Information about Asia and the Pacific Asia y el Pacífico
Equitable and Sustainable Pensions

Chapter 8. Is the Egg Basket Worth Its Price? The Fiscal Implications of Pension Privatization in Eastern Europe

Benedict Clements, Frank Eich, and Sanjeev Gupta
Published Date:
March 2014
  • ShareShare
Information about Asia and the Pacific Asia y el Pacífico
Show Summary Details
Jan Drahokoupil and Stefan Domonkos The authors are grateful to Nicholas Barr, Ivan Lesay, Teodoras Medaiskis, Ringa Raudla, András Simonovits, and Edgars Volskis for their valuable comments and help with data collection. The authors also benefited from stimulating comments from the participants of the IMF Fiscal Affairs and European Departments Conference “Designing Fiscally Sustainable and Equitable Pension Systems in Emerging Europe in the Post-Crisis World,” Vienna, Austria, March 18, 2013. The authors assume full responsibility for remaining imperfections.


Eastern European countries were among the pioneers in introducing mandatory privately funded schemes into their pension systems. Their approach to pension privatization, however, changed significantly after 2008. As a consequence of the global economic recession, the region saw a number of pension “reform reversals.” 1 However, the crisis was only one of the factors that contributed to the transformation of the political and economic context of pension reforms. The first wave of reforms was also accompanied by a learning process about the nature of reform transition costs as well as their benefits.

The second wave of reforms, after 2008, produced a wide variety of outcomes. 2 At one extreme, Hungary, a pioneer of pension privatization, de facto nationalized its mandatory funded pension scheme. Poland and, eventually, the Slovak Republic, reduced the size of their mandatory private schemes substantially by lowering the percentage of contributions diverted to these schemes from their pay-as-you-go (PAYG) pillars. The Baltic states temporarily reduced or suspended contributions to their mandatory private schemes. In contrast, the Czech Republic actually implemented pension privatization in this period, albeit at a more modest level than had been common in the first wave of reforms. Privatization is thus still on the agenda in this new economic and political environment; however, the learning process has changed the rationale for privatization, as presented by its proponents. A diversification argument has gained prominence, seemingly replacing the argument that the introduction of funded schemes can resolve the pressure of demographic aging. Accordingly, the policymakers of the post-2008 era should follow the advice of Don Quixote’s servant Sancho Panza that “it is the part of a wise man to keep himself today for tomorrow, and not venture all his eggs in one basket” (Whitehouse, D’Addio, and Reilly, 2009, p. 10).

This chapter discusses the policy lessons learned from the two waves of pension reform in eastern European member states of the European Union (EU). It focuses, in particular, on the changing approach to the fiscal implications of pension privatization. The next section provides an overview of the two waves of pension reform as well as the political and economic factors that conditioned the outcomes. The focus then turns to the problem of financing the funding gap related to the diversion of contributions from the public pension system to private schemes, a key issue that was sidestepped in the first wave of reforms, largely because of a lack of understanding of the problem. The funding gap was underestimated, mainly because of a mistaken argument that explicit debt can be ignored because it replaces implicit debt. The third section addresses the actual solutions to the funding gap problem, showing that it remained largely unresolved in the first wave of reforms, despite the learning process that accompanied the implementation of pension privatization. The final section then discusses the new rationale that informs the second wave of reform, the “diversification argument.” Its growing importance follows from a learning process in which many of the myths that influenced the policy debates in the first wave of reforms were dispelled. However, the diversification argument itself can be seen as old wine in new bottles because its underlying rationale is largely based on one of the myths of the first wave of reforms—that is, the assumption that prefunding can hedge against the macroeconomic shock induced by demographic aging. Once this and other misunderstandings are put aside, what remains as a rationale for pension privatization in the second wave is not so much a positive economics argument but a mistrust of the state and collective provision of social insurance. The comparison of the costs and benefits of diversification thus suggests that the second proverbial egg basket (mandatory private pension accounts) comes at a rather steep price.

The Eastern European Love Affair with Private Pensions: From Pension Privatization to Reform Reversals

In the late 1990s to middle of the first decade of the 2000s, the majority of eastern European countries pursued pension privatization, reforming their PAYG pension systems by following the model laid out by the World Bank in its influential 1994 report Averting the Old-Age Crisis (World Bank, 1994). The World Bank model was a three-pillar pension system comprising a publicly financed first pillar based on the PAYG principle, a mandatory fully funded second pillar based on private individual accounts, and a voluntary fully funded third pillar. 3 The second pillar was to be created by diverting contributions from the first pillar (privatization). The second pillars created in the first wave of reforms were often not based on cross-party consensus, but in all cases they initially survived the incumbency of their opponents. 4 The Czech Republic and Slovenia were the only countries that did not implement the World Bank–style pension reform in this first wave. They pursued only parametric reforms in their state-run PAYG schemes. An overview of pension-reform trajectories in all eastern European member states of the EU can be found in Table 8.1.

