Introduction
Latin American and South European countries share a common legacy over the past 40 years not only in terms of democratization but also of Bismarckian welfare states (Ferrera, Reference Ferrera1996; Huber and Stephens, Reference Huber and Stephens2012; Cruz-Martínez et al., Reference Cruz-Martínez, Vargas-Faulbaum and Velasco2024). In the field of pensions, by the mid-20th century, countries in both regions had developed public pension systems with relatively high levels of coverage and based on the Pay-As-You-Go (PAYG) principle where current contributions from workers are used to pay current pensions.
Faced with similar pressures of population ageing and government fiscal imbalances to which the pension system contributed significantly, they embarked on reforms, yet their paths have differed (Mesa Lago and Muller, Reference Mesa Lago and Muller2002; Chulia, Reference Chulia, Immergut, Anderson and Schulze2009; Carrera et al., Reference Carrera, Angelaki and Carolo2010; Huber and Stephens, Reference Huber and Stephens2012; Natali and Stamati, Reference Natali and Stamati2014; Barrientos, Reference Barrientos2018; Petmesidou, Reference Petmesidou and Greve2018; Cruz-Martínez et al., Reference Cruz-Martínez, Vargas-Faulbaum and Velasco2024). Latin American countries opted for the full or partial privatization of their public pension systems, yet subsequent reforms have challenged the public–private mix (Kay, Reference Kay1999; Madrid, Reference Madrid2003; Rofman and Oliveri, Reference Rofman and Oliveri2012; Mesa Lago, Reference Mesa Lago2014; Baba, Reference Baba2015; Busquets and Pose, Reference Busquets and Pose2016; Carrera and Angelaki, Reference Carrera and Angelaki2021, Reference Carrera and Angelaki2024). Meanwhile, countries in Southern Europe opted for a less radical path, entailing different degrees of reform of their public pillars and the introduction of supplementary private ones (Chulia, Reference Chulia, Immergut, Anderson and Schulze2009; Natali and Stamati, Reference Natali and Stamati2014; Petmesidou and Guillen, Reference Petmesidou, Guillen, Petmesidou and Guillen2014).
In understanding reform variation, we argue that by focusing on the role of political institutions and policy legacies, it is possible to identify reform mechanisms. Our analysis focuses on Argentina, Chile, and Uruguay – in Latin America – and Spain, Italy, and Greece – in Southern Europe – and the reforms implemented since 1990. While cross-regional comparisons raise some questions regarding the comparability of cases and generalization of findings, we agree as argued by Read (Reference Read, Simmons and Smith2021) that it is possible to establish a rationale for comparison, allowing us to gain insights on a phenomenon studied in varying political circumstances, broaden our understanding of the regimes and help conceptual development.
On a broader level, our analysis aims to contribute to the literature on cross-regional comparisons of social policy reforms (Pierson, Reference Pierson1994, Reference Pierson1996, Reference Pierson2001; Hacker, Reference Hacker2004; Streeck and Thelen, Reference Streeck, Thelen, Streeck and Thelen2005; Béland, Reference Béland2007; Huber and Stephens, Reference Huber and Stephens2012; Béland and Powell, Reference Béland and Powell2016; Ortiz et al., Reference Ortiz, Durán-Valverde, Urban and Wodsak2018; Anderson and Weaver, Reference Anderson and Weaver2025) as well as the literature on pension reforms (Chulia, Reference Chulia, Immergut, Anderson and Schulze2009; Weaver, Reference Weaver2010; Datz and Dancsi, Reference Datz and Dancsi2013; Angelaki and Carrera, Reference Angelaki and Carrera2015).
The role of political institutions and policy legacies in reforming pensions
The study of pension reforms in the two regions has resulted in the emergence of an important body of literature shedding light on the factors that explain reform variation, while embracing methodological plurality (Queisser, Reference Queisser1998; Bonoli, Reference Bonoli2001; Mesa Lago and Muller, Reference Mesa Lago and Muller2002; Madrid, Reference Madrid2003; Bonoli and Shinkawa, Reference Bonoli and Shinkawa2005; Natali, Reference Natali2008; Orenstein, Reference Orenstein2008, Reference Orenstein2013; Chulia, Reference Chulia, Immergut, Anderson and Schulze2009; Ebbinghaus, Reference Ebbinghaus2011, Reference Ebbinghaus2015; Kohli and Arza, Reference Kohli, Arza, Binstock and George2011; Busquets and Pose, Reference Busquets and Pose2016; Carrera and Angelaki, Reference Carrera and Angelaki2022, Reference Carrera and Angelaki2024; Barrientos, Reference Barrientos2024).
