Inequality and Fiscal Policy

Chapter 15. Equity Considerations in the Design of Public Pension Systems

Benedict Clements, Ruud Mooij, Sanjeev Gupta, and Michael Keen
Published Date:
September 2015
  • ShareShare
Show Summary Details
Benedict Clements, Csaba Feher and Sanjeev Gupta 


Public pension systems are key instruments of social policy: they prevent poverty after retirement and limit the difference between pre- and postretirement consumption levels, thereby influencing income distribution (Figure 15.1). Public pension spending is also one of the largest items in advanced and emerging economies’ budgets (Figure 15.2); the financing of public pensions can require high taxes and constrain other public expenditures. Pension spending is projected to increase more in the future, rising another 1 and 1¾ percent of GDP in advanced and emerging economies, respectively, between 2014 and 2030 on a weighted average basis (IMF 2015). These schemes influence workers’ behavior toward labor supply, tax compliance, and savings. If designed well, they can improve welfare and equity—but design shortcomings can also impose an unnecessary fiscal burden and welfare losses on current and future generations. The unavoidable trade-offs between the size and redistributive features of these systems versus other public expenditures and fiscal sustainability present governments with difficult policy choices.

Figure 15.1Public Pension Spending and Its Impact on Inequality

Source: IMF staff calculations.

Note: Data labels in the figure use International Organization for Standardization (ISO) abbreviations.

Figure 15.2Public Pension Expenditures

(Percent of general government expenditures)

Sources: IMF (2015); and IMF staff calculations.

To address the impact of aging on pension systems’ long-term fiscal sustainability, advanced and European transition economies introduced various parametric and structural pension reforms during the past 25 years. But sustainability issues still remain in light of demographic developments, including rising life expectancy. In contrast, emerging economies outside Europe, along with low-income countries, face different problems: low pension coverage, special occupational schemes relying heavily on budget transfers, or promises leading to large pension deficits despite favorable demographic conditions. Fiscal constraints and equity objectives cannot be considered in isolation. This chapter provides an overview of equity considerations in the design and reform of pension systems.

The main findings from economic theory and international experience, further elaborated in this chapter, are the following:

First, although the choice of policy objectives and the actual design of old-age income support systems are fraught with difficult choices, certain policies are unequivocally equity enhancing. Most important among these are the expansion of coverage and the uniform treatment of economic sectors (for example, the inclusion of farmers in mandatory social insurance) and types of employment (for example, permitting part-time and informal workers to participate in pension insurance schemes).

Second, design features and reforms entail trade-offs between intergenerational and intragen-erational redistribution objectives; fiscal constraints and fairness; the resources allocated to public pensions and to other expenditures; and primary objectives (that is, consumption smoothing and prevention of old-age poverty) and externalities (such as labor supply responses to high social security contributions). It is crucial that policymakers be aware of these trade-offs, have a clear policy stance to be communicated to the population, and be familiar with the welfare and fiscal consequences of particular design features.

Third, pension policy is formulated under various constraints (fiscal, political, administrative, and the economy’s level of market development) that may require policymakers to settle for second-best options.

Finally, pension policy cannot fully compensate the elderly for inequities suffered before retirement; indeed, pension policy can only be successful if its objectives and instruments are aligned with other welfare systems and are supported by labor market regulations and public education and health policies.

The remainder of the chapter is structured as follows: First, pension systems’ basic objectives and instruments are discussed. Next, an overview of pension systems’ equity features is presented, followed by the introduction of the most common types of public pension arrangements. The chapter’s final sections present the equity impact of various pension reform options, followed by conclusions.

Pension Systems’ Policy Objectives

Public pension systems—and, in a broader sense, old-age income support—have two basic objectives: to protect participants against income poverty and to limit the decline in consumption after retirement. These objectives can be achieved through savings and insurance schemes as well as welfare transfers. Earnings or contributions to the system are important determinants of benefits in insurance schemes, but old-age welfare payments are typically unrelated to a given worker’s employment history, earnings, and contributions. Thus, although earnings-related benefits can provide for consumption smoothing, discretionary social welfare payments are better suited to poverty alleviation. Although both objectives are important, fiscal constraints or societal preferences may lead governments to partially or fully forgo one of them as part of their pension policy. In these cases, consumption smoothing may be delegated to privately managed financial sector providers on a voluntary or mandatory basis, while poverty alleviation may become part of social assistance outside the pension system (Figure 15.3).

Figure 15.3Structure of Public Pension Schemes in 41 OECD and Selected Non-OECD Countries

Source: Organisation for Economic Co-operation and Development (OECD) (2013).

Note: Argentina, Australia, Austria, Belgium, Brazil, Canada, China, the Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, India, Indonesia, Ireland, Israel, Italy, Japan, Korea, Luxembourg, Mexico, the Netherlands, New Zealand, Norway, Poland, Portugal, Russia, Saudi Arabia, the Slovak Republic, Slovenia, South Africa, Spain, Sweden, Switzerland, Turkey, the United Kingdom, the United States.

In developed economies and most European emerging economies, mandatory pension schemes reach near-universal coverage of the elderly population (Figure 15.4), aiming at both poverty alleviation and consumption smoothing. The distributional impact of these schemes is significant: in the absence of public pensions, poverty among the elderly would be four times as high in European Union member states than what is observed today (Clements, Eich, and Gupta 2014). Developed economies’ main instrument for achieving both policy objectives is mandatory participation in social insurance schemes, requiring affiliates to forgo a share of their labor income in exchange for savings and social insurance entitlements. Exceptions exist, however; in Australia, Denmark, Israel, Mexico, the Netherlands, and New Zealand, the state only provides a universal or targeted basic pension approximately corresponding to the poverty-alleviation objective, while encouraging contributions to regulated pension savings or insurance schemes. This arrangement results in lower public pension expenditures and pension benefits as well as greater individual choice regarding intertemporal savings and consumption patterns. At the same time, if voluntary participation in contributory pension schemes is limited—for instance, because of myopia, income poverty, or labor informality—the pressure to increase the basic pension or to augment it with additional social transfers may rise.

