roles of the state, market and family in defining and responding to social needs. The varieties of capitalism approach emphasizes how diverse systems of production offer different types of comparative advantages, which help sustain divergent models of capitalism. Both typologies, therefore, not only assume that developed countries can be categorized into different types of welfare regimes and production systems, but that each welfare or production model follows a qualitatively different development trajectory, producing different types and degrees of outcomes as compared to other models, even in times of economic crisis.
This chapter seeks to test divergent theories of welfare development and varieties of capitalism by examining the extent to which different types of welfare states and production regimes exhibit markedly different socio-economic outcomes in the face of external pressures. We seek to raise a series of questions about divergent theories of national development within the context of crisis management and recovery. According to the literature, distinct welfare arrangements and production systems should produce qualitatively different outcomes, even in times of crisis, but is there enough empirical evidence to support such claims? It is also postulated that countries categorized under the same type of welfare and production model should display similar outcomes, but can we expect to find any national outliers within any of the distinct regime-clusters? Moreover, were some welfare states and market economies better able to return to their pre-crisis levels more than others? Or did all countries move in a broadly similar direction and experience a general worsening of outcomes in the post-crisis period?
To answer these questions, we begin the following section with a discussion of path dependent development and continuity in national distinctiveness. These concepts correspond with two main arguments found in the literature about divergence in national models of welfare and capitalism. Next, the two path-dependent arguments that distinguish between different welfare production systems are described in some detail. This is a followed by a brief discussion of the theory of convergence, which stands in contrast to theories emphasizing national diversity. We then discuss the sample of countries, socio-economic indicators and time period included in our datasets. In what follows, we summarize and analyze cross-national data on nine indicators. Finally, we end with some concluding remarks about cross-national divergence and convergence in socio-economic outcomes and whether the type of welfare production regime various countries had made much of a difference in the way their economies performed during the recent crisis.

Path dependence
The literature on path dependency emphasizes multiple pathways of national development, the persistence of institutional arrangements and continuity. In simple terms, path dependence means as time progresses meaningful change becomes more difficult to implement. In the political science literature, path dependence is used to refer to political trajectories that are resistant to change due to the self-enforcing nature of institutions [1]. In comparative political economy, path dependence is usually a reference to the persistence (inertia) of historically entrenched institutions, or the idea that variations in the size, structure and outcomes of welfare regimes are a consequence of the individual paths each one follows. In such a context, welfare states are viewed as diverse and resistant to change. Their development trajectories have become 'locked-in' over time making shifts away from them difficult, even in the face of common external pressures, which others view as driving all countries toward the same practices and outcomes. This has led to a debate over whether welfare and production systems are converging toward a similar model due to common pressures, or whether the path-dependent nature of institutional development continues to exist and reinforce the diversity of national trajectories. In general, support for the latter argument can be found in comparative studies based on typologies of the welfare state and market economy, which are reviewed below.

Worlds of welfare
Esping-Andersen's three worlds of welfare capitalism model is perhaps the most famous typology of welfare states [2]. For many years, comparative welfare scholarship has used his typology to categorize the welfare state according to three dimensions: decommodification (or the extent to which market dependency is reduced through state entitlements), the modes of social stratification (the idea that regime types manifest variable degrees of inequality) and the relationship between the state, market and/or family in the production and management of social wellbeing (a balance that typically varies across countries). This classification scheme is supposed to allow us to identify dimensions of variation within welfare regimes, dimensions that are explicit and narrow enough to be operationalized in an assessment of three ideal welfare cases-the typical liberal, conservative-corporatist and social democratic welfare regime-types.
The liberal welfare state provides a minimum level of social protections for its citizens. This is done in order to maximize market forces. Liberal regimes emphasize personal responsibility in welfare provision and deservingness for market relief. As a result, benefits are set low, social assistance is means-tested, and wage bargaining systems are decentralized so that the private sector can expand to its full potential. Liberal regimes try to ensure all who can work do so and obtain their income through participation in the free market. Social policy is therefore aimed at maximizing the market and individual independence. In such a model, a system of administrative surveillance monitors and enforces eligibility determinations for social assistance. Consequently, less than those who are eligible for relief end up receiving benefits. In general, liberal welfare states are said to have lower levels of public sector employment, direct job creation, social protection expenditures, long-term unemployment, union density and collective bargaining and higher levels of involuntary part-time work, marginally attached workers and short-term job tenure. The United States is considered the prototypical example of the liberal welfare regime [3].