Table 8.1Pension Reform in Central and Eastern European Countries
Country• Pension system in 2008Main Reform Tendency during the Global
(year second pillar was• Total pension contributionsEconomic Crisis
launched)as percentage of wage, contributions to the second pillar
• Type of public first pillar
Bulgaria (2002)• Three-pillar pension systemNo changes in contribution rates, but
• 22% for regular employees, 5% + up to 1% of accumulated savings in second pillar as management feepart of private occupational pension funds nationalized; 1.8 percentage point increase in overall contribution rate.
• Conventional defined benefit
Croatia [EU member as of• Three-pillar systemNo changes in contribution rates; those
July 1, 2013] (2002)• 20%, 5%. Point systemborn between January 1, 1952, and January 1, 1962, who voluntarily joined the second pillar in the past could opt to move back to being full members of the first pillar only.
Czech Republic (2013)• PAYGs ystemIntroduction of a voluntary second pillar
• 28%, 20% of which are for old-age pensions only, complemented by third pillaras of January 1, 2013, with 3 percentage point transfer from statutory social security (first pillar) contributions plus 2 percentage
• Conventional defined benefitpoints additional contributions by savers.
Estonia (2002)• Three-pillar systemContributions returned to 6 percent
• 22%, 6%after temporary lowering between 2009
• Conventional defined benefitand 2011.
Hungary (1998)• Three-pillar systemDe facto nationalization of the second
• 33.5%, 8%pillar and transfer of assets accumulated
• Conventional defined benefitin second-pillar accounts to the state in a one-off transfer in 2011; those who chose to stay in the second pillar can make voluntary payments to their pension accounts, but no statutory social security (first-pillar) contributions are transferred into second-pillar accounts.
Latvia (2001)• Three-pillar systemContributions to second pillar set temporarily
• 20%, 8%at 2 percent of covered income,
• Notional defined contributionincreased to 4 percent in January 2013.
Lithuania (2004)• Three-pillar systemContributions to second pillar set at
• 26.35%, 5.5%2 percent of covered income, with further
• Conventional defined benefitdecrease in 2012 (1.5 percent); an increase to 2.5 percent took place in 2013.
Poland (1999)• Three-pillar systemContributions to second pillar set at
• 19.52% for retirement insurance only, 7.3%2.3 percent; increasing to 2.8 percent beginning January 2013; contribution rate to
• Notional defined contributionsecond pillar should stabilize at 3.5 percent of covered income by 2017.
Romania (2008)• Three-pillar systemAfter a temporary freezing of second-
• 29%, 2%pillar contribution rates, the rates reached
• Point system3.5 percent in 2012, increasing to 4 percent as of 2013.
Slovak Republic (2005)• Three-pillar systemContribution rates decrease from 9 percent
• 18% for old-age insurance only, 9%of covered income to 4 percent beginning September 1, 2012; membership in
• Point systemthe second pillar is made optional with default option being not to join.
Sources: Authors.Note: PAYG = pay-as-you-go.
Sources: Authors.Note: PAYG = pay-as-you-go.

The economic crisis of 2008 marked a turning point in the evolution of pension systems in eastern Europe. Public deficits and demands on state spending increased as revenues fell. Reduced fiscal space made financing the missing revenue flowing into the second pillars particularly challenging. Therefore, most countries pursued some form of reform reversal,5 but the range of outcomes also indicated a commitment to privatized pillars. By the end of 2012, only Estonia had returned to the full level of precrisis second-pillar contribution rates. 6

Fiscal Constraint Effects

Variations in immediate policy responses can be linked to the differences in fiscal constraints. These responses varied because the crisis had very different impacts across eastern Europe (Myant and Drahokoupil, 2012). Early reformers with larger second pillars also faced higher financing needs to cover the gaps created by the reforms because the number of those who entered the mandatory funded pillar had grown significantly since the initial introduction of the funded schemes. With the exception of Hungary, the levels of sovereign indebtedness in eastern Europe were modest. The fiscal space in the Baltic states, however, was limited by large output contractions and their commitment to defending fixed exchange rates through “internal devaluation”—that is, cuts in public sector wages and public expenditure (Kattel and Raudla, 2013). The Baltic countries maintained their second pillars, but they had to lower contributions temporarily but significantly to be able to deal with the fiscal consequences of the crisis.

Among the central and eastern European countries (CEECs), 7 Hungary was particularly vulnerable to the crisis given its dependence on credit from abroad. The government experienced significant problems financing its debt after October 2008 and was compelled to seek funding from the IMF and the EU. These economic constraints led decisively to the de facto nationalization of the second pillar in 2010. 8 This move allowed the government to stabilize finances through revenue increases—the one-off transfer of accumulated assets amounted to 10 percent of GDP. The influx helped to balance the budget at a delicate time when the government had decided to terminate a Stand-By Arrangement with the IMF. 9

Other CEECs did not face such constraints, but they all chose to pursue austerity policies to address their growing deficits (Myant, Drahokoupil, and Lesay, 2013). The broad commitment to pension privatization had begun to unravel throughout the CEECs. Poland decided on a permanent reduction of second-pillar contributions largely because it was approaching the self-imposed constitutional debt ceiling of 55 percent of GDP. The Slovak Republic had more room to maneuver, given its low indebtedness and quick return to growth, but the socialist government elected in 2012 decided to permanently reduce the second pillar as part of a concerted attempt to balance the budget. 10 Finally, the Czech Republic actually jumped on the pension privatization bandwagon when others were reversing various reforms, 11 but the second pillar it introduced was modest and voluntary.

Other Effects on Reforms

The change in pension privatization policies cannot be fully attributed to the fiscal effects of the economic crisis. The crisis was one of several factors that contributed to the transformation of the political, economic, and ideational context shaping pension reforms. First, the initial wave of reforms stimulated a learning process in international policymaking networks and among experts in eastern Europe. The information about the actual costs and benefits of these programs was poorly distributed in the first wave because of numerous reform “myths” (Barr, 2000; Orszag and Stiglitz, 2001). The propositions of the World Bank’s Averting the Old-Age Crisis were subjected to a wide range of criticism (for example, Barr, 2000; Fultz and Ruck, 2000; Orszag and Stiglitz, 2001; Barr and Diamond, 2008). Thus, following the first wave of reforms, pension privatization was less often seen as a solution to demographic aging. Nor was it assumed any longer to have automatic growth-stimulating effects.

In eastern Europe, the issues of first-pillar stabilization and pension privatization were finally seen as two separate problems, which had not often been the case in the first wave of reforms. The maturation of the first-wave reforms also contributed significantly to the learning process about the actual implications of pension privatization. New data made it possible to evaluate the performance of pension funds using their rates of return, fees charged, and investment strategies. This maturation also exposed the problem of transition costs that had been poorly understood in the first wave of reforms and had been left largely unresolved. Transition costs could no longer be ignored because significant austerity measures seemed necessary to continue financing these reforms. 12 The ability to deal with transition costs became a major concern for the architects of pension privatization in the Czech Republic.