In discussing the role of institutions Pierson (Reference Pierson1996) argued that institutional settings that provide power concentration to the executive and few veto points allow for the introduction of significant reforms. Building upon this argument, Kay (Reference Kay1999) in his study on the politics of social security reforms in Argentina, Brazil, and Uruguay stresses the role of political institutions in providing actors the opportunity to act as veto players in the reform process. Similarly, Brooks (Reference Brooks2009) and Madrid (Reference Madrid2003) stress the role of institutions, namely the president’s degree of control over the legislature and the strength of organized labour in explaining reform outcomes. Readapting the institutional approach, other scholars note the adoption of pension reforms, despite the existence of a significant number of veto players, by reference to political competition leading policy makers to adopt a more consensual approach or to withdraw their proposals (Bonoli, Reference Bonoli2001; Immergut et al., Reference Immergut, Anderson and Schulze2009).
Scholars have also studied policy legacies, thus placing emphasis on how (existing) policies impact politics and policy development (Pierson, Reference Pierson1996). Policy legacies can refer to the legal structure set out by a given policy and the types of actors it generates (Ewig and Kay, Reference Ewig and Kay2011). Policies can impact through the transformation of state capacities, the goals and capabilities of different groups, and the lock-in effects created (Béland, Reference Béland2010). For example, the Toledo Pact in Spain transformed pension policy into a multiple veto player system due to the participation of different societal actors in the permanent commission set out by the Pact, thus explaining the less radical nature of reforms (Chulia, Reference Chulia, Immergut, Anderson and Schulze2009).
Scholars like Weaver (Reference Weaver2010) and Jacobs and Weaver (Reference Jacobs and Weaver2015) have also argued that policies may create negative feedback that can ultimately undermine them. Studying pension re-reform in Argentina, Arza (Reference Arza2012a) shows how the low levels of support for the private system facilitated its elimination, while Borzutzky (Reference Borzutzky2019) in her study on Chile stressed the role of the No+ AFP movement in the elimination of the private pillar. Yet, as argued by Kay (Reference Kay1999), interest group strength, as measured in terms of resources and union density, cannot by itself provide an explanation for reform processes, thus suggesting the need to examine further the role of political institutions in providing interest groups with opportunities to veto policies.
In terms of ideology, while some scholars have attributed the expansion that took place in Latin America in the early 2000 to left-wing governments (Huber and Stephens, Reference Huber and Stephens2012), others like Niedzwiecki and Pribble (Reference Niedzwiecki and Pribble2017) have showed that social policy reforms in Argentina and Chile in mid-2010 by right-wing governments have not resulted (as would have been expected) to deep spending cuts, thus stressing the role of institutions and policy legacies. The same authors argue, though, that in the context of economic and political crisis right-wing governments (with weak opposition) have more leeway in adopting unpopular reforms. Overall, while we do not disregard the role of ideology – and thus do not consider the reform proposals as politically neutral – we believe that the reform outcome is largely shaped by the interplay of the domestic institutional setting with policy legacies.
Regarding economic crises, the divergent reforms paths of Italy and Greece during the 1990s stress the fact that while they make the need for reform more evident, the outcome is still shaped by the interplay between domestic actors and their ability to form (or fail to do so) wider coalitions (Featherstone et al., Reference Featherstone, Kazamias and Papadimitriou2001; Natali, Reference Natali2004; Bonoli and Palier, Reference Bonoli and Palier2007). Similarly, in Latin America, Brooks (Reference Brooks2009) and Angelaki and Carrera (Reference Angelaki and Carrera2015) argue that while crises led governments to prioritize short-term funding needs over long-term macroeconomic goals and adequacy issues, government strength played a key role in achieving reforms.
Economic crises have also been discussed in relation to the role of international organizations and the pressure they exercise in implementing structural adjustment policies, including pension reforms (Brooks, Reference Brooks2009; Angelaki and Carrera, Reference Angelaki and Carrera2015). For example, scholars have studied the World Bank-led campaign for pension privatization during the 1990s with a focus on the diffusion of the privatization paradigm (Madrid, Reference Madrid2003; Orenstein, Reference Orenstein2008, Reference Orenstein2011, Reference Orenstein2013; Brooks, Reference Brooks2009; Beland and Orenstein, Reference Beland and Orenstein2013).
The 2009 global economic crisis revived the interest in the role of international organizations in pension reform. However, these reforms caught international organizations in the middle of significant ideational changes. Indeed, analyses have shown that since the early 2000s there was an ideational shift within the World Bank and the IMF moving away from promoting pension privatization as a one size fits all solution in favour of designs that are sustainable, extend coverage, and provide a social protection floor (Heneghan and Orenstein, Reference Heneghan and Orenstein2019). Yet, in both Eastern Europe and Latin America analyses have shown that unlike in the 1990s international organizations were not actively involved in pension re-reforms discussions in specific countries and that domestic factors such as the power of government and the role of societal veto actors were more determinant in shaping reform outcomes (Datz and Dancsi, Reference Datz and Dancsi2013; Baba, Reference Baba2015; Wilson Sokhey, Reference Wilson Sohkey2017).
