Summary

Table of
Contents

About
the
Author

Order this
Book

UIP
Bookstore

Chapter 1

During the past fifteen years, debate about and actual change in the scope and structure of national pension systems have grown to unprecedented levels. This intense focus on pension systems and institutions is occurring throughout the world, in fully developed and developing economies alike. It is being driven by a number of factors that vary in mix from one part of the globe to another. These include the need or desire to restructure entire economic systems, reinvigorate ineffective pension institutions, or improve social protection in concert with improving economic conditions. Debate is often occasioned by the hope that alternative pension structures will improve macroeconomic performance and help respond to changing demographics, or the desire to reflect changes in social philosophy about the relative importance of individual and collective provision for retirement.

One of the most prominent features of the current debate is sharp criticism of the pay-as-you-go, public pension programs that are the primary means of providing retirement income in many industrialized countries. For decades, these programs were widely viewed as valuable social and economic institutions. Today, they are often accused of costing too much, reflecting outmoded social philosophies, and having undesirable consequences for the economy. The previous consensus in support of these pay-as-you-go, public programs has broken down.

In the face of an increasingly polarized debate, the member organizations of the International Social Security Association (ISSA) asked the Association to help lead the search for a new consensus. The logical way to begin this search is with a careful review of the various arguments put forward by both critics and defenders of the traditional public pension arrangements. The review must examine the economic and social impacts of various pension approaches, assess how effective each is likely to be in ensuring adequate retirement incomes in an uncertain world, and identify the role that variations in social, cultural, and political traditions play in determining whether or not a particular kind of institution can prosper in a given environment. It should supply the foundation for an ongoing dialogue among the supporters of all approaches with the objective of developing a new consensus about the range of appropriate structures for national pension systems and the merits of following different approaches.

This volume is the initial step in the ISSA review process. It includes nine chapters that examine a variety of pension issues from an economic perspective. These chapters explore different aspects of the impact of pensions on the economy, the fiscal dynamics of different public pension approaches, and the challenges involved in ensuring that pensions provide adequate incomes. The review has begun with economic concerns because these are usually the basis for the strongest criticisms of existing pension systems and the major impetus for change in those systems.

The most common criticisms of pay-as-you-go pensions involve their economic impact, and that particular set of concerns is reviewed in the first five chapters of this volume. Pension systems, however, are not created because of the impact they might have on the macroeconomy. They are designed, first and foremost, to be mechanisms that provide retirement income to the aged population. This role is examined in detail in the issue briefs on the fiscal dynamics of public pension approaches and their adequacy as sources of retirement income.

For the most part, this volume leaves the social and cultural issues, which are also of great importance in the debate, for subsequent examination. These issues include the role of national pension programs as social institutions and the conditions under which different kinds of public and private institutions are likely to succeed or fail. These other topics are extremely important. The role that public pensions play in ensuring cohesion in a modern society is likely to be as important as any economic effect that they may have. In addition, history shows that serious problems with the operation of public pensions are frequently as much the result of institutional weaknesses as of design flaws. Designs that appear to work fairly effectively in one institutional setting can easily prove to be a disaster in another setting. These issues provide an agenda for future analyses as a part of this important ISSA initiative.

Summary of the Issue Briefs

The logical starting point for this discussion is the question of why public pension programs exist at all. What purposes are they supposed to serve? What impact should we expect them to have? How should they be designed to achieve their purposes?

Both supporters and critics of the traditional pay-as-you-go pension systems agree that governments ought to require working-age people to make provisions for their retirement; they disagree about the most desirable mechanisms for achieving this. The agreement that some form of government intervention is necessary demonstrates a shared belief that free markets would not work properly to provide all citizens with adequate financial protection in retirement in the absence of government intervention.

One reason for government intervention is the desire to alleviate poverty, particularly among those no longer expected to work. As economies develop, extended family linkages weaken, and governments traditionally accept the responsibility of ensuring a minimum living standard for the aged. In many countries, public pension programs are the most important tool for discharging this responsibility, since they are effective in supplying at least a modest level of income to most aged and do so in a manner that preserves dignity and self-respect. Almost invariably, however, the scope and structure of the public pension program go far beyond the type of government effort that would be required just to provide a “safety net” to assure minimum living standards. It is this expanded scope that requires additional explanation.

The most common argument for this greater government role is that many working individuals who could adequately provide for their own retirement needs are myopic. In the absence of a government mandate, they would not have the foresight or discipline to save adequately for retirement. By the time they realized their mistake, it would be too late. In effect, the government acts paternalistically to enforce a mandate that people may resent when they are young but will grow to appreciate as they get older.