The conservative-corporatist welfare regime is based on two different kinds of fragmentation in the provision of welfare. The first one entitles narrowly defined groups with their own specific benefits, with occupational and employment status playing an important role in the type or level of services provided. The second one provides welfare provision according to a perceived need for assistance using a measurable criterion. In general, traditional families are the main targets of welfare services, as social policy rewards breadwinners, and provides greater benefits to larger families. In this way, the conservative-corporatist welfare state promotes and sustains the position of the traditional family in society. Both types of fragmentation (occupation/employment and family) aim to preserve existing social structures and hierarchies. In this way, social benefits are not universal (equal for all), they tend to depend on past employment contributions, and financing is made possible through employer-employee payroll taxes rather than general tax revenues. Market forces are constrained to the extent that firms consider the interests of different shareholders in their business calculations. This helps maintain the 'social market' (diverse stakeholder economy). Conservative-corporatist welfare regimes generally feature medium levels of public sector employment, direct job creation, social protection expenditures, long term unemployment, union density and collective bargaining and lower levels of involuntary part-time work, marginally attached workers and short-term job tenure [3]. Germany is considered a primary example of the conservatist-corporatist welfare regime.
The social democratic welfare state tends to provide entitlements that are more universal in scope, egalitarian in their distributive goals and generous in their benefit levels. In this regime type, social rights provide high levels of decommodification and defamilialization in order to sustain social solidarity rather than reinforce hierarchical divisions or market forces. As the ideal welfare model, it supports the individual and family over the life course in order to make it easier for people to transition between life roles while dealing with challenges associated with work and/or parenthood. Such welfare states provide a range of cradle-to-grave protections for their citizens, including such things as child allowances, early childhood education, training policies, job protection, paid maternity leave, day care for preschoolers, wage replacement benefits, retirement pensions and home care for seniors. Social expenditures are typically financed through higher rates of taxation, with redistribution of additional income achieved via wealth taxation. Social democratic welfare states are said to have the highest levels of public employment, direct job creation, social protection expenditures, long-term unemployment (due to generous benefits), union density and collective bargaining and the lowest levels of involuntary part-time work, marginally attached work and short-term job tenure. Sweden is viewed as the main example of a social democratic welfare regime [3].
The southern European welfare regime was added to the three-fold typology above due to some developed countries exhibiting characteristics that distinguished them from other regime types [4,5]. The distinctive regime properties of the southern welfare state include familism, strong kindship networks, Catholicism, agricultural production and a fragmented system of welfare provision. Generally, the southern welfare state is said to have medium or lower levels of public employment, direct job creation, social protection expenditures, long-term unemployment, involuntary part-time work, marginally attached work, short job tenure, union density and collective bargaining. Spain is conventionally regarded as an example of the southern European welfare regime [3].
Significant debate has occurred over the extent to which countries actually fit into one of the ideal typical welfare regime models. Supporters of the typology view it as a valid and reliable methodological approach for categorizing mature welfare states and explaining cross-national similarities and differences between them. Others are much more skeptical. They criticize the typology [6] for its alleged inexhaustiveness (may be more than three or four ideal types), inexclusiveness (more anomalous cases than admitted), methodological soundness (wrong criteria and unsuitable operationalizations, variables and methods), lack of explanatory power [7] and usefulness in comparative analysis [8]. More sympathetic critics, however, admit significant differences in national political contexts exist among countries, particularly variations in institutional configurations, societal cleavage patterns, pressure of socio-economic forces, actor-constellations and degree of dependence on the state, market and family. These recognitions are generally based on contributions that delineate differences in government expenditures, ideological orientations, institutions, and formal policy elements [9,10]. The issue is that these aspects of welfare development fail to capture the outcomes of policies and institutions in practice. It is therefore important to take into account socio-economic effects in order to show how people fare under different welfare regimes and where regimes-types diverge and overlap in terms of real-life impacts [11]. While difficulties certain exist in making causal connections between different welfare arrangements and social outcomes, an outcome-centred approach nevertheless offers a unique vantage point to examine similarities and differences both between and within welfare regimes, which policy-focused and state-centred research has so far failed to adequately capture [12]. This is not to deny, of course, there may be a deep and important connection between policy, institutions and socio-economic outcomes. For instance, it may be that variations in social policy and levels of expenditure and the way institutional arrangements have developed in different countries are directly correlated with noticeable differences in their welfare outcomes and that countries, with highly diverse and unique configurations of capitalist political-economic institutions, cluster around a range of redistributive effects. In this way, variations in the socioeconomic outcomes observed, then, should be explainable with reference to the characteristics of the regime type in question. Moreover, countries associated with a particular regime-type should all exhibit similar outcomes and follow similar development paths.