Greece represents the only exception regarding the role of international organizations in pension reforms. There, in the aftermath of the 2009 crisis, the troika of the ECB, IMF, and the European Commission forced the country to implement economic reforms as part of their bailout programme. Nonetheless, the fiscal constraints and those emanating from the system structure (and the transition costs associated with moving from a PAYG system to a funded one) have impacted the reform content and have led to measures aimed at the significant reduction of the public PAYG system and not directed at adopting privatization. Furthermore, Greece was already on a reform path since the early 1990s, even if with limited success.
Overall, while the literature on the role of international organizations in the diffusion of pension privatization focuses on the ideational level (Madrid, Reference Madrid2003; Weyland, Reference Weyland2005; Orenstein, Reference Orenstein2011, Reference Orenstein2013; Beland and Orenstein, Reference Beland and Orenstein2013) scholars agree on the prevalent role of domestic actors and structures in explaining the varying degrees of privatization (Brooks, Reference Brooks2009; Ebbinghaus, Reference Ebbinghaus2015). While we do not disregard the role of international organizations, our analysis focuses on the interplay of the domestic institutional setting and policy legacies to explain reform variation, where we argue that the locus of activity that shapes pension reform outcomes takes place.
This analysis of the extant literature shows that pension reforms are complex phenomena that result from the interaction of various factors at the domestic level, where policy legacies and institutions stand out. We thus expect these two macro variables to play a significant role in explaining reform outcomes. Our analysis focuses on pension reforms, and as such, we seek to understand the pathways leading to different outcomes. Based on Hall (Reference Hall1993), we distinguish between structural (or paradigmatic) and parametric reforms. Structural reforms entail a shift in the goals of the system through the introduction or elimination of a dominant pillar (i.e. the pillar that provides most of the retirement income) or a change in the funding principle (shift to Notional Defined Contribution/NDC), thus affecting the degree of solidarity/individualization. When it comes to parametric reforms, we distinguish between major and minor ones. The former expands the role of the current dominant pillar by increasing contributions or by adding a complementary pillar (e.g. a non-contributory one), while the latter is generally limited to changes in retirement age or the number of contribution years to qualify for a pension.
The key institutional factors that we consider are the level of support for the executive in the legislature and the ability of the executive to make decisions with his cabinet, which could also include the use of emergency legislative powers. We consider the support for the government in the lower chamber of Congress in the case of presidential systems in Latin America and in the main chamber of Parliament in the case of Southern Europe. Therefore, a weak institutional setting is given by situations in which the executive does not have support in congress or parliament and/or does not have recourse to emergency powers, and thus needs to negotiate with the opposition to pass reforms. By contrast, a strong setting is where executives face a less fragmented legislature or can make use of emergency powers.
Regarding legacies, and building upon Niedzwiecki and Pribble (Reference Niedzwiecki and Pribble2017), we consider the role of societal actors with an interest in the pension system, such as trade unions, grassroots movements, and the pension industry. We consider their relevance by using a mix of quantitative and qualitative sources. In general, a strong legacy is given in situations where these actors participate in the reform process or can block the process.
We expect paradigmatic change to take place when a strong institutional setting combines with a weak legacy, where actors are not able to (meaningfully) influence the process. When a strong institutional setting combines with a strong legacy (actors that can block the reforms through various strategies like protests), the outcome is minor parametric reform. Conversely, when a weak institutional setting combines with a strong legacy, we expect major parametric reforms as a weaker government will be more open to negotiate reforms. When the institutional setting is weak and combined with a weak legacy, we expect no change. This would be represented by the case of a government with no majority able to pass measures and weak societal actors that are also unable to propose or influence any reform process. Therefore, it is logical to expect that reforms will not be initiated in this case. Table 1 illustrates our expectations for reform outcomes.
Table 1. Theoretical expectations for pension reforms
We test our expectations by examining the cases of Argentina, Uruguay, Chile, Italy, Spain, and Greece. Our research design follows a cross-regional comparative analysis framework by selecting cases that are broadly similar but vary in the outcome of interest, that is a most similar system (MSS) research design (Przeworski and Teune, Reference Przeworski and Teune1970: 34). While the countries under study do not belong to the same region, we believe that they share some common traits (initial establishment of PAYG systems based on occupational lines; aspects of corporatist labor relations; mismanagement of the schemes; similar fiscal and demographic pressures; dualism; fragmentation; high informality in the labor market, high levels of urbanization and lower participation of women) that renders their comparison possible and allows us to gain a better insight in pension reforms (Minas et al., Reference Minas, Jacobson, Antoniou and McMullan2014; Giuliani, Reference Giuliani and Trigilia2022; Barrientos, Reference Barrientos2024). Furthermore, countries in both regions implemented significant reforms –yet of different nature – since the early 1990s.