Second, the World Bank changed its views on pension privatization. The Bank’s active promotion of pension privatization had a major influence on the first wave of reforms and became an important basis for the political support for reformers (Müller, 2001; Orenstein, 2008; Lesay, 2009), but the consensus on the issue within the Bank began to unravel well before the crisis. By 2008, its pension privatization advocacy campaign was effectively over (Orenstein, 2011). The IMF, which had had a more critical stance on private pensions, also did not promote privatized pensions in its assistance programs after 2008. The withdrawal of the World Bank from the privatization agenda made pension privatization essentially a domestic affair, with party politics significantly shaping the policy reactions. In this new economic and ideational environment, pension privatization became an initiative of the liberal right (Drahokoupil and Domonkos, 2012).13

Third, requests by countries that had introduced second pillars to exempt the transition costs of privately funded schemes from the Maastricht criteria were rejected by the EU several times in 2004–05 and 2010. 14 The EU’s Maastricht criteria obligated new member states to maintain their public deficit and public debt at less than 3 percent and 60 percent of GDP, respectively. These obligations significantly limited fiscal freedom before the global economic crisis. The narrower fiscal space, in combination with growing transition costs, made debt financing difficult. Therefore, dealing with the funding gap became a matter of distributing the costs in the short term. The funding gap became a priority on the policy agenda for these reasons. The actual solutions to dealing with transition costs in the first wave of reforms had often been misrepresented. Because of the crisis, policymakers could no longer evade the question of transition costs and their financing, as done before 2008.

The Funding Gap: Lessons About Implicit and Explicit Debt

Pension privatization in eastern Europe was legislated in states that already had mature PAYG systems. The working-age population and current pensioners thus had already accumulated pension rights that represented future liabilities on state budgets. PAYG systems pay out pensions from present revenues. Pension privatization diverted part of this revenue to the second pillar. The subsequent funding gap in the first pillar resulted from the loss of resources that had previously been designated to finance the pensions of those with accumulated pension rights. In this chapter, the term “funding gap” refers to the annual budget revenue shortfall attributable to the introduction of mandatory funded pillars minus the decrease in claims on the PAYG scheme, also caused by the introduction of the second pillar. From a policymaking perspective, the transition costs of pension privatization include the funding gap and its financing costs—the actual transition costs are thus likely to be significantly larger than the funding gap. After introducing the second pillar, the number of people in the noncontributing workforce increases, whereas those who could, in theory, reduce the burden on the state by receiving part of their pension from the second pillar will not retire until several decades after the introduction of the reform. 15

A stylized depiction of the development of the annual funding gap can be seen in Figure 8.1. The funding gap in the Slovak Republic, which introduced its second pillar in 2005, was expected to peak in 2030, when the yearly costs would have reached approximately 2.5 percent of the country’s GDP. After this point, the number of retired workers who rely on the second pillar would have started to grow, thus reducing the annual funding gap. Contributions lost and benefits spared because of the introduction of the second pillar were expected to even out about 2052 (Ódor and Novysedlák, 2005).

Figure 8.1Development of the Funding Gap

Source: Based on Simonovits (2003, p. 156).

Some proponents of pension privatization question the relevance of the funding gap, claiming that these costs merely amount to the transformation of the implicit debt of the PAYG system into explicit debt. 16 This notion indicates either a misunderstanding of the difference between implicit and explicit (real) debt or a conflation of implicit debt and real deficits.

Implicit Debt

Implicit debt is a theoretical construct that refers to liabilities, pension benefits in this case, due in the future (Cheikh and Palacios, 1996; Holzmann, Palacios, and Zviniene, 2004). In a PAYG system, these liabilities are financed by current revenues. The expected revenue streams could thus be understood as implicit financing, a counterpart to implicit debt. Real deficits can occur if current revenues do not match current liabilities. In the long term, therefore, a PAYG system can generate real deficits, or surpluses, if implicit debt and implicit financing do not balance. A conflation of these concepts has contributed to a misunderstanding of the fiscal implications of making the transition from PAYG to a funded system in the first wave of reforms. The debate often ignored the idea that, although pension privatization reduces implicit debt by making it explicit—with all related consequences—privatization also reduces implicit financing by permanently lowering the public budget’s revenue stream from social security contributions.

It is important to note that both implicit debt and implicit financing are theoretical concepts based on predictions about an uncertain future. Both depend to a large extent on pension legislation, which provides the necessary tools for balancing state-run PAYG schemes by adjusting outlays and contributions. Instead of lowering implicit liabilities, increasing implicit financing, or doing both in the PAYG system, the proponents of pension privatization advocate partial replacement of an imbalanced public system with a privately funded scheme. This change is accompanied by an immediate increase in explicit indebtedness: pension liabilities accrued up to the point of pension privatization have to be honored by the state, but simultaneously, social security contributions are being transferred to individual accounts in the private scheme. Full privatization relieves the state of dealing with possible future imbalances in the pension system by transferring pension insurance risks to individuals. At the same time, it also binds the state to financing the pensions of an entire generation of retirees through means other than social security contributions from working-age cohorts. The decrease of contributions caused by pension privatization amounts to a shock theoretically equivalent to the death of the next generation that would have replaced the current workers— a scenario that is not equivalent to the actual demographic shock (discussed below). In the World Bank–type reform, the transition costs actually “pay” only for transforming a mature PAYG system into a mixed system with a funded pillar.

Simulation studies on second-pillar reforms and their reversals in eastern Europe show that the income replacement rate (levels of pensions relative to wages) is the key variable that influences the long-term sustainability of national pension systems (Égert, 2012). In Hungary, introducing a second pillar could actually lower the costs of the country’s pension system in the long term, but this savings effect is driven by the indirect effect of downsizing the first pillar (that is, lowering pensions), which has an excessively generous replacement rate. In principle, obviously, pension privatization is not needed to decrease first-pillar replacement rates.

Explicit and Implicit Debt

More informed understandings of transition costs were often based on the assumption that implicit and explicit debts are equivalent—and, correspondingly, that second-pillar pension savings are not newly created savings but only explicit valuations of implicit claims in the form of accumulated pension rights. Explicit debt generated by financing the funding gap thus should not count as new debt because it is equivalent to the implicit debt arising from the pension rights accrued in the social security system (the first pillar). In accounting terms, it is factually correct that the funding gap is not a new cost because it is produced by making the implicit debt (or part of it) explicit. Moreover, assuming an economy with a single interest rate and no transaction costs, pension privatization is cost-neutral, as explained in greater depth by Simonovits (2003) in his account of the no-pain-no-gain scenario (see also Kubíček, 2008). The level of public debt increases as a result of the transition costs, but the costs of servicing that debt are equivalent to the returns on the second-pillar savings. In the no-pain-no-gain scenario, the outcome of privatization is a higher level of public indebtedness. And the operation of making the implicit debt explicit does not cost society or the public purse anything.