Still, the question of whether and to what extent measures of socio-economic outcomes between different countries vary significantly with the regime-type has been relatively under researched in the literature. The little research that does exist has tended to focus on the micro level, the pre-crisis period and few cases and indicators. What is needed is an empirical assessment of outcomes across a large number of country cases during a period of external shock like the 2007-2008 crisis, which put intense pressure on public budgets, forcing many countries to restructure their welfare states [13,14].
External and internal pressures do not necessarily determine the trajectories of welfare states, neither do they dictate their capacities to achieve desired goals in a uniform way. But they can restrict the possibilities and choices welfare states have at their disposal. Such modifying pressures toward conformity can include demographic changes, changing household and family patterns, the growth of non-standard employment arrangements, structural unemployment, technological changes, international competition, mobility of capital, international trade, participation in a common currency, globalization and economic crises [15]. These pressures can come in different forms and welfare states may react to them differently. But exactly how they affect the socio-economic outcomes of different welfare regimes remains an under-researched area.
Arguably, periods of economic crisis impose challenges and constraints that are somewhat different and perhaps more important than some of the other pressures on the welfare state. However, economic crises also tend to affect countries differently because of the diverse political-economic capacities and varied vulnerabilities of welfare regimes to such external shocks. Certainly, the global financial crisis of 2007 marked a 'stress test' for many welfare states [16] as more than ever before policymakers were forced to consider cuts in welfare provision and enact flexibilityenhancing measures in order to improve financial solvency and economic performance [17]. Adding to this challenge was the fact that the recent financial crisis, much like the recession of the early 1990s, occurred when many countries were already suffering from an economic downturn [13,14].
Given the significant pressures the recent crisis placed on many welfare states, it is worth examining whether the boundaries between them have been fundamentally redrawn. For instance, did the crisis produce similar socio-economic outcomes in different regime types? Or, have there been notable and systematic differences in outcomes and patterns of continuity across welfare models in the years following the worst moments of the crisis? To what extent do the levels and distribution of welfare, as measured by certain outcomes, separate countries from each other? Countries within a regime cluster should not only be different enough as a regimetype to distinguish them from other countries, but also have outcomes in common. In this sense, within-regime coherence is just as important as between-regime differences [18].

Varieties of capitalism
Similar to the typology of welfare states, the varieties of capitalism framework popularized by Hall and Soskice [19] highlights systematic differences in the structural characteristics of countries. However, unlike welfare regime theory, it uses market and strategic coordination to differentiate national systems. According to the varieties of capitalism theory, domestic business coordinating capacity is a crucial factor in understanding the divergent structures, strategies and outcomes of countries. Configurations of national institutions are supposed to function in a complementary manner in order to produce different comparative advantages in innovation and production systems at the national level. As such, the business firm is placed at the center of the varieties framework, with actors formulating their decisions within the framework of institutions such as corporate governance systems, training regimes, wage setting and employee representation, labor legislation and the character and level of coordination between firms. These firms are connected to specific institutional arrangements. They have discretionary powers to make different decisions, but they tend to do so in ways that complement specific production models and competitive advantages. The concept of institutional complementarity is essential to the varieties approach. Complementarity is present when the existence or efficiency of one institution increases the returns from another [19].