Figure 15.4Pension Coverage by Region

Low-income countries are typically younger demographically and have higher levels of informality in the labor market. Mandatory pension coverage is often limited to civil servants, public employees, the armed forces, members of the judiciary, and legislators (Bloom and McKinnon 2013). The total cost of these schemes today is relatively small, but they will need to expand to accommodate population aging and the disappearance of traditional intrafamily and communal income support mechanisms.

The Concept of Equity in Pension Systems

There is no universally applicable, “correct” answer to what makes a pension system equitable. One way to assess whether a pension system is equitable is to determine how closely the distribution of retirement benefits reflects income distribution in active working years. A system paying benefits strictly based on contribution performance may be viewed as equitable since it treats every dollar of contribution in the same way. A flat, general revenue–financed basic pension can be viewed as equitable, too, because it redistributes in favor of people with low lifetime incomes and insufficient pension entitlements or savings. These two approaches are captured by the concepts of vertical and horizontal equity (Table 15.1).

Table 15.1Factors of Equity in Public Pension Systems
Horizontal EquityVertical Equity
Intragenerational EquityStrong contribution-benefit link along the entire income distribution; uniform internal rate of return within cohortsContribution-benefit link weak for low-income scheme members; internal rate of return negatively correlated with contribution performance
Intergenerational EquitySame relationship between lifetime contributions and benefits for subsequent cohorts; uniform rate of return for subsequent cohortsInternal rates of return negatively correlated with affluence; better-off generations receive lower relative benefits and may reduce underfunding
Source: Authors.
Source: Authors.

With regard to intragenerational equity, horizontal equity requires that similar contributions result in similar benefits, whereas vertical equity requires consideration of individuals’ needs (McDaniel and Repetti 1993). A horizontally equitable pension scheme will promise the same total pension benefit to people who have made similar contributions; in other words, it will provide similar internal rates of return—the theoretical interest rate that equates lifetime contributions to expected pension benefits discounted to the time of retirement (Santos and Domínguez 2011). A life-expectancy-adjusted, uniform internal rate of return across the income distribution is horizontally equitable, whereas vertical equity requires internal rates of return that are differentiated according to income or other characteristics.

Intergenerational equity within a pension scheme refers to the manner in which the internal rate of return on contributions compares across generations: intergenerational equity requires that the burden of financing the pension system and the benefits to be paid out be spread fairly across successive generations. It is important to recognize, however, that what this kind of fairness means may be subject to different interpretations. On the one hand, it could be argued that it would be fair to shift a larger absolute burden onto successive generations, because they are expected to have higher real incomes than preceding generations. This approach would reduce the internal rate of return realized by later generations and can be interpreted as seeking intergen-erational vertical equity. On the other hand, fairness could be interpreted as meaning that all generations should face the same relative burden so that the proportion of income that will have to be transferred to the state to finance the pension system remains stable.

The question of intergenerational equity also arises when major paradigmatic shifts happen in the pension system. The introduction and later expansion of pay-as-you-go, defined-benefit pension schemes in postwar Europe and the United States was one such case (Figure 15.5). Early retiree generations often reaped a windfall while the cost of actuarial imbalances was (and still is) borne by subsequent generations of taxpayers and beneficiaries1 (Clements, Eich, and Gupta 2014). Similarly, in a shift from an unfunded to a funded pension system, as occurred in several Latin American and European emerging economies, the transition requires financing the legacy liabilities of the prereform system.2

Figure 15.5Intergenerational Equity in Public Pension Schemes

(Percent of lifetime income)

The Impact of Pension Design Features on Equity

Intragenerational equity is affected by numerous factors (Table 15.2) but coverage stands out as one of the most important features. Partial versus universal coverage results from two factors: eligibility and willingness to participate. Eligibility to participate is a crucially important design feature, because, as opposed to other determinants of income redistribution within the pension system, it is a matter of equal opportunity to partake in the benefits of a pension system, whatever the system’s internal distributional characteristics may be. Partial eligibility to participate may result from the exclusion of certain forms of employment (such as part-time and informal employment) and types of employers (small enterprises and the self-employed) from social security. Partial coverage may also result from the incapacity or unwillingness to participate because of low per capita incomes, large informal sectors, or limited monetization. Weak enforcement may also undermine coverage when data administration, collection, or enforcement capacity is inadequate (Auerbach, Genoni, and Pagés 2007).

Table 15.2Pension System Design Features and Equity Characteristics
Horizontal EquityVertical Equity
Intragenerational EquityUniform scheme rules, gender-specific annuities, linear accrual schedules, no minimum vesting or eligibility criteriaUniversal coverage, minimum pension provisions, progressive adjustments to the calculation base, progressive income taxation of pensions, benefit ceilings
Intergenerational EquityFull funding, benefits adjusted to longevity changes, stable income tax systemInflation-indexed minimum pensions, means-tested basic pensions, debt financing of pension deficits
Source: Authors.
Source: Authors.

Today, low coverage is a problem typical of low-income and non-European emerging economies. This situation may change, however. In European transition economies, where labor market informality, inactivity, and unemployment rates have grown significantly in comparison with the pretransition era of notional full employment, the share of the elderly meeting minimum eligibility conditions will decline in the coming decades (Schwarz and Arias 2014). If governments wish to protect everyone against old-age poverty—including people without sufficient contribution histories—then the contributory principle of the system will become weaker, potentially undermining contribution compliance at higher income levels. Contributions will become increasingly viewed as net lifetime taxes instead of actuarially fair insurance premiums.