A second argument is that the government mandate is required to protect the prudent members of society from free-riders. If, in the end, people believe that the government will ensure that all of the aged have access to a minimum living standard, some may make a conscious decision not to save on their own. To avoid having to pay both for themselves and for any imprudent neighbors, the prudent members of society force everyone to contribute.

A third argument focuses on the possibility of reducing the uncertainty involved when each individual is required to make his or her own retirement arrangements. Government interventions can reduce the difficulty of preparing for retirement in the face of uncertainty about the pace of future economic activity, the path of future investment returns and inflation rates, and the length of one’s life.

Several observations about the structure of public pension plans flow from a review of these arguments. First, while the arguments suggest that some form of mandatory program is desirable, they do not suggest that the program must offer full earnings replacement for middle- and upper-income retirees.

Second, the arguments presume that many working-age people would not voluntarily make adequate provision for retirement. One key to successful implementation of a public pension program, therefore, is the willingness and ability of the government to enforce collections from reluctant individual and organizational contributors. One sometimes hears suggestions that compliance problems could be solved simply by linking benefits more closely to contributions. While such a change might have a beneficial impact, expecting it to produce a major increase in compliance would seem to ignore the basic assumptions about human behavior that motivated the creation of the pension program in the first place.

A third observation involves the likely impact of a public pension program on participant behavior. Public pension programs are designed to make it easier for people to retire at an “appropriate” age. The assumption is that, in their absence, people would have saved too little and thus been forced to work too long. It should be expected, then, that the implementation of a mandatory pension program will cause many participants to retire earlier than they otherwise would have, thereby reducing the labor force participation of the aged. To some degree, such an impact is the intended result.

Chapter 3: The Economic Cost of Pension Programs

The focus turns next to the fundamental economic issues underlying the public pension debate. Chapter 3 provides a careful examination of the factors that determine the actual economic cost of supporting the retired population. Chapters 4, 5, and 6 examine the likely impact of public pension systems on savings, labor force behavior, and international competitiveness.

One source of confusion about the economic impact of a public pension program can be traced to a failure to distinguish between the actual cost to the economy of the pension program and the social insurance contributions that are levied to finance those costs. Chapter 3 explores how changes in demography and public policies affect the economic cost of supporting the retired population. A later chapter explores how different economic and demographic environments cause the contribution rates required to finance pensions to rise or fall, even when the actual economic cost of supporting the retired has not changed.

The economic cost of supporting the retired population is best measured as the fraction of each year’s total national economic activity that is devoted to supplying the goods and services the retired consume. This assumes that whatever part of the economy’s capacity is used for this purpose cannot be used for some other purpose, such as producing consumer goods for the rest of the population or making new investments to increase future productivity.¹

This economic cost is financed through some combination of transfers from the labor earnings of those who are not retired (usually in the form of pension contributions) and allocations of a portion of each year’s returns to invested capital (usually in the form of earnings on assets owned by individual retirees or by pension funds). Different approaches to pension finance often involve different allocations of these costs between contributions and returns on assets. Confusion can occur when one approach appears to be cheaper than another because it involves lower pension contributions from earnings. If the lower charge to labor is offset by a higher charge to capital income, the total cost to the economy is the same even though it may be distributed differently.

The share of total economic activity devoted to the consumption of the retired—the actual economic cost of their support—is influenced by a variety of economic, demographic, and public policy developments. Perhaps the easiest way to understand how these various elements interact is to focus on the behavior of three key ratios: (1) the aggregate consumption ratio, which is the fraction of economic activity that is devoted to producing consumer goods and services for domestic use, (2) the retiree dependency ratio, which is the fraction of the population that is retired, and (3) the living standards ratio, which is the ratio of the average consumption of the retired population to the average consumption of all persons. When multiplied together, these three ratios will produce the ratio of retiree consumption to total economic activity, which is the economic cost of supporting the retired.

The relationship between changes in any of these three ratios and the corresponding change in the economic cost of the retired population is direct and proportional. Anything that causes one of these ratios to rise by a given percentage will increase the economic cost of supporting the retired by the same percentage. By the same token, the cost of supporting the retired can only be reduced if changes that reduce at least one of these key ratios are introduced.

As populations age, and if no other changes are made, the retiree dependency ratio will rise and the economic cost of supporting the retired will increase proportionately. The two most common adjustments that are discussed as ways of offsetting some of this cost increase are increasing the statutory retirement age, which would reduce the retiree dependency ratio, and reducing retirement benefits, which would lower the living standards ratio.