However, in addition to ignoring transaction costs and assuming that pension fund returns at least match the implicit returns on the PAYG system, the no-pain-no-gain model is based on the unrealistic assumption of perfect rationality and information. Such an assumption allows implicit and explicit debts and savings to be perceived as equivalent: an increase in the level of public indebtedness would thus have no effect on the ability of the state to borrow—and the newly created savings would have no impact on the savings behavior of households. The increase in the public debt level would be discounted by the decrease in the level of implicit debt, the amount of which would be known because of the availability of information on future changes in the first pillar (pension levels, retirement age, and so on). In other words, the negative reaction of the financial markets to higher public debt would be mitigated by a decrease in implicit debt. However, it is unlikely that the creditors in international financial markets would accept the long-term budget neutrality of the transformation of implicit debt into explicit debt (Simonovits, 2003). Financial market actors in the real world have proved not to conform to this assumption. The increase in explicit debt resulting from pension privatization had a negative impact on country risk ratings, while the assumed decreases in implicit debt did not lead to rating improvements (Cuevas and others, 2008). The refusal of the EU institutions to attribute perfect rationality to the market and thus to exempt the transition costs from the Maastricht criteria was thus correct from this point of view.

In the real world, future liabilities of the state toward its citizens (that is, implicit debt) are very different from current accumulated debt. Implicit debt is largely dependent on the creditor’s own legislation. Ultimately, the size of the implicit debt is subject to unilateral government decisions. Implicit liabilities are thus more akin to a political pledge than an actual quantifiable financial category (Holzmann, Palacios, and Zviniene, 2004). By contrast, the funding gap translates into a current and real liability that is often accumulated against foreign entities (nonresidents). Influencing that liability through national legislation might equal de facto sovereign default on debt. Experience has shown that the markets, indeed, price explicit and implicit debt very differently, with increased explicit debt levels being punished with little regard to a theoretical lowering of the sum of future obligations. The accumulation of explicit sovereign debt has thus involved the risk of increasing the dependence of CEECs on international financial markets, leading to worsening credit ratings and higher debt-service costs. Therefore, from the perspective of the state, retaining debt in its implicit form appears to be the preferable option.

Costs and Benefits

The assessment of transition costs has to take into account the benefits and costs of pension privatization that are ignored in the simple no-pain-no-gain scenario. Unfortunately, the assessment that informed the first wave of reforms was not balanced and realistic. Policymakers at that time were too optimistic about the gains from possible higher returns in the funded pillar (in comparison with the implicit returns in the PAYG pillar). The simple no-pain-no-gain scenario assumes a single interest rate, but the average return on investment can in reality be assumed to be higher than the interest paid on public debt, provided that the pension funds do not invest in state bonds.17 In practice, however, government bonds made up a large part of pension fund portfolios, making privatization appear to be a pointless accounting exercise. As shown in Figure 8.2, government bonds and bank deposits represented, on average, two-thirds of the total assets of the pension funds in 2011 and reached almost 80 percent in the Slovak Republic and Romania. This outcome obviously defeated the expectation that privatization would result in higher returns.

Figure 8.2Share of Government Securities and Bank Deposits in Total Second-Pillar Assets in 2011

(Percent of total assets)

Source: Data provided by World Bank staff.

Furthermore, even when the circular transaction of pension funds buying government bonds is avoided, the assumption that portfolio returns will be higher than the implicit returns in the PAYG system (that is, real wage growth) cannot be taken for granted. This assumption is derived from a study of the long-term performance of the U.S. stock market (Kotlikoff, 1995), which cannot be easily generalized outside the United States. A number of countries can be found for which long-term portfolio returns do not compare well with real wage increases (e.g., Loužek, 2006). The experience of converging economies in eastern Europe also showed that returns in pension fund portfolios did not tend to perform well in comparison with wage growth. Figure 8.3 provides a comparison of wage growth and average returns in the period before the 2008–09 financial crisis.

Figure 8.3Average Real Wage Growth and Real Rates of Return in the Second Pillar, 2002–07


Source: Data provided by World Bank staff.

First, the returns were influenced by high wage inflation in the run-up to the crisis in the Baltic states, on the one hand, and, on the other, by initial high rates of return in Hungary, Poland, and the Slovak Republic caused by an appreciation of the value of state debt in the initial period. In 2008, pension fund portfolios recorded substantial losses. 18 The period that followed was characterized by large year-over-year fluctuations.

Second, the comparison of returns in funded and PAYG schemes cannot be disentangled from the transaction costs of running a funded pension system. These costs need to be subtracted from the returns. Both PAYG and funded schemes incur management costs, but they are relatively low in PAYG systems. Large administrative costs are an inherent problem in a system with individual accounts and individual choice of pension provider. These fixed administrative costs also undermine the rationality of smaller second pillars that were introduced in the second wave of reforms in many countries. The experiences of the United Kingdom and Chile, both of which took a hands-off approach to regulating management fees, show that the fees were a drag on investment returns, consuming about 20 percent of the wealth accumulated during the savers’ careers (Diamond, 1998; Murthi, Orszag, and Orszag, 2001). In eastern Europe, management fees varied widely by country and by investment product. 19 Although the overall tendency has been a gradual decline in the percentage share of fees on net asset value and contributions, fees have remained relatively high in several countries. 20 In 2007, yearly administrative charges amounted to as much as 2 percent of total assets in Hungary, and 1.5 percent of total assets in the Slovak Republic and Poland (Tapia and Yermo, 2008). It might be argued that administrative fees significantly decrease with the maturation of the mandatory funded system, but the Chilean experience shows that even almost 30 years after the inception of the funded scheme, annual administrative charges still were about 0.7 percent of total assets (Tapia and Yermo, 2008).

Third, arguments in favor of pension privatization in the first wave of reforms included the assumption that the reform would spur output. The expectation was that capital from the increased savings would be channeled into more productive segments of the national economy. However, even in theory, this expectation is based on rather exacting conditions, including the following: savers would not lower their other savings as a reaction to owning shares in pension portfolios, 21 savings accumulated in individual pension accounts would not be invested in newly issued national government bonds, and the funds would not be invested abroad or in commodities (Barr and Diamond, 2008). The eastern European experience defied these assumptions.