The varieties approach is based on a comparison between two ideal types of capitalism-the liberal market economy and the coordinated market economy. In a liberal production regime, market-based mechanisms are maximized, with firms competing and coordinating according to supply and demand factors shaped by exchange rates of products and market actors. In a coordinated regime, nonmarket-based relations play an important role, as firms engage with each other in more collaborative ways, achieving coordination via partner networks and other institutionalized arrangements. Liberal economies are defined by an outsider corporate governance system, which promote general skills whereas coordinated economies are framed by an insider system of corporate finance, instead favoring vocational training programs and specific skills for long-term employment relationships. The varieties framework complements welfare-regime explanations insofar as liberal market economies and liberal welfare states overlap, while coordinated economies can be embedded within social democratic and conservative welfare regimes [20].
A third type of production system, the Mediterranean economy, is presented as a hybrid between the market and coordinated varieties of capitalism. It is characterized by specific kinds of non-market coordination in the area of corporate finance and more liberalized labor relations. Mediterranean economies are said to have a large agrarian sector and more extensive forms of state intervention in the economy. They sometimes have greater employment protections but lower decommodification rates and low product market competition with less short-term pressures for profits because of their centralized financial systems that can devalue national currencies. They also tend to have more limited qualifications that curb a high skill and wage strategy. The Mediterranean economy is said to overlap with the southern European welfare state [19].
Institutional divergence is emphasized in the varieties framework because of the recognition that there is no 'one size fits all' approach to capitalism. Different institutional configurations, themselves products of political development, culture and historical processes, support different production, market and investment strategies, resulting in different types of comparative advantages. Another important element in the varieties school is the recognition that firms make discretionary choices that are shaped by national institutional factors, which in turn influence, constrain and support those decisions. The varieties approach is therefore mainly concerned with the ways in which different types of economies co-exist, given their respective strengths and weaknesses in specific areas.
The varieties approach has been widely debated. It has been criticized for being too parsimonious for identifying only two types of economies or exaggerating the differences between them. Others claim it faces problems explaining changes over time [21,22]. There could be important processes common to both types of production regimes, which the varieties perspective not only fails to explain but also directs attention away from [23]. For example, it has been suggested that the varieties school cannot adequately account for economic restructuring processes as well as it can describe the institutional sources of comparative advantage [24]. Research has shown, for example, how different institutional arrangements have increasingly morphed into liberal governance systems in order to adapt to short term investor interests [25]. Other research suggests, however, that while internal and external pressures have not yet channeled different institutional configurations into a single converging and parallel model, there have been similar patterns of policy adoption across countries such as the transition from income provision to the promotion of market participation [26].
The global financial crisis has raised new questions about the persistence of cross-national differences identified by the varieties perspective. Sympathetic critics claim that convergence toward neoliberal and austerity policies in North America and Europe has diminished the differences within varieties of capitalism, but that path dependent development has nevertheless given way to divergent trends [27]. From this perspective, typical varieties of capitalism can change in degree 'from within' , without necessarily transforming to another type of production system [28]. However, very little research has been conducted to assess the outcomes of different market systems over the crisis period as compared to institutional and policy trends [29]. This relatively neglected area can be used to uncover evidence of convergent or divergent trends between different production regimes.
The varieties approach has been used with some success to explain the empirical pattern of continuity and change in outcomes in different types of economies [30]. Scholars have used the typology to identify the micro processes of the relationship between institutional structures and certain policy outcomes. For example, one study sought to explain why trade union density and collective bargaining coverage have decreased in mainly liberal market economies but not so much in coordinated market economies [31]. Others have used the approach to make empirical claims about trends in labor market policies and outcomes such as employment protection, unemployment compensation and wages [32].
Most of the research in this area, however, has focused on a few socio-economic indicators, a minimum number of country cases, and the period before economic stagnation combined with recession put intense pressure on public budgets and increased the likelihood of negative labor market effects. More research is therefore needed to determine whether the labor market performance of the three archetypal forms of distinct capitalist systems conform to the expectations of the varieties of capitalism paradigm.

Convergence
Welfare states and varieties of capitalism are depicted as being more or less resilient to internal and external pressures. A number of studies have tried to show, for example, how the core institutional and policymaking characteristics of different welfare and production systems have remained firmly in place since the 1980s, even in the face of common pressures [33]. However, another body of comparative research has argued the opposite that different countries not only experience the same pressures but also in ways that lead them to converge toward a single organizational structure or pattern of output. The convergence theory holds that sociopolitical and structural changes, including globalization and economic crises [15] are homogenizing pressures that push all countries toward the same institutional logics, policies and/or outcomes. In this way, convergence is a process whereby differences between countries become less discernible over time.