The three Latin American countries are regarded as pioneers in the region in terms of the development of social protection, establishing well-developed PAYG pension schemes by the mid-20th century (Mesa-Lago, Reference Mesa-Lago1978) and grouped under the same model in various typologies (cf. Cruz-Martínez et al., Reference Cruz-Martínez, Vargas-Faulbaum and Velasco2024). While Costa Rica, Brazil, and Cuba also developed public systems with high coverage levels, their larger rural populations would render their inclusion more complex and could potentially lead to less generalizable findings.
Italy, Spain, and Greece, are classified under the south-European corporatist welfare regime (cf. Leibfried, Reference Leibfried, Ferge and Kolberg1992; Ferrera, Reference Ferrera1996) characterized by an initially lower level of social expenditure as a percentage of GDP, but with large and generous public PAYG public pension systems. These characteristics make them similar enough to one another as well as closer to the pioneers of Latin America, while they are also different enough from the conservative countries of northern Europe to constitute a separate group (Minas et al., Reference Minas, Jacobson, Antoniou and McMullan2014). Thus, including countries from the conservative welfare model in this analysis, such as Germany or France, (and Brazil or Costa Rica in the Latin American context), would not allow us to control for the differences they would introduce. In turn, this would prevent us from isolating how the institutional setting and policy legacies affect reform outcomes and reaching more generalizable findings.
Pension reforms in Latin America and Southern Europe
Chile
Chile replaced in 1981 (under the Pinochet dictatorship) its public PAYG system with a mandatory pillar of private accounts managed by private administrators (Administradores de Fondos de Pension/AFP) (Orenstein, Reference Orenstein2013). However, since the early 2000s, experts have highlighted adequacy issues for low and middle-income workers due to the lack of contributions because of unemployment and/or working in the informal economy (Arenas de Mesa, Reference Arenas de Mesa2019). (Table A1 in the Supplementary Material provides further details on the operationalization of the reforms and the key independent variables of policy legacies and institutional setting.)
During the first mandate of President Bachelet (2006–2010), a Pensions Commission was established that proposed significant changes. However, the 2008 reform was limited to enhancing adequacy through the introduction of a basic pension and an additional pension for those with either no or low income from the private pillar, both funded out of public revenue, while leaving the private pillar untouched. This major parametric reform is attributed to a weak institutional setting given the slim majority of the government coalition in both chambers, which was not enough to pass a reform given the Constitution requirement of a qualified majority for such reforms (Cifuentes and Williams, Reference Cifuentes and Williams2019). This, combined with the strong policy legacy, as the powerful pension industry succeeded in blocking measures that would negatively affect the system structure (Bril-Mascarenhas and Maillet, Reference Bril-Mascarenhas and Maillet2019).
The No+ AFP movement that emerged in 2015, advocating for the elimination of the private pillar and the creation of a new public one, criticized the reform as not going far enough. However, while numerous reform attempts took place during the second Bachelet administration (2014–2018) and the centre-right Pinera administration (2018–2022), they were blocked (Carrera and Angelaki, Reference Carrera and Angelaki2021). As discussions around pension reform continued, the leftist government of Gabriel Boric presented in November 2022 a bill that proposed wide ranging changes such as changing the role of the AFPs (and potentially eliminating them), an increase in the state pension and a new employer contribution of 6 per cent (de Chile, Reference de Chile2022). Negotiations took almost 2 years, with the most contentious issue being whether the new contribution should go entirely to the private pillar (a position favored by the right and the pensions industry) or use part of it to compensate pensioners receiving very low pensions. After much negotiation, a reform was passed in 2025 setting out how the new employer contribution would be divided (4.5 per cent to the private pillar, 2.5 per cent to a newly created social insurance pillar to compensate poor pensioners and 1.5 per cent to fund a government bond to increase current pensions) (Infobae, 2025). The reform is the most significant since 1981, but still leaves the private pillar largely untouched and indeed reinforces its role, given the extra 4.5 per cent employer contribution going to this pillar. We thus classify it as a major parametric, being the result of a weak institutional setting as the government does not have a majority in either chamber and a strong policy legacy given by the relevance of the No+AFP movement (Castro and Gandara, Reference Castro and Gandara2022).
Argentina
In 1994, Argentina proceeded under the neoliberal Presidency of Carlos Menem (Partido Justicialista/PJ) to a structural reform by replacing its public PAYG, earnings-related system with a mixed one (Arza, Reference Arza2012a; Carrera and Angelaki, Reference Carrera and Angelaki2021). While the government benefited from its majority in both chambers to pass the reform, the latter was adopted following lengthy negotiations with the strong labor movement to which some PJ deputies had links (Madrid, Reference Madrid2003; Brooks, Reference Brooks2009). The strength of the labor movement is explained by legislation that, among others, recognizes the union with the most members as the only one to negotiate wages and allows them to run their own health care providers for their members (Murillo, Reference Murillo2005: 197). The main concessions included the maintenance of the (reformed) public pillar, the introduction of a public tier in the second pillar, and the unions being allowed to run their own pension administrators (AFJP) in the newly created private tier. However, both the labor movement and other political and societal actors remained critical of the new system, a stance that was further strengthened after the 2001/2002 economic crisis (Madrid, Reference Madrid2003; Brooks, Reference Brooks2009; Datz and Dancsi, Reference Datz and Dancsi2013).