Box 15.1Actuarial Balance, Actuarial Fairness, Actuarial Neutrality

A pension system’s distributional features are also closely linked to how assets and liabilities compare at the individual’s, the cohort’s, or the entire scheme’s level. A pension scheme is actuarially balanced if its total liabilities equal its assets, regardless of the time period being assessed. An actuarially fair scheme matches expected pension entitlements to lifetime contributions, whereas actuarial neutrality implies that the marginal net benefit earned by working one year longer remains constant (Queisser and Whitehouse 2006). An actuarially balanced scheme is equitable in an intergenerational sense but being in actuarial balance does not necessarily imply actuarial fairness or neutrality. Likewise, a system that is actuarially fair does not need to be either neutral or balanced; indeed, the actuarial imbalances of most public defined-benefit schemes are the result of benefits that were actuarially unfair, that is, overly generous, compared with contributions made. Actuarial imbalances may contribute to intergenerational inequity, while the lack of actuarial fairness may impose horizontal inequity in exchange for greater vertical equity for members of the same generation.

In addition to coverage, the other main systemic features influencing equity are regulatory homogeneity, actuarial fairness, risk sharing, mode of financing, and taxation of benefits. Regulatory homogeneity refers to the extent to which the system differentiates between schemes and schemes’ members according to sector, gender, length of service, and age-earning profiles. For the concept of actuarial fairness, please refer to Box 15.1. Risk sharing is the manner in which the risk of underfunding expected benefits is distributed between beneficiaries and the underwriter of pension promises. Mode of financing refers to the difference between the incidence of revenues and benefits.

If a system consists of several schemes (for instance, specialized by economic sector or geographic region), their regulations are homogeneous if the horizontal and vertical equity characteristics are similar across the schemes. Homogeneity does not ensure either actuarial fairness or fiscal sustainability; it simply implies that people with similar characteristics are treated similarly (in an actuarially fair or unfair manner). For instance, if private sector employee pensions are based on lifetime average wages while the public sector scheme uses final salaries to establish benefits, then the system is heterogeneous, reducing horizontal equity across schemes and their members. Similarly, if farmers pay lower contributions in return for entitlements similar to the entitlements received by nonagricultural workers (as in Austria, Finland, France, Germany, Poland, and Moldova), the heterogeneity of the internal rates of return on contributions translates into horizontal inequity across people belonging to different schemes (Choi 2009).

At the level of individuals, if a system is actuarially fair, it is horizontally equitable, too, and provides no intentional, ex ante redistribution across members who have varying characteristics; redistribution is constrained to the type of ex post redistribution that is implicit to annuities. In defined-contribution schemes,3 redistribution is limited to what the payout products can accommodate. If benefits are paid as a lump sum or as phased withdrawals, then payouts are determined by the individual’s account balance at retirement, which, in turn, is defined by the amount and timing of contributions and the net investment returns earned. If annuities are competitively and fairly priced, and there are no regulations enforcing ex ante redistribution among annuitants to impose vertical equity, there will be little room for additional redistribution across, for instance, people of different genders, retirement account balances, or marital status.

The most common divergence from actuarial fairness is with regard to gender. Women, on average, live longer than men but their contribution histories tend to be shorter because of child rearing and other unpaid services they provide within the household (Clements, Eich, and Gupta 2014). Women also tend to earn lower wages. These factors—lower average wages, shorter service histories, and pension entitlements paid out over a longer period of retirement—would lead to monthly pensions significantly below men’s. However, public pension systems may compensate for these circumstances through higher pension accrual rates, recognition of noncontributory years spent at home with children as service time, and the disregarding of women’s longer life expectancy at retirement. These measures improve women’s welfare and imply sacrificing horizontal equity for the sake of vertical equity.

Pension rules may also diverge from actuarial fairness and horizontal equity in other ways. In defined-benefit schemes, income does not enter directly into the pension formula but goes through some adjustment to form the basis of benefits. The adjustment involves valorizing past earnings to wage levels observed at retirement and compressing the distribution of the pension calculation base. The result of these progressive pension calculation rules is that higher earners realize a lower return on their contributions than do low earners (Figure 15.6). It is important to note that compressing the distribution of benefits (relative to the distribution of the contribution base) may not, in itself, make the system internally more redistributive—improved vertical equity would require that at least part of the savings thus achieved be reallocated to low earners to increase their pension levels.

Figure 15.6Vertical Equity in Public Pension Systems

(Ratio of income replacement rates for low and high earners)

Source: Organisation for Economic Co-operation and Development (2013).

Note: EU27 = the 27 members of the European Union (before Croatia joined in 2013); OECD34 = the 34 members of the OECD. The figure compares two theoretical workers who earn 50 percent and 150 percent, respectively, of the economy-wide average wage throughout their lives, and shows how much higher the low earner’s replacement rate is compared with that of the high earner. Note that the replacement rates compare both workers’ pensions to their own wages; thus, low earners’ higher replacement rates will still translate into lower absolute pension levels.

With regard to risk sharing, a pension scheme can be defined benefit or defined contribution. In a defined-contribution scheme, the risk that assets will prove insufficient to generate expected retirement wealth is borne by the individual contributor; in defined-benefit schemes this risk is borne by the scheme sponsor or underwriter (Broadbent and Palumbo 2006). The underwriter for private schemes is the sponsoring employer or the financial intermediary underwriting the pension insurance policy. For public schemes, it is the community of current and future taxpayers as represented by the state. If the incidence of these incremental taxes or reduced services differs from that of pension benefits, then publicly managed, underfunded, defined-benefit schemes redistribute income from current or future taxpayers to current pensioners. In countries with large occupational pension sectors (such as Denmark, the Netherlands, the United Kingdom, the United States, and others), explicit or implicit government guarantees may also exist. Because these guarantee schemes are funded by contributions from the pension industry but may enjoy an ultimate state guarantee, guarantees issued to private pension schemes might imply redistribution across schemes, as well as across scheme members and taxpayers outside the schemes.

