1 Introduction

Since the turn of the century, social pension programmes have evolved as a new instrument for old-age provision in many Eastern, South-Eastern, and Southern Asian countries (Asher and Bali 2015; Barrientos 2015). This is remarkable given that some of these countries introduced social insurance systems only in the 1990s, partly following the International Labour Organization’s (ILO) considerable efforts to establish social-insurance-based pension schemes around the globe (Deacon 2015). In recent years, many Eastern, South-Eastern, and Southern Asian countries have seen encompassing welfare state reforms (Hujo 2014; Haggard and Kaufman 2008; Asher and Kimura 2015). While much research has focused on the development of social pensions in these countries (Gliszczynski 2015; Leisering 2019; Barrientos 2015; Böger and Leisering 2020), it is important to keep in mind that social pensions have not replaced the older and often still much more comprehensive contribution-based systems. Against this backdrop, this chapter focuses on these contribution-based pension systems. The trajectories of the contribution-based pension systems in many countries in these regions follow a similar pattern. Although they are facing considerable problems in expanding effective coverage and providing a decent pension after retirement, these countries have undertaken considerable political efforts to maintain or even expand those systems. We aim to explain these dynamics through a modular and mechanism-based approach and thereby specifically seek to demonstrate the potential of mechanism-based analysis in comparative research. To this end, we compare two different types of contribution-based pension systems: pension systems that largely follow a Bismarckian logic and pension systems that largely rely on national provident funds (the latter have only received scant attention in the comparative social policy literature so far; but see Dixon 1989; Lindeman 2002; Kaseke et al. 2011). Generally, provident funds can be understood as a social policy legacy from British colonial rule. Most share the following characteristics: (1) They are compulsory for those in formal employment with larger employers. (2) They operate as defined-contribution schemes managed as part of the public sector. (3) Contributions are paid by both employees and employers. (4) By design, they accumulate interest based on defined rates. (5) Instead of regular payments, most provident funds rely on lump-sum payments for employees that enter retirement age (Kaseke et al. 2011). Thus, in contrast to Bismarckian pension schemes, they generally involve no redistributive elements and are not a pay-as-you-go system but only individual accounts. Moreover, lump-sum benefits have been criticised for not guaranteeing a regular income for the whole duration of retirement.

We have chosen the pension systems of South Korea, Vietnam, Sri Lanka, and Malaysia for our analysis, thus following a most different systems design for our comparison of contribution-based pension systems. Two of the countries have a Bismarckian social insurance scheme (South Korea, Vietnam), one with a rather high (South Korea) and one with a rather low level of coverage (Vietnam). The two other countries have a national provident fund (Malaysia, Sri Lanka), again one with a rather high (Malaysia) and one with a rather low level of coverage (Sri Lanka). The aim of this chapter is to show that similar mechanisms and outcomes can be found in these different contribution-based pension systems and thus highlight the set of key complex causal mechanisms that are at work across the different variants of contribution-based pension systems.

The four countries differ with regard to a number of additional factors. These include, first, their income level. Second, the countries also differ with regard to their types of political system and frequency of government changes. While South Korea and Sri Lanka are characterised by party competition, the Socialist Republic of Vietnam has always been governed by the Communist Party, and Malaysia experienced its first government change only in 2018. The national time frames of the study range from the initial establishment of a pension system for workers outside of the state sector up until today. Although this implies a relatively long time frame for countries with a national provident fund (which was established in Malaysia in 1951 and in Sri Lanka in 1958), Bismarckian pension systems covering workers in the private sector were introduced more recently in South Korea (1986/1988) and Vietnam (1995). Yet, irrespective of these different time periods, none of the selected countries have been able to achieve universal coverage, and there are significant differences between legal and effective coverage in each country. Data from the ILO illustrates this point: According to the most recent World Social Protection Report 2017–2019 (ILO 2017), legal coverage of mandatory contributory systems for old-age provision as a percentage of the working-age population is 70.9% in South Korea, 48.6% in Malaysia, 32.9% in Sri Lanka, and 33.1% in Vietnam. In contrast, effective coverage (indicating active contributors to a pension scheme in the working-age population between 15 and 64 years of age) was 53.7% in South Korea, 28.1% in Malaysia, 18.9% in Sri Lanka, and 20.6% in Vietnam.Footnote 1 Although this data is not fully comparable, it nevertheless shows that important coverage gaps remain.

The puzzle for researchers is to explain a permanently low effective coverage rate in spite of ongoing political efforts to maintain or even expand these contribution-based systems in all countries. Our analysis of the pension systems in the four countries reveals six complex causal mechanisms that explain important dynamics that are at play. In a nutshell, (1) the outcompeting mechanism explains the expansion of social policy as a result of party competition in democratic settings. Conversely, (2) the gaining acceptance spiral mechanism reveals how autocratic regimes introduce or expand social policies in order to increase their legitimacy. (3) The evasion mechanism explains the interaction between national policy-makers and international organisations and how national policy-makers announce reforms in line with recommendations from international organisations but do not implement them. As these political mechanisms are not limited to the realm of contribution-based pension systems—and, in the case of the evasion mechanism, not even to the realm of social policy—we define these as general political mechanisms. In addition to these three mechanisms, we identify three more that are characteristic of contribution-based pension systems: (4) The double benefit mechanism explains how money accumulated in a pension fund not only serves old-age provision but also helps to finance other government purposes, thus leading to an expansion of the pension system. (5) The crisis management by going further mechanism explains how a strategy of expanding social policies to include new groups can help to maintain the system’s financial sustainability in times of crisis. (6) The alarmed middle classes mechanism explains how fears among privileged groups of higher costs or lower benefits lead to an only moderate expansion of the pension system.

The chapter is structured as follows. We start by presenting the causal effects and key causal mechanisms that we briefly identified above. The third section gives a historical overview of the pension systems of South Korea, Vietnam, Sri Lanka, and Malaysia, and traces the causal mechanisms that are at play. The fourth section provides our conclusion. Our analysis relies on a broad assessment of the secondary literature as well as on national documents and policy reports from international organisations.

