Committee on Ways and Means, 2-11-99 Testimony

Statement by Stephen J. Kay
Economic Analyst, Latin America Research Group
Federal Reserve Bank of Atlanta, Atlanta, Georgia

Testimony Before the House Committee on Ways and Means

Hearing on Social Security Reform Lessons Learned in Other Countries

February 11, 1999

My name is Stephen Kay. I am an economic analyst in the Latin America Research Group at the Federal Reserve Bank of Atlanta. Prior to joining the Federal Reserve Bank of Atlanta last year, I spent several years conducting research on social security reform in Argentina, Brazil, Chile, and Uruguay. I am here as a private citizen, and the views outlined below are my own, and do not reflect the views of the Federal Reserve Bank of Atlanta, or the Federal Reserve System.

Chile's 1981 pension privatization has garnered a great deal of international attention, receiving praise for its relative transparency and simplicity, and its role in promoting the development of Chile's capital markets. However Chile's new system of individual savings accounts has experienced its share of difficulties, which also merit our attention. The new system has been criticized for its high operating expenses, commission charges, and transition costs, as well as its low rates of compliance and its distributional impact on women. By studying the strengths and weaknesses of Chile's new private system, we are better able to understand the potential risks and rewards of defined contribution social security systems that are based upon individual savings accounts.

Background

Prior to recent reforms, South America's social security systems were in varying degrees of disarray: aging populations and massive evasion by both employers and workers meant that fewer contributors were supporting more pensioners, surpluses had been wasted on bad investments, benefits were highly inegalitarian and financed regressively, deficits were mounting, administrative performance was poor, and payroll taxes were among the highest in the world.⁽¹⁾ Social security systems were also organized into multiple sub-systems, each with its own administration and benefits structure.

By the end of the 1980s (Latin America's "lost decade" of economic development) there was consensus in the region that reform was necessary; however intense political conflict arose over the direction of reform. Reforms in the Southern Cone of South America ranged from partial privatizations in Argentina (1993) and Uruguay (1995), to a short-lived privatization effort in Brazil.

Chile became the pioneer of privatization when the Pinochet dictatorship implemented the world's first-ever social security privatization in 1981. Under the tripartite "pay-as-you-go" (PAYG) model used in most of the world (including the United States), a combination of payroll taxes on workers and employers, and government contributions (when necessary) are used to fund social security benefits. In Chile's new defined-contribution system, workers are required to contribute 10% of their salaries to individual investment accounts, where funds are invested by private pension fund companies in closely regulated portfolios. An additional 3 percent of a worker's salary goes toward commission fees and a disability and survivors' insurance premium. Pension fund companies must guarantee profitability relative to the average profitability in the pension fund industry. The self-employed are not required to join a pension plan (only about 10% do), and the military and police have kept their relatively generous defined-benefit PAYG systems.

In Argentina and Uruguay, democratically-elected governments found little popular support for a Chilean-style reform, and consequently both reforms differ from Chile's privatization in two significant respects. First, both systems maintain a universal public pay-as-you-go benefit, while Chile's PAYG system is being gradually phased-out. Second, membership in the private system is optional for Argentine and Uruguayan workers (although workers in Uruguay must make contributions to a pension fund on earnings between $800 and $2500). In Chile the new private system was optional when it was introduced (there was a financial inducement to join), and it has been mandatory for all workers entering the labor force since the reform.

Administrative Costs

Chile's new private system is plagued by high administrative costs, which are in part passed on to workers in the form of high commissions. Commissions have come down from their peak of 8.69% of taxable salary in 1984 to around 2.96% in 1997 (these include a disability and survivors' insurance premium of around 0.7%). High costs are expected during the start-up phase of a new pension fund industry. However in 1997, commission charges alone still accounted for around 18% of a worker's total contribution. Although its founders expected that competition would lead to lower commission costs, this has not occurred, and expenses generated by marketing have helped keep costs high. The recent trend in the industry is toward greater concentration, as three firms (out of a total of eight), controlled 73% of all affiliates in 1998.

Marketing and operations costs have been high as Chilean pension funds have engaged in expensive sales campaigns to capture workers from competitors. 28% of affiliates in Chile switched pension funds in 1996. Marketing costs absorb between 30% and 40% of all operating costs. Roughly a third of these administrative costs are generated by salespersons seeking to persuade workers to switch funds.(2) The government has recently enacted measures aimed at reducing the number of transfers by making the process of switching pension funds more complicated. Chilean regulators are also considering a number of steps to lower commissions, including a plan to allow lower group-rate commissions (which are currently prohibited), and a proposed rating system where pension funds would be ranked according to charges and service provided. In 1998, commission fees as a percentage of contributions dropped an average of 8%.