Defined-benefit schemes permit more redistribution (or divergence from horizontal equity) because these schemes do not record financial assets; instead they record information that at the time of retirement is translated into cash benefits according to rules that may or may not be actuarially fair. The “price” at which past contributions buy retirement benefits is set by the legislature at the time of retirement and with as much regard for horizontal and vertical equity as social policy objectives require. From a public policy perspective, defined-benefit schemes, and within those, universal coverage public schemes, are more accommodating of vertical equity considerations. Privately underwritten defined-benefit schemes (such as pension insurance products offered by insurance companies or occupational pension schemes underwritten by employers) have less room for intergenerational redistribution than do publicly managed defined-benefit social insurance schemes. There are various reasons for this, including the funding requirements for privately managed defined-benefit schemes, smaller risk pools, limited access to additional resources, and the accounting differences between public and private schemes.

With regard to the mode of financing, a pension scheme may be fully funded or pay-as-you-go financed. Pay-as-you-go schemes are extreme cases of underfunding: they function without reserves or with reserves that are dwarfed by the scheme’s accrued liabilities (Figure 15.7). Defined-contribution schemes are fully funded, by definition, while defined-benefit schemes can function fully funded or underfunded (pay-as-you-go with or without reserves). Whether a defined-benefit scheme is fully funded or pay-as-you-go financed is immaterial from an intergen-erational equity perspective as long as the scheme’s resources are sufficient to meet its current obligation.4 If, however, the scheme’s current resources are insufficient, then expenditure- or revenue-side adjustment measures become necessary, leading to intergenerational redistribution, typically toward earlier from later generations.

Figure 15.7Public Debt versus Implicit Pension Liabilities

(Percent of GDP)

Taxation can alter a pension system’s equity characteristics significantly. Pensions are deferred compensation and should be treated similarly to other income, but they rarely are (Figure 15.8). Taxes may be levied at the contribution, accumulation, or payout phase, but the choice of when to tax pension wealth has distributional consequences, depending on the progressivity of income tax schedules. Since the early 1970s, a discernible trend has emerged toward subjecting social insurance benefits, including pensions, to income taxation, although concessions remain widespread. In the majority of Organisation for Economic Co-operation and Development countries, tax allowances and tax relief, such as deductions from the taxable income base or lower tax rates (for instance, in Australia, Canada, Chile, Estonia, Ireland, and others), are extended to pension benefits. In the United States, between 15 and 50 percent of public pension income is tax free, depending on the recipient’s total income (OECD 2013).

Figure 15.8Effective Tax Rates Levied on Workers and Pensioners

Source: OECD (2013).

The equity impact of taxation of pensions depends on whether the personal income tax regime is progressive and whether tax rules change between the contribution period and retirement period. If the income tax rate is flat and the rate remains unchanged over time, the choice of when to tax pensions is immaterial from an equity perspective. A progressive income tax regime can help compress the benefit distribution relative to the distribution of gross incomes although the extent of compression will be typically higher if the tax is levied when the contribution-liable income is earned5 rather than when it is paid out in the form of pensions. Total lifetime taxes will be lower if pensions are taxed at the payout phase, and the individual’s welfare may also improve, depending on how the progressivity in the tax system and the higher marginal utility of marginal income in old age compare.

Equity in Common Types of Pension Schemes

The most common arrangement among public pension schemes is a combination of universal coverage, pay-as-you-go financing, and earnings-related, defined-benefit pension benefits. These schemes tend to provide significant intragenerational redistribution to ensure vertical equity while their reforms, which aim to improve fiscal sustainability in the face of aging populations, lead to intergenerational redistribution. Intragenerational redistribution is partly intentional, such as in the case of redistribution in favor of women; partly unintentional but in line with social policy preferences (for instance, from single to married people or to people with children); and partly unintentional and undesired, such as the channeling of pension wealth from less educated to more educated people (on account of the latter’s higher life expectancy). With regard to fiscal sustainability, the liabilities of these schemes are not automatically adjusted to be commensurate with their assets (expected revenues). Thus, unless the benefit rules and contribution rates were established at their long-term equilibrium, the combination of a no-policy-change scenario with aging will result in a worsening financial position, necessitating intergenerational redistribution from future workers to those of today.

With regard to adequacy, these schemes can perform well and provide people who have low lifetime earnings or short contribution histories (or both) with benefits high enough to meet poverty alleviation objectives. Being able to do so is a direct consequence of these schemes’ ability to diverge from actuarially fair horizontal equity. Defined-benefit schemes typically involve redistribution both within and between cohorts, but the size of the redistribution is difficult to assess given the multiple nonlinearities introduced by minimum benefits, contribution ceilings, and the compression of benefit distributions through partial recognition or indexation of contribution histories. The extent of redistribution also depends on whether it is measured on an individual or household level (Gustman and Steinmeier 2000) and whether the income position of households is established based on actual or potential earnings.