Finally, private pillars were expected to increase fiscal revenues by motivating workers in the shadow economy to formalize their status because contributions they paid would not be redistributed, but would instead be placed in workers’ own private accounts (Ferge, 1999; Müller, 2008). This rather optimistic and challenging assumption also proved unrealistic for the CEECs. No empirical evidence of which the authors are aware suggests that it would have been proved true in eastern Europe.

Dealing With the Funding Gap

In theory, transition costs in the form of interest on new debt should be financed by a “solidarity tax” (Sinn, 2000) on the earnings of those who entered the second pillar, to meet the criteria of intergenerational justice (that is, those who will benefit from the second pillar should pay for the transition). 22 In practice, however, intergenerational justice was a less important concern for reformers. Pension reform was thus largely financed from taxes on general income and consumption and from the sale (privatization) of assets produced by current and retired workers. Moreover, transition costs included not just the interest on new debt but the funding gap itself. Transition costs thus could have been covered through a variety of means: issuing state bonds, increasing taxes, cutting spending in the general budget or the PAYG scheme, or by using exceptional revenues such as receipts from privatization of state-owned enterprises.

The full funding gap thus represented a real liability that states needed to face—on top of the long-term costs of demographic aging and short-term fiscal challenges. The size of the funding gap often proved to be higher than originally expected by the first-wave reformers. The reformers underestimated the number of individuals who would enter the second pillar. 23 In addition, as discussed above, pension privatization was expected to be at least partly self-financing and even to generate additional revenues by increasing output and formal employment. This supposition has proved to be excessively optimistic. By late in the first decade of the 2000s, the contributions diverted from the first pillar amounted to more than 1 percent of GDP annually in several first-wave reform countries. 24 An overview of the funding gaps created by pension privatization in eastern Europe is provided in Figure 8.4. The funding gap experienced so far is only a fraction of the deficit that will occur once the baby boomer generation retires. 25 Most of the pension expenditures associated with the retirement of the 1950s cohorts will have to be covered from the PAYG system. Yet the younger cohorts entering the labor market in the coming years will not only be smaller but will also have the possibility (or obligation) of opting out of the PAYG system.

Figure 8.4The Funding Gap: Actual Developments

(Percent of GDP)

Sources: Estonia:, and World Bank World Development Indicators (WDI); Hungary: Hungarian Central Administration of National Pension Insurance and Central Statistical Office; Latvia: Latvian Financial and Capital Market Commission, and WDI; Lithuania: personal communication with Prof. Teodoras Medaiskis, and WDI; Poland: Epstein and Velculescu (2011); Slovak Republic: Slovak Social Insurance Agency, Ministry of Finance of the Slovak Republic, and Bureau of Statistics of the Slovak Republic.

1 The fluctuation of the funding gap in Hungary is influenced by changes in financial reporting between 2007 and 2008.

2 Hungary’s funding gap decreases to zero following the quasi-nationalization of the second pillar.

None of the first-wave reformers provided a credible financing plan to cover the cash shortfall for the time until the World Bank–type pension system reaches maturity. The policies laid out to fund the transition costs were either politically unfeasible (i.e., cutting pensions payable from the PAYG system in Poland) or insufficient for covering the expenses of the several-decades-long transition (i.e., privatization in the Slovak Republic). As a consequence, debt financing proved to be a major option among all the reformers. This was not an option of choice but, for the reasons mentioned above, an option of last resort and least resistance. The reformers were also working under the assumption that the debt resulting from financing the funding gap would be discounted from total debt levels because it resulted from making the implicit debt explicit. Relying on debt financing became more difficult after the global financial crisis.

The region’s reformers appeared to go through a gradual learning process about the funding gap. This process led to more consideration of how the reform was to be financed as well as of the overall size of the second pillar in the late-reformer nations. 26 In early-reformer Hungary, the dominant argument was that the second pillar would be a self-financing reform, because the savings accumulated in the funds would increase the country’s capital availability, which would, in turn, lead to higher economic growth. In addition, it was widely believed that the mandatory private scheme would be instrumental in combating the shadow economy because pension contributions would be accumulated as savings in individual accounts (Ferge, 1999; Müller, 2008). Polish reformers expected that the funding gap would be covered by cuts in the first pillar and income from privatization, but none of these plans were fulfilled. The major reformer of the mid-2000s, the Slovak Republic, used income from the privatization of the national gas company, equivalent to approximately 4.4 percent of the country’s 2005 GDP, to cover the funding gap in the first five years following reform (Drahokoupil and Domonkos, 2012). The latest of the region’s reformers, Romania and the Czech Republic, embarked on reforms that allowed considerably fewer cohorts to opt for the reformed system (see Table 8.1). The ability to deal with the funding gap was a major concern for Czech reformers, who implemented reform in the context of austerity politics. The decrease in the cash flow of the Czech PAYG system was supposed to be covered by a value-added tax (VAT) increase.27 The second pillar (operational since January 2013) has remained voluntary, and includes the requirement that those who opt for membership in the plan pay an additional 2 percent of contributions into their pension accounts.

In the CEECs, the learning process seems to have created a consensus among pension privatization supporters that a financially sustainable second-pillar system should restrict any annual funding gap to no more than 1 percent of GDP. This rule of thumb informed pension privatization in the Czech Republic and reform reversals elsewhere in the CEECs.28 Policymakers in the rest of the region seem to be ready to accept higher transition costs and to finance them through current revenues. The next section takes a closer look at what benefits these costs actually pay for.

The Diversification Argument

Experience with pension privatization in the first wave of reforms has generated a learning process, which has dispelled some of the reform myths and created awareness of the funding-gap problem. As a consequence, the immediate budgetary impact of the pension reforms was more carefully evaluated and the use of pro-second-pillar arguments shifted. The idea that pension privatization protects against the fiscal shock caused by demographic aging became less frequent; instead, a diversification argument, emphasizing the weaknesses of the public PAYG system, gained prominence. Accordingly, the main advantage of the second pillar is that it diversifies the sources of pension income for individuals (Bezděk and others, 2005; Burdová, 2010; Égert, 2012). The frequent use in the respective debates of a quote from Cervantes’s Sancho Panza about not putting all one’s eggs in one basket indicates the influence of Organization for Economic Cooperation and Development studies that included the Quixote story (e.g., Whitehouse, D’Addio, and Reilly, 2009; as pointed out by Vostatek, 2012).