Upon retirement, workers face a number of options. The accumulated funds may be used to purchase an annuity indexed against inflation, or pensioners can elect to receive a "programmed pension," paid directly by the pension fund company, based on the accumulated funds in an amount that is reassessed every year based upon the fund's investment performance. Workers may also elect to withdraw funds in a lump sum, as long as they leave enough capital to purchase an annuity in the amount of 110% of the minimum pension. These options give workers greater control over their funds, but they pose the risk that workers may outlive their income (in the case of a programmed withdrawal), or spend a large portion of their retirement income at once and be left with a pension just slightly above the minimum. The annuities market in Chile has not functioned well, with no provision in place for group contracts. Since these options are complex, workers are exposed to intense selling pressure by insurance agents who may charge as much as 4% of the value of the contracts.(3)

In the Chilean system, government guarantees against inflation are provided for life annuities which can be purchased upon retirement (until workers purchase annuities they are exposed to the risk that inflation could diminish their capital).(4)

Retirement pensions granted in the next few years are not indicative of future benefits because between 60% to 75% of the funds accumulated in these individual accounts will come from government "recognition bonds". These bonds are paid to pensioners upon retirement in recognition of past contributions made to the old public system, and carry a real interest rate of 4%.(5)The true test will come in another thirty years, when workers who have contributed exclusively to the new system begin to retire.

Returns and Pensions

Since its founding, Chile's private system claims to have achieved an impressive average annual return on investment. The figure usually cited shows that the pension funds have produced a real average annual return of 11% since 1981, but these are gross returns, without consideration of investor-paid commission expenses. Once commissions are factored in, the real average return is considerably lower. For example, while the simple real average annual return on invested pension funds between 1982 and 1995 was 12.7%, this figure does not incorporate the commission charges that workers pay on contributions. If you consider the amount workers contributed and deduct commission charges, an individual's real average annual rate of return over this period would be 7.4%. The disparity between these two figures illustrates how commissions affect the rate of return. Over shorter periods of time, the impact on the rate of return is even greater since contributors may earn negative or very low returns for several years (imagine the impact of an 18% load). (6) Furthermore, fixed commissions are a source of regressivity in the new system because these fees consume a proportionately greater percentage of the contributions of low-income workers.(7) The high cost of pension fund accounts for poor people led the World Bank to suggest that poor people might be better served by saving for retirement in bank savings accounts.(8)

Each pension fund company offers just one investment fund, and these funds are required to deliver returns comparable to their competitors over a twelve-month period, or make up the difference from their reserves. Funds are required by law to deliver a real return that is no less than 50% of the industry average, or 2% below the industry average, whichever figure is lower. As a result, there is little divergence among returns as each fund seeks to emulate the returns of its competitors. In order to encourage pension funds to take a longer term view and avoid the "herd effect" of uniform returns, the government is considering extending the time period over which relative fund pension fund performance is measured. However any movement away from uniform portfolios has to be weighed against the risk of workers receiving lower returns by choosing poorly-managed funds.(9)

While the system has achieved high annual returns thus far, their sustainability is uncertain. Chile's high returns resulted from specific macroeconomic circumstances in the 1980s. The economy had hit a low point in 1982 when real GDP fell by 14%. After the banking crisis in 1981-83, real interest rates were very high. Pension funds invested heavily in government debt instruments, and when real rates fell, they realized large capital gains. Pension funds increased investment in equities in the 1990s, and high real returns came mostly from the impressive performance of the stock market.(10) Stock prices increased in part because of pension fund demand(11), and conversely, were adversely affected in 1998 by the decision of pension funds to cut their holdings of Chilean shares in half (from 28% at the end of 1997 to 14% at the end of 1998). Pension fund returns over the past four years have averaged under 2% before commissions.