However, previous studies have paid more attention to institutions, policies and the ideas influencing policy-making than to socio-economic outcomes [34]. A more up-to-date and comprehensive study is therefore needed to determine the extent of convergence in the post-crisis period.

Country cases, indicators, time period, limitations and measurements
In what follows, we compare the outcomes of different welfare and produc-  [2,19].
We use the following nine indicators from the OECD database in order to compare the outcomes of 16 countries (clustered into four regime-types) before and after the recent crisis: public sector employment (employees in general government as a percentage of total employment), direct job creation (temporary work, and in some cases, regular jobs in the public or non-profit sector, offered to unemployed individuals), social protection expenditures (as a percentage of total expenditures), long term unemployment (share of unemployed as a percentage of the unemployed population), involuntary part time work (share of involuntary part-timers as a percentage of part-time employment, in dependent employment), marginally attached workers (those aged 15 and over, neither employed nor actively searching for work, but are willing to take up a job, as share of labor force), short-term job tenure (percentage of employees working for their employers for less than 1 year), trade union density (the ratio of wage and salary earners who are union members, divided by the total number of wage and salary earners) and collective bargaining coverage (number of employees covered by the collective agreement, divided by the total number of wage and salary earners) [35].
These nine indicators [36][37][38][39][40][41][42][43][44] were chosen for a variety of reasons. First, they were readily available for two points in time-one pre-crisis (2007) and one post-crisis (2017-2018)-for all the selected countries. Second, they are decent measures of socio-economic outcomes as they encompass a range of factors integral to macro-economic stability and social well-being. Third, they are easily accessible to a very broad audience. Fourth, they are measured regularly by the OECD, so they can be periodically updated to show further divergent or convergent trends beyond the most recent year for which data was available. Fifth, the choice of indicators was guided by an interest in processes connected to marginalization and social exclusion, given the type of welfare regime and variety of capitalism is supposed to be consequential for determining the relative inclusion and exclusion of certain groups in society. While some indicators do not capture the socio-economic position of large numbers of people, they are nevertheless important measures of social cohesion, which is an important social policy goal in all welfare and production regimes.
Making cross-country comparisons, however, is complicated by the differential impacts of the global financial and public debt crises. They did not affect all countries to the same extent and policy reactions and recovery processes have been very much unequal. In part, this is because of significant variations in financial and trade linkages and the practices and quality of other institutional factors such as exchange rate regimes and capital controls. The debt crises, for example, were prevalent in European countries, especially the Mediterranean region. Different regulatory settings, economic conditions, demographics and political factors all help determine both the level of vulnerability of welfare regimes and market economies to external shocks and the nature, direction and robustness of policy responses. This is one of the reasons why it is difficult to provide a single start date for the crisis or attribute crisis mitigation and recovery processes to the designs and capacities of specific types of welfare states and production models. Care should therefore be taken before generalizing across cases.
The pre-and post-crisis periods can be used as a test case to compare the socioeconomic outcomes of different welfare production systems. For this purpose, several indicators can be utilized to shed light on the extent of cross-national outcome convergence and divergence, which can be measured by comparing the: overall direction of change in the indicators (e.g., numeric outcomes improved or deteriorated), numerical values of the indicators (e.g., values very similar or different to each other), percent changes from one period to another, average country group scores, and average country group percent changes from one period to another ( Table 1).
Public sector employment decreased in 12 out of 15 countries. The share of employment in general government ranged between 10.49 and 28.83% in 2017. In line with divergent expectations, the SDWS-CMEs had the highest rates of public employment (averaging 27.05%), although they all experienced a small decline. Perhaps more surprisingly, the LWS-LMEs, CWS-CMEs and SWS-MEs all had similar rates and group averages, making them somewhat indistinguishable. However, there were small differences in the rates of the CWS-CMEs.
Direct job creation for the unemployed increased in seven countries, decreased in five countries and remained the same in three countries. Percent change increases ranged from 5.26 to 4100%. Percent change decreases ranged between 6.12 and 100%. While more countries had created public jobs for their vulnerable members since 2007, the 16 country group average in 2017 had declined by a change of 1.48%. Unexpectedly, the group average of the CWS-CMEs had decreased by a change of 52.71% while the group average of the SWS-MEs had increased by a change of 241.54%. Within-regime differences in rates were quite common.