Diversification refers to a technique of reducing risk by investing in a variety of assets with less than perfect correlation, thus reducing the risk associated with the performance of an investment portfolio. A multipillar system does indeed reduce the reliance of an individual on one form of retirement insurance and thus diversifies an individual saver’s exposure to risks. However, as discussed in the previous section, this diversification comes at the high price to the government of increasing explicit debt or of direct spending to cover the funding gap. The operation thus weakens the fiscal position of the state, effectively undermining not just public pension insurance, but potentially other forms of social insurance upon which an individual might rely.

Moreover, for pension systems, the benefits of diversification as a risk-reducing mechanism are limited. First, private pension pillars are not immune to regulatory risks, as often assumed. Second, and perhaps more important, diversification cannot protect against the main challenge faced by pension systems, namely, the effects of macroeconomic shocks, particularly that stemming from demographic aging.

Proponents of pension privatization, however, appear to refer primarily to protection against what could be called regulatory risks—that is, reliance on the state to provide old-age insurance. The introduction of a funded pillar is believed to decrease savers’ exposure to the risk of any future government defaulting on its implicit pension liabilities by arbitrarily changing the benefit formula of a defined-benefit (DB) PAYG system and thus lowering pensions—because pensions from the second pillar are provided by private entities based on a contractual obligation. Monthly pensions from funded systems are determined by standard actuarial calculations, taking into account life expectancy at the age of retirement and accumulated wealth in the individual account. Moreover, it is often argued by the advocates of second-pillar reforms (e.g., Urban, 2012) that, from the citizen’s point of view, owning explicit debt in the form of government bonds should be preferred to a mere promise from the government. Explicit liabilities stem from a legal relationship in which the state does not have the prerogative to unilaterally change the size of the debt. By contrast, so it is claimed, governments can default rather flexibly on part of their implicit debt, for example, by changing the benefit formula in a DB system.

The problem with this argument is that all pension arrangements are vulnerable to bad government. Although a complete default on liabilities to pensioners relying on the PAYG system is very unlikely, even under extreme conditions, a partial default on government liabilities may easily take place, for either explicit or implicit liabilities. A good example is an inflation tax or a change in the taxation of interest earned on state securities (Holzmann, Palacios, and Zviniene, 2004). Therefore, although the introduction of the second pillar limits the degree to which governments can alter their pension liabilities by amending social security legislation, members of the funded pillar are exposed to partial default of the state through other mechanisms—even when pension portfolios are not exposed to government bonds, a situation that was not avoided in eastern Europe. Taking this into account, it can be concluded that fiscally stable states are crucial for both public and private pension pillars.

Turning to the second point, diversification as an investment strategy applies at the individual level when the goal is to eliminate unique risk. Nevertheless, both the first and second pillars are exposed to similar macro-level challenges. Because macroeconomic shocks are essentially systemic risks, portfolio diversification is not a useful remedy. In fact, the way the diversification argument is used in eastern Europe is based on skepticism about the state’s ability to pay adequate pensions in the future. The most important question remains whether different systems manifest a different level of resilience to the shock of demographic aging. The diversification argument appears to be a new embodiment of one of the old myths of pension privatization, namely, that it allows insurance or hedging against the demographic shock. However, a possible output shock caused by demographic aging will hit the economy as a whole, and therefore represents a threat to the standard of living of pensioners regardless of the type of scheme in which they have been enrolled (Barr, 2012). If demographic aging leads to a decrease in aggregate output, both the implicit return in the public PAYG pillar and the explicit returns gained by pension funds investing in financial markets are likely to suffer.

The essential problem of demographic aging from the perspective of economic theory is not insufficient budget revenue, but a decrease in aggregate output (Barr, 1979; Eatwell, 1999; Barr and Diamond, 2006). If the aggregate output produced by a small workforce is not enough to sustain aggregate consumption at a desired level, then a shift to the second pillar cannot in itself be a remedy to the adverse consequences of this output shock on the living standards of the elderly (Barr, 2012). Therefore, a mere outsourcing of pension insurance to private entities cannot solve the problem. The only measures that will efficiently combat the adverse effects of demographic trends on GDP are those that address demographics directly (for example, increasing labor participation, fertility, immigration, or retirement age) or that improve the productivity of the relatively smaller workforce (for example, investment in education and new technologies).

Future pensioners cannot avoid the impact of a decline in aggregate output resulting from a demographic shock by merely joining the second pillar. Although it is true that a funded pillar is not dependent on current budget revenue, it is not resistant to the adverse effects of demographic aging on the economy. The funded system will also be affected by the aging shock, but the transmission mechanism will be different from the one that acts in a PAYG system. If a country runs a PAYG system, then decreasing aggregate output will lead to a decline in aggregate wages and pension contributions. To maintain the balance of the PAYG system, the state will have to decrease pensions.

In fully funded pension systems, the retirement of a generation larger than the next generation causes either inflationary pressure—because pensioners’ consumption exceeds the desired savings of the workers—or a reduction in asset prices because the supply of assets by the retiring generation exceeds the demand for assets by workers. The two mechanisms would negatively affect second-pillar pensions by lowering either their real or nominal levels. However, if aggregate output does not decrease despite demographic aging, then pensioners’ growing aggregate demand for goods will not outstrip the aggregate supply of goods. Similarly, growing aggregate supply on the assets market will be matched with the growing aggregate demand for assets by workers earning more than they had in the past (Barr, 2000).

The centrality-of-output argument (Barr, 2000; Barr and Diamond, 2006) explains why pension privatization in itself cannot serve as a solution to demographic shocks in a closed economy. However, the second pillar could, in principle, permit investments in pension portfolios in economies not yet affected by adverse demographics. It could thus serve as a device to allow the population of aging countries to prepare for the consequences of a demographic shock by purchasing assets in “young” nations, 29 that is, investment in less-developed and least-developed regions of the world. 30 However, such investment goes hand in hand with greater political risks.31 At least in the current era, favorable demographic development and political risks—including political instability, sovereign risk, the weak rule of law, inadequate shareholder rights, and restrictions on profit repatriation—are positively correlated. Empirically, investment in emerging markets might lead to higher returns, but it also entails higher volatility of investment outcomes (Bebczuk and Musalem, 2009). 32


Pension privatization in eastern Europe has generated a learning process about the actual implications of privatization and the problems associated with running privatized pension pillars based on individual accounts. Many of the problems of second pillars, as they were designed in the first wave of reforms, could be resolved by better regulation, which could, for instance, reduce the administrative costs of managing individual accounts. Furthermore, policymakers can use regulation to avoid the circular transaction through which second-pillar savings are invested in state bonds that actually finance the transition deficit of the first pillar. The main lesson from the first wave of reforms, however, seems to be about the nature of the costs involved in making the transition from the PAYG system to a mixed system. Because the difference between explicit and implicit debt matters a great deal in the real world, the consequences of the transition from a PAYG system to a funded one are much more severe than assumed in many of the theoretical models that underpinned the first wave of pension privatization. Policymakers have become well aware of these costs. In the longer term, policy toward the second pillar has been conditional on the calculation of the long-term benefits that privatization brings along with its concomitant costs. Arguments about the costs and benefits have become more balanced given that many of the arguments used in the first wave of reforms turned out to be myths.

However, the reform outcomes of the second wave are not necessarily optimal. In Hungary, nationalized pension savings were not used exclusively for the repayment of government debt. The new consensus on the desirability of smaller second pillars—which appear to be more sustainable—is difficult to justify given the large fixed administrative costs of running individual accounts with individual choice of pension provider.

Moreover, the prominence of the diversification argument in the recent discussion on reforms suggests lessons are still to be learned. As discussed above, the merits of the diversification argument are questionable in a number of respects. What is more, the argument, as typically used in the region, often starts from the need to address future fiscal challenges for public budgets caused by demographic aging. It is thus based on the great myth of the first wave of reforms—that privatization can address the problem of an aging population. More generally, diversification as an insurance tool does not apply to the main challenges faced by retirement insurance in the 21st century.

In fact, the other rationale behind the diversification argument appears to be beyond the analytical apparatus of positive economics or the social sciences: it is the normative bias against collective insurance solutions and the state in general. A deep mistrust of the state and distaste for collective solutions may make pension privatization seem worth the price (see Vostatek, 2012, on the Czech policy discourse). Paradoxically, pension privatization as an ideological quest to reduce reliance on the state is likely to be a self-fulfilling prophecy: it produces unnecessary fiscal pressures that are likely to reduce the capacity of the state to deliver social insurance and compensate for market failures. For this reason, the normative rationale for diversification is not found by these authors to be a wholly credible one.


“Reform reversals” is used to refer to policy adjustments discussed in this chapter because the term has become widespread after its introduction by the World Bank. However, the term is used as a policy description and does not imply any possible value judgment.

By late in the first decade of the 2000s, the overwhelming majority of central and eastern European countries had national pension systems composed of three parallel schemes, also known as three pillars. In the three-pillar model, the publicly run pay-as-you-go scheme is known as the first pillar. Mandatory private accounts allowing savers to partially opt out of the public system are known as the second pillar. Finally, voluntary private accounts are generally referred to as the third pillar of the pension system.

In the World Bank model, the first pillar was supposed to provide a flat subsistence pension, but none of the central and eastern European countries implemented the pure model. The implementation of pension privatization often went hand-in-hand with linking pensions from the public PAYG scheme to past covered income.

In Hungary, a coalition led by the conservative right-wing Fidesz came to power in 1998, only six months after the three-pillar model had been introduced by the social democrats. The government stopped the planned increase in the second pillar contribution rates from 6 percent to 8 percent of gross wages, but it did not reverse the reform. In the Slovak Republic, when the opponent of pension privatization, the social-democratic party Smer, took power in 2006–10, it implemented only minor changes in regulation of the second pillar. In Poland, the government of the Democratic Left Alliance, a party that voted against privatization in 1999, did not reverse the reform when it took power in 2001, nor did the conservative-nationalist coalition government led by the Law and Justice Party beginning in 2005.

Bulgaria and Croatia, which were the exceptions, apparently prioritized the financing of second rather than first pillars, as indicated by particularly low aggregate income replacement rates in their first pillars.

Romania, another late reformer, continued with its initial plan of increasing contributions to the second pillar, following a break in 2009. The contributions should grow from 2 percent of covered earnings to 6 percent by 2016. Nevertheless, as of 2013, the country is still half a percentage point behind the original schedule, with contributions to the second pillar reaching 4 percent instead of the 4.5 percent initially planned.

The CEECs comprise the Czech Republic, Hungary, Poland, the Slovak Republic, and Slovenia.

The savings accumulated in the private funds were automatically transferred to the state in June 2011. Although savers had the option of staying in the private system, the conditions set by the new law were so unfavorable that only 3 percent chose to remain (Simonovits, 2011). In January 2012, all pension contributions, including those paid by savers who remained in the second pillar, began to flow into the public PAYG system.

Contrary to the initial plans, however, the pension wealth redirected to the state budget was not used exclusively for the repayment of government debt. Approximately 1,200 billion forints (42 billion euro), 43 percent of the pension wealth accumulated in private accounts, was used for purposes other than servicing Hungarian state debt. These expenditures included the acquisition of the country’s biggest oil refinery and covering the revenue shortfall caused by a governmental tax-cut program.

The Slovak left-wing government that took office in April 2012 decreased second-pillar contributions from 9 percent of covered income to 4 percent in September 2012. However, the Slovak reform reversal looks toward a gradual increase in the contribution rates to the second pillar. From 2017 onward, contributions should grow by a quarter of a percentage point per year until they reach 6 percent in 2024.

A right-wing coalition gained an exceptional majority in May 2010, making the reform breakthrough possible.

In this context, the circular transactions in which sovereign bonds issued to cover transition costs were bought by pension funds attracted attention in Poland and Hungary.

The change of party positions on pension privatization can be attributed to the learning process about its actual implications and to a better understanding of the difference between implicit and explicit debt discussed in the next section (Drahokoupil and Domonkos, 2012).