Regulations on minimal profitability and requirements that each pension fund company can only administer one fund have restricted diversity among fund portfolios as pension funds have largely produced similar returns.(12) Until now, pension fund companies have been limited to offering just one investment portfolio, even though individuals have different tolerances for risk according to their ages. For example, an individual close to retirement might want to have a portfolio largely composed of bonds, while a younger worker's asset mix would be invested heavily in stocks. In July, a World Bank report criticized Chile's pension fund managers for not offering a wide enough asset mix. The danger of this was brought home last fall, when the average pension fund had lost 9.9% between January and September (the year-end rally in financial markets cut the average loss to 1.1%), a development that no doubt caused hardship for individuals planning to retire last fall. The government now plans to allow pension funds to offer a fixed-income fund that will only be open to workers with ten or fewer years left until retirement. This fund will not be available to all workers, reportedly because of fears that this would generate an exodus from existing pension funds.(13)The government is also considering permitting the introduction of other investment portfolios aimed at younger workers.

Predictions for future returns, which will determine pension benefits, range between 3% and 5%. Pension benefits will vary dramatically depending on which, if any, of these predictions hold true. For example, a 3%, 4%, or 5% annual return on retirement savings in Chile would lead to benefits representing 44%, 62%, and 84% of pre-retirement earnings for men.(14) One study estimated that for Chile to achieve the system's goal of a 70% earnings-replacement rate, the system would have to have annual returns of around 4.5% (leading numerous analysts to suggest that the 10% workers contribution might not be sufficient to generate a 70% earnings-replacement rate(15)

Impact on Women

Perhaps the most striking distributional consequence of moving to a system of individual accounts concerns the treatment of women. In the old PAYG system, the disparity in benefits between men and women was smaller. When compared to PAYG systems, the private social security systems in South America disadvantage women by strictly linking benefits with earnings, and placing men and women in separate actuarial categories. Because they tend to earn less, spend more years of their lives in unpaid labor, and have greater longevity, women purchasing annuities upon retirement will systematically receive lower benefits than men.

In a forthcoming study, economist Alberto Arenas de Mesa and sociologist Veronica Montecinos project the rates of return that women would need in order to achieve the same pensions as men. For example, assuming identical wages and years of contribution, a woman retiring at age 65 and purchasing an annuity would receive approximately 90% of what a man would receive. When we consider the actual disparities in income profiles and years of contribution, the differences are even more striking. The authors cite a 1992 study that found that a typical woman retiring at age 60 and purchasing an annuity after earning a 5% annual rate of return would receive a replacement rate of 57% of her former salary, while a man retiring at age 65 would receive 86%.(16)

Non-Compliance

Many believed that the private system would reduce evasion because workers have a greater incentive to contribute to their own personal retirement accounts than to a PAYG system. However, evasion persists. Only 60 to 62% of workers are covered by the new system, figures that are similar to the old system.(17)In August of 1998, just over half of workers covered by the new system (55.3%) made contributions to their accounts.(18) ^{The compliance rate also varied by income
level. Funds which catered to lower-paid workers received contributions from 45-55% of
their worker-contributors, while those serving higher paid workers had a compliance rate
of 80-90%.(19)} To further illustrate the problem of non-compliance, as of December 1995 over 35% of the 5.4 million contributors to the private system had accumulated less than $500 in their accounts, while more than half had less than $1228 in their accounts.(20)

Part of the compliance problem may be related to a moral hazard incentive in the program for workers not to comply. The government provides a subsidy to workers who contribute for at least twenty years but do not accumulate enough capital to earn a minimum pension. This provides an incentive for individuals to evade by contributing just enough to qualify for a minimum pension, but no more, thereby shifting the funding burden to the taxpayers. The minimum pension is approximately 30% of the average salary, and is currently around $120 a month for those under the age of 70. Government subsidies for the minimum pension will rise dramatically in the future, as some estimate that the number of affiliates who will not save enough to receive a minimum pension could be as high as 70%.(21)

Capital Markets, Growth, and Savings

According to economist Nicholas Barr, the effects of a funded pension system on national savings, and hence economic growth is "arguably the most controversial area" of this debate, and he suggests that the "experience of countries in the West is inconclusive both theoretically and empirically."(22) Increased pension fund savings are likely to be offset by a decline in government savings as payroll taxes are diverted to private accounts. Individuals expecting a larger retirement benefit may also elect to save less in other ways.