Shifting some or all of the responsibility for managing public pension plans from the public sector to the private sector has occasionally been advocated as a mechanism for reducing the cost of supporting the retired. Whether such a change has the desired effect depends entirely on whether it serves to decrease one of these key ratios. For example, if the shift is accompanied by changes that increase retirement ages or reduce the relative incomes of the retired population—or will be more effective at keeping retirement ages from drifting down or relative incomes from drifting up—it may well be an effective mechanism for reducing costs. If the shift is not accompanied by changes in dependency ratios or in the relative living standard of the retired, however, it will have no impact on the actual economic cost. Indeed, such a shift can actually increase the cost of supporting the retired if it produces higher retirement incomes. Higher incomes for the retired will most likely lead to an increase in their living standards relative to the rest of the population, causing the cost of supporting retired persons to rise.

Others advocate economic policies which they believe will accelerate economic growth as part of a strategy to deal with the rising cost of an aging population. While such policies may be desirable for other reasons, it is not at all clear that faster economic growth should be expected to reduce the economic cost of supporting the retired. If faster economic growth translates into more rapidly rising living standards of the working-age population without having the same impact on the living standards of the retired population, the relative cost of supporting the retired will drop. On the other hand, rising living standards among the working-age population may cause them to prefer earlier retirement and to offer less resistance to gradually rising pension contribution rates. Either reaction could mean that faster economic growth actually has the effect of increasing the cost of supporting the future retired population.

In summary, the economic cost of supporting the retired population is best measured by looking at the resources devoted to their consumption. Public pension benefit payments are a major source of support for the consumption of this population. For this reason, the best measure of the economic cost of a pension program is the benefits it provides. If it is important to prevent too great an increase in the costs of supporting the aged population, the evaluation of alternative policies to constrain these costs should focus on how effectively each will be in keeping pension benefit payments from rising.

Pension Impacts on the Economy

Chapter 4: Pensions and Savings

The relationship between pension finance and saving behavior has attracted the attention of economists for several decades and has produced a considerable volume of statistical studies. The primary issue has been whether pay-as-you-go pension systems reduce aggregate national savings and/or whether greater reliance on advance funded pension plans would increase national savings. If a consistent relationship can be found, pension policy changes could be used to increase aggregate national saving and produce a somewhat higher level of per capita economic activity.

Careful reviews that assess the whole body of recent analyses usually conclude that no consistent evidence exists which links the introduction of pay-as-you-go pension programs to declines in national saving rates. The linkage may well exist, but either it is too small to show up in the data or its impact is obscured by other factors.

In a similar vein, there is evidence (based largely on data from the United States) that the accumulation of assets in retirement accounts will cause total household saving to increase, though by less than the increase in the balances in the retirement accounts themselves. This positive effect from advance funded pensions appears to be either overshadowed or offset by the behavior of other components of national savings, however. Among these are the fiscal operations of government, business finance strategies, and customs and practices used to finance housing and other major consumer expenditures. At least among OECD countries, there is essentially no correlation between the rate at which pension assets have grown and the total savings rate in the economy.

All of this suggests that if higher saving is an important national goal, pension policy may have a role to play, but it is unlikely to have a discernable impact by itself. Pro-savings pension policies would have to be accompanied by other interventions such as a pro-saving tax code, government fiscal surpluses, and policies which discouraged consumer lending.

Several recent studies have found that the development of financial markets can provide an independent impetus to economic growth. Although the results are still controversial, they suggest that using advance funding for at least a portion of the pension system may have a beneficial economic impact independent of any impact on aggregate savings. The growing pension plans could provide a market for new financial instruments and help financial markets to develop.

Chapter 5: Pensions and Labor Supply

Public pension plans may also cause a reduction in productive economic activity through their impact on labor supply, the topic explored in this issue brief. One possibility is that mandatory pension plans may unduly discourage individual work effort both by lowering net pay during people’s prime working years and by encouraging retirement when people reach the age at which benefits become available. Another is that they may encourage an artificial movement of people into less easily taxed, and often less productive, sectors of the economy if they force younger workers to set aside more for their retirement than these workers would prefer to do.

Studies of worker behavior offer some confirmation of the first concern. Pension contributions (and other earnings taxes) seem to have little effect on the work effort of those who are the primary source of support for themselves or their family. For those who have alternative sources of support, however, these studies find that social insurance contributions and other taxes on earnings do tend to reduce work effort at least somewhat.