While the labor movement remained an important actor, the private tier did not acquire significant strength. Coverage was also low (45 per cent of the economically active population in 1994 contributed to the system, declining to 37 per cent in 2003 only to reach 45 per cent again in 2008) while most workers affiliated to the private system had insufficient savings and contribution years to qualify for a full pension from the public pillar (Rofman and Oliveri, Reference Rofman and Oliveri2012). It therefore comes as no surprise that a 2008 survey showed that 90 per cent of respondents believed that pension provision should be in the hands of the state (Arza, Reference Arza2012a).
The process to dismantle the private pillar culminated in 2008 through the elimination of the private pillar that allowed members to switch back to the public one. To understand this reform outcome, it is important to note the country’s exclusion from international markets due to the 2001 sovereign debt default (Angelaki and Carrera, Reference Angelaki and Carrera2015). However, such an option would not have been possible without a favorable institutional setting: the government enjoyed a comfortable majority in both chambers of Congress, while President Cristina Kirchner maintained a high level of power concentration through using emergency powers delegated by the Congress. In addition, the reform was supported by the strong labor movement which opposed the private system (Carrera and Angelaki, Reference Carrera and Angelaki2021).
Overall, both the 1994 and 2008 structural reforms were adopted thanks to the strong institutional setting. However, and contrary to our initial expectations, this was combined with a strong policy legacy due to the role of the labour strong movement. We argue that the unions, on both occasions, tried to maximize their narrow interests (Etchemendy and Collier, Reference Etchemendy and Collier2007). They supported privatization in 1994 to expand their role in the new private system, but as the private pension system failed to amass significant savings and became unpopular in the early 2000s, they subsequently supported its elimination.
Uruguay
Following several failed reform attempts, Uruguay introduced a mixed system with a private pillar in 1995 under the Sanguinetti government. The reform was negotiated with political and societal actors to ensure their support, but also to avoid attempts to reverse it in the future and thus included some concessions such as the maintenance of the public pillar and the participation in the new private second pillar of workers above a minimum income (Kay, Reference Kay1999; Brooks, Reference Brooks2009). Negotiation was an important element in the reform process as the Uruguayan constitution allows organizations to put proposals to change legislation in a ballot, a feature successfully used in the early 1990s to block reform attempts (Kay, Reference Kay1999; Mesa Lago and Muller, Reference Mesa Lago and Muller2002). Furthermore, the Sanguinetti government had a slim first minority in both the Chambers. In this context, the government justified the limited structural reform as necessary to save the system from collapse (given the increasing pension spending) and to reinforce its solidarity and effectiveness (Brooks, Reference Brooks2009: 244).
Mandatory membership in the private pillar was limited to those earning above a minimum income, thereby maintaining the public pillar while ensuring adequate funding. The (relatively) small size of the workforce and the decrease in labor informality since the early 2000s also resulted in the pension industry concentrating around four AFAPs that managed in 2020 assets around 30 per cent of GDP (FIAP, 2020).
Concerns about coverage and adequacy of future pensions led the centre-left Frente Amplio government to initiate a minor parametric reform in 2008, following lengthy negotiations with political parties, social partners, the pension industry, and pensioners. Measures included a reduction in the years to qualify for a contributory pension and the possibility for those not voluntarily affiliated to the private system to disaffiliate. The reform is the outcome of a strong institutional setting, given the government majority in both chambers and a strong pension industry, given the increasing levels of savings and strong unions and pensioners represented in the social security administration bank (Banco de Prevision Social – BPS) (Busquets, Reference Busquets2013).
Nonetheless, sustainability issues accentuated by the country’s demographic prospects (low birth rates) and the seven special regimes not captured by the 1995 reform, coupled with increasing public pension spending, resulted in pensions being high on the agenda during the 2019 presidential elections (Comisión de Expertos en Seguridad Social, 2021). The elected centre-right administration of Lacalle Pou set up a Pension Commission in 2020 to discuss reform options. Facing divisions within its own coalition and needing to get some support from the centre-left, the government underwent discussions with various stakeholders, and several concessions were made. A reform bill was finally passed by the Chamber of Deputies in 2023 (Infobae, 2023). The (major parametric) reform marks a departure from the status quo as it includes the consolidation of all schemes into the main public one and the mandatory affiliation to the second private pillar for all workers. Yet, given the transition period enshrined in the law, the effects of the reform will be felt in the medium to long term.