2 The Development of Pension Systems: Causal Effects and Mechanisms

The trajectories of Eastern, South-Eastern, and Southern Asian pension systems indicate that contribution-based pension schemes are being maintained irrespective of the problems of using these as a basis to build a coherent pension system that covers most individuals. The persistence of these systems we define as the continuation effect. To specify this broader effect, we distinguish more specific causal effects that characterise the development of social policy in the respective countries: The ratchet effect indicates that contribution-based schemes once introduced are not politically abolished, transferred to other schemes, or fundamentally retrenched. The expansion effect describes the extension of the pension system to additional groups; the moderate expansion effect is a variant of this that refers to an expansion that is limited by political dynamics. Finally, we identify an embellishment effect, a special form of interaction between national governments and international organisations, by which national governments signal agreement to avoid conflict but do not effectively implement the measures in question. All effects result from different complex causal mechanisms that can be combined to explain the development of pension policy in the four countries. We thus rely on the set of mechanisms outlined above to show how a modular explanation can be achieved (Chap. 1). We first focus on three more general political mechanisms that we identified throughout the study of the four country cases before introducing three mechanisms that are more specific to contribution-based pension systems. Note that this does not imply that all mechanisms play a role in all four cases. Rather, the mechanisms can be combined in different ways to explain the trajectories of the different pension systems in the countries in question.

2.1 Outcompeting Mechanism

The outcompeting mechanism explains the expansion of the pension system. It can occur wherever there is competition in elections, especially party competition. The relationship between party competition and social policy is a long-standing topic in the comparative social policy literature, which in this respect often focuses on differences between parties to explain social policy developments (e.g. Schmitt and Zohlnhöfer 2019). Yet, party differences might also blur if several parties aim to expand social policies. If several parties in a country have the prospect of winning the elections, the elites of these parties choose strategies that intensify political competition. There might be different reasons why the elites of competing parties decide to introduce or expand a social insurance scheme, among them economic growth, increased tax revenues, or hopes of a future increase in state revenues, not to mention broader political or societal changes that alter existing cleavage structures (Wong 2004). In any case, these party elites will only propose introducing or expanding a social insurance system if they also think that there will be a large number of potential voters who will welcome this policy, whether for personal economic reasons, as an expression of their country’s advanced development, or out of enthusiasm for a normatively convincing political programme.

One variant of the outcompeting mechanism involves two or more strategically interacting parties. Since Party A may know that it makes sense for Party B to promise the introduction or expansion of social insurance institutions, and Party B may know that this also applies to Party A, both parties will outbid each other in their promises to extend social benefits (Derthick 1979). However, this is only successful if the parties can hope that voters will bear the costs of the expansion strategy later on or that rising costs will not be considered as mistakes or failures for which a certain party will be held accountable. This condition may be neglected, however, when party elites can afford to shorten the time horizon of their strategic considerations to the time of the election campaign and do not feel bound by norms of good governance.

2.2 Gaining Acceptance Spiral Mechanism

This mechanism can play a role in non-democratic countries and lead to the same expansion effect as in the case of the outcompeting mechanism in democratic ones. In the literature, it is widely accepted that not only democratic governments adopt or expand social policy measures but that autocratic governments also have reasons to do so. In particular, it has been highlighted that autocratic governments might adopt social policy measures to gain the support of particular groups as “a special form of co-optation” (Knutsen and Rasmussen 2018; see also Mares and Carnes 2009). Autocratic governments do not rely on repression and co-optation alone for survival. They also maintain stability by cultivating belief in their legitimacy (Gerschewski 2013). Autocratic regimes might opt for introducing or expanding social benefits if they see scope for additional public expenditure and view social policy as being a more effective means of lending themselves legitimacy than the pursuit of additional economic growth. Once established, these social policies can set in motion an acceptance spiral that draws on further social policy expansion. This acceptance spiral might become even more important if improving the living conditions of the population is a key ideological pillar of that autocratic regime, as in many socialist regimes (Wurster 2019). Moreover, even if social policy reforms do not significantly increase acceptance among the population, governments may still have reason to further expand social policies. Lack of acceptance may lead party elites to outbid each other in their reform efforts to bolster their support among the population.

2.3 Evasion Mechanism

The evasion mechanism (Kuhlmann and Nullmeier 2021) sheds light on the role of international organisations in national reform processes. This aspect has been increasingly researched in comparative social policy in recent years (Orenstein 2008; Deacon 2015; Leisering 2019), with mixed findings on the overall role that international organisations actually play in national social policy reform (see also Chap. 8, Chap. 9). In fact, while international organisations might be deeply involved in social policy reform, it is not unusual for the adopted policies to primarily reflect national considerations. And even in this case, both national and transnational actors may well assess the cooperation as having been highly successful. This we define as the embellishment effect. If, for instance, an international organisation urges a country to restructure its social security systems and to introduce or expand a social insurance scheme, national actors might employ strategies of circumventing recommendations by international organisations. They typically do not do so openly because of international pressure for reform. What national actors do instead is either stress that their planned policies are in line with international recommendations on social insurance schemes or develop policy compromises that make vague references to the recommended social insurance schemes. In a mixture of rational calculation and the normative belief that social policy legislation is a matter of national sovereignty, political elites resist international organisations’ attempts to shape the direction of reforms while avoiding open dissent so as not to risk offending these international organisations. To accomplish this, national policy actors obscure the fact that they are not (fully) following the respective recommendations and are not adopting the social insurance scheme in national policy accordingly so as not to endanger important contacts with the international organisations.

