Skip Navigation

Mr. Chairman and Members of the Committee, I am pleased to be with you this morning to discuss the lessons from the experience of other countries that have reformed their social security systems at least in part through privatization.

The retirement of the baby-boom generation in the United States will put our Social Security program under financial pressure, and a debate is now proceeding about how to pay for retirement in a financially sound way. Many recent proposals would allow workers to invest some portion of their earnings in personal retirement accounts. The amounts accumulated in those accounts would replace some of Social Security's benefits. Because some of a worker's retirement income would come from savings in his or her account rather than from government transfers, such plans would partly privatize Social Security.

Other countries face the same demographic and financial pressures as the United States. In fact, for many countries, the pressures are much more severe and immediate. Some countries have already responded to those pressures by privatizing their public pension systems to some extent, and their experience can offer lessons for the design of privatized pension systems. The economies and pension systems of those countries differ considerably from those of the United States, however, and comparisons should therefore be made cautiously.

The Congressional Budget Office (CBO) released in January a paper that reviews the experience of five countries--Chile, the United Kingdom, Australia, Argentina, and Mexico--that have introduced individual accounts to fully or partly replace their public retirement system.⁽¹⁾ Such plans are defined contribution plans--that is, retirement income depends in part on the uncertain returns on contributions to the accounts. Some other countries have relied on more traditional measures to close the financing gap, such as changing benefit rules and retirement ages or increasing payroll taxes, but those countries were not included in our analysis.

All five countries already had some type of old-age income support system before reform. Those systems relied primarily on "pay-as-you-go" financing, in which taxes collected each year mainly or entirely finance the benefits paid to retirees in the same year. For example, in the United Kingdom (U.K.), a payroll tax finances the government's expenditure for pensions (and other benefits) in the same year. Before reform, three of the other countries also generated most of the revenue for their pension systems by earmarked taxes on wages.

By contrast, a system with personal retirement accounts can prefund retirement income by requiring people to accumulate savings during their working years. For example, Chile's system requires workers to invest in personal retirement accounts from which workers may withdraw money only after they retire. Moving from a pay-as-you-go system to a prefunded private system, however, imposes a financial burden on transitional generations.

All five countries encountered the same set of issues in privatizing their systems, and those issues are also relevant to efforts to privatize the U.S. Social Security system.

• Policymakers have to decide who will pay for the transition between the pay-as-you-go system and a prefunded system. The transitional generation must continue to support retirees under the old system while saving for their own retirement. That issue is obviously not unique to privatization and must be faced in any reform of Social Security that moves toward a prefunded system.
  
  
• Some countries have required workers to shift to a new system of private accounts, and others have allowed workers to choose whether to join the new system or stay in the old pay-as-you-go system. Allowing choice can mean that the pay-as-you-go system lingers on and may (as in the United Kingdom) entail some additional administrative problems. But it can also help workers accept the change, particularly older workers who have substantial accrued benefits.
  
  
• Policymakers must decide whether to offer minimum benefit guarantees and how generous the guarantees should be. Without such guarantees, some people risk not having adequate retirement income. Making such guarantees, however, imposes a contingent liability on future taxpayers.
  
  
• Countries must decide how to regulate investment choices in the retirement system and how the retirement funds may be used. Regulation may be needed to limit fraud and risk--both the risk to retirees if investments turn sour and the risk to taxpayers if the plan guarantees minimum benefits. Regulations about how the retirement funds may be used, such as conditions for withdrawal and whether annuities would be mandatory, are also important. However, regulations also limit an individual's choice about investment and retirement.

TYPES OF PRIVATIZATION PLANS

The countries we examined followed one of three major models in privatizing their pension systems. Chile, Mexico, and Argentina used a model in which workers establish private retirement accounts. The United Kingdom allowed its workers to choose between the old pension system and the new system. Australia based its system on employers' contributing to retirement accounts for workers.

The Chilean Model

Chile, a pioneer in privatization, replaced its pay-as-you-go system with a system based on private retirement accounts in 1981. New workers had to establish private accounts. Workers already in the old system could choose to remain there or switch to the new system and earn a more attractive expected return on future contributions. To encourage switching, the government compensated workers who did so with "recognition bonds" that would be paid into a worker's account at retirement. Workers with sufficient years in the system were guaranteed a minimum retirement income of about 25 percent of the average wage. Obligations to existing workers were financed with general revenue and debt (the recognition bonds).