The latest reform signals a combination of a weak institutional setting given the fragmentation in the ruling coalition and the need to agree on the reform with the opposition and other stakeholders. On the legacy front, the relevance of the strong union and pensioner movement led the government to consider their views, which was reflected by their participation in the Pension Commission discussions (Infobae, 2024). Nonetheless, the institutional feature of referenda has been successfully used by the opponents in the labor movement, and a proposal to revert the latest changes and to also eliminate the private pillar was voted on in the 27 October elections but failed to be passed affirmatively (Infobae, 2024).
Italy
In Italy, the generous provisions of the public PAYG system combined with unfavorable economic and demographic developments since the early 1980s impacted its financial sustainability and made reform necessary. Following a series of failed reform attempts in the 1980s, a series of strong governments of a technocratic nature in the early 1990s, and the use of delegated decrees allowed the adoption of significant reforms (Carrera et al., Reference Carrera, Angelaki and Carolo2010). This was combined with the weakened position of the labour unions following the industrial changes during the late 1970s, leading to a significant reduction in the unionization rate and more weight of the pensioners within the unions (Ferrera and Jessoula Reference Ferrera and Jessoula2009).
The Amato technocratic government used the political fragmentation of the traditional parties to its advantage to introduce sweeping changes, aided by recourse to delegated decrees (Squarzon, Reference Squarzon1992). The 1992 reform included changes in the formula calculating pensions for new entrants and increases in the number of years necessary to qualify for a seniority pension. This was complemented by a further reform in 1993, introducing voluntary private pensions. The 1995 reform of the Dini technocratic government was once again adopted following the agreement with the labour unions, who succeeded in ensuring that the new method to calculate pensions based on the amount of contributions (Notional Defined Contribution/NDC) would apply only to new entrants. Both the 1992/1993 and 1995 reforms are structural, given the changes in eligibility rules, years to calculate pensions, the introduction of voluntary private pensions, and most notably, the introduction of a new method to calculate pension benefits.
Further reforms have been adopted since the mid-2000s to address issues such as inequities between private and public workers and increase the take-up of private voluntary pensions. In 2007, the Prodi government passed a reform that automatically transferred workers’ severance payment funds (Trattamento fine del Rapporto/TFR) to private occupational schemes instead of them remaining with their employer, unless workers opted out. The result has been a significant expansion of the private occupational sector, with assets under management over 15 per cent of GDP in 2023 (COVIP, 2021). In 2011, the technocratic Monti government, in the aftermath of the global financial crisis, passed a reform that shortened the phasing-in period foreseen in the Dini reform, while linking retirement age with changes in life expectancy.
Overall, during the 2000s, minor parametric reforms have taken place in a different context from those previously adopted. In 2007, the relatively weak Prodi government agreed to the reform with the unions movement (Baccaro, Reference Baccaro2009). By contrast, the Monti Government 2011 reform can be explained by the combination of a strong institutional setting, given the caretaker nature of the government, and a diminished role of the unions that needed to accept the measures as part of a package of emergency measures.
Spain
The democratization process that started in 1975 led to the expansion of social protection, regarded as a catch-up process with Europe (Guillén and Alvarez, Reference Guillén and Alvarez2004). The election of the socialist government in 1982 opened the reform path through the introduction of cost-containment measures (increase in the number of contributions required to obtain a pension, change in the calculation formula, stricter rules for survivor and disability pensions), followed by the introduction of private pension plans in 1987.
The rise in unemployment levels in the early 1990s, coupled with the political developments (the socialists losing majority and forming a minority government with Catalan and Basque nationalists), led to the adoption of the Toledo Pact, bringing together political parties, the labour movement, and academics (Chulia, Reference Chulia, Immergut, Anderson and Schulze2009). The Toledo Pact (which concluded its work in 1995) set the framework for an incremental reform in the years to come via a permanent commission that would seek the input of different societal actors (Chuliá, Reference Chuliá2000). This “transformed the pension policy area into a multiple partisan veto player system” (Chulia, Reference Chulia, Immergut, Anderson and Schulze2009) by strengthening the role of unions as the main societal veto actor in pension reform (Natali and Rhodes, Reference Natali, Rhodes, Arza and Kohli2007). Thus, mostly minor parametric reforms have taken place since the mid-1990s and have been negotiated among different stakeholders. For example, the reform implemented by the conservative government in 1996 included cost-containment measures, yet it also included measures to improve the situation of the less advantaged as a concession to the trade unions.