The availability of pensions for older workers also seems to reduce work effort, especially among those whose health has deteriorated. Not surprisingly, there is a tendency for more generous pensions to have a more dramatic effect on work effort, although a change in the age at which pensions first become available would probably have a more powerful impact on retirement behavior than would a modest change in the amount paid at a given age. This means that if population aging forces pension retrenchments, a reduction in the monthly benefit payable beginning at a given age is as likely simply to produce lower retirement incomes as it is to lead to an increase in the average retirement age.

Since mandatory pension programs are established to require people to make more adequate provision for retirement, it must be assumed that their successful implementation will allow people to retire earlier than they would have without access to pension income. Current studies of labor supply impacts tell us that people do retire earlier than they would have without access to pension income. Since most societies have never articulated the social, political, or economic criteria for judging the impact of their policies, however, these studies are unable to tell us whether the actual impact is more or less than is desirable or if the benefits paid are too generous or not generous enough. We are also unable to say whether some discouragement of work effort among secondary earners is a reasonable price to pay for the gains that are secured when a pension plan is established.

Levying pension contributions creates an incentive for people either to hide from the tax collector in informal labor markets or to classify themselves as self-employed, since ensuring compliance among the self-employed has always been a challenge. By itself, reclassifying oneself as self-employed may have little impact on aggregate economic activity, although by reducing tax compliance it can create fiscal problems for the government and, depending on how closely benefits and contributions are linked, for the pension plan as well. Aggregate economic activity will suffer, however, if people seek out fringe employers in the informal labor market who operate in somewhat less productive sectors of the economy but who can offer a higher net wage by not paying pension contributions.

The situation can be particularly troublesome if the link between pension benefits and contributions is weak. In these cases, workers may be able to spend much of their careers in informal employment (or self-employment), escape the full payment of contributions, and still draw full pensions. In addition to any economic losses, such situations will cause major financial problems for the pension plan.

Pension systems in which the link between contribution payments and benefit receipt is as direct and clear as possible introduce fewer compliance disincentives and are better insulated from the financial problems associated with any remaining compliance problems. A close linkage should also help reduce any other labor market distortions. Some compliance problems should still be expected no matter how clear and close the linkage, however. If a clear and close linkage between contributions and benefits were sufficient by itself to produce near perfect compliance, there would be no need to have mandated the program in the first place.

Chapter 6: International Competitiveness

The recent slowing of economic growth in the industrialized world—and particularly in Western Europe—has caused concern that overly generous social security systems may be undermining international competitiveness.

It is generally acknowledged that well-designed social security programs can enhance international competitiveness by, for example, helping to smooth transitions from one industrial structure to another or facilitating worker movement to new employment opportunities. Poorly designed programs can discourage work effort, however, and even the best-designed programs can also be expensive. The question is whether the positive impact of these programs is being offset by the impact that financing them has on the cost of doing business in a particular country.

Economic theory suggests that the costs of running social security programs should not cause any particular international competitiveness problems where product markets, labor markets, and foreign exchange markets are allowed to operate fairly freely. In such an environment, any increase in pension contributions (or other social security charges) will translate into reductions in worker take-home pay rather than increases in the cost that businesses incur to hire labor, regardless of whether the contributions are collected initially from the employer or the employee.

Both labor market and international competitiveness problems can arise for an extended period of time, however, if some combination of government policies and labor market rigidities prevent the normal market adjustments to increases in social security contributions. When either government policies or private labor agreements prevent real wages from falling, an increase in employer contributions can cause business costs to rise and lead to increased unemployment. This situation can also cause international competitiveness problems if, in addition, either government policies or private capital movements prevent exchange rates from adjusting to trade imbalances, at least for a while.

Analysis of the relative competitiveness of different countries supports the view that social security programs can be beneficial to a country’s competitiveness, but that high employer charges to finance them can be a problem. When OECD countries are ranked by their relative competitiveness, the more highly ranked tend to be those that spend a higher fraction of their GDP on social security. At the same time, higher employer contribution rates tend to be associated with somewhat lower competitiveness scores.

Choosing Among the Pension Approaches

Chapter 7: Setting Pension Contribution Rates

The economic cost of a pension program is best measured by looking at the relationship between aggregate benefit payments and total economic activity. However, the contribution rates needed to finance a given level of benefits may be higher under one approach to providing pensions than under another. The seventh chapter looks at the mathematics of pension contribution rates. It analyzes how economic, demographic, and institutional variables interact to cause differences in pension contribution rates, even when the actual economic costs of the pension program are the same.