2.4 Double Benefit Mechanism

Pension schemes are often characterised by a double benefit mechanism. The double benefit of pension funds lies in disbursing old-age provision to the population on the one hand and accumulating a considerable amount of money for the state on the other. This leads to the said ratchet effect of the pension system. Governments ensure the public financing of the pension fund because they want to use the money to finance the public debt or promote the country’s economic development (Holliday 2000; Koreh 2017; see also Kuhlmann and Nullmeier 2021, 2022).Footnote 2

The size of the fund whets rational actors’ appetite for accessing it for other purposes. The prospect of sizeable amounts of discretionary funds also provides a strong incentive for the government to push for expanding the system if this entails additional contribution payments and thus increases the fund. Capital assets increase in particular if contributions have already been paid but no payments have yet been made due to a statutory minimum savings period or waiting period. Under these conditions, the government as well as opposition parties will find any expansion of coverage highly reasonable, as expanding the insured population becomes a viable policy option because of the attractiveness of further increasing the possibilities of state financing. The impetus for enlargement will even be stronger if the government has to win elections. In this case, the promise of new or more encompassing social benefits may correspond with securing the financing of public debt or enhancing the national capital stock. The other side of the coin is that abolishing contribution-based schemes becomes highly unattractive for a government that has once enjoyed the opportunity to tap such a reservoir to finance its public debt or other government expenditures.

The money from the pension fund also allows it to serve other purposes. For example, policy-makers can choose to use money from the pension fund to relieve pressure on a tight labour market by facilitating early retirement (Trampusch 2005). This constitutes a special variant of the double benefit mechanism.

2.5 Crisis Management by Going Further Mechanism

This mechanism can become relevant when a government is faced with a situation in which a social security system covering only a limited number of persons runs into massive financial trouble either because the number of insured persons is falling, or the contributions of insured persons are too low in relation to the promised benefits, or high inflation rates are diminishing the fund’s real value, or the funds are simply too small to generate good investment rates in the financial market. In these cases, policy-makers can weigh several policy options, such as dismantling or transforming the system, cutting benefits, raising contribution rates, or providing massive state subsidies, for instance, in the form of tax breaks. If the government considers expanding the circle of insured persons to be another viable political option, this will become the preferred choice to emerge from rational calculation because it involves low burdens for both the group of already insured persons and the state budget. In this situation, rational government actors will increase the degree of inclusion to ensure the stability of the system, thereby leading to an expansion effect. Thus, the crisis of the old system with low coverage rates becomes the driving force for including other segments of the population.

2.6 Alarmed Middle Classes Mechanism

In countries with a large group of informal workers, formal workers can regularly be considered as members of the middle classes. As a privileged part of the population, these formal workers are among the first groups to be included in a contribution-based pension scheme (e.g. Huber and Stephens 2012), and they do not want to lose these privileges. If a contributory insurance system has been introduced in a country for one group, extending this system of social protection to other parts of the population (such as farmers, fishermen, the self-employed, or informal workers) may appear to be a legitimate concern in the eyes of the groups already included in the system. The middle classes will typically consider these demands argumentatively convincing and worthy of recognition on normative grounds, as equal rights for all combined with some form of performance justice, a characteristic especially of social insurance institutions, are widely recognised.

These middle classes therefore initially approve government plans to expand social protection schemes. However, when the conditions of expanded coverage begin to take shape in detail, it often becomes clear that this expansion will also have material consequences for those already insured. Especially if the new groups are considered to be financially weaker, this might stoke fears that expanded coverage could entail higher contribution rates or reduced benefits for all in the long term. Moreover, if the government grants subsidies to include new, financially less powerful groups into the system, the middle classes might perceive this as unfair. Such a discrepancy between initial normative approval of an expansion and the fears and concerns related to the enactment of the actual reform can incite anxiety among the middle classes. This constellation will allow for a rather moderate expansion of social protection to new groups only if the governing parties expect such an expansion to improve their prospects of electoral success.

3 Country Studies

In the following sections, we first describe the trajectories of the national pension systems in the four countries under study (see also the overviews in Kuhlmann and Nullmeier 2021 for Vietnam and Sri Lanka, and Kuhlmann and Nullmeier 2022 for South Korea and Malaysia). We then explain the development of these systems by referring to the causal effects and mechanisms that were introduced in the previous section.

3.1 South Korea

3.1.1 Overview of the Pension System

Although the president of the former military regime, Park Chung-hee, had announced the introduction of a compulsory national pension system as early as in 1973, a general pension insurance was introduced only in 1988, following the country’s democratisation. It then expanded quite quickly. Since its inception, the South Korean pension scheme has been based on the principle of defined benefits and can be attributed to the Bismarckian type, which implies that contributions are paid by employers and employees. While compulsory insurance initially applied only to employees working in companies with more than ten employees, this changed in 1992 to also capture employees working in companies with more than five employees. Between 2003 and 2006, the compulsory pension scheme was extended to additionally include employees working in companies with fewer than five full-time workers. The pension system was also opened to new groups: It was extended to the rural self-employed, farmers, and fishermen (2.1 million people; Yang 2017, 121) in 1995, and to the urban self-employed in 1999. Notably, it was not until 2008 that the pension scheme issued the first full pension payments. This owed itself to a minimum insurance period of 20 years. Additional pension schemes exist for different occupational groups, such as civil servants (this scheme was established in 1960), military personnel (established in 1963), and private school teachers (established in 1974) (Hwang 2006; Kwon 2014). Moreover, large companies also have company pension schemes. Together with Germany, Austria, Slovakia, and the USA, Korea is one of five OECD countries whose system of old-age security does not provide a minimum pension (first tier) but only social assistance with or without special regulations for older people (OECD 2017, 88).

Since 1999, the country’s pension law has stipulated that the entire population shall be included in an old-age provision scheme with a standard retirement age of 61, rising to 65 until 2034 at a current contribution rate of 9%. Different types of mandatory membership exist for workplace-based insured and individually insured persons. While the contributions of the former group are paid by both the employer and the employee, those in the latter group pay all contributions themselves. Moreover, individuals can be insured voluntarily and in some cases exempted from compulsory insurance (e.g. in the event of unemployment, business closure, temporary leave, livelihood difficulties, and in the case of students in tertiary education). To increase the level of old-age pension coverage, subsidies for low-income earners were introduced in 2012 (Duru-Nuri programme). Current studies suggest that this has increased the number of insured people over the following four years by 2% (Yoo et al. 2016). Yet, given that coverage rates continue to be low, especially among people younger than 35 years of age, extending the subsidy to additional groups is a matter of debate (Young et al. 2016). Effective coverage rates are at 68.2% of the economically active population between 18 and 59 (Young et al. 2016), which can be considered a relatively high rate compared to other Asian pension systems (Hujo and Cook 2012).

3.1.2 Mechanisms

First and foremost, the expansion of the South Korean pension system can be explained by two of our mechanisms: The gaining acceptance spiral mechanism, in the late phase of the dictatorship, and the outcompeting mechanism, which characterised party competition during the democratisation process. The military dictatorship established a lifetime employment system and a seniority wage system. In the course of the country’s democratisation process, the lifetime employment system gradually gave way to a more flexible labour market. The South Korean labour market thus became more flexible, and workers were compensated for growing employment insecurity by introducing social policy measures (Peng 2012), specifically by an expansion of health insurance and old-age pension insurance. However, despite the strengthening of civil-society organisations, labour unions, and the single-member district electoral system, there remained major obstacles to a vigorous social policy (Huber and Niedzwiecki 2015; Yang 2017).

Overall, however, the democratisation process gradually enhanced the importance of social policy legislation (Shim 2019). In the years that followed, pension insurance was extended to more and more groups of the population, especially to the conservative party’s agricultural clientele. An attempt at a neo-liberal restructuring of old-age provision failed during the Asian crisis of 1996–1998. The poor performance of the old government during the crisis also made it possible for the former opposition leader Kim Dae-jung to win the presidency. Only after parts of the middle classes, including the labour unions, protested the far-reaching reform plans of his successor, Roh Moo-hyun, was Dae-jung able to wage a successful anti-welfare election campaign. Yet, in the subsequent presidential election, the new conservative candidate, Park Geun-hye, returned with an aggressive pro-welfare programme that the centre-left candidate tried unsuccessfully to outbid. Since 1988 (with the exception of 2008), party competition in presidential elections has served as a driving force for the expansion of the welfare state, even though many electoral promises were not implemented (Yang 2017).

The development of the pension system can also be explained by the double benefit mechanism and the ensuing ratchet effect: Creating a capital stock through a funded pension insurance scheme had already played an important role in South Korea’s political decision-making processes before the pension scheme was actually introduced (Hwang 2006, 57–64), especially among executives (Kim and Choi 2014) who promoted a welfare state concept that focused more on economic than on social objectives (Kim 2008). Within this framework, the central aim of social policy was to promote the national economy (see also Holliday 2000). The creation of a pension fund came with an enormous accumulation of capital, especially given the fact that a minimum insurance period ensured that no payments had to be made for 20 years. In essence, this resulted in a Bismarckian system that was based on capital funding instead of a pay-as-you-go system. The fund was ultimately administered autonomously but in close coordination with the state (Kim and Stewart 2011; Yang 2017, 140). Today, the National Pension Fund is the third largest pension fund in the world, with a capital base equivalent to 516 billion euros. Between 1988 and 2020, it generated an average return on investment of 4.62%, with 34.4% currently invested in financial assets and development projects outside South Korea. The fund owns shares in state-owned companies, is deeply involved in the financial sector, and is subscribed to a shareholder-activism philosophy (Choi et al. 2018; Nomura 2011). In the light of its sheer volume, the role of the pension fund for economic purposes can hardly be overestimated. It can support the country’s national foreign and financial strategy even if social policy goals have gained significance over the years (Kuhlmann and Nullmeier 2022).

The alarmed middle classes mechanism can explain the moderate but limited expansion of the system to new groups. In fact, the social policy reforms under the presidencies of Kim Dae-jung and Roh Moo-hyun came under increasing pressure from parts of their own clientele. After a short phase of opening up the pension scheme to include non-standard workers (Durazzi et al. 2018), this strategy towards greater inclusion soon lost the support of the labour unions (Yang 2017). A crucial factor was the employers’ reaction to the reforms under Kim Dae-jung. To reduce labour costs, they massively expanded non-standard employment (Park 2008), which revived the major labour unions’ focus on an insider strategy. This abandoning of a universalist policy at the beginning of the 2000s can only be explained by dualisation (Peng 2012) and the split within the working class (Yang 2017, 154–83). The high-earning working class as a central component of the middle class defended its social privileges and was increasingly sceptical about extending social security to other groups of the population. Their own economic situation was improved more markedly by occupational social benefits than by public benefits (Yang 2017, 213). The situation was different at the beginning of the reform movement under Kim Dae-jung. Political plans to universalise social security had been supported by the working class in large companies because extending social protection complies with the norm of universal social rights and the norm to improve the situation of disadvantaged informal workers. However, detailed information about the redistributive effects of expanded coverage, especially the decrease in expected pension benefits for the middle classes, led the latter to a switch from norm compliance to self-interested rational choice:

Labor unions and civic groups did not openly question the legitimacy of the redistributive mechanism when they were involved in the making of the moderate pension reform bill. However, when it became apparent that the average income of the urban informal sector was so low that the pension benefits for those currently covered (i.e., corporate employees in the formal sector) would be significantly reduced as a result of the expansion, labor unions and middle-class NGOs turned their back on it. The average income of urban self-employed and informal sector workers (i.e., workers at small businesses with four or less and day and casual workers) was about half the average income of formal sector corporate employees. Accordingly, coverage expansion entailed a significant reduction of average income for contributors, which would in turn reduce pension benefits. (Yang 2017, 141)

Under the presidency of Roh Moo-hyun (2003–2008), opposition to costly reforms in social policy intensified, particularly among the unions of large companies. However, not just opposition by privileged workers but also the routine avoidance of contributions by the self-employed was part of the middle-class resistance against a universalist pension insurance with redistributive effects.