The 2008 global financial crisis and the continuous negative demographic trends opened a new reform round led by the conservative PP since 2011, which now enjoyed a single majority in Parliament. The minor parametric reform was implemented unilaterally outside the Toledo Pact negotiation process and entailed parametric measures. A further reform in 2013 introduced a sustainability factor linking retirement age to life expectancy from 2019 (and revised every 5 years) as well as a new benefit indexation mechanism taking into consideration the total budget constraint of the system. The 2013 reform was also unilaterally adopted and, as a result, has been criticized by both the opposition and the trade unions (Conde-Ruiz and Gonzalez, Reference Conde-Ruiz and Gonzalez2013); (Natali and Stamati, Reference Natali and Stamati2014).
The latest reforms introduced between 2021/23 by the socialist government revoked some of the previous measures, such as the sustainability factor and the revalorization index. Serious concerns have been raised in relation to the impact of the measures on the sustainability of the system and its intergenerational equity, despite the government’s optimistic stance (de la Fuente, Reference de la Fuente2024). Nonetheless, following the 2020 recommendations of the Toledo Pact commission, reforms also included measures to promote the setup and take-up of private occupational pension funds (Ministerio de Inclusión, Seguridad Social y Migraciones, 2023; Natali et al., Reference Natali, Janta, Spasova, Bannister Curran, Fassino, Copping, McCarthy, Kominou, O’Brien, Webb and Hussain2024). This major parametric reform (given the move to expand private pensions) is explained by a combination of a strong institutional setting and a strong policy legacy. The socialist government has a minority in Parliament, yet it has successfully exploited the fragmentation of the opposition to its advantage to ensure it can get legislation passed (Field, Reference Field, Field and Martin2022, 134). This combined with a strong policy legacy given the multiple veto actors such as unions and the opposition, who have a role in the Toledo Commission.
Greece
Greece (much like Italy) entered the 1990s in a twin context of economic and political crisis. Against this background, pension reform assumed a central role in the fiscal adjustment process undertaken by the New Democracy centre-right government, as increases in spending (pensions already accounted for 15 per cent of GDP) would otherwise have been unsustainable. Nonetheless, the reforms of 1990/1992 were limited to minor parametric measures, leaving structural deficiencies untouched. In understanding this outcome, we argue that single-party governments were faced with the strong opposition and mobilization capacity of the social partners who have succeeded in blocking (or significantly watering down) measures considered to negatively affect their members (Carrera et al., Reference Carrera, Angelaki and Carolo2010; Angelaki, Reference Angelaki2016).
The advent of the 2009 crisis and the agreement of the socialist government with the troika of the EC, ECB, and IMF on a bail-out programme, put pension reform back on the agenda. The reform was part of the broader effort of the government to undertake measures that would allow the country to re-access international markets. The 2010 reform (part of the bail-out agreement) resulted in the separation of the social assistance from the social insurance functions through the introduction of a basic (flat-rate) quasi-universal component, and a proportional PAYG one. The law also foresaw the introduction of a safeguard clause, whereby if actuarial analysis falls somewhat short of the final objective to reduce increases in future pension costs by 2.5 percentage points of GDP (taking 2009 as reference year) the government will introduce further measures to lower pension expenditure. While the narrow time frame did not allow any time for consultation with the social partners, the latter described the bill as an effort to dismantle social security (Angelaki and Carrera, Reference Angelaki and Carrera2015).
Additional measures introduced in 2012 from the national unity government under L. Papademos, former ECB Vice-President, that consisted of the socialists (PASOK), the conservatives (New Democracy) and the Popular Orthodox Rally (LAOS) a right-wing populist party. Measures foresaw the merging of auxiliary pensions and the transition from defined benefit to a notional defined contribution system based on individual pension accounts, while mandatory occupational pension schemes were introduced in 2013. Overall, the 2010/2012 reforms are of a structural nature and can be explained by a strong institutional setting that allowed more leeway to the government in the context of the crisis and the weakened role of social partners compared to the pre-crisis period (cf. Natali and Stamati, Reference Natali and Stamati2014).
The reforms undertaken over the past 15 years indicate a shift towards a multi-pillar system; yet, the Greek occupational pension market remains underdeveloped, despite the increase in the number of funds established, as shown in the low level of penetration (0.65 per cent between 2011 and 2021 compared to the European Economic Area average of 16.68 per cent) (Chrysanthopoulou et al., Reference Chrysanthopoulou, Mavrommati and Migdalas2023).
Discussion and conclusion: institutions and legacies still at play
Our cross-regional analysis examined the role of the institutional setting and policy legacies as key factors in understanding pension reform. Table 2: illustrates our findings.
Table 2. Pension reform outcomes expected and actual cases
Our analysis largely confirms that significant path departure reforms are possible when a strong institutional setting is combined with a weak policy legacy, as shown by the cases of Italy in the early 1990s and Greece in 2010/12. This finding lends support to earlier institutional approaches (Pierson, Reference Pierson1996; Kay, Reference Kay1999; Madrid, Reference Madrid2003) claiming that strong governments with few veto players can implement significant reforms. The Italian reforms of 2007 and 2011 – while adopted by strong governments – were not of structural character as major challenges had already been tackled in previous reform rounds.