Contribution rates under pay-as-you-go, defined benefit pension plans are set so that the aggregate receipts from all workers are sufficient to finance the aggregate benefit payments to all retirees. In these plans, contribution rates are sensitive to demographic changes but fairly insensitive to economic developments. A decrease in the number of contributors relative to the number of pensioners will force rates to rise, since the cost of financing a given quantity of pensions is spread among a smaller number of contributors. On the other hand, a change in prevailing earnings levels often has relatively little impact on the contribution rates needed to finance these plans, since any change is likely to cause more or less equal percentage changes in receipts and expenditures.²

Under the individual savings approach that is characteristic of most defined contribution plans, contribution rates need to be set so that each individual can accumulate financial assets of an amount sufficient to finance their desired level of retirement income. Since the financing plan focuses on the balance in each individual’s account, declining birth rates (and the changes in the contributor/pensioner ratio that they induce) have no direct effect on the contribution rate calculation.³ Contribution rates needed to obtain a specified level of retirement assets are determined instead by the interaction of interest rates (or, more generally, capital returns) and the rate of wage growth. An increase in the interest rate makes it easier to accumulate the necessary balance in retirement accounts and allows the system to operate with lower contribution rates. In contrast, an increase in the rate of earnings growth raises the amount of retirement assets that the pensioner must accumulate in order to finance a pension that will preserve the relationship between retirement income and pre-retirement earnings. This forces contribution rates to rise.

The relationship between contribution rates under funded, defined benefit pension plans and these particular economic and demographic variables is very similar to the response under individual savings approaches. In funded, defined benefit plans, contribution rates must be set so that sufficient assets are accumulated over the working life of each entering cohort to finance their retirement pensions. A major difference between these two approaches involves the ability of a group plan to smooth the impact that temporary changes in economics or demographics might otherwise have on each individual’s pension. (This topic is explored more fully in the ninth chapter.)

Contribution rates under pay-as-you-go and funded plans are equally sensitive to changes in life expectancy at retirement. An increase in retiree life expectancy increases the amount that each worker must accumulate under an individual accounts approach, since the accumulation must last longer. It has the same impact on the amount that must be accumulated for an entire cohort of retiring workers under a funded, defined benefit plan. It also causes the ratio of contributors to pensioners to fall, thereby forcing contribution rates up in the pay-as-you-go approach. In all cases, the desirable impact on individual life prospects is accompanied by an undesirable impact on pension contribution rates.

The relationship between the pension contribution rates required under pay-as-you-go and funded approaches is fairly straightforward and predictable as long as one is operating in a simple world in which nobody dies before reaching retirement age and pension plans have no administrative costs. In such a world, relative contribution rates depend on only two numbers: the rate of population growth and the gap between the interest rate and the rate at which wages are growing. If the amount by which the interest rate exceeds the wage growth rate is larger than the population growth rate, the funded approaches will have lower contribution rates. If the gap is smaller than the population growth rate, the pay-as-you-go approach will have lower contribution rates.

One reason why the funded approaches have attracted more attention in recent years is that population growth rates have been falling and, at least in OECD countries, interest rates have been rising relative to wage growth rates for the last couple of decades. Assuming that both trends will continue into the future, the funded approaches will produce a given average pension with a lower contribution rate.

The relative attractiveness of the three basic approaches—pay-as-you-go defined benefit, funded defined benefit, and defined contribution—may change, however, when more realistic assumptions about mortality structure and administrative costs are introduced into the comparison. Defined benefit pension plans incorporate certain insurance features not found in defined contribution plans. When a worker dies in the years just prior to retirement, the pension obligations under a defined benefit plan are reduced, allowing the plan to be financed with a somewhat lower contribution rate. In contrast, preretirement mortality has no impact on the contribution rates required under the individual savings, defined contribution approach, since each account is financed on the assumption that its owner will survive.

Managing the accumulated financial assets causes administrative costs under advance funded plans to exceed those under comparable pay-as-you-go plans. Moreover, among advance funded plans, administrative costs are consistently higher under a system of individual accounts than under a group defined benefit plan, due to the loss of some economies of scale and to differences in marketing expenses. Finally, defined benefit plans automatically pay benefits in the form of life annuities, whereas holders of individual accounts must purchase annuities in order to achieve the same degree of assurance that their income will last for as long as they live. When purchased separately, annuities introduce another set of marketing and administrative costs.