Social protection expenditures increased in 10 out of 15 countries. Percent change increases ranged from 2.72 to 66.76%. The group averages of all regimetypes except the LWS-LMEs increased in 2017. Contrary to expectations, the group averages of the SDWS-CMEs and LWS-LMEs were very similar in 2017 (36.06 and 33.56%), despite these two regime-types representing opposite ends of the social expenditure spectrum in the worlds of welfare literature. Within-regime percentages were very similar among the LWS-LMEs and SDWS-CMEs, with most of them in the early to mid-30% range. However, the rates among the CWS-CMEs varied considerably (ranging from 16.66 to 46.14%). Similarly, the CWS-MEs all had very different rates (2.92, 8.37, 16.05 and 22.36%) ( Table 2).
The share of individuals out of work for 1 year or more increased in 12 out of 16 countries. Percent change increases ranged between 2.54 and 104.41%. In 2018, the LWS-LMEs and SDWS-CMEs experienced an increase in their long-term unemployment rates and both regimes had very similar group averages (21.98 and 19.5%). While most CWS-CMEs saw their rates decrease in the post-crisis period, their group average (39.25%) was higher than that of the LWS-LMEs and SDWS-CMEs. There were significant within-regime differences in terms of rates among the LWS-LMEs, CWS-CMEs and SWS-MEs, showing once again divergent theories fail to account for dissimilar outcomes within particular regime-clusters.
The share of involuntary part-timers increased in 7 out of 13 countries. Percent change increases ranged from 19.75 to 157.41%. Consistent with divergent theories, the LWS-LMEs and SWS-MEs had some of the highest rates (23.07 and 64.42%), while the CWS-CMEs and SDWS-CMEs had the lowest (8.48 and 11.25%). The sample group average increased by a change of almost 20%.
The share of marginally attached workers decreased in eight countries, increased in six countries and remained the same in two countries. Country group averages were nearly indistinguishable in 2018, blurring the dividing lines between countries. There were notable differences among the LWS-LMEs, CWS-CMEs and SDWS-CMEs, particularly Australia, the Netherlands and Finland whose rates were more than twice that of their group members ( Table 3).
The share of employees with less than 1 year job tenure decreased in nine countries and increased in six countries. Percent change decreases ranged between 3 and 22.43%. Percent change increases were from 0.92 to 89.42%. Short-term job tenure rates were similar in the LWS-LMEs and SDWS-CMEs (close to 20% points), contrary to expectations. In the literature, the prevalence of flexible employment is said to be greater in the liberal economies. While short-term job tenure decreased in most places, at least 11 countries from four different regimes had rates between 16 and 22%, meaning nearly one-fifth of all their employees were working for their employers for less than 12 months. These similar rates make the regime-clusters indistinguishable as well. There were also within-regime differences among the CWS-CMEs and SWS-MEs.
Trade union density decreased in 15 out of 16 countries. In eight countries, there were percent change decreases ranging between 15 and 30%. Consistent with expectations, the LWS-LMEs and SWS-MEs had the lowest union density figures, with the CWS-CMEs and SDWS-CMEs coming in second and third place respectively. There were considerable outliers in all the regime-types except the SDWS-CMEs whose rates were similar.  Table 2.

Long-term unemployment (LTU), involuntary part-time workers (IPTW) and marginally attached workers (MAW) in worlds of welfare and varieties of capitalism.
Collective bargaining coverage decreased in eight countries, increased in five countries and remained the same in three countries. Group averages in 2017 were consistent with assumptions in the literature, although the SDWS-CMEs, CWS-CMEs and SWS-MEs had very similar averages in 2007. Within-regime differences were notable in all the regime-types except the SDWS-CMEs whose rates were similar.

Conclusion
Overall, there was unidirectional convergence toward negative outcomes in many countries in the post-crisis period. Many countries have not returned to their precrisis levels as several indicators show, including public sector employment (12 out of 15 countries experienced a decrease), direct job creation (decreased in 5 out of 15  Table 3. Short-term job tenure (STJT), trade union density (TUD) and collective bargaining coverage (CBC) in worlds of welfare and varieties of capitalism.