A quarterly magazine of the IMF

September 2002, Volume 39, Number 3

The Latin American countries are at the vanguard of global pension reform. Eight have reformed their pension systems in the past 20 years, and additional reforms are now being considered throughout the region. Did the earlier reforms work? What should new reforms aim for? And are the ideas driving the reforms sound?

Public pension systems in many developed countries face crises that are due largely to demographic factors—retiring baby boomers, shrinking working-age populations, and lengthening life spans. Public pension systems in many developing countries, including those in Latin America, also face crises. In these countries, though, while demographics do play a role, the principal problem is that benefits under the old publicly managed, defined-benefit, pay-as-you-go systems were disproportionately generous relative to contributions, encouraging early retirement and discouraging labor force mobility. The only options open to policymakers for sustaining these systems were to substantially raise taxes or cut benefits—neither of which was politically feasible.

Some Latin American countries therefore undertook systemic reforms. Although reforms differed across countries, most countries jettisoned all or a significant part of their public social security systems in favor of privately managed, individual-account-based, defined-contribution, funded systems (similar to the 401(k) plans in the United States, with the added twist that they are mandatory) largely along the lines suggested in Averting the Old Age Crisis, a study published by the World Bank in 1994. The new systems were supposed to improve sustainability, ease fiscal pressures, reduce old-age poverty, and strengthen financial markets—all of which were expected to stimulate economic growth.

The reformed systems have had impressive beginnings in terms of the volume of assets under the management of the private pension funds, gross returns, and number of participants (Table 1). Some analysts report that the benefits of the reforms have included lower implicit debt-to-GDP ratios, higher participation rates in terms of affiliates to the new systems, and increased equity in terms of rates of return on pension contributions obtained by participants across wage profiles and genders. Private pension funds—because of their large size relative to the overall economy—have also been driving financial sector reforms. While each country's experience is different, many of these benefits, as well as the question of whether the new regimes represent a net improvement over the old systems, are still very much a subject of debate.

Although there has been little criticism in Latin America of the decision to abandon the public systems—surprisingly, given the raging debate on the same issue in the developed world—the new private defined-contribution systems have come in for a fair amount of criticism, particularly in four areas: coverage, fund performance, governance, and fiscal costs. These criticisms have reignited the debate about the various approaches to pension reform. But the lack of information—or the misinformation—about pension systems has sometimes impeded a balanced assessment of these approaches. In reviewing the criticisms of Latin America's reformed pension systems, we hope to clarify some of the misconceptions associated with pension reform.

Coverage

There is a concern that Latin America's private pension systems, given their limited coverage, may not be able to achieve the stated goal of broad-based reduction of old-age poverty. These systems are mostly accessible only to formal sector workers, and only a fraction of these actually contribute, even though many workers are affiliated with the new systems. Therefore, many Latin Americans are still not covered by a formal system of retirement income security. At present, about one-third, on average, of the countries' economically active population is covered. In the two decades since Chile reformed its system, coverage in that country has barely increased. Stagnant or shrinking formal sectors in many Latin American countries are making the problem worse, as the number of potential contributors decreases. Even among those eligible to contribute, the ratio of active contributors to those nominally registered in the systems is declining. In pension systems that tie benefits to contributions, a limited number of contributors means that only a fraction of the economically active population will receive an adequate pension after retirement.

One way to increase coverage is to make participation mandatory for individuals who are currently excluded—such as the self-employed. However, it is not clear if there are adequate incentives for the self-employed to comply, and enforcing compliance is administratively expensive and difficult. Another option is to increase the share of the formal sector in the economy, thereby enlarging the pool of workers required to save for retirement. This approach, however, requires reversing the recent expansion of the informal sector, something that is complex and likely to take a long time. A third approach is to provide old-age income security through, for example, grants, social pensions, and social assistance to those who have not contributed to, and are not covered by, the new pension systems. But, while universal and non-means-tested old-age benefits may be effective in addressing old-age poverty, they are usually expensive and inefficient, and means-tested programs are difficult to administer. Other policy options include lowering barriers to participation in pension systems, educating workers about saving for retirement, and redesigning minimum benefit guarantees. Better investment of pension assets would also provide greater incentives for workers to contribute to the pension system.

Implementation of some or all of the above steps will contribute to increasing coverage. But the introduction of a defined-contribution system, even though it provides incentives to workers to save for their retirement, may not automatically increase coverage. The size of the informal sector is a major determinant of the scope of any pension system, and expanding the formal sector is complex. A well-designed defined-contribution pension system can reduce labor market distortions by clearly linking contributions and benefits. But pension reform is only one component of the set of policies needed to reduce old-age poverty.