These effects can cause the costs of running a defined benefit plan, particularly a pay-as-you-go plan, to be substantially lower than the costs of operating a defined contribution plan producing a similar retirement income. Depending on the particular demographic and economic conditions prevailing in a given society, the effects of early mortality and increased operating costs can amount to the equivalent of a 1.5 to 2.5 percentage point reduction in the annual rate of return earned on individual accounts.

Chapter 8: Changing from One Approach to Another

This chapter explores two important kinds of transitions that societies make in pension policies. First is the transition from having no public pension plan at all to creating a public pension plan. Second is the transition from one pension approach to another after a number of years.

The single most important reason for creating a pension program is to help ensure adequate incomes in retirement. This objective will be achieved much more rapidly under some approaches than under others, however, because the different approaches phase in at different speeds. Noncontributory programs phase in the most rapidly. Whether the program is income tested or universal, its adoption can quickly improve the income of all of the aged, including those retired at the time the program is initiated. Contributory, defined benefit approaches will be of no benefit to those already retired at the time they are instituted, but can be used to ensure adequate retirement incomes to those who are close to the retirement age as well as all who will follow. Defined contribution plans are the slowest to mature. They require over half a century to approach their full potential as a source of retirement income and are of limited benefit to anyone beyond the midpoint of their work career at the time they are instituted.

Because of the differences in the speed of maturation, the defined contribution approach is frequently not the option selected when a country is setting up its first public pension arrangements.⁴ Over time, however, preferences about national pension policy are likely to evolve as systems mature and the pressure to assure adequate retirement incomes eases, and as economic, demographic, and social conditions change. For example, many countries which began by following one approach have subsequently broadened their strategy by combining several different approaches to form a mixed, or “multi-pillar,” system. Contributory approaches are added to older, noncontributory approaches, and privately managed approaches are added where older systems were largely run by government. The new approaches frequently supplement and occasionally partially replace the earlier programs.

Many of these transitions occur gradually and relatively painlessly. Often, as income levels rise, the newer approaches are simply added in lieu of an expansion of the older ones. The added programs may be voluntary or they may be mandated.

The one transition that cannot be achieved gradually and painlessly is the transition in which a publicly managed, pay-as-you-go, defined benefit plan is replaced by a system of privately managed, advance funded, defined contribution accounts. The challenge is that the transition requires paying off the express and implied liabilities for future benefits under the pay-as-you-go system, while at the same time financing the new, defined contribution program. This involves duplicate payments which can easily be as much as 5 or 6 percent of a country’s gross domestic product every year for several decades.

This kind of transition is difficult to justify on the basis of narrow cost calculations. Even if the transition payments are financed entirely through more government borrowing, under current economic conditions the additional cost just to service the new government debt is likely to be more than the cost of keeping the pay-as-you-go pension system in financial balance. For this reason, phasing out a social security system is usually not a very effective way of dealing with government fiscal problems, since it will likely make them worse for quite a few years into the future. Moreover, at least in principle, any positive economic impact that might be associated with such a change could have been achieved much more easily and with equal effectiveness by changing the benefit structure in the public system or by introducing a degree of advance funding to a system that had previously operated on a pay-as-you-go basis.

This type of transition may be justified, however, if it is the only effective way to solve certain political or institutional issues. It may provide the only politically acceptable way to reduce future benefit commitments, making it the most practical strategy for reducing the cost of supporting an aging population. It may also provide the most effective way for assuring that future benefit commitments don’t grow beyond levels that the society can afford, or that assets accumulated to help pay retirement benefits are actually used for that purpose. In some societies, such a transition may be deemed the only practical way of assuring a reasonable quality of service to pensioners, or may be favored on general ideological grounds.

Chapter 9: Mid-Career Economic and Demographic Risks

Chapter 2 notes that one rationale for setting up a mandatory pension program is to give individuals a more predictable source of retirement income than they could obtain on their own. As has already been discussed, the amount that needs to be set aside each month (the contribution rate required) in order to accumulate a given volume of assets grows larger or smaller as investment returns fall or rise, but people in the earlier stages of their careers cannot predict future investment returns accurately. Moreover, the amount of assets that must be accumulated in order to assure a regular retirement income at a particular percentage of preretirement earnings depends on how fast earnings levels grow during the person’s career and how long the person can expect to live once they have retired, which also are difficult to predict at the beginning of a career.