Although some have claimed that privatization explains the meteoric rise of Chile's national savings rate (which climbed from around 15% of GDP in the 1980s to 27% in 1995), recent studies negate this claim.(23) One study argues that Chile's rapid economic growth is the primary reason for the increased savings rate, with the pension reform contributing to growth in the savings rate equivalent to 3% of GDP.(24) Arrau argued that increased corporate savings resulting from Chile's 1984 tax reform played a large role in boosting Chile's savings rate, and that the private pension system's direct contribution to the increase was around 1% of GDP.(25) Another study concurred, concluding that "A very popular perception in Chile and in international circles is that the introduction of a privately-administered pension system based on individual capitalization has been the driving force behind the growth in national savings.YHowever the principal sources of increased savings in Chile are elsewhere: in private enterprise and the government."(26)

There is agreement that shifting to a funded pension system has contributed to the deepening of Chile's domestic capital markets, which in turn has had a positive impact on economic growth.(27) In a country like the U.S., with its well-developed capital markets, the same process may not occur. As economist Sebastian Edwards put it, "It is not clear that these mechanisms that have benefited Chile will be there in other, more developed countries."(28)

Transition Costs

In a defined contributions system with individual accounts, workers stop paying social security contributions to the government, but the government will continue to owe benefits to individuals belonging to the old PAYG system. This shortfall in revenue can only be financed through cutting other areas of government spending, raising taxes, cutting benefits, extending the retirement age, and/or by issuing debt. The Chilean reform did many of these things. Prior to privatizing, the Chilean government raised the retirement age to 65 for men and 60 for women and eliminated special early retirement programs. The government ran budget surpluses, privatized state-owned industries, and issued bonds that were purchased by the new pension funds. Since 'recognition bonds" (recognizing past contributions to the old PAYG system) are not issued to workers until retirement, the transition costs are incurred over an extended period of time. In 1996, the transition costs were 3.7% of GDP, and are expected to range between 3% and 4% of GDP over the next five years.(29) If welfare pensions, minimum pensions, and the deficit in military and police pension funds are included, the 1997 figure would be 5.5% of GDP.(30) As the transition costs grew in the 1980s, they consumed relatively greater percentages of social spending, while expenditures on health and education were cut (social spending has increased since Chile's return to democracy).(31)

Conclusions

Prior to their recent reforms, social security systems in the Southern Cone of Latin America suffered from financial and administrative problems that we in the United States have not encountered, and debates over reform were informed by fundamentally different political and economic realities. By most measures, governments in Latin America had failed to provide adequate social security coverage. Privatization has offered an alternative strategy that has been pursued, to varying degrees, throughout most of the region. This statement has outlined some of the costs, risks, and distributional consequences associated with the Chilean reform. Continued study of the pension reforms in Chile and the rest of Latin America would contain valuable lessons for all of us as we proceed along the path of reform.

Endnotes

1. .For a description of the ills of Latin America's social security systems see Mesa-Lago, Carmelo. 1994. Changing Social Security in Latin America. Boulder: Westview.

2. Fidler, Stephen. 1997. "Lure of the Latin Model," Financial Times, 4-9-97.

3. Gillion, Colin, and Alejandro Bonilla. 1992 "Analysis of a National Private Pension Scheme: The Case of Chile," in International Labour Review. Vol. 131, No.2, 1992, p.171-195. Also see Vittas, Dimitri. 1995. "Strengths and Weaknesses of the Chilean Pension Reform." Mimeo. Financial Sector Development Department, World Bank.

4. On the risk of inflation and defined-contribution savings accounts, see Barr, Nicholas. 1992. "Economic Theory and the Welfare State: A Survey and Interpretation" in Journal of Economic Literature. Vol. 30 (June 1992), p.741-803 and Barr, Nicholas. 1987. The Economics of the Welfare State. Stanford: Stanford University Press.

5. Arenas de Mesa, Alberto, and Veronica Montecinos. Forthcoming (1998). "The Privatization of Social Security and Women's Welfare: Gender Effects of the Chilean Reform."Forthcoming in Latin American Research Review.

6. Shah, Hemant. 1997. "Toward Better Regulation of Private Pension Funds." Paper presented at the World Bank, LAC Division, 1-16-97.

7. See Arenas de Mesa, Alberto. 1997. Learning from the Privatization of the Social Security Pension System in Chile: Macroeconomc Effects, Lessons and Challenges. Dissertation. University of Pittsburgh, Pittsburgh (March) 1997. Arenas de Mesa quantifies how commission charges affect low and high income earners.

8. Bridge News, 1998. "Chile Press: World Bank Study Says Asset Mix in AFPs Lacking." Story #14616, 7-23-98.

9. Queisser, Monika. 1998. "The Second-Generation Pension Reforms in Latin America" in OECD, in Maintaining Prosperity in an Ageing Society, OECD Ageing Working Papers 5.4.. P.58.