Mexico and Argentina generally followed the same model as Chile, with some modifications. In Mexico, for example, all workers have been required since 1997 to join the new system and save in private accounts. At retirement, however, workers who have contributed to both systems may choose to receive benefits from either system (but not both). Argentina has both benefits that are financed on a pay-as-you-go-basis (similar to those in Social Security) and private retirement accounts. People who choose to contribute to private accounts receive an additional pension that reflects their contributions to the old system (like the recognition bonds in Chile).

The U.K. Model

The United Kingdom, when it began its reforms in 1986, followed a different model. Its existing retirement system already had a privatizing option; that is, people whose employer offered a pension were allowed to opt out of part of the government's pay-as-you-go system. Those who did so received a rebate on their payroll taxes. The reform simply extended that option by allowing workers who set up a personal pension plan to opt out as well. Transition costs are financed out of general revenue (possibly including debt) and by reduced benefits in the government system.

The Australian Model

The third model is that of Australia, which chose to base its reformed system on employers by requiring most of them to contribute to workers' retirement funds. Unlike the other four countries, Australia never had a Social Security-like system funded by earmarked contributions. Instead, the government used general revenues to pay for a means-tested pension that was not regarded as an entitlement. Because the old system lacked a specific entitlement, it did not require the government to compensate workers for any benefits accrued under the old system. However, if the reform succeeds in replacing the government pension, it will be true in Australia, as in the other countries, that one generation will pay for their parents' as well as their own retirement.
 

DESIGN ISSUES

The experiences of the countries that have already begun their reforms highlight the importance of the design of the new pension systems. Our analysis revealed three issues: the need for additional information if a complex system is to work; the need to regulate investment choices; and the need to regulate withdrawals from the accounts.

Information Requirements of a Complex System

The reform in the United Kingdom demonstrates the difficulties that can arise if the new system offers workers a large array of choices and decisions to make but does not ensure that the worker has sufficient knowledge to make informed decisions. In the U.K. case, figuring out whether they should stay in their employer-based plans or switch to the newly available private accounts was difficult for many workers. If they switched, they would lose accrued benefits in the employer plans but would gain a more attractive return in the private accounts. Under pressure from sellers of the private accounts--including, apparently, some fraud--some workers made poor decisions. The United Kingdom responded to that problem with more careful regulation. Sellers of private accounts now have to provide enough information to enable workers to make a reasonable decision.

Regulation and Risk

Regulation of investment choices within the private accounts differs among the five countries. Such regulation could be important to protect either retirees or taxpayers, who in many cases are on the hook to finance a minimum benefit guarantee if investments in the accounts prove to have been unwise. One would expect, therefore, that systems that guarantee a minimum benefit would tend to have more regulation, though that is not always the case.

Neither the United Kingdom nor Argentina has a contingent minimum benefit. A worker whose investments went sour (and who had worked long enough to qualify) would have to rely on a basic pension that was not means-tested. The basic pension therefore does not depend on how successful the worker's investments are. The possibility of poor returns in the private accounts does not explicitly impose any risks on taxpayers. Of course, taxpayers still have to pay for the basic pension.

By contrast, the basic pension is means-tested in Chile and Mexico. Workers in those countries can choose their investment portfolio. (Australia also has a means-tested pension, but employers generally choose the portfolio.) Consequently, workers in Mexico and Chile have an incentive to invest in risky assets offering high expected returns--the worker reaps all the benefits if the gamble pays off and can rely on the basic means-tested pension if it does not. Taxpayers in those countries thus have a greater interest in ensuring that returns on the private accounts do not fall too low. (Means-tested pensions can also have other disadvantages: for example, they can reduce incentives to work and save.)

The taxpayer thus bears part of the risk of poor investment choices in Chile, Mexico, and Australia but not in the United Kingdom or Argentina. One would therefore expect the United Kingdom and Argentina to have little regulation and the others to regulate investment choices more closely. As expected, regulation of investment choices is minimal in the United Kingdom, consisting mainly of the ordinary "prudent man" fiduciary standard, and is quite stringent in Chile and Mexico. The odd couple are Australia and Argentina. In Australia, taxpayers bear some of the risk of the accounts, but regulation is as light as in the United Kingdom. In Argentina, by contrast, taxpayers do not bear that risk, but regulation is as heavy as in Chile, which has in other respects also been a model for Argentina.

Regulation of Withdrawals

In Australia, workers can "game" the system by withdrawing all their money from the accounts at retirement and spending it, for instance, by paying down their mortgage or buying a new house. Housing receives special treatment under the rules for the means-tested pension. Currently, most people qualify for the pension. If that practice continues, the reform will have made almost no difference in the government's costs for retirement. Australia's experience suggests the importance of establishing rules that govern when, how, and for what purpose funds may be withdrawn from the accounts. Many proposals for reform in the United States, for example, prohibit lump-sum withdrawals and require workers to purchase an annuity at retirement. Having such rules would avoid the problem Australia encountered.
 

ADMINISTRATIVE COSTS

Most analyses of the administrative costs associated with proposals to privatize pension systems examine the cost of managing private accounts. That is, of course, only one part of the cost of a proposal; both the current social security system and any reformed system also impose administrative and accounting costs on employers and workers. CBO is now conducting a more detailed study of administrative costs in a privatized system.

Comparing the administrative costs of managing private accounts for the five countries is quite difficult. Some plans take out administrative costs as an initial payment at the time of investment, and other plans charge an annual fee. The different fee mechanisms preclude any direct comparison, particularly since most of the reforms are recent and the plans have not matured. Nevertheless, a couple of lessons have emerged.

First, fees and commissions of individual accounts appear to be close to what managed mutual funds charge for individual accounts in the United States. In Chile, account fees and commissions are about 1 percent of the assets held in Chilean pension accounts. A 1 percent charge is quite common for managed mutual funds in the United States. The large accounts in Australia that give limited choices to workers seem even less costly, with fees approaching those that index funds charge in the United States (about 1/3 percent of assets). In addition to managing investments, systems with individual accounts need to collect and maintain data in more detail and collect it more frequently than a large-scale public system without individual accounts. Such systems therefore tend to be more expensive than, for example, the U.S. Social Security system.

The second lesson is that design choices seem to affect management costs. In Chile and the United Kingdom, for example, funds are marketed directly to individuals, which leads to relatively high sales costs and little bargaining power for purchasers. In addition, workers in Chile can switch funds several times a year, and workers in the United Kingdom can contribute sporadically and to several small accounts. All those factors increase total administrative costs. In Australia, by contrast, companies representing many individuals and contracting on a more stable basis face much lower fees.
 

NATIONAL SAVING

All of the reform plans hoped to reduce strains on the government's financing of retirement and, by encouraging private saving, increase the national saving rate. That is an important goal because the only way that real resources can be put aside for retirement is through saving and capital investment in plant and equipment and human capital (education and training).

Because of limited information on what the governments and workers would have done had the pension systems not been reformed, estimating the reforms' exact impact on national saving is difficult. In the United Kingdom, the fiscal tightening associated with pension reform indicates that the government offset little if any of the additional private saving in personal retirement accounts. In Chile, a fall in government saving probably offset only a portion of the increased private saving. As a result, Chile's national saving rate may have increased by 2 percent to 3 percent of gross domestic product (GDP). In Australia, estimates indicate that under certain behavioral assumptions, the reform might increase national saving by about 1.5 percent of GDP in the long run. The saving effect of reforms in Mexico and Argentina cannot yet be ascertained; however, the gains in national saving are probably less in Mexico and Argentina than in Chile.

Another important lesson from the countries we studied is the difficulty of funding a retirement system controlled by a national government. Several of the countries intended to fund or partially fund their systems over time. However, in each case the good intentions were overcome by demographic pressures and the ease with which trust funds can be deployed for other purposes. A motivating force for privatization may have been the failure of the national governments to establish and maintain a cache of assets in a trust fund as we commonly understand it.
 

CONCLUSION

The aging of the population is not unique to the United States--many countries are experiencing growing retirement populations supported by fewer workers. Those facts mean, in part, that the traditional pay-as-you-go pension and health care programs for retirees will be strained. Other countries have undertaken, and the United States is considering, reforms to those programs to help ensure future benefits.

Judging the desirability of reform--indeed, judging the results of other countries' reforms--depends critically on at least two related questions: Can the reform help economic growth? And can the reform reasonably be expected to work?

The first question is critical. It is ultimately the size of the economy that determines our ability to support a growing elderly population with fewer workers. Increasing national saving should enhance productivity and thereby economic growth. Increased saving results from funding a heretofore unfunded system with real assets, not with increases in government debt.

The second question addresses considerations of practicality, ease and cost of administration, protection against severe losses, and the extent of regulation.

Our comparisons of the five countries suggest that:

• None of the five countries successfully maintained permanent prefunding of their government-run, defined benefit pension system.
  
  
• Prefunding through privatization offers an opportunity to increase national saving and economic growth.
  
  
• Administrative concerns, including cost, do not appear to be insurmountable, but the details are important.

1. See Congressional Budget Office, Social Security Privatization: Experiences Abroad, CBO Paper (January 1999).