Notional defined contribution (NDC) and point systems are both strictly earnings related and provide benefits that reflect relative contribution performance and life expectancy at retirement within (point systems) as well as across (NDC) cohorts. NDC schemes are relatively new and were first introduced in Latvia, Poland, and Sweden, then followed by others such as those in Italy and the Kyrgyz Republic. NDC systems have equity features similar to those of traditional defined-contribution schemes; at the same time, however, they do not observe the same asset-liability matching constraints and allow for more government intervention to ensure fiscal sustainability. Because assets and returns are notional, governments can adjust credited returns; however, once contributions and notional returns are credited to individual accounts they become fixed monetarily, allowing little room for ex post revaluation of entitlements for the sake of reducing unfunded liabilities. A pure NDC scheme imposes no interpersonal redistribution across the members of the same cohort but may imply intergenerational redistribution, depending on how notional rates of return and annuity factors are changed, by the government, over time.6 Point systems (as in Germany and France) record relative contribution performance within cohorts but leave room for policymakers to decide how much a point is worth in monetary terms, from cohort to cohort, depending on fiscal or other considerations. Thus, point systems can maintain intragenerational horizontal equity while allowing expenditure-side adjustments for the sake of intergenerational equity.7

Fully funded defined-contribution arrangements are common among privately managed voluntary pension schemes but were relatively rare among mandatory, universal schemes before the structural pension reforms of the 1990s and 2000s in Latin America (for instance, in Bolivia, Chile, Colombia, El Salvador, Mexico, Panama, and Peru) and Eastern Europe (for instance, in Bulgaria, Croatia, Estonia, Hungary, Poland, and the Slovak Republic). Under this arrangement, there is usually no ex ante interpersonal redistribution, either within or between cohorts, at least during the accumulation period. This basic model may be altered if rate-of-return guarantees are issued by an entity separate from the scheme’s members. If a horizontally equitable pension scheme is defined as one that imposes neither net taxes nor transfers, then a defined-contribution scheme is a perfectly equitable design from an interpersonal perspective. It is also important to note, however, that these schemes may not be able to address vertical equity considerations and meet the poverty alleviation function of public pension schemes. This shortcoming is especially the case in countries with large informal sectors and low pension coverage. Thus, such schemes may need to be supplemented with welfare transfer programs to reduce poverty among the elderly (as, for instance, has been done in Chile, Israel, El Salvador, Mexico, and Paraguay).

Noncontributory basic pension schemes are traditionally viewed as part of the pension system although conceptually they are closer to conditional welfare transfers (with age used as a proxy for incapacity to generate labor income) or guaranteed minimum income schemes. These schemes are operated with vertical equity and, in particular, poverty alleviation, in mind. The incidence of benefits and the sources financing them are weakly related. Noncontributory schemes are the most transparent approach to redistribution aiming at vertical equity; they can also create the political economy conditions for making contributory schemes horizontally more equitable and imposing hard budget constraints on them. Eligibility and the targeted income level play an important role in this regard. Age, usually augmented by residence criteria to forestall benefit tourism, is the defining eligibility criterion of these pension benefits, which either are universally available (“basic pensions,” as in Estonia, Ireland, and the Netherlands) or may be targeted to the low-income elderly, with eligibility based on the level of other pension income or total income. Targeting may be achieved by ex ante means testing (for example, in Belgium, Denmark, and Korea) or by self targeting; in the latter case, a tax claw-back progressively reduces net benefit receipts as total taxable income rises (as in Australia).

Two policy constraints apply to benefit levels: first, basic pensions should be sufficient to keep retirees out of poverty in the absence of any other income—otherwise the pension system cannot meet its social policy objective; second, they should not provide strong disincentives for participating in contributory arrangements and voluntarily saving for old age. These two considerations also suggest that caution must be exercised when introducing noncontributory basic pensions. In particular, in countries in which the contributory scheme’s benefit distribution is compressed around the minimum pension or the contributory minimum pension falls close to the basic pension level, a noncontributory scheme may aggravate noncompliance and harm the fiscal sustain-ability of contributory social insurance schemes (Holzmann, Robalino, and Takayama 2009). Under such conditions, basic pensions may ultimately crowd out contributory public schemes as compliance falls in response to highly different internal rates of return on contributions.8

The Impact of Pension Reforms on Equity

In an ideal case, longevity-adjusted implicit returns on contributions are homogeneous, contribution histories are long, contributions are high enough to ensure that benefit levels are adequate to prevent poverty, and fiscal considerations pose no constraints to meeting pension obligations. In reality, no contributory public pension scheme meets this description. Fiscal sustainability, and the goals of equal treatment of all contributions (horizontal equity) and protection of the vulnerable against poverty (vertical equity) mutually constrain each other. In the presence of fiscal constraints, prioritizing poverty alleviation requires heterogeneous returns—that is, it requires intragenerational redistribution to improve vertical equity. The same fiscal constraints combined with uniform returns to contributions along the entire income distribution may result in inadequate benefits for low earners, leading to poverty or higher social transfers outside the pension system. Horizontally equitable benefits that are adequate even for low earners can lead to overgenerous pensions and unsustainably high pension expenditures. As a consequence of demographic aging these trade-offs become more pronounced, forcing governments to initiate pension reforms.

Pension reforms fall into three broad categories: parametric, structural, and paradigmatic. Parametric reforms maintain the system’s objectives and main instruments but adjust its parameters such as retirement ages, benefit calculation rules, and pension indexation. Structural reforms keep the system’s objectives (intended coverage, the relative importance of poverty alleviation, and consumption smoothing) mostly unchanged but modify the main channels of achieving them by altering the division of responsibilities across the state, the private sector, and the individual, or amending the nature of the system’s constituent elements (such as introducing mandatory funded pension schemes or shifting the poverty alleviation function into a separate scheme). Paradigmatic reforms amend the system’s basic objectives by perhaps introducing universal coverage, establishing a poverty alleviation pillar in a system that previously only aimed at horizontal equity, or substantially revising the extent of mandated intertemporal and interpersonal redistribution (the size of the system).

Parametric Reforms

Parametric adjustments are the most common reform, intended to enhance the system’s long-term fiscal viability, modify its distributional characteristics, or address its unintended externalities—but without changing the system’s objectives or structure. With the exception of changing pension indexation rules and the tax treatment of benefits, both of which affect the current stock of beneficiaries, parametric reforms typically focus on future retirees’ eligibility and benefit levels, influencing intergenerational equity. Parametric measures may alter retirement ages, contribution rates and bases, and benefit calculation rules, as well as the valorization of past earnings (in defined-benefit schemes) and the indexation of pensions.