When a public pension program has been instituted, it promises (either explicitly or implicitly) that pensions will be available to those who make the required contribution payments. People measure the reliability of these promises in large measure by the degree to which a pension in the amount promised at the beginning of an employment career is actually available at the end of the career.

Pension promises are rarely realized precisely. Invariably, adjustments to earlier promises are required to reflect developments occurring during an individual’s employment career. Chapter 9 explores how unexpected economic and demographic changes are likely to affect pension promises under each of the various type of public pension systems.

Pay-as-you-go, defined benefit plans generally have less risk of substantial benefit change than advance funded, defined contribution plans for two reasons, one related to the financing principle they employ and the other to the way benefits are set. Because of the way they are financed, the benefits provided under a pay-as-you-go plan are not nearly as sensitive to unforeseen economic developments, particularly changes in the rate of wage or price growth or changes in the rate of return on investments. This advantage is partially offset by the fact that pay-as-you-go approaches are more sensitive to changes in the rate of growth of the working-age population than are funded approaches. Historical evidence suggests, however, that sensitivity to economic changes is likely to be a more serious source of unpredictability than is a sensitivity to changes in the growth rate of the working-age population. The two financing approaches are equally vulnerable to changes in mortality experience among those who are retired.

When unforeseen problems do arise, prospective retirees under defined benefit plans are usually not required to absorb the entire impact of any necessary adjustment. Under defined benefit plans, unforeseen changes in economic and demographic conditions lead initially to an imbalance between receipts and expenditures rather than to a change in the promised benefits. Sooner or later, the imbalance is eliminated through some combination of benefit and contribution rate adjustments. Typically, the impact of correcting the imbalance is spread among current retirees, future retirees, and other contributors, with each absorbing a fraction of the total impact. In contrast, under the funded, defined contribution approach, all unanticipated economic changes are reflected fully in changes in the retirement assets and future retirement income of each participant.

Chapter 10: Post-Retirement Risk

Once individuals are retired they face two more kinds of uncertainty: unanticipated inflation and uncertainty about their own lifespans. These are the subject of the last chapter.

Traditional pay-as-you-go public pension programs all but eliminate both of these risks by paying benefits in the form of life annuities and adjusting the benefit amounts periodically to reflect changes in price or wage levels. The major risk that retirees bear is that benefit adjustments will be delayed or altered in the face of adverse economic conditions or unanticipated financial problems in the pension plan.

Dealing with these two sources of uncertainty is a somewhat bigger challenge under traditional defined contribution approaches, where retirees must support themselves for the rest of their lives by drawing down the stock of financial assets accumulated during their working lives. Where financial markets are reasonably well developed, both kinds of uncertainty can be reduced through the purchase of variable annuities. Absent some form of government intervention, however, the private market for variable annuities is likely to have two shortcomings: costs will be higher than under collective insurance schemes due to adverse selection problems, and the annuities will probably be indexed to financial market returns rather than to inflation or wage growth.

Governments can effectively remove these two problems by mandating that all retirees purchase annuities, thereby all but eliminating the risk of adverse selection, and by selling bonds whose principal and interest are indexed to prices, allowing the sale of private annuities whose benefits are indexed to inflation. However, both of these actions have drawbacks. In addition to possible philosophical objections, the disadvantage to mandating the purchase of annuities is that it may permanently harm cohorts who happen to reach retirement age at a time when the value of investment portfolios is temporarily depressed. The disadvantage of issuing indexed bonds is that they create for the government a pay-as-you-go liability which, though smaller than that associated with a pay-as-you-go public pension system, will be less amenable to modification in times of economic distress. Governments find it much easier to postpone or alter pension benefit adjustments than to postpone payments on bonds that they have issued.

General Observations

First, certain elements of the economic critique that has been directed at traditional defined benefit, pay-as-you-go social security programs deserve to be taken seriously, but much of the criticism is either not supported by a careful review of current economic knowledge or seems overstated. Advance funding of public pension programs may yield economic benefits if pursued as one part of a plan for developing efficient financial markets. At the same time, there appears to be far less justification for assuming that advance funding will increase a nation’s saving rate, and there is no reason to believe that it will reduce the economic cost of dealing with an aging society by itself.

If an increase in national savings is desired, advance funding of pensions might form one part of a much larger strategy, but the strategy needs to focus on the uses to which any additional funds supplied to capital markets are put as much as it does on using pension reform as a way to supply more funds. Moreover, taken by itself, even a large increase in national savings is likely to produce only very modest increases in real earnings levels.