Fund performance

The savings out of which pension benefits are paid to retirees from private pension systems equal their contributions plus the returns on those contributions less fees charged by pension funds. Contributions are usually stipulated in reform legislation. Therefore, accumulating adequate savings requires good returns and low fees. Critics feel that Latin America's new pension funds will not be able to provide retirees with adequate savings. First, many pension funds do not yet invest in well-diversified portfolios (Table 2). Second, contributors have little choice in how their retirement savings are invested and virtually no say in how they are managed. And, third, the administrative costs of private pension systems continue to be high.

Table 2 Improving returns Pension fund portfolios in Latin America are not diversified enough. (percent of assets as of December 2001)   Govt. bonds Bank deposits Corp. bonds Equities Invest- ment funds Foreign securities Other Argentina Bolivia Chile Colombia El Salvador Mexico Peru Uruguay 69.5 73.2 35.0 48.7 80.7 89.4 13.8 61.2 10.9 10.5 17.5 13.5 17.7 00.0 24.4 19.2 01.7 13.4 21.0 24.2 01.6 08.2 30.8 08.5 10.2 00.0 10.6 05.1 00.0 00.0 24.8 00.0 3.6 0.0 2.4 0.7 0.0 0.0 0.5 0.0 01.8 00.0 13.4 00.0 00.0 00.0 04.8 00.0 02.3 02.9 00.1 07.8 00.0 02.4 00.9 11.1    Source: International Federation of Pension Fund Management Companies (FIAP).

Relaxing regulations on asset allocations by pension funds increases the possibilities for diversification, reduces risks, and can increase returns. Although arguments can be made for limiting investments to government bonds during the early years of a new pension system, experience has shown that, over long investment horizons, a diversified portfolio with a large share of investments linked to the real sector is essential for good returns. In many Latin American countries, a shift away from an emphasis on investments in government bonds toward a more balanced portfolio of domestic and international assets (including equities) would also provide greater opportunities for wealth creation and permit further development and strengthening of financial and capital markets, one of the stated objectives of pension reform.

Regulations governing the assessment of pension fund performance need to be redesigned. Currently, assessments are based on comparing the return of one fund with the average return of all funds in the pension system. With a growing concentration of pension fund assets in a few funds in most countries, such regulation creates incentives for funds to "herd" around the leaders and hold portfolios that are nearly identical to each other and yield similar returns, thereby narrowing the choices available to contributors. Using the pension industry's average performance as a benchmark skirts the crucial issue of how well the benchmark itself is performing, compared with other investment opportunities. Moreover, if pension reform is to strengthen financial markets, sound benchmarks must be developed for equity and bond portfolios.

Contributors to private pension systems in Latin America have little choice in terms of the portfolio allocations that might be best for their stage of life or of their risk preferences. At present, in most countries, each fund manager offers only one fund to all contributors (and the holdings and returns of most funds are similar), whereas, outside Latin America, the trend in defined-contribution systems is to allow workers to tailor their portfolios to their individual needs. Of course, increased choice does not automatically enable contributors to make intelligent and well-informed decisions. Giving contributors more types of investments to choose from should be coordinated with educating them about the importance of saving, investment diversification, and the power of compounding over time.

Policymakers in Latin America also urgently need to lower the costs of fund management. Commissions, which, in most pension funds, are based on contributors' wages and charged up front for the life of the contribution, are higher in Latin America than elsewhere. The way commissions are structured, combined with the erratic work history of the typical contributor, affects poorer workers disproportionately. Lowering administrative costs will require some basic changes in most countries' pension funds—for example, from some current models where both collection and management of assets have been decentralized to models with centralized collection. However, centralization requires improved governance, better regulation, and reduced political interference.

Pension reform is not carried out, as many believe, simply with the "stroke of a pen." It is an ongoing process. Efforts to improve pension fund performance must be continuous, and the design of a pension system, however appropriate it may be when the system is introduced, often needs to be fine-tuned after the system is up and running.

Governance

Critics of Latin America's pension reforms also claim that governance of the reformed systems is weak—weak governance contributed to the shortcomings of the old public pension systems—and that large investments by pension funds in government bonds are essentially a back door to financing government deficits. Advocates claim that the public sector has poor incentives to manage pension systems well and that privately managed pension systems are therefore inherently better. There are at least two channels through which governments influence the performance of private pension funds.

The first is the quality of pension fund management, which Latin American governments have, in general, tried to monitor through regulation, implementation, and supervision. Although the quality of supervision is a function of a country's overall level of institutional development, in most of Latin America—unlike some other parts of the world—the oversight of regulators has generally been effective, and responses to mergers and takeovers by pension fund managers, sound. Contributors' pension fund assets have largely been protected from any deterioration of operational soundness and fraudulent behavior on the part of pension fund managers through safe custody of assets and failure-resolution mechanisms. As the pension fund industry evolves, it is critical that the role of regulators be further strengthened and oversight maintained.

The second channel is governments' macroeconomic management and fund regulation. Governments create the macroeconomic environment and maintain the integrity of the financial systems in which pension funds operate. Even small regular contributions can grow into significant amounts over the long time horizon of a pension system. However, such accumulations retain their value only if government policies contribute to macroeconomic stability. Governments also need to resist the temptation to dip into the large pool of pension assets by requiring large investments in government bonds or investments based on patriotic and other nonfinancial considerations.

A common misperception is that private pension systems can be insulated from the poor governance that has afflicted the publicly managed systems. However, there is little reason to believe that a government that administered a public system poorly would regulate a private one well. Of course, the specific avenue through which governance affects the pension system is different for public and private systems. That said, governance of pension funds is as much of an issue for the new systems as for the old. Ultimately, governance of the pension system can be only as good as that of the country overall. Privatization of pension systems does not diminish the importance of good public policy.

Fiscal costs

Another problem is that pension obligations under the old systems are putting a strain on government finances, prompting critics to ask whether and when the fiscal benefits of reform will be realized. After defined-contribution systems were introduced in Latin America, participation in them became compulsory for all new workers in many systems; those who had joined the labor force before pension reforms took effect were often given the choice of staying in the old system or moving to the new one. Incentives were sometimes provided for workers to move, while governments were still obligated to make pension payments to retirees who had been enrolled in the old systems. Thus, during a long transition period—almost 50 years in some countries—governments have to service their liabilities under the old systems while not having access to the new inflows, which are being directed to individual accounts. These liabilities, which amount to significant, though declining, fractions of GDP, are being financed out of fiscal revenues as well as by requiring private pension funds to invest in government bonds. In many instances, these liabilities are not new costs but simply an existing commitment made explicit by the change in the pension system. In general, the loss of revenues in the short run has a long-term counterpart in lower future spending.

Another area of concern is the likely accumulation of new fiscal liabilities under the reformed systems. Most governments have instituted some type of guarantee of a minimum level of pension (usually a fraction of the average wage), which they might one day be called upon to honor. Also, workers who have met the requirements to qualify for the minimum pension guaranteed by the government may be less likely to continue making contributions to the system (Packard, 2002). Investing contributions well and providing incentives to workers to continue to participate in the system are essential if future fiscal liabilities are to be reduced.

Introduction of a system based on individual accounts does not immediately lead to lower fiscal outlays for governments, as many believe. In fact, on a cash-flow basis, things may even get worse before they get better, although long-run savings are likely to be significant. Also, as discussed in Orszag and Stiglitz (2001), the common assertion that governments are likely to face less pressure to bail out private pension systems than they face with public defined-benefit systems is a myth. If governments do a poor job of regulating private pension fund managers and problems occur with one or more institutions, the many individuals investing in the national retirement system are likely to demand compensation from the government.

More reform needed

Pension reforms in Latin America respond to a very real need. Movement from a fully public system to one involving a greater role for private pensions is unavoidable and can yield important benefits. Properly structured and managed, mandatory national private pension systems can help alleviate old-age poverty. If such systems are funded—as all the reformed systems in Latin America are—they also mobilize substantial long-term resources that can contribute to reducing government liabilities for pensions, enhance financial and capital market development, and, potentially, lead to faster economic growth. The link between contributions and benefits is likely to be actuarially fairer in defined-contribution systems and provides workers with the right labor market incentives. Countries that postpone reforms—or that choose radically different options—risk finding themselves with worse problems later.

But more needs to be done, and done better. Contrary to initial expectations that a once-off reform would be enough, there will always be the need for adapting systems to changes in countries' situations and populations. Coverage needs to be expanded, regulatory reforms implemented, and governance improved. Finally, as the global economic picture begins to look less rosy, and Latin American economies face more challenges, managing the costs of transition to the defined-contribution systems will be more demanding.

References:
  Estelle James, Dimitri Vittas, and James Smalhout, 2001, "Administrative Costs and the Organization of Individual Account Systems: A Comparative Perspective," World Bank Policy Research Working Paper No. 2554 (Washington).
  Peter Orszag and Joseph Stiglitz, 2001, "Rethinking Pension Reform: Ten Myths About Social Security Systems," in New Ideas About Old Age Security: Toward Sustainable Pension Systems in the 21st Century, edited by Robert Holzmann and Joseph Stiglitz (Washington: World Bank).
  Truman G. Packard, 2002, "Pooling, Savings and Prevention: Mitigating the Risk of Old Age Poverty in Chile," World Bank Working Paper No. 2849 (Washington); http://econ.worldbank.org
  World Bank, 1994, Averting the Old Age Crisis: Policies to Protect the Old and Promote Growth, World Bank Policy Research Report (New York: Oxford University Press).