Retirement age increases have been among the most common reforms since the early 1990s as a response to the dual challenge of declining labor force participation rates among older segments of the population and the increasing share of elderly people. Practically all OECD countries have increased retirement ages, and many have also tightened early retirement opportunities to close the gap between statutory and effective retirement ages. The equity impact of later retirement is determined by the actuarial fairness of the retirement age increase, whether the labor market can accommodate the marginal labor supply of older workers, and whether equity is viewed as a matter of pension wealth (the total value of all benefit payments) or the level of periodic (monthly) pension benefits.

The impact of retirement age increases on total pension wealth is indirectly related to income. Since lifetime income and wealth are strongly correlated with life expectancy, a change in the retirement age implies redistribution from less to more well-off beneficiaries. Although this redistribution can be partly compensated for by progressive benefit calculation rules, retirement age increases intensify intragenerational redistribution on a pension wealth basis. At the same time, if the marginal labor supply can be accommodated, monthly pensions will increase on account of longer service periods (Rawdanowicz, Wurzel, and Christensen 2013). This effect does not benefit all contributors to the same extent, however, because better educated people have a higher probability of late-career employment and their age-earning profiles make it more likely that their additional earnings will increase their average wage entering the pension formula. Retirement age increases should, therefore, be supplemented by measures that improve, through continuing education and training, older workers’ productivity, and by regulations that reduce age discrimination and allow for flexible working conditions (Heywood and Siebert 2009).

Less generous benefit indexation influences a pension system’s intergenerational equity features by changing the relationship between total benefits received and contributions made. In theory, indexation rules can be amended without changing overall pension wealth—total benefits received are similar whether combining less generous indexation with higher accrual rates (higher first pensions) or increasing lower pensions at a higher rate. Leaving pension wealth unchanged does not imply that total lifetime welfare remains unchanged, however; that depends on individuals’ intertemporal consumption preferences and their subjective discount rates. Indexation of public pensions can have an unintended distributional impact on households’ total income, too: the poorer the pensioner household, the higher the share of public pension benefits in total income (Brown and Weisbenner 2013). If the real value of benefits does not increase (as occurs under price indexation) while alternative sources of old-age income (such as private pensions, returns on savings, and assets) increase, lower indexation will have a greater negative impact on total old-age income at the lower end of the benefit distribution.

Reforms that reduce average starting pensions can also alter systems’ equity characteristics. Longer benefit assessment periods, especially calculating benefits based on lifetime earnings instead of on final salary, improve horizontal equity. High earners’ age-earning profiles are not only higher but the profiles’ steepness is also greater than that of lower earners. Thus, high earners’ pensions are higher relative to their own career average wage than are lower earners’ pensions. Increasing the minimum contributory period for eligibility has ambiguous welfare consequences, depending on the labor supply response of people with service lengths between the old and the new criteria. Stricter eligibility criteria may reduce coverage. However, the system’s horizontal equity for its eligible members may increase—longer contribution histories would make it possible to move away from accrual schedules that reward short contribution histories with relatively higher replacement rates.

Modifying the accrual schedule’s shape and the average replacement alters both vertical and horizontal equity. A uniform downward shift of the accrual schedule has no impact on horizontal equity. At the same time, it may reduce vertical equity and require compensating antipoverty measures within or outside the pension system to meet a government’s antipoverty objectives. Conversely, a uniform upward shift of the accrual schedule, by virtue of providing a larger segment of contributors with an adequate benefit, will permit greater horizontal equity without increasing poverty among pensioners with low career earnings. The shape of the accrual schedule—whether marginal years of contribution are rewarded with lower, equal, or higher marginal replacement rates—can be an instrument of redistribution from people with different contribution histories. Moving from concave to linear and then to convex accrual schedules reduces the system’s capacity for ensuring vertical equity; at the same time, it provides stronger incentives for people to work longer.

Structural Reforms

Structural reforms maintain the extent of the mandate (the amount of present consumption forgone for the sake of retirement income) but change the mode of financing, risk sharing, or the distribution of the underfunding risk. Recent structural reforms have focused mostly on partially privatizing public pension schemes and on changing the risk-sharing characteristics of the system. The introduction of mandatory, privately managed defined-contribution schemes (as in Bulgaria, Croatia, Hungary, Poland, Romania, and others) increased the systems’ overall horizontal equity by strengthening the link between contributions and future benefits. These reforms also reduced vertical equity and exposed people with short contribution histories and low career earnings to the risk of old-age poverty, especially among those people with little or no pretransition contribution histories. Structural reforms, similarly to parametric ones, can improve a system’s intergenerational equity by gradually reducing unfunded pension liabilities and the portion of future cohorts’ tax burdens that are financing the funding gap accumulated in the past.

Another common type of structural reform is the introduction of uniformity into pension schemes and the treatment of different types of labor contracts, such as urban versus rural residents and formal versus informal employment. Such efforts are under way in China, Japan, Thailand, and a number of other African and South Asian countries where the workers outside the public sector are either not covered by mandatory pension schemes or the schemes’ regulations are significantly different. These reforms improve horizontal equity and can have positive indirect effects on labor mobility and compliance.