10. See Vittas (1995).

11. On the impact of pension fund demand on stock prices, see Uthoff, Andras. 1997. "Reformas de los Sistemas de Pensiones y Ahorro: Ilustraciones a Partir de la Experiencia Chilena," in Ahorro Nacional: La Clave Para Un Desarollo Sostenible En America Latina. Madrid: Instituto de Relaciones Europeo-Latinoamericanas (IRELA).

12. Diamond, Peter, and Salvador Valdés-Prieto. 1994. "Social Security Reforms," in Barry P. Bosworth, Rudiger Dornbusch, and Raúl Labán eds., The Chilean Economy: Policy Lessons and Challenges Washington D.C.: Brookings Institute.

13. Ruiz-Tagle, Jaime. 1998. "Jubilo o Tristeza? El Futuro del Nuevo Sistema de Pensiones en Chile," in Problemas Actuales y Futuros del Nuevo Sistema de Pensiones en Chile. Santiago: Fundación Friedrich Ebert.

14. The figures on the rate of return and benefits are from Gillion and Bonilla (1992). Cheyre predicts a 5% annual return (see Cheyre, Hernan 1991. La Previsión en Chile, Ayer y Hoy. Santiago: Centro de Estudios Públicos).

15. See Gillion and Bonilla (1992 p.186) and Barr (1994 p.213).

16. Arenas de Mesa, Alberto, and Verónica Montecinos. 1999. "The Privatization of Social Security and Women's Welfare: Gender Effects of the Chilean Reform," forthcoming in Latin American Research Review, volume 34, No. 3. Arenas de Mesa and Montecinos project that a woman working 70% of a "normal" 45-year career (having remained outside the labor force due to family responsibilities), would receive 32% to 46% of her working salary (assuming 5% annual returns). In order to receive a 70% replacement rate, she would have to have an average annual return between 7.7% and 10%.

17. See Fidler (1997).

18. Salomon Smith Barney. 1998. "Private Pension Funds in Latin America-1998 Update." Salomon Smith Barney Industry Report, December, 1998.

19. These figures are from 1990. See Barr (1994 p.212).

20. See Shah (1997).

21. See Ruiz-Tagle (1998).

22. Barr, Nicholas, 1994."AIncome Transfers: Social Insurance" in Barr, Nicholas, ed., Labor Markets and Social Policy in Eastern Europe: The Transition and Beyond. New York: Oxford. See page 214.

23. For an analysis and summary of recent work on the relationship between pensions and the savings rate in Chile, see Uthoff (1997).

24. Gavin, Michael, Ricardo Hausmann, y Ernesto Talvi. 1997. "El Comportamiento del Ahorro en América Latina: Panorama y Consideraciones Globales," in Ahorro Nacional: La Clave Para Un Desarollo Sostenible En America Latina. Madrid: Instituto de Relaciones Europeo-Latinoamericanas (IRELA).

25. See Arrau, Patricio 1996. Nota Sobre El Aumento del Ahorro en Chile. CEPAL: Santiago. The 1% figure is cited in Titelman, Daniel. 1997. "Impacto De Los Fondos De Pensión En El Proceso De Ahorro Interno" in 1er Seminario Internacional Sobre Fondos De Pensiones. Buenos Aires: Asociación Internacional de Organismos de Supervisión de Fondos de Pensiones.

26. Agosin, Manuel R., Gustavo Crespi T. y Leonard Letelier S. 1997. Análisis sobre el aumento de ahorro en Chile. Office of the Chief Economist, Inter-American Development Bank, Red de Centros de Investigación, R-309. My translation.

27. See Holzmann, Robert 1997. On Economic Benefits and Fiscal Requirements of Moving from Unfunded to Funded Pensions. Santiago: ECLAC. Also see Barr (1994 p.214). Barr reports that according to some commentators, "this is the only substantial benefit of the pension reform which could not have been achieved by redesigning the old PAYG system."

28 See Fidler (1997).

29. See Asociación Internacional de Organismos Supervisores de Fondos de Pensiones, 1996. Reformas a los Sistemas de Pensiones. Buenos Aires: SAFJP. See p.85.

30. Interview, 2-3-99, Alberto Arenas de Mesa, Ministry of Finance, Chile.

31. Vergara, Pilar. 1994. "Market Economy, Social Welfare, and Democratic Consolidation in Chile," in Smith, William et al eds., Democracy, Markets, and Structural Reforms in Latin America. New Jersey: Transaction - North/South Center. See page 254. Also see Gillion and Bonilla (1992 p.192-195).