When it comes to the role of international organisations, South Korea is a special case, as it is one of three countries (the other two are Slovenia and Venezuela) that have resisted the World Bank’s push since 1994 (and intensified after the Asian crisis in 1997) to privatise old-age provision (Yang 2004; Orenstein 2008, 45). Orenstein (2008, 155–56) ascribes this to the presidency of Kim Dae-jung (1998–2003), who disagreed with the plans of his predecessor Kim Young-sam. Nevertheless, there is some evidence of the evasion mechanism being at work here: A privatisation plan was drawn up by the government in 1995, which moved in the direction of the World Bank’s ideas in order to obtain a loan worth billions. But then these plans were modified more and more within the government and finally dropped. This refusal to privatise was successful, even though the literature usually classifies South Korea as a “productivist welfare state” (Holliday 2000; Rudra 2008; see also London 2018). The Asian crisis, with its high unemployment and company collapses, even among the large ones, the chaebols, led to a policy shift that departed substantially from the privatisation blueprint to continue a unified insurance system. The government was able to pit the recommendations of the ILO (Hagemejer and Schmitt 2012) against the World Bank programmes and thus demonstrate compliance (thus creating an embellishment effect), albeit compliance with the ideas of a competing international organisation (Kim 2008).

3.2 Vietnam

3.2.1 Overview of the Pension System

The Vietnamese social insurance system was established in 1961 and included a non-contributory defined-benefit pension scheme for state sector employees (Long 2012, 205–6). In the course of doi moi—the Vietnamese process towards economic liberalisation which had started in 1986—the Vietnamese welfare regime was modified to a considerable degree (see London 2018). The existing pension system was converted into a pay-as-you-go defined-benefit scheme in 1995 (Long 2012, 206) and became compulsory for all formal workers, thereby capturing employees that have a contract for at least three months and that work in firms with at least ten employees (ILO 1996b; Long 2012, 206). Vietnam Social Security (VSS) was established as a state agency subordinate to the Vietnamese government and tasked with managing the social insurance fund (VSS 2018). The current pension system comprises persons who are insured under the pension system both before and after the 1995 reform (Long 2012, 206).

The Social Insurance Law of 2006 specified the foundations of both compulsory and voluntary social insurance, including old-age benefits (Socialist Republic of Vietnam 2006). It stipulated that social insurance was open to all employees who have a contract of at least three months’ duration, irrespective of the size of their workplace. The voluntary pension scheme was designed as a separate system for self-employed workers and people residing in the rural area, particularly farmers, and was implemented in 2008 (Nguyen and Chen 2017, 237). The Social Insurance Law of 2014 laid down important revisions that sought to expand the number of insured persons (Castel and Pick 2018, 18). This became especially clear as from 2018 on workers with a contract of one month duration were included in the social insurance scheme as well. Moreover, the pension system was also opened to specific groups of part-time workers and some foreign citizens legally employed in Vietnam (International Social Security Association 2019, 283).

The current contribution rates are 8% of gross monthly earnings for employees and 14% for employers; for the self-employed, the contribution rate is 22% of their declared earnings. The general retirement age is 60 for men and 55 for women. To receive a monthly old-age pension, people need to have paid contributions for at least 20 years (International Social Security Association 2019, 283); otherwise, they receive a lump sum. Early retirement with reduced pension payments is also possible (Socialist Republic of Vietnam 2014). Available data from the ILO suggests that both legal and effective coverage rates have remained at relatively low levels: While legal coverage for old age as a percentage of the working-age population is estimated at 33.1%, effective coverage is estimated at 20.6% (ILO 2017).

3.2.2 Mechanisms

The expansion of the pension system to include employees working in the private sector can be explained by the gaining acceptance spiral mechanism in the context of economic liberalisation: In the 1990s, the government decided that the pension system should no longer be limited to state sector employees but expanded to private-sector employees as well, who played a key role in the country’s economic liberalisation process (Goodkind et al. 1999, 147). Apart from the political rationale, financial and demographic considerations reflecting system requirements also played a role, which can be explained by the crisis management by going further mechanism. In fact, extending the pension system to new groups was perceived as crucial to maintaining the system’s financial sustainability in a society that was both ageing and shrinking (Goodkind et al. 1999, 147–48).

In addition to the social policy objectives associated with the expansion of the pension scheme, the double benefit mechanism can explain why expanding the system is considered an attractive political option (see also Kuhlmann and Nullmeier 2021). Here, it is important to note that the pension fund cannot be considered only as a financing source for old-age pensions. For the Vietnamese government, it also presents an opportunity to “obtain funds at the lowest cost”, which thus “give[s] fiscal space and may sustain growth” (World Bank 2012, 12). In fact, in 1998, three years after the pension reform, the financial management of the fund had been specified so as to allow the VSS to invest in government bonds and bonds issued by commercial state-owned banks as well as in larger state projects and enterprises, given the government’s approval (ILO 2000d, 5).

It can be assumed that using the money from the pension fund for other purposes was especially attractive in the first years of the fund, when not many pensions were paid out due to the minimum insurance period of 20 years, which led to quite large reserves in the first years. According to a report from the World Bank, in 2010, “VSS had accumulated US$5.78 billion of reserves, equivalent to about 5 percent of GDP” (World Bank 2012, 7). Today, the social insurance fund is the biggest welfare fund in Vietnam. Most of the money from the fund is invested in government bonds, bonds by commercial banks owned by the state (VSS 2018) and in bank deposits. In 2018, 90% of the social insurance funds were invested in government bonds (Viêt Nam News 2018). While this is considered a safe investment strategy, it provides only limited returns (World Bank 2012, 11; VSS 2018). A solution to this problem that has been discussed is diversifying the investment strategy by investing in “national key industries” as well as in “infrastructure development, electricity, transportation, urban development” (VSS 2018). This would imply that the pension fund would not only finance the government debt but also actively promote the country’s economic development. Another variant of the double benefit mechanism can be identified when it comes to early retirement: Especially in the 1990s, a popular option for state enterprises was to phase out older workers (ILO 2000d, 7). Thus, the pension fund can be seen as performing important stabilising functions in the area of labour-market policy to solve employment problems.

