Questions and Answers about the Chilean Pension System
Statement before the U.S. Senate Special Committee on Aging, May 18, 2004

by Jacobo Rodríguez [Cato Institute].

1. What percentage of retirees draws a minimum pension from the government? How is that figure expected to change over time as personal accounts build up?

As of March 2002, the government had supplemented 33,029 pensions, including 11,759 old-age pensions, out of over 400,000 pensions, in its role as the financial guarantor of last resort in the new private system. Because the new system has tougher requirements to qualify for the minimum pension and is far more efficient than the old one, the cost to the Chilean taxpayer of providing a general safety net is lower than under the old system. Indeed the cost to the government of supplementing these pensions has been about $33 million. Projections about the percentage of pensions that will receive a government subsidy range from about 10 percent to close to 50 percent, but if returns continue to be above 4 percent in real terms and workers contribute to their accounts regularly, the government contribution will continue to be minimal.

2. Chile has been criticized for having high administrative costs? Do you believe this criticism is accurate? What has the rate of return been net of administrative costs?

The often-cited figure of 18-20 percent represents administrative costs as a percentage of current contributions, which is not how administrative costs are usually measured. This figure is usually obtained by dividing the commission fee, which is on average equivalent to 2.3 percent of taxable wages, by the total contribution (10 percent plus the commission). This calculation fails to take into account that the 2.3 percent includes the life and disability insurance premiums (about 0.7 percent of taxable wages on average) that workers pay, which are deducted from the variable commission, and thus overstates administrative costs as a percentage of total contributions. Also, if, for instance, the mandatory contribution were lowered to 5 percent of total wages instead of 10 percent, then administrative costs measured as a percentage of the total contribution would increase from 18.69 percent to 31.51 percent (2.3/(2.3 + 5)), even if those costs measured in absolute terms or as a percentage of assets under management remained the same.
When administrative costs are compared to the old government-run system, the criticism is not accurate. Chilean economist Raúl Bustos Castillo has estimated the costs of the new system to be 42 percent lower than the average costs of the old system. However, comparing the administrative costs of the old system with those of the new one is inappropriate, because the underlying assumption when making that comparison is that the quality of the product (or the product itself) being provided is similar under both systems, which is certainly not the case in Chile.
Furthermore, the Congressional Budget Office reported in 1999 that, “In Chile, the country with the longest experience with private retirement accounts, [administrative costs] can be equivalently expressed as 1 percent of assets, which is similar to costs of mutual funds in the United States.” The CBO report goes on to say that, “It is difficult to convert a charge on contributions to a charge on assets (typical for a U.S. mutual fund). The calculation depends on the rate of return and the length of the investment horizon and therefore does not yield a single figure.” Chilean economist Salvador Valdés has estimated the average annual cost of the AFP system to be equivalent to 0.84 percent of total assets under management over the life cycle of the worker, which is lower than the average cost of the mutual fund industry in Chile but higher than other savings alternatives.
To the extent that such administrative costs are still considered too high, that is the result of government regulations on the commissions the AFPs can charge and on the investments these companies can make. The existence of a “return band” prevents investment product differentiation among the different AFPs. As a result, the way an individual AFP tries to differentiate itself from the competition is by offering better service to its customers. One way to provide better service would be to offer a discount on the commission fee to workers who fit a certain profile—e.g., workers who have maintained their account for an extended period of time or who contribute a certain amount of money to their accounts; however, government regulations do not allow that. Those regulations state that the AFPs may only charge a commission based on the worker’s taxable income and expressed as a percentage of that income.
Another reason administrative costs are not as low as they could be is that AFPs have a monopoly in the administration of pension savings accounts. Mutual funds, banks, insurance companies, and individuals themselves are not allowed to manage those accounts. The existence of this monopoly (which is part of the fragmentation of the financial services industry in Chile across product lines) prevents the establishment of one-stop financial supermarkets, where consumers can obtain all their financial services if they so choose. Such supermarkets would substantially reduce administrative costs by eliminating the duplication of commercial and operational infrastructure.
The average rate of return net of administrative costs for the average retirement savings account has ranged from 7.18 percent to 7.50 percent, depending on the type of account, from 1981 to April 2001, according to the Chilean government agency that regulates the industry.

3. Some people say that women and low-wage workers will disproportionally end up receiