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

Chapter 20. Macroeconomic Implications of Pension Reform in Brazil

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
Joana PereiraThis chapter updates and builds on IMF (2012). The author is thankful for comments provided by the Brazilian authorities.


The long-term challenges facing Brazil’s pension system are well documented. Recognizing these challenges, the authorities have sought to advance reforms. Much of the debate has focused on the fiscal implications of the outlook for and possible future adjustments to the pension system. However, different reform options can have very different macroeconomic implications, including for savings and growth, and have different effects on the intra- and intergenerational distribution of wealth. To illustrate these differential impacts and to inform the debate on the issue, this chapter simulates the general equilibrium effects for Brazil of various pension reform options that have been used in countries around the world. All options examined help address the system’s long-term funding gap and are conducive to raising real private savings and growth in the long term. However, the analysis finds that reforms that involve lower mandatory contributions or higher retirement ages have larger effects on output through a boost in labor supply. Meanwhile, reforms focused on reducing benefits would promote growth mostly through a larger impact on private savings.

The Brazilian Pension System

Current Structure and Fiscal Position

The Brazilian public pension system currently comprises three defined-benefit schemes:1 a mandatory private sector regime (Regime Geral de Previdência Social, RGPS), currently covering some 25 million beneficiaries and disbursing the equivalent of about 7 percent of GDP; a mandatory public sector regime (Regimes Próprios de Previdência Social, RPPS), with close to 1 million beneficiaries and disbursements of some 2 percent of GDP;2 and a noncontributory branch for rural workers, disabled people, and other low-income families, which disburses the equivalent of 0.7 percent of GDP (Table 20.1). Benefits are financed from the current proceeds of a payroll tax, paid by employees, of 8 percent to 11 percent of wages; a 20 percent contribution tax by employers (which also finances other social insurance benefits, such as for sickness and maternity); and two other specific taxes.3 Both contributions and benefits are capped in the RGPS, but in the RPPS only future participants will be subject to such rules (see below).

Table 20.1Structure of the Brazilian Pension System
RGPS (private sector workers)RPPS (civil servants)Social Assistance (means tested)
Zero PillarSubsidized contributory regime for urban workers and special regime for rural workers. Includes survivor and disability pensions. Eligibility based on age: 60 for men; 55 for women.Noncontributory regime for the poor, including disability pensions.

Eligibility based on age: 65.
First PillarContributory regime. Includes survivor and disability pensions. Eligibility based on years of contributions (35 for men; 30 for women) or age (65 for men; 60 for women).
Third PillarClosed occupational pension plans, managed by single or multiple employer pension funds.Closed pension funds for civil servants.
Open pension plans, managed by insurance companies.
Sources: Gragnolati and others (2011); and author’s elaboration.Note: RGPS = Regime Geral de Previdência Social; RPPS = Regimes Próprios de Previdência Social. Based on the most recent active rules. Transition rules apply to older cohorts of participants.
Sources: Gragnolati and others (2011); and author’s elaboration.Note: RGPS = Regime Geral de Previdência Social; RPPS = Regimes Próprios de Previdência Social. Based on the most recent active rules. Transition rules apply to older cohorts of participants.

Pension spending in Brazil is very high by international standards (Figure 20.1), considering the relative youth of the Brazilian workforce. Indeed, both the RGPS and RPPS are running deficits—each close to 1½ percent of GDP in 2012—as a consequence of relatively generous replacement rates, a low average retirement age,4 and current indexation rules. The indexation of minimum pensions to the minimum wage is a particularly large driver of overall pension costs.

Figure 20.1Brazil’s Pension Spending in an International Context, 2011

Sources: Clements and others (2013); and IMF staff calculations.

It has been estimated that the pension system faces a net present value (NPV) funding gap of close to 25 percent of GDP during the 2010–30, rising to 100 percent through 2050.5 Under current rules, the financing needs of the social security system should undergo a modest rise in the coming 20 years, when the population is still relatively young. After that, the funding gap will increase sharply as the old-age dependency ratio continues to rise steeply (to more than 60 percent in 50 years, from 10 percent in 2012).

Distributional Features

Important reductions in old-age (and overall) poverty have been accomplished in Brazil through the pension system, particularly since early in the first decade of the 2000s (Gragnolati and others, 2011). Currently, about 40 percent of total pension spending goes to beneficiaries receiving the minimum pension (two-thirds of RGPS beneficiaries), which has more than doubled in real terms since 2000. As a consequence, poverty rates in the 60 and older age group fell to less than 5 percent as of 2008, and there is now practically universal coverage of the elderly population at the benefits phase. In contrast, however, coverage rates in the contribution phase are still estimated to be less than 50 percent (Gragnolati and others, 2011)—albeit comparatively large within Latin America—in part because the system discourages labor formalization at the lowest income level (through high contributory rates coupled with high noncontributory benefits).

When compared with other social assistance programs in Brazil, the pension system consumes a relatively high share of federal resources, even when exclusively considering the part allocated to minimum pensions. Other forms of social support, including unemployment benefits, generate an annual cost of less than 2 percent of GDP, while covering as many as 13 million low-income households through the Bolsa Família program alone (which costs less than ½ percent of GDP). This fact, coupled with the notoriously high contribution rates, imply a strong distributional bias toward the elderly in Brazil (Gragnolati and others, 2011). Within the pension system, in turn, resources are redistributed from richer to poorer regions (and from urban to rural workers), but the high replacement rates indicate an undesirable allocation of resources to wealthier pensioners (Rocha and Caetano, 2008).

The RPPS has traditionally offered especially advantageous conditions, including very high replacement rates (still equal to 100 percent for participants who started service before 2003, compared with an average of the best 80 percent of monthly salaries during the working life in RGPS),6 a short entitlement period (only 10 years of civil service to qualify for an RPPS pension), and the indexation of pension benefits to the salaries of active civil servants instead of inflation. This explains why deficits in the two subsystems are of the same order of magnitude, even though the RGPS has much wider coverage.

Recent Reforms

Important changes to the RPPS were first enacted in 2003, including steps for the ongoing establishment of a dual-pillar system (2012 reform). Faced with mounting pension costs (a 1998 pension reform had a relatively limited impact on curbing deficits) and rising external risk premiums, a reform of the pension subsystem for civil servants was introduced in 2003 to improve its long-term fiscal prospects. The reform introduced a number of parametric changes: an 11 percent tax on pension benefits,7 lower replacement rates (harmonizing the rules with RGPS for new civil servants), and a penalty for early retirement of 5 percent of benefits per year (before age 60 for men and 55 for women).8 It also set the stage for the creation of a fully funded pillar for public servants, which was finally approved by the Senate in March 2012 and implemented in February 2013.

The 2012 reform introduces a defined-contribution pillar to the RPPS.9 Benefits and contributions for new civil servants will be subject to the same ceilings as those in the RGPS, while participants have the option to enroll in a complementary defined-contribution scheme (Previdência Complementar) if they wish to receive a pension greater than the ceiling. Previously active civil servants may choose to stay in the old system or switch to the new two-pillar system. Participants can choose how much to put into a retirement account, knowing that the employer will match their investments by up to 8.5 percent of the portion of their salaries that exceeds the RGPS’s ceiling. Upon retirement, they receive the returns from this investment.10

During a long first stage of implementation, the 2012 reform will generate a net cost driven essentially by the loss of contributions to the pay-as-you-go (PAYG) pillar. The state will also be making transfers to the individual pension accounts on behalf of employees. However, because the reform affects only the RPPS subsystem and the contribution ceiling is relatively high—only one-third of civil servants earn salaries above the base—the transition cost is expected to be contained (about 0.1 percent of GDP per year).

As new generations of civil servants retire and receive lower disbursements, the government will reap the benefits of the 2012 reform. In net terms, the authorities expect an improvement in the balance of the RPPS from 2033 onward and a neutral fiscal impact in the mid-2040s, with gains rising to 0.4 percent of GDP per year in the long term. IMF staff estimates point to an overall impact of about 10 percent of GDP in NPV terms in the long term.

The introduction of a funded pillar into the RPPS is advantageous. The reduction of replacement rates for higher earners is expected to encourage long-term private savings and thereby support the development of financial markets. The progressivity of the RPPS system is also enhanced, as is equity with private sector workers. Finally, the relatively small transition cost is important for the sustainability of the reform—especially given that the fiscal framework in Brazil is anchored by a primary surplus target—in light of international experience in which costly pension transitions have, at times, led to some unwinding of pension reforms. Brazil’s reform may thus be a stepping stone for further improvements to the system.

The next section comments on the macroeconomic effects that this reform is expected to generate, paying particular attention to the role that the choice of financing strategy may have on the prospect for increasing savings at a national level. It is followed by a section that discusses two other types of reforms that have commonly been adopted to address long-term pension system imbalances—benefits cuts and increases in retirement age. Before concluding, the chapter briefly discusses the equity implications of the various reforms.

Macroeconomic Implications of the 2012 Reform

The chapter now assesses the broader macroeconomic implications of the recent reform. The analysis uses the IMF’s Global Integrated Fiscal and Monetary (GIMF) model parameterized on data for the Brazilian economy.11 The GIMF is a non-Ricardian, dynamic stochastic general equilibrium model that features overlapping generations, finite horizons (myopia), and endogenous labor and capital markets, allowing for a meaningful discussion of the short- and medium-term impacts of pension reforms.

The baseline is an economic environment reflecting fiscal trends in Brazil before the collapse of Lehman Brothers and the ensuing global financial crisis. In particular, data as of 2007 were used to parameterize initial levels of government spending, revenue decomposition, and transfers (including pensions), thereby abstracting from the impact of the crisis on these variables. Net public debt is assumed to be 40 percent of GDP in the initial steady state.

The 2012 reform is introduced as a shock, first to contribution rates and later to pension benefits. By capping mandatory contributions to the PAYG pillar, the government will effectively be lowering average mandatory contribution rates for public servants. Based on the estimated transition cost shown in Figure 20.2, the analysis proxies that change by the shift in labor taxes that, in the model, would produce such a cost (up to its peak in 2035). In other words, the analysis assumes that contributions to defined-benefit schemes are generally perceived by participants to be a tax, in contrast to an optional defined-contribution plan.12 After 2035, the average contribution rate is fixed and the fiscal trajectory thereafter is dictated by the reduction in pension benefits for new entrants.13 As will be shown, the quantitative impact of the reform is small in broad macroeconomic terms, but this outcome is only a consequence of the circumscribed scope of the reform (the limited number of affected beneficiaries). The results do suggest a high elasticity of private savings and growth rates to the implied fiscal savings for this particular reform.

Figure 20.2Estimated Fiscal Impact of the 2012 Reform

(Percent of GDP)

Source: Ministry of Finance.

Macroeconomic Impact When the Reform Is Financed by Public Debt

For illustrative purposes, the exercise first analyzes the effects of the 2012 reform under the assumption that it is financed by public debt. The resulting path of the relevant fiscal variables, in deviations from the baseline scenario, is shown in Figure 20.3. Primary balances worsen in the first 20 years and improve subsequently, as in the estimated net costs and benefits shown in Figure 20.2. Implicitly, the initial increase derives from the reduction in the average contribution rate of 0.2 percentage points by 2027. Pension benefits start falling on that date, and the system matures with disbursements 0.4 percent of GDP lower than the baseline. Because public investment does not change and the impact on GDP is small (see below), net public savings mirror the dynamics of the overall deficit.

Figure 20.3The 2012 Reform with Debt Financing: Dynamics of Fiscal Variables

(Deviations from steady state; periods correspond to years)

Source: IMF staff calculations.

Labor supply increases, pushing up real GDP growth (Figure 20.4). The drop in compulsory contribution rates reduces a labor market distortion, raising marginal incentives to work and thereby increasing the return on capital.14 Real investment is thus higher, although with some recoil in the medium term because real interest rates rise during that period.15

Figure 20.4The 2012 Reform with Debt Financing: Impact on Macroeconomic Variables

(Deviations from steady state; periods correspond to years)

Source: IMF staff calculations.

As expected, the reform creates incentives to increase private savings. Faced with higher net income during their working lives and foreseeing lower pension transfers in the future, individuals accumulate savings during the next 30 years, incidentally in the form of optional contributions to the second pillar of the pension system. The subsequent drop in government transfers reduces disposable household income and thus the ratio of private savings to GDP, but in NPV terms, private savings increase. Because agents are myopic and a share of the population is liquidity constrained, consumption is not perfectly smoothed. Furthermore, private savings undershoot in the medium term because the long-term decline in pensions is not fully internalized by the currently active population.

National savings, however, stay roughly constant during the transition period. Initially, higher household savings are simply traded off for government debt accumulation. However, as pension benefits drop permanently, so do private savings, the public debt ratio, and interest payments. Therefore, national savings rates increase slightly in the very long term, owing almost exclusively to lower public debt service. With the investment ratio mostly constant over the entire period, the current account improves permanently after the transition period.

Although a low impact from multipillar reforms on total savings has been observed in a number of emerging market economies (Figure 20.5), the existing evidence is far from conclusive.16 In countries such as Chile, Peru, and Latvia, national savings rose in the aftermath of reform, but in other cases—including Colombia, Mexico, and Uruguay—it either remained unchanged or dropped slightly. The relationship between these types of reforms and savings is nonetheless hard to pin down in the long term because savings rates depend on myriad factors. In practice, reform packages often include parametric changes, such as increases in the retirement age, that have adverse effects on private savings (see below). Furthermore, an important determinant of the impact on savings is the financing strategy for the transition cost, as argued in the next paragraphs. Finally, the relatively short interval since most multipillar reforms were introduced—particularly in central and eastern Europe—makes it hard to fully assess the impact on household savings and labor incentives. Indeed, reforms of pension systems in the 1990s in advanced economies such as Sweden have been associated with increases in household savings, but these behavioral changes have only been observed gradually over long periods.

Figure 20.5Trends in Gross Domestic Savings

(Percent of GDP)

Sources: World Bank Independent Evaluation Group (2006); and World Bank World Development Indicators (2011).

*Latest available data point.

Macroeconomic Impact When the Reform Is Financed by Government Savings

Given the primary surplus fiscal target used in Brazil, one may alternatively assume that the transition cost will be financed by a reduction in government consumption.17 Labor taxation and pension benefits still follow the same path as before, but government consumption now mirrors their combined budget impact, such that the primary balance remains unchanged. Thus, changes to the government’s overall debt balance solely reflect the small variation in interest rates.

As before, the fall in contribution rates promotes labor supply, investment, and real growth; however, the impulse to national savings turns positive at all dates (Figure 20.6). Tighter fiscal balances (lower weight of the public sector) over time render the impact on aggregate labor supply stronger in the longer term, which, in turn, brings up real GDP by more than double the amount found when the reform is financed through debt. At the same time, the decline in government consumption keeps government savings close to the baseline level. National savings will then rise in the short to medium term, arguably in the form of increased savings in household retirement accounts.

Figure 20.6The 2012 Reform with Constant Primary Surplus: Impact on Macroeconomic Variables

(Deviations from steady state; periods correspond to years)

Source: IMF staff calculations. Note: CPI = consumer price index.

Macroeconomic Implications of Alternative Pension Reform Options

Reflecting existing high costs, which will be exacerbated by the demographic transition, further adjustments to the social security system will be needed. Efforts will likely need to focus not only on the RPPS—toward an equalization of regimes for public and private sector workers—but also on the RGPS. Although less generous, the private sector subsystem covers a much wider range of the population and is therefore bound to be most affected by population aging. Furthermore, as discussed above, the NPV of the 2012 reform is modest when compared with the actual pension gap.

The chapter now presents some illustrative simulations of the macroeconomic effects of possible alternative approaches. For the purposes of this analysis, the focus is on parametric changes to the PAYG systems that reduce their financing needs, other than those associated with an expansion of the defined-contribution pillar. However, the simulations from the previous section suggest that an expansion of the defined-contribution pillar could be beneficial (by further reducing the contribution ceiling in the first pillar and increasing the importance of pension savings accounts), should the government identify fiscal space to finance the transition cost.

With a medium- to long-term horizon in mind, the analysis uses average pension spending in Group of 20 countries as an indicative benchmark for Brazil. Convergence to such an average would result in a decline in social security disbursements in Brazil of about 2 percent of GDP, practically eliminating the projected social security deficit, barring aging pressures. These simulations assume that such reforms could be phased in during the next 20 years. This gradual implementation would be associated with higher sustainability of the reforms over time, which Brazil can afford because it still enjoys the demographic dividend of a young labor force and a low overall old-age dependency ratio. Two types of general, stylized instruments are considered for convergence to such a benchmark—lowering benefits and increasing the retirement age—and in both cases it is assumed that the government keeps its primary surplus target unchanged.18

A number of specific policies could be adopted to achieve a direct reduction in benefits. In the short term, addressing the generosity of survivor benefits is in order. Survivor pensions have a replacement rate of 100 percent and are not contingent on wealth, income, age, or remarriage status.19 In total, survivor benefits account for about one-quarter of total pension spending by the federal government (RGPS and RPPS). Furthermore, a change in the minimum pension indexation rules could be considered. Replacement rates could gradually be aligned with international practice—average net replacement rates in the Organization for Economic Cooperation and Development stand at about 60 percent, while they approach 100 percent in Brazil. Revisiting the formula of the fator previdenciário20 (and extending the concept to the RPPS) is a possible avenue for doing so on a permanent basis given that it would automatically adjust benefits to life expectancy. Although originally meant to discourage early retirement, the introduction of the fator has instead mostly contributed to a reduction in pension spending.

Average retirement ages could be increased directly by hiking the minimum retirement age or by penalizing early retirement (for instance, extending the minimum years of contributions required to qualify for a full pension). The government could also consider eliminating the compulsory retirement age of 70 years. It has been estimated—based on the current level of pensions and life expectancy at average retirement age—that the average retirement age would need to be increased by three years to produce savings of 2 percent of GDP. Assuming a 20 percent lower labor force participation rate for workers older than the current average retirement age, such an increase corresponds approximately to a 7 percent increase in the workforce. As a first approach, existing gender differences in the qualification criteria could be reduced. Originally meant to compensate for maternity, earlier female retirement ages will become increasingly costly with the rise of female labor force participation in Brazil, and can be replaced by better targeted maternity support programs.

The decline in benefits raises private savings,21 investment, and labor supply (Figure 20.7), although with a modest gain in real GDP of 0.8 percent in 20 years.

Figure 20.7Decrease in Pension Benefits: Impact on Macroeconomic Variables

(Deviations from steady state; periods correspond to years)

Source: IMF staff calculations.

When the reform is announced, current workers and beneficiaries internalize the permanent decline in future pension benefits. Thus, consumption immediately drops and savings rise, putting downward pressure on interest rates. Labor supply increases—because consumption and leisure are complementary goods—which, together with the lower interest rate, encourages higher private investment. As transfers decline, so does disposable income, and consumption will continue converging to a lower level. For the same reason, the private savings rate eases in the long term, but it is still permanently higher than in the baseline. In all, the national savings rate increases on the back of higher private savings and a permanent (albeit small) reduction in public debt service.

By contrast, an increase in retirement ages depresses savings in the short to medium term (Figure 20.8), but has a large positive impact on investment and output growth (a 6 percent increase). Agents foresee a shorter retirement period at the time the reform is announced, and immediately decrease savings. Interest rates go up, but the substantial rise in labor supply improves returns on capital so much that private investment rises in equilibrium. This increases demand pressures, contributing to higher interest rates. In the medium to long term, output rises significantly, as does household disposable income. Thus, private saving rebounds in real terms, although its ratio to GDP is permanently lower than in the baseline. With a slightly higher deficit, the national saving rate falls permanently in this case, although mostly as the result of GDP growth.

Figure 20.8Increase in Retirement Age: Impact on Macroeconomic Variables

(Deviations from steady state; periods correspond to years)

Source: IMF staff calculations.

Equity Implications of the Various Pension Reform Options

Population aging will continue to drive the reform agenda in Brazil, and will require that the pension system be less generous in the future. Although future retirees would be less favored in all options considered in this chapter, the analysis has also shown that economic conditions, and thereby wealth, would improve relative to a no reform scenario (although with more working hours during a lifetime). It is important to note that the choice of reform strategy also carries different distributional effects, both within and across generations.

The 2012 reform will contribute to a more equitable future system. Should it be financed with public savings, it will also enhance intergenerational equity. Replacement rates in the RGPS and RPPS will be almost identical, benefiting both equity and mobility across the public and private sectors for the next generations. Furthermore, the cap on first-pillar benefits and contributions under the RPPS is both intra- and intergenerationally redistributive because it (1) affects high-income earners only, and (2) reduces the contribution rates of young and future workers, who will also receive lower pensions. If the transition cost is financed by current public savings (for example, with a reduction in government consumption), the burden of adjustment is spread across all cohorts alive in the coming 35 years, including current retirees. In contrast, if it is financed with debt, today’s young and future cohorts will be left paying for the full cost of the reform.

Across-the-board benefit cuts are normally inequitable, both within and across generations. To mitigate such an effect, reforms should target the wealthiest for the largest benefit revisions. The comparatively high replacement rates in Brazil, specifically among higher-income percentiles, suggest ample space for doing so. A case in point is survivor pensions, which consume a significant amount of resources and are not conditioned on either income or wealth. Conversely, the revision of minimum pension indexation rules, which is desirable from a sustainability point of view, will most directly affect the poorest pensioners. Finally, equitably increasing retirement ages calls for a similar alignment of statutory retirement ages with life expectancy across the various groups (by generation, gender, profession, and so forth). Other equity concerns, not related to retirement, can be addressed through complementary programs.

For both direct benefit cuts and increases in retirement ages, saving the fiscal windfall from the reform would enhance intergenerational equity by lowering debt, which translates into lower taxes in the future, when pensions would also fall.


Current levels of pension spending in Brazil are high by international standards, particularly given that the country is now enjoying the peak of its demographic dividend. The generosity of the system is believed to provide disincentives for private savings, investment, and aggregate labor supply. If unadjusted, pension spending pressures in Brazil will compromise fiscal sustainability without substantially squeezing discretionary spending or resulting in further increases in already high—by emerging market standards—income and consumption taxes.

This chapter’s simulations suggest important effects on macroeconomic variables, such as savings and growth, from different parametric adjustments that have been used in other countries and that might possibly be considered in Brazil. The chapter discusses the macroeconomic impact of different parametric reform options, starting with the recently approved introduction of a defined-contribution scheme for the public sector subsystem. The analysis finds that pension reforms increase real private savings and growth, although the elasticities to implicit fiscal savings are quite different across the different options. Reforms that involve an increase in retirement ages or a decline in average contribution rates are supportive of higher growth through their positive impacts on labor supply and investment, even if the impact on savings is not necessarily higher than in options that focus primarily on reducing benefits.


    ArnoldJ.2011Raising Investment in BrazilOECD Economics Department Working Paper No. 900 (Paris: Organization for Economic Cooperation and Development).

    AttanasioO. and S.Rohwedder2003Pension Wealth and Household Saving: Evidence from Pension Reforms in the United KingdomAmerican Economic Review Vol. 93 No. 5 pp. 1499521.

    ClementsB.J.D.CoadyF. EichS.Gupta A.KangurB.Shang and M.Soto2013The Challenge of Public Pension Reform in Advanced and Emerging Market EconomiesIMF Occasional Paper No. 275 (Washington: International Monetary Fund).

    DisneyR.2005Household Saving Rates and the Design of Social Security Programmes: Evidence from a Country PanelCESifo Working Paper No. 1541 (Munich: CESifo Group).

    GragnolatiM.O.H.JorgensenR.Rocha and A.Fruttero2011Growing Old in an Older Brazil: Implications of Population Aging on Growth Poverty Public Finance and Service Delivery (Washington: World Bank).

    International Monetary Fund (IMF)2012Brazil: Selected Issues PaperIMF Country Report No. 12/192 (Washington: International Monetary Fund).

    KumhofM.D.LaxtonD.Muir and S.Mursula2010The Global Integrated Monetary and Fiscal Model (GIMF)—Theoretical StructureIMF Working Paper 10/34 (Washington: International Monetary Fund).

    RochaR. and M.Caetano2008O Sistema Previdenciário Brasileiro: Uma Avaliação de Desempenho ComparadaTexto Para Discussão Nr.1331 (Brasilia: Instituto de Pesquisa Econômica Aplicada).

    World Bank Independent Evaluation Group2006Pension Reform and the Development of Pension Systems An Evaluation of World Bank AssistanceIndependent Evaluation Group Studies (Washington: World Bank).

In addition, a network of private pension funds (mostly defined contribution) is growing. Participation in these schemes is voluntary, and the government plays only a regulatory and monitoring role. An optional defined-contribution pillar was introduced in 2013 in the public sector, and is discussed in the chapter.

Applies to the federal government’s subsystem only. Consolidated state and local government programs consume a similar amount of resources and cover 2.5 million beneficiaries.

The 20 percent employer contribution is not explicitly paid for by the employer (the government) in the RPPS. The Contribution to Social Security Financing (Contribuição para o Financiamento da Seguridade Social, COFINS) and the Social Contribution on Net Profits (Contribuição Social sobre o Lucro Líquido, CSLL), paid by businesses, are the social security system’s remaining financing sources.

The average retirement age is 53 for RGPS beneficiaries who contributed for a minimum of 35 years (for men) or 30 years (for women). Among retirees who contributed for shorter periods and among RPPS retirees, the average retirement age is close to 60.

The estimates are based on the authorities’ actuarial projections of RGPS and RPPS financing needs as of April 2012, as well as of the fiscal impact of the latest reform (see the “Macroeconomic Implications of Alternative Pension Reform Options” section of this chapter) and IMF staff projections of public pension increases (Clements and others, 2013).

Furthermore, benefits in the RGPS are subject to an adjustment factor (fator previdenciário), based on age and length of contributions, that was introduced in 1999 to account for changes in average life expectancy. However, contributions are uncapped in the RPPS (except for civil servants who joined after February 2013), while they are subject to a ceiling in the RGPS.

Levied on the portion exceeding 60 percent of the RGPS’s benefits ceiling, for all RPPS participants.

The minimum retirement age is 53 for men and 48 for women, provided that participants contribute for the full 35 years for men and 30 for women (to either the RGPS or RPPS).

The 2012 reform applies to federal government employees only. Some large states already had a similar system in place, while several other government agencies are considering adopting it.

The defined-contribution scheme is to be administered by a newly created Fundação de Previdência Complementar do Servidor Público Federal (Funpresp), divided into three branches for servants in the executive, judiciary, and legislative branches, respectively. Members of the Funpresp’s Executive Board and Financial Committee are appointed by the government, but the institutions enjoy administrative independence and are subject to a private legal regime (like public enterprises).

A detailed outline of the GIMF model can be found in Kumhof and others (2010).

The analysis also assumes that the copayments by public employers to the optional pension savings accounts are perceived to be part of the tax rate reduction, and participants would take that into account when targeting a desired pension savings amount.

The GIMF features two types of agents: a group of liquidity-constrained households (LIQ agents), who do not have access to capital markets, and intertemporal optimizers (OLG agents), who can borrow and save. This section assumes that reductions in PAYG benefits affect only OLG agents because, in reality, only the highest earners will be affected.

Labor and capital are complementary factors of production in the model.

With both higher labor supply and capital accumulation, real GDP increases above its long-term trend during the first years, which ends up putting pressure on prices. Monetary policy therefore adjusts by hiking real interest rates temporarily.

The World Bank’s Independent Evaluation Group (2006) notes a generally small impact on national savings in the short to medium term. A number of papers also note a very high (low) substitutability between pension savings with a high (low) actuarial component and other kinds of financial wealth (e.g., Attanasio and Rohwedder, 2003, and Disney, 2005), hinting at the low overall impact of these reforms. However, Arnold (2011) points to the generosity of the PAYG pension system as one of the main causes of low savings in Brazil.

Likewise, when the reform produces a net benefit (after 2035) it is assumed that government consumption rises commensurately. This assumption does not significantly affect the short- to medium-term macro-impulses, but it leads to a lower government saving rate and higher private savings in the very long term.

The analysis adjusts government consumption so that the target is met. Public investment or taxes could be adjusted instead, although the effects on real GDP would be harder to identify. For simplicity in the simulation, the analysis also assumes that pension reductions are evenly spread across liquidity-constrained and -unconstrained agents (see footnote 13).

In the RPPS, the replacement rate falls to 70 percent above a threshold equivalent to the RGPS’s contribution ceiling.

See footnote 6.

In this scenario, the longevity risk is transferred from the government to individuals, who would now rely more on own savings to provide for retirement. Doing so effectively would require developing a well-established annuities market by insurance companies, supported by the appropriate regulation and financial education.

    Other Resources Citing This Publication