Finally, when it comes to the role of transnational factors in shaping the Vietnamese pension system, evidence from the ILO suggests that the evasion mechanism and the resulting embellishment effect plays a certain role (Kuhlmann and Nullmeier 2021). The Social Protection Development and Training Project from 1995 to 1999 provides an early example of how the ILO was involved in Vietnamese social policy-making and how national reform trajectories differed from ILO recommendations. The aim of the project was to create a sustainable social security system for all Vietnamese employees (ILO 2000b, 1). In the course of the project, the ILO made a number of recommendations on future reforms of the Vietnamese pension system, including harmonising the pension age for men and women at the level of that of men, abolishing early retirement subsidies and lump-sum benefits (unless conditions for receiving a regular pension are not fulfilled), granting pension credits for maternity and childcare, maintaining adequate income for survivors, and keeping contribution rates low. Moreover, the ILO welcomed expanding the system to include new groups (including voluntary insurance) (ILO 2000d, 36–39).

The project’s executive summary to the government, however, suggests that Vietnamese policy actors did not follow ILO recommendations. The ILO even criticised that “not only have […] recommendations not been adopted, except for maternity pension credits, but the opposite policies have been followed” (ILO 2000d, v). A case in point is that early retirement was further facilitated, which—together with a low female pension age—helped to keep people outside of the labour market (ILO 2000d, 30). ILO recommendations for the following years included the drafting of a social security act and pushing voluntary pension schemes in order to expand old-age provision to include more groups (ILO 2000b, 35).

Although the Vietnamese pension reforms differed from the reform path suggested by the ILO, both the ILO and Vietnamese policy actors expressed only minor criticism and stated that the project had been highly successful overall (ILO 2000c, 14–24). Several actors, among them the Ministry of Labour, Invalids and Social Affairs (MOLISA), stated that future assistance was necessary to improve the Vietnamese pension system (ILO 2000c, 15).

Some pension reforms that were adopted after this project are clearly in line with ILO recommendations, such as the first comprehensive Social Insurance Law in 2006, which strengthened the role of social insurance within the country. Moreover, Vietnam established a voluntary pension scheme. On the other hand, many elements of the pension system that had been criticised by the ILO are still in place, such as early retirement and different pension ages for men and women. Lump sums are only available to retired persons with fewer than 20 contribution years, yet the majority of people currently receiving them are people who leave employment with social insurance and “cash out” their contributions (Castel and Pick 2018, 16). In fact, many employees do not necessarily consider their social insurance contributions as retirement provision but rather as unemployment protection. Moreover, trust levels towards the pension system are not very high to begin with, which is why opposition among formal workers to abolish lump-sum benefits is very strong and the ILO recommendation to abolish lump-sum benefits is currently considered politically unfeasible (Castel and Pick 2018, 17).

3.3 Sri Lanka

3.3.1 Overview of the Pension System

The pension system in Sri Lanka has a dualised structure (Karunarathne and Goswami 2002, 95), mainly consisting of pay-as-you-go schemes (defined benefits) for public-sector workers and mandatory saving schemes (defined contributions) for employees in the private sector. A pension scheme for the public sector has been in place since 1901 (Public Service Pension Scheme, PSPS), and a provident fund for public-service workers who are not eligible for the pension scheme was established in 1942 (Public Service Provident Fund, PSPF). The first provident fund for workers in the private sector was established in 1958 (Employees’ Provident Fund, EPF) as a mandatory defined-contribution scheme run by the state. It is based on employers’ and employees’ contributions at fixed rates, which are then accumulated in individual accounts.

While the EPF constitutes the largest social security scheme in Sri Lanka (Employees’ Provident Fund 2021), additional funds and savings schemes exist as well. In 1981, the Employees’ Trust Fund (ETF) was established for employees in the public sector, university employees, and private-sector employees (Karunarathne and Goswami 2002). Moreover, the self-employed, as well as migrant workers, can seek coverage on a voluntary basis (International Social Security Association 2019, 244). The ETF is funded by employers’ contributions only (with a contribution rate of 3%), with the initial aim to “promote employee ownership of equities” (Karunarathne and Goswami 2002, 98). Moreover, three additional voluntary schemes have been created for informal-sector workers. They are the Farmer’s Pension and Social Security Benefit Scheme (1987), the Fisherman’s Pension and Social Security Benefit Scheme (1990), and the Pension and Social Security Benefit Scheme for Self-Employed Persons (1996) (Rannan-Eliya and Eriyagama 2003b, 3).

The current EPF contribution rates are 8% for the employee (with the possibility of additional contributions) and 12% for the employer (International Social Security Association 2019, 244). Employees are entitled to their benefits—which are paid out as lump sums—when they reach the statutory retirement age of 55 (men) and 50 (women), respectively. In certain cases, it is possible to withdraw benefits earlier; these include marriage (for women only), permanent disability, and migration (Employees’ Provident Fund 2017, 220). Moreover, since 2015 members are permitted to withdraw money for housing loans or medical treatment. Overall, coverage rates are relatively low in Sri Lanka, with legal coverage at 32.9% of the working-age population and effective coverage rates at 18.9% (ILO 2017).

3.3.2 Mechanisms

Why do policy-makers in Sri Lanka stick to the provident fund? Evidence from the secondary literature suggests that the double benefit mechanism is very important when it comes to understanding this ratchet effect (see also Kuhlmann and Nullmeier 2021). The EPF is Sri Lanka’s largest savings stock. Whereas, as in many other former British colonies, it was initially established as a “second-best alternative to the national pension system already existing in the United Kingdom” (Rannan-Eliya and Eriyagama 2003a, vi), it also plays an important role in financing the government’s structural fiscal deficit (Karunarathne and Goswami 2002). In 2016, 93.1% of the EPF’s investments were in government securities (Employees’ Provident Fund 2017). Moreover, money from the EPF not only plays a role in old-age provision but also serves other (social) policy purposes, a fact that the fund explicitly acknowledges. A case in point is in particular pre-mature withdrawals for a housing loan. According to the EPF’s self-description:

The EPF is not only a helping hand or a shoulder to lean upon in the winter of life but a great partner throughout, for it will provide you with the option of obtaining a housing loan […] Thus, the EPF will help you realize your dream of a home before retirement. (Employees’ Provident Fund 2021)