One danger in focusing too much attention on how pensions are financed is that too little attention will be focused on a set of issues that is much more central to the challenge of dealing with the costs of an aging society. The economic cost of a retirement program is best measured by the benefit payments it makes; these benefit payments are also the major route by which the program affects the economy. If an aging society is causing these costs to rise to undesirable levels, adjustments will most likely require raising retirement ages and reducing retirement benefits. Changing the way pensions are financed changes the distribution of the costs, but doesn’t necessarily change their magnitude.

The debate over social security and international competitiveness should be viewed as a debate about the balance between the gains from enhanced worker security and the losses from higher employer costs. It is a debate that deserves more careful analysis and less rhetoric than it has received thus far. In theory, free markets will ensure that the burden of social security contributions is borne by workers, regardless of whether contributions are initially levied against the employer or the employee. If this theory were always correct, the cost of social security programs would never affect international competitiveness. Scattered evidence suggests, however, that this theory does not necessarily describe what happens in the real world. Well-designed social security programs may actually work to increase international competitiveness, but part of the gain may be lost if they involve higher-than-average charges levied against employers’ payrolls.

It is difficult to know what to make of the debate over the impact of social security on labor supply, largely because nations fail to address, either analytically or politically, the question of what the desired impact is supposed to be, and there is little if any evidence to suggest that different approaches to mandatory pension programs would have different impacts. Perhaps the best that can be said presently is that the picture is mixed. All public pension programs probably cause people to retire earlier than they otherwise would have, but that is why the programs were created in the first place. Unfortunately, no one seems to know how to address the issue of what the actual impact of pension programs on the work behavior of the aged ought to be.

Social insurance contributions, along with all other forms of taxes on labor income, probably cause some reduction in the work effort of members of the family other than the primary earner, and may encourage tax cheating in situations where the line between employment and self-employment is blurred or where enterprises are operating at the margin of the economy. A loose link between contributions and benefits can also cause financing problems for the pension plan. There is no evidence to suggest, however, that how a plan is financed (whether funded or pay-as-you-go) will alter the degree to which labor supply is reduced or tax cheating is encouraged. Each seems to be one of the costs that must be accepted as part of the price to be paid for the benefits secured through a public pension program.

Second, the economic critique has focused on the perceived shortcomings of the macroeconomic effects of the defined benefit model without paying sufficient attention to the relative merits of each pension model as an efficient device to supply retirement income. As a mechanism for supplying retirement income, the defined contribution model suffers from several marked shortcomings. The size of the retirement income stream produced is less predictable, while ensuring that retirement incomes last an entire lifetime and keeping benefits up to date with prevailing wage or price trends is more difficult. Individual defined contribution accounts have also proven expensive to administer, artificially increasing the economic cost of the retirement income system.

In a political system capable of exercising sufficient self-discipline, the defined benefit model also provides a more predictable source of retirement income and can be administered at a substantially lower cost. Neither of these advantages should be sacrificed unless there is good reason to believe that they will be more than offset by other kinds of gains.

A final observation goes beyond the material covered in the first phase of this review, but is necessary in order to complete the picture. As noted above, pay-as-you-go, defined benefit pension programs have many advantages when operated in the context of a political system capable of exercising sufficient self-control. History shows, however, that some political systems have not been able to operating such plans responsibly. They have allowed benefit promises to rise to levels that exceeded their society’s willingness and ability to pay, creating serious fiscal problems for governments. Eventually these promises must be retracted, upsetting the retirement income expectations of people caught in the transition.

For reasons that go beyond the current review, this sort of political problem appears more common in some parts of the world and in some cultures than in others. Where it is a serious problem, the defined contribution approach is an attractive alternative because its benefits are more effectively insulated from political interference. Whether there are other institutional arrangements that would more effectively protect defined benefit plans from excessive political promises is currently unknown.

It is also too early to know how effectively the new systems based on the defined contribution model will be insulated from irresponsible behavior. Politicians are not the only people who are prone to promise more than they can deliver. The defined contribution model requires sophisticated oversight and regulation to ensure that one set of problems resulting from public sector political dynamics is not simply traded for a different set of problems derived from the dynamics of private sector operations.

The Future

Public pensions are critical institutions for a substantial portion of the population of many countries, and are destined to become even more important as population aging continues throughout the world. If conducted with care for the important viewpoints and dimensions, the debate over how these institutions should be structured will lead to a new and broader consensus about the role and shape of public pension programs. In turn, the emergence of a new consensus will ensure that changes introduced in the name of reform will strengthen these programs and make them better able to serve the millions who rely upon them.

Notes