At the same time, our analysis showed that significant reforms have also been implemented in a context of strong institutions, yet combined with a strong policy legacy given by a strong labor movement, as in Argentina, where the strong labor movement supported both structural reforms to maximize their narrow interests. This finding highlights the importance of considering how powerful veto actors can use their strength not just to block reforms but also to further their interests (Etchemendy and Collier, Reference Etchemendy and Collier2007; Natali and Rhodes, Reference Natali, Rhodes, Arza and Kohli2007).
As expected, most of the cases of major parametric reforms have taken place when a weak institutional setting combined with strong societal actors, leading governments to negotiate or make concessions. Our analysis also seems to confirm Anderson and Weaver (Reference Anderson and Weaver2025) argument that cooperative institutions and broader coalitions can help overcome the stalemate. Minor parametric reforms have generally been the consequence of a strong institutional setting and a strong policy legacy, as illustrated by the case of Uruguay in 2008.
Overall, our analysis has broader policy implications of interest to scholars and practitioners alike. First, we have shown that significant reforms are less the result of ideology but more of government strength and its willingness to consider the role of different stakeholders. Even a weak government, as in Uruguay in 1995, succeeded in adopting a structural (albeit limited) reform, facing a strong labor movement by arguing about the need to save the system from collapse and agreeing to maintain a rather significant first pillar. Second, we have also shown (cf. Niedzwiecki and Pribble, Reference Niedzwiecki and Pribble2017) that not only centre-right (as in Argentina in 1994) but also centre-left governments (as in Greece in 2010 or Uruguay in 1995) in the context of severe crisis gain more leeway to implement structural reforms.
Third, while much of the earlier literature on pension reforms (Madrid, Reference Madrid2003; Orenstein, Reference Orenstein2008; Brooks, Reference Brooks2009) seemed to imply an irreversible path towards neoliberal retrenchment and individualization, our analysis shows a more mixed picture. Key to these outcomes has been the role of unions and grassroots movements. Meanwhile, in southern Europe, while public pillars have been reformed, there has been privatization through the backdoor after negotiation and concessions to societal actors given by the expansion of complementary private pensions, as shown by the cases of Greece in 2009/2012, Italy in the 1990s and the recent Spanish reforms of 2023. This finding confirms insights from analysis of more developed pension systems about the role of societal actors not just as “blockers” of reforms, but also in achieving positive policy change (Anderson and Weaver, Reference Anderson and Weaver2025).
Our analysis also has implications for feminist approaches to social policy (Jefferson, Reference Jefferson2009). The heavier reliance on private pensions in both regions in the future makes relevant the role of care credits, designed to compensate breaks in working years for child-raising or elderly-care responsibilities (mostly undertaken by women). While recent reforms in Spain and Chile have made provisions for this, more work is needed by policymakers and practitioners to address this risk. The interaction of occupational pensions and care credits is an important element not only for the protection of pension rights of carers but also for supporting gender equality (Natali et al., Reference Natali, Janta, Spasova, Bannister Curran, Fassino, Copping, McCarthy, Kominou, O’Brien, Webb and Hussain2024; Anderson and Weaver, Reference Anderson and Weaver2025). The introduction of non-contributory pensions also contributes to closing the gender pension gap (Arza, Reference Arza2012b).
Recent reforms in Uruguay (2023 and Chile (2025) also have implications for understanding how policies may address labor market inequalities. In the case of Chile, the legislated increase in low pensions funded partly by the new employer contribution will favor those with insufficient savings in their private pension due to periods of unemployment and/or informality, but also non-standard workers (such as those in the gig economy who are subject to intermittent periods of work). This mechanism is also likely to benefit women, given their more intermittent labor careers due to caring responsibilities.
We conclude with four main insights from our analyses. First, our cross-regional analysis contributes to the literature on comparative pension and social policy reform by showing that institutions and legacies help explain different reform outcomes in different regions. Second, it is of interest to explore further the ability of actors not only to block reforms, but also to enable positive changes. Third, while our cross-regional comparison allows us to provide meaningful insights, we should not underestimate the issues of dealing with multiple cases and complex causality (King et al., Reference King, Keohane and Verba1995). In this sense, further cross-regional analyses could explore more systematic comparative methods such as fuzzy-set qualitative analysis (Ragin, Reference Ragin2008; Carrera and Angelaki, Reference Carrera and Angelaki2021, Reference Carrera and Angelaki2024). Finally, the analysis has provided a range of insights on how practitioners and scholars interested in pension policy change could address recommendations for reforms to ensure that they can be implemented and address the identified needs.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/ics.2025.10062.
Acknowledgment
The authors would like to thank the anonymous reviewers and the journal editors for their constructive critique on an earlier version of the paper.
Open access