When it comes to transnational factors, there is also evidence of an embellishment effect, which can be explained by the evasion mechanism (Kuhlmann and Nullmeier 2021). In fact, the ILO has advocated for a contribution-based social insurance pension scheme in Sri Lanka at least since the 1980s (ILO 1980), as provident funds generally had some “intrinsic deficiencies” (ILO 1980, 16) when compared to social insurance schemes as promoted by the ILO. Most importantly, lump sums instead of periodic payments provided only an “inadequate” level of social protection, and there was the “impossibility of maintaining the real value of the Provident Fund contributions” (ILO 1980, 16). The ILO therefore recommended transforming the EPF into a contribution-based social insurance scheme with periodic payments. In 1982, a project on introducing a social insurance scheme was launched (ILO 1984). While the final report outlined several options for introducing such a pension scheme, a gradual conversion of the EPF into a pension scheme, thereby transferring all assets and liabilities, was considered to be the most fruitful option (ILO 1984, 45–48, 1991, 6). In 1989, the Sri Lankan government decided to replace the EPF with the Employees Pension Scheme and create the National Social Security System (ILO 1991, 1). Yet, the ILO’s “Report on the Conversion of Provident Fund to Social Insurance System Prepared for the Government of Sri Lanka” (ILO 1991) indicates that the reform had been “delayed” (ILO 1991, 1) due to opposition by two domestic actors: While the Ministry of Labour and Vocational Training was concerned about the structure of the new benefits, the Ministry of Finance worried that the money collected through the EPF could no longer be used that easily for government financing, which again points to the role of the double benefit mechanism in the development of the pension scheme (ILO 1991, 1). Moreover, it is important to note that other domestic actors, such as trade unions and employer organisations, adopted a rather sceptical stance as well. The ILO was asked to “review the situation and suggest a basis on which implementation of the cabinet decision can proceed” (ILO 1991, 1), which it laid out in its 1991 report. However, until today, the EPF continues to be the main instrument for old-age provision in Sri Lanka, and a pension scheme in line with ILO recommendations has yet to be implemented. Moreover, the ILO no longer recommends converting the EPF into a pension scheme in official documents but only to convert lump-sum benefits into periodic payments (ILO 2008, viii).

3.4 Malaysia

3.4.1 Overview of the Pension System

While in Sri Lanka, the Employees’ Provident Fund (EPF) was introduced after independence, in Malaysia, it was established by the British colonial government in 1951 (thus making it one of the oldest provident funds in the world; see Dixon 1989). The EPF is a mandatory defined-contribution scheme for workers in the private sector as well as for some workers in the public sector that are not covered by a public-sector scheme. Generally speaking, the Malaysian pension system is highly fragmented. Apart from the EPF, there are a number of different old-age provision systems, such as the Civil Service Pension Scheme, stemming from colonial times (Darmaraj and Narayanan 2019), and the Armed Forces Fund, established in 1972 (Asher 2012). The EPF can nevertheless be considered the central institution within the Malaysia pension system. In 2016, it was the 15th largest pension fund in the world (Price et al. 2018, 7). Over the years, the EPF was continuously expanded to cover all workers in the private sector, irrespective of the size of their workplace (Asher 2012, 106), and it was also opened to self-employed workers in 1977 (Employees Provident Fund 2019, 18). In recent years, one aim of the Malaysian government has been to extend the EPF to all Malaysian citizens, including housewives and other groups who do not belong to the group of formal workers. People who decide to contribute voluntarily to the EPF can receive an additional government subsidy to supplement their voluntary contributions (i-Saraan) (Employees Provident Fund 2019; Price et al. 2018).

Each individual member has two accounts comprising 70% (account 1) and 30% (account 2) of the contributions, respectively. At the age of 55, members can withdraw the accumulated money, and both accounts are merged. From the age of 50 on, members can withdraw the money accumulated in the second account. Earlier withdrawal from this account is moreover possible for purposes related to housing, education, health needs, or the Islamic pilgrimage to Mecca (hajj). For approved investments, it is also possible to withdraw money earlier from the first account (Ramesh 2005, 193–94; Price et al. 2018, 15). It is not obligatory to retire at the age of 55 and withdraw the money from the EPF.

The monthly contribution rate depends on employees’ age and income. In 2018, the general employees’ contribution rate was 11%. The employers’ contribution rate varied between 12% and 13% depending on the employees’ income. For employees older than 60, the regular contribution rate is halved (Price et al. 2018, 13). Moreover, additional voluntary contributions to the first account are possible. According to current data, roughly half of the Malaysian labour force contributes to the EPF, indicating more or less universal coverage of formal private-sector workers, while coverage gaps remain when it comes to self-employed workers and workers in the informal sector (Price et al. 2018, 42; see also ILO 2017).

3.4.2 Mechanisms

Despite international pressures to convert the EPF into a pay-as-you-go pension scheme, the EPF has remained in place, and coverage has been continuously expanded to the point of almost achieving universal coverage in the late 1980s—at least when it comes to formal workers (Ramesh 2005, 193). One reason for this ratchet effect of holding on to the EPF is that its functions go beyond old-age provision. In fact, the double benefit mechanism plays an important role (Kuhlmann and Nullmeier 2022).

Like in Sri Lanka, the provident fund was created as an alternative to social insurance schemes, the establishment of which was considered too ambitious at the time. As some authors have argued, the introduction of national provident funds during British colonial rule reflects an acknowledgement by British colonial administrations that at least some social policy measures had to be introduced to address the social situation in the colonies, particularly with regard to the “increasingly urbanized and industrialized workforce” (Kaseke et al. 2011, 145; McKinnon et al. 1997; see also Chap. 6). Taking this into consideration, it is plausible to at least explain the introduction of the provident fund in Malaysia by means of the gaining acceptance spiral mechanism, although the immediate British influence of course vanished after independence.