Paradigmatic Reforms

Paradigmatic reforms change pension systems’ basic objectives. The most common type of paradigmatic reform is the expansion of partial coverage to explicitly target the entire population, as in China, India, Nepal, and Thailand, for instance. This type of reform may involve making contributory schemes universally accessible or introducing noncontributory social pensions that aim to reduce old-age poverty. These reforms are typical in low-income countries and non-European emerging economies where previous pension systems only covered civil servants, public employees, and urban dwellers employed in the formal sector. In theory, governments may also decide to give up the consumption-smoothing objective of a pension system in the face of fiscal constraints and focus, instead, solely on poverty alleviation. The fiscal cost of terminating contributory schemes that have large legacy liabilities keeps governments from reducing earnings-related schemes to a basic pension; at the same time, economic crises (for example, hyperinflation, loss of the revenue base because of war or civil strife) may compel policymakers to pursue such marked paradigm shifts, as was the case in Georgia and Kosovo.

An important issue regarding paradigmatic changes is the dialectic nature of reforms: the compounded impact of parametric and structural reforms may amount to a major paradigm shift. For instance, compressing the benefit distribution because of fiscal constraints (upper limit) and poverty alleviation objectives (lower limit) may result in a quasi-flat benefit structure that is incompatible with earnings-related contributory financing. Likewise, making all components of a pension system strictly related to earnings (regardless of the choice of defined-benefit or defined-contribution risk sharing) while constraining total pension spending for reasons of fiscal sustainability can make achievement of the poverty alleviation objective impossible.

Conclusions and Policy Implications

Historic, economic, and social contexts determine the interpretation of equity that is acceptable in a given society. Despite the absence of a universally applicable “mix” of horizontal and vertical equity or levels of inter- and intragenerational redistribution, policymakers need to be aware of the equity impacts of pension system design features; otherwise they will be unable to efficiently select the means of achieving their policy objectives. Despite these difficulties in determining what constitutes an equitable pension system, some clear principles emerge from international experience.

First, some policies are clearly equity enhancing. Expanding coverage—or keeping coverage from declining—is the most important of the policies contributing to improved vertical equity. Coverage can be directly increased by dismantling regulatory barriers to participation, introducing administrative structures that make it easy to register and comply, providing financial incentives through the tax treatment of contributions or targeted matching contributions, ensuring that explicit and implicit pension contracts are honored, and finally, by expanding the obligation to participate. Factors beyond pension policy, such as better control over informality, or higher transparency of public finances, can also help improve coverage. In Eastern European emerging economies, where informality has undermined compliance and fiscally driven parametric reforms may have devalued pension promises, efforts to maintain compliance are necessary. Equal treatment of economic sectors, industries, types of contracts, ethnic groups, and so on can unequivocally improve horizontal equity. Homogeneity of pension rules across schemes with similar characteristics can equalize internal rates of return, provide the same level of protection against poverty, and reduce labor market distortions arising from the existence of different schemes.

Second, most design features and reform options require trade-offs. In a fiscally constrained policy space, the trade-off is between horizontal and vertical equity. Under circumstances of broad and deep poverty requiring significant redistribution toward the poor to achieve vertical equity, the trade-off is between horizontal equity and fiscal cost. If there is a manifest social preference for a strong contribution-benefit link, then the choice is between fiscal cost and the extent of poverty alleviation. Making benefit indexation less generous, for instance, to address fiscal constraints, will leave consumption smoothing unchanged but may compromise a system’s ability to provide poverty indexation. Likewise, introducing a noncontributory minimum pension to cover people without other pension entitlements will either increase total pension expenditures or will require that contributory pension benefits be reduced. This will compress the total benefit distribution around the basic pension and reduce the system’s horizontal equity. Another example of policy trade-offs is permitting women to retire earlier than men: ensuring that their benefits are adequate despite their shorter contribution periods will reduce horizontal equity or increase pension expenditures. Similar choices are unavoidable in the design or reform of pension systems; therefore, the gains and losses inherent in these policy choices need to be carefully analyzed.

Third, pension policy is not made on the basis of pure economic analysis but under political constraints. Sociocultural circumstances and traditions may hinder the equal treatment of certain social groups, including women; a society’s self-image as a meritocracy may stand in the way of redistribution toward people with insufficient contribution histories; powerful groups—such as the armed forces, high ranking civil servants, and legislators—may be disinterested in reforms that reduce their privileges; and resistance from older workers to partially revising their pension entitlements may necessitate slowing down reforms. Likewise, the available instruments to achieve the preferred policy objectives may be constrained, too, by limited administrative capacity and financial market development. The absence of a sufficiently developed domestic capital market, for example, can hinder the development of a competitive and efficient private pension sector; the lack of electronic databases of past contributions may prevent the transformation of a defined-benefit scheme into an NDC one; and the absence of reliable death and residence registries may hamper efforts to exercise effective control over benefit uptake.

Fourth, the fiscal and welfare consequences of various reform options need to be analyzed, and the goals of pension policy should be communicated, to achieve broad and lasting support for the policy direction chosen. Pension policies take years, even decades, to fully implement, and their reversals are costly both fiscally and to the credibility of structural reforms. Transparency of pension finances, and their consequences for intergenerational equity, should be ensured by the development and regular publication of indicators capturing public pension systems’ long-term financial position and the impact of reforms on future pension spending and revenues.

Finally, pension policy effectiveness hinges, to a large extent, on other policy areas. These areas include, among others, labor regulations, access to and quality of public education and health care, and tax policy and administration. Retirement age increases need to be accompanied by policies promoting late-career employment; the expansion of contributory pension schemes is only feasible if tax administration capacity can accommodate the expansion; improving women’s labor force participation rates and, consequently, their pension coverage and average benefits cannot happen without flexible work arrangements and access to affordable child care facilities. Welfare in old age is a reflection of individuals’ working careers, income, and assets; their decisions regarding the intertemporal reallocation of consumption; and the social policies, including pension regulations, that may alter the welfare consequences of preretirement life events. Public pension systems can lessen the consequences of inequities suffered in education and the labor market caused by ethnic, gender, or other discrimination. But the pension system cannot, in itself, be expected to correct all these ills. Indeed, a pension system’s ability to prevent poverty and a marked drop in consumption is a gauge not only of the system’s success but of the effectiveness and integrity of a variety of public policies.