A formal request by countries that had introduced second pillars (including Hungary, Poland, and Sweden) to exempt the transition costs of private funded schemes from the criteria was rejected by EU institutions in 2004–05 (see Eurostat, 2004). In 2005, the European Council allowed transition costs to be exempted from the Maastricht debt criteria for a five-year transitory period and only at a declining rate (European Council, 2005, Art. 3.4). In 2010, the European Commission rejected another request of nine member states, most of them eastern European countries. However, it allowed for some discretion in starting the procedure against countries violating the criteria, but no explicit promises or amendment of criteria were made (for details, see Kovacheva, 2010; Kovacheva and Marini, 2010; Economic and Financial Committee of the European Union, 2012).

Assume an economy in which, beginning in a given year, the mature PAYG system is entirely replaced by a fully funded system available for the younger part of the workforce. Once the last cohort relying purely on the PAYG pillar dies, the fully funded system becomes a mature pension scheme. In this hypothetical example, the pension reform would no longer increase the funding gap in the year in which the first cohort that exclusively relies on the funded scheme for its pensions enters into retirement.

This perspective was also prominent among policymakers who were supportive of the second pillar and whom the chapter authors interviewed in 2011–12.

It can be argued, however, that in a closed economy, gains by households from higher interest rates on their pension portfolios will equal losses on their nonpension portfolios that will (indirectly) also include government bonds (Kubíček, 2008).

The year-over-year losses as of October 15, 2008, amounted to 30.5 percent in Estonia, 35.0 percent in Hungary, 48.4 percent in Lithuania, and 12.4 percent in the Slovak Republic (World Bank, 2008).

In 2005, administrative charges in Poland were estimated to be about 18 percent of accumulated savings (World Bank, 2005). Polish regulation at the time was comparatively restrictive, but it gradually became even more restrictive.

Management fees were adjusted to what seemed to be relatively moderate levels in the Czech Republic, Poland, and the Slovak Republic. In the Czech second pillar, management fees vary greatly among the various types of funds, ranging from 0.3 percent of net asset value (NAV) annually in the government bond fund to 0.6 percent of NAV and 10 percent of the annual return in the dynamic fund. The Slovak second pillar has relatively low management fees, reaching 0.3 percent of NAV annually. However, this fee is further increased by 1 percent of annual contributions and 10 percent of portfolio returns. In Poland, asset-based management fees range from 0.023 percent per month to 0.045 percent per month, depending on the size of the fund, but they cannot exceed 15.5 million zloty per fund monthly. Furthermore, the management fee also has a contribution-based component of up to 3.5 percent of yearly contributions. By contrast, second-pillar management fees in Estonia can amount to as much as 2 percent of NAV and 3 percent of contributions per year (World Bank, 2011).

A perfectly rational household should treat its second-pillar portfolio as a mere explicit valuation of implicit PAYG claims that were lost because of privatization. As a consequence, it should not lower its other savings. If the assumption of perfect rationality is relaxed, households might lower savings held outside the pension funds in reaction to the new policy because they are likely to consider pension portfolios to be new savings (Kubíček, 2008).

The theoretical no-pain-no-gain scenario implies that governments need to cover only interest on the new debt to deal with the transition costs in a sustainable way. In reality, for the reasons discussed above, governments had little fiscal space or actual willingness to increase the level of public debt to finance pension privatization. And in practice, the actual transition costs were much higher, as discussed above.

This was a particularly important issue in Poland (Égert, 2012). In many eastern European countries a large number of middle-aged and elderly workers chose to join the multipillar system although the reforms were primarily aimed at younger cohorts. For example, in the Slovak Republic, the compulsory saving period before accumulated pension wealth could be annuitized was set at 10 years and think tanks close to the right-wing reform government considered ages 45 to 50 to be the limit for when it would still be sensible to enter the second pillar (SME, 2004). Nevertheless, as of June 30, 2006, almost 90,000 members (approximately 6 percent of all savers) ages 45 and above were enrolled in the second pillar.

The size of the funding gap was especially large in Poland, which undertook a World Bank–type pension reform in the late 1990s and made joining the second pillar compulsory for new labor market entrants.

In the CEECs, the retirement of the 1950s cohorts will take place mostly throughout the second half of the 2010s, but the retirement of the early baby-boomer cohorts have already started to retire.

In Slovenia, the funding-gap problem was a major reason the country did not pursue World Bank–type reform. The minister of finance thus joined the labor unions in opposing the failed 1998 pension privatization proposal.

The lower bracket of the VAT rate rose from 10 percent to 14 percent effective January 2012. Nevertheless, it should be noted that the government did not establish a specific subaccount in the budget that would collect a portion of the revenue from VAT designated to cover the funding gap. Therefore, the border between general budget revenue and revenue collected with the declared purpose of financing the second-pillar reform is blurred. The Czech deputy minister of finance admitted this in November 2012 (Urban, 2012).

In Poland, the amount of social security contributions excluding interest transferred to the second pillar is estimated to be approximately 0.5 percent of 2012 GDP. In the Slovak Republic, the argument that pension reform should not divert more than 1 percent of annual GDP to the second pillar was already present among reformers early in the first decade of the 2000s (Záborský, 2003). In 2013, after the decrease in the second-pillar contribution rate from 9 percent of covered income to 4 percent, the social security contributions transferred to the second pillar will amount to approximately 0.5 percent of GDP.

It should be added that, until now, fund managers in the CEECs’ second pillars demonstrated considerable home bias, and investments outside the home country were normally allocated to OECD member states.

According to the United Nations (2006), in less-developed and least-developed countries, the share of children ages 14 and younger is expected to remain greater than 20 percent, while the share of people ages 60 and older will stay at less than 20 percent until the 2050s. By contrast, in more-developed nations, the share of those ages 14 and younger in the total population will reach 15.5 percent, while the elderly (60 and older) will constitute 33.5 percent of the population.

The term “sovereign risk” is used to include more than just the risk of sovereign default. This term also encompasses the risk that a sovereign would introduce foreign currency regulations or a lower standard of shareholder rights, threatening the profitability of investments.

Demographic diversification may also involve exposure to a possible exchange rate shock at the time of retirement (Barr, 2000). Dissaving pensioners’ need to exchange the currency of the country in which second-pillar funds were invested may lead to depreciation in the exchange rate of such currency and a decrease in the real value of their pensions.

    Other Resources Citing This Publication