Policy-makers in Malaysia quickly realised the fund’s great economic potential (Ramesh 2005, 192), using it as an “essentially bottomless source of long-term investment capital with which to finance state-led industrial and infrastructural development” (McKinnon 1996, 50). In fact, EPF balances have grown considerably throughout the years and accounted for around 60% of GDP in 2016 (Price et al. 2018, 28). In the early days of the EPF, all funds had to be invested in government securities or bank deposits, and until the 1980s, 80–90% of the money from the EPF was in Malaysian government securities. Since the 1990s, the EPF’s investment portfolio has become more diversified, now also including privatisation programmes and joint-venture projects, equities, and debentures of public companies, as well as investments in the housing market (Ramesh 2005, 203). Importantly, this not only fuelled the Malaysian capital market but also enabled the government to finance big infrastructure projects (Price et al. 2018, 25) that can be considered important for the country’s economic development. Moreover, the EPF now also invests in foreign markets, which resulted in a share of foreign investments of 30% in 2017 (compared to only 1% in 2006). These overseas investments followed a solid investment strategy but they also reflect the fact that the EPF had become too large for the domestic financial market (Price et al. 2018, 28; see also Asher and Bali 2015). In 2018, the EPF’s rate of return on investment was 6.57% (Employees Provident Fund 2019, 16). The EPF explicitly acknowledges not only its function for old-age provision but also its key role in promoting the country’s economic development:

In its initial years, the EPF was able to park all of its assets locally in Malaysian Government Securities, Loans and Bonds, Equities, Money Market Instruments and Property. In so doing, it served to help finance a large number of major government as well as private sector projects that have contributed towards shaping the nation’s development. More recently, it has had to look at suitable investment options outside of the country, as the local capital and money markets have not kept pace with the growth of EPF’s assets. (Employees Provident Fund 2013, 3)

Some authors suggest that the EPF’s role in state financing is even more central than the EPF’s role in old-age provision: “All in all, the achievements of the EPF and the CPF [the provident fund in Singapore, JK/FN] are considerable, despite the fact that they do little in the areas of income protection and health care. They played a significant role in financing economic development projects in the early stages of industrialization” (Ramesh 2005, 207). In this regard, the gaining acceptance spiral mechanism can also play a role in explaining the expansion of the EPF, as the country’s economic success—to which the EPF contributes decisively—is crucial to the legitimacy of the Malaysian political system (McKinnon 1996).

For decades, the ILO has advised Malaysia to restructure the provident fund and to establish a pension scheme based on social insurance principles (McKinnon 1996, 48). A key reason was that the ILO perceived lump-sum benefits as inadequate and was concerned that many “will fritter their money away […] Clearly if it is accepted as desirable to compel workers to save for their old age, it is equally desirable to ensure that, when the time for retiring comes, the worker is protected against the rapid dissipation of his resources through misfortune, through his own folly or through the wiles of others” (ILO 1960, 62). There have been several projects—initiated by the government of Malaysia and involving Malaysian stakeholders—on how the EPF could be converted into a pension scheme (ILO 1996a, 2000a). An ILO report (ILO 2000a, 93–97) summarises that Malaysian stakeholders engaged with the suggestions for converting the EPF into a pension scheme; however, a summary of the discussion held at a national seminar also shows that Malaysian stakeholders stress the need for further studies and discussions with regard to key questions (such as retirement age or such a scheme’s relation to the civil service pension scheme), all of which would require more time. In practice, this has resulted in a suspension of the suggested reform, which is in line with the evasion mechanism. In fact, especially given the EPF’s economic importance, it is rather unlikely that policy actors will adopt recommendations by international organisations to convert their pension system:

EPF managers believe pragmatically that they have the best option. The duality of the developmental role of the EPF is at the heart of this issue, and it is this more rounded appreciation of that institution’s impact that suggests that the Malaysian government’s rejection of the ILO’s social security-focused advice will be repeated in the case of the World Bank’s equally unbalanced concentration on issues relating to investment efficiency. (McKinnon 1996, 48)

4 Summary and Conclusion

The inherent problems of contribution-based pension systems are frequently discussed in the social policy literature on Eastern, South-Eastern, and Southern Asian countries. At first glance, it might therefore seem quite noteworthy that, these well-known problems notwithstanding, many countries do not only stick to their contribution-based pension system but even undertake political efforts to expand them. Focusing on the cases of South Korea, Vietnam, Sri Lanka, and Malaysia, our analysis identified similar causal effects and causal mechanisms. Taken together, they provide a modular explanation for this general observation of a maintenance and an expansion of contribution-based pension systems at the policy level, despite limitations when it comes to expanding their effective coverage. Notably, for different variants of contribution-based pension systems and for countries with different starting points as well as different political, societal, and economic backgrounds, the set of causal mechanisms that we identified provides a useful and productive instrument for a modular explanation of the trajectories of these pension systems (Kuhlmann and Nullmeier 2021). This chapter identified six key causal mechanisms that can explain these dynamics, although to a different extent. When it comes to the general political mechanisms, the analysis was able to identify the evasion mechanism in all four cases. By contrast, the outcompeting mechanism—which does not capture reform dynamics in autocratic regimes—could only be identified in the South Korean case, while the gaining acceptance spiral mechanism played a role in the cases of South Korea, Vietnam, and, at least to a certain extent, Malaysia. With regard to the institution-specific mechanisms, the double benefit mechanism figured prominently in all four cases. While the analysis showed that all countries—albeit to different degrees—use the money from the pension fund for national investments, the Vietnamese case also presents an example in which money from the pension system is used to tackle problems in the labour market. In contrast, evidence for the crisis management by going further mechanism could be found solely in the Vietnamese case and the alarmed middle classes mechanism only in the South Korean case. The fact that we find more cases for some mechanisms than for others means neither that these mechanisms are theoretically superior to the other mechanisms nor that these mechanisms are generally more likely to explain social policy developments. The six mechanisms are a first step in demonstrating how comparative studies can work with causal mechanisms. Examining individual cases allows us to develop a set of complex mechanisms that may then be employed to study other cases. However, whether the mechanisms derived from such case studies can explain additional cases is a matter that must be decided in empirical research.