    AuerbachPaulaMaria EugeniaGenoni and CarmenPagés. 2007. “Social Security Coverage and the Labor Market in Developing Countries.IZA Discussion Paper No. 2979Institute for the Study of LaborBonn.

    • Search Google Scholar
    • Export Citation

    BloomDavid E. and RoddyMcKinnon. 2013. “The Design and Implementation of Public Pension Systems in Developing Countries: Issues and Options.IZA Policy Paper No. 59Institute for the Study of LaborBonn.

    • Search Google Scholar
    • Export Citation

    BroadbentJohn and MichaelPalumbo. 2006. “The Shift from Defined Benefit to Defined Contribution Pension Plans—Implications for Asset Allocation and Risk Management.Paper prepared for the Working Group on Institutional Investors Global Savings and Asset Allocation established by the Committee on the Global Financial System.

    • Search Google Scholar
    • Export Citation

    BrownJeffrey R. and Scott J.Weisbenner. 2013. “The Distributional Effects of the Social Security Windfall Elimination Provision.Journal of Pension Economics and Finance12(4): 41534.

    • Search Google Scholar
    • Export Citation

    ChoiJongkyun. 2009. “Pension Schemes for the Self-Employed in OECD Countries.Employment and Migration Working Paper No. 84OECD PublishingParis.

    • Search Google Scholar
    • Export Citation

    ClementsBenedictFrankEich and SanjeevGupta eds. 2014. Equitable and Sustainable Pensions: Challenges and Experience. Washington: International Monetary Fund.

    • Search Google Scholar
    • Export Citation

    GustmanAlan L. and Thomas L.Steinmeier. 2000. “How Effective Is Redistribution under the Social Security Benefit Formula?NBER Working Paper No. 7597National Bureau of Economic ResearchCambridge, Massachusetts.

    • Search Google Scholar
    • Export Citation

    HeywoodJohn S. and W. StanleySiebert. 2009. “Understanding the Labour Market for Older Workers.IZA Discussion Paper No. 4033Institute for the Study of LaborBonn.

    • Search Google Scholar
    • Export Citation

    HolzmannRobertDavid A.Robalino and NoriyukiTakayama eds. 2009. Closing the Coverage Gap: The Role of Social Pensions and Other Retirement Income Transfers. Washington: World Bank.

    • Search Google Scholar
    • Export Citation

    International Labour Organization. 2014. World Social Protection Report 2014–15: Building Economic Recovery Inclusive Development and Social Justice. Geneva: International Labour Organization.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund. 2015. Fiscal Monitor: Now Is the Time—Fiscal Policies for Sustainable GrowthWashingtonApril.

    McDanielPaul R. and James R.Repetti. 1993. “Horizontal and Vertical Equity: The Musgrave/Kaplow Exchange.Florida Tax Review1(10): 60722.

    • Search Google Scholar
    • Export Citation

    Organisation for Economic Co-operation and Development (OECD). 2013. Pensions at a Glance. Paris: OECD Publishing.

    QueisserMonika and EdwardWhitehouse. 2006. “Neutral or Fair? Actuarial Concepts and Pension-System Design.OECD Social Employment and Migration Working Paper No. 40OECD PublishingParis.

    • Search Google Scholar
    • Export Citation

    RawdanowiczLukaszEckhardWurzel and AneKatherine Christensen.2013. “The Equity Implications of Fiscal Consolidation.OECD Economics Department Working Paper No. 1013OECD PublishingParis.

    • Search Google Scholar
    • Export Citation

    SantosTania and ImmaculadaDomínguez. 2011. “Financial Solvency of Pension Systems in the European Union.International Conference on Economic Modeling – EcoMod2011 ProceedingsUniversity of AzoresPortugal.

    • Search Google Scholar
    • Export Citation

    SchwarzAnita and Omar S.Arias eds. 2014. The Inverting Pyramid: Pension Systems Facing Demographic Challenges in Europe and Central Asia. Washington: World Bank.

    • Search Google Scholar
    • Export Citation

    van der WalDirk. 2014. “The Measurement of International Pension Obligations.DNB Working Paper No. 424De Nederlandsche BankAmsterdam.

    • Search Google Scholar
    • Export Citation

The authors would like to thank Mauricio Soto and Baoping Shang for their insightful comments.


The case of Italy demonstrates the impact of pension reforms on intergenerational equity: In the early 1990s, and then in the 2000s, the retirement age was increased, reducing the benefit-receipt period and total pension wealth. Contribution rates were increased and a solidarity tax was imposed on high pensions. Furthermore, the manner in which new pensions are calculated relates lifetime contributions much more closely to benefits then the prereform system did.


Transition costs emerge because pension liabilities mature over a very long horizon while contribution revenues start to decline as soon as a funded reform becomes effective. The revenue shortfall may be financed by higher current taxes, additional debt (higher future taxes), or lower public expenditures in other areas. Regardless of the chosen financing strategy, the level or distribution of lifetime taxes paid and services or transfers received will be different from what they would have been before the reform.


Strictly speaking, this only holds for defined-contribution schemes that operate without performance guarantees.


Theoretically, pay-as-you-go contribution rates can be established at their long-term equilibrium level with reserves built up when the system is young (has few pensioners compared with contributors), and drawn down as the system matures.


Technically, this would mean that income tax is payable on total income, without deducting pension contributions.


In an NDC regime, the only instruments available for adjusting the growth of future pension liabilities are the credited returns and the annuity factors.


Governments, in practice, can and do adjust pension parameters in line with fiscal constraints. The design of the public system determines the relative ease and transparency of revenue- and expenditure-side adjustments.


For further discussion, please see Schwartz and Arias 2014.

    Other Resources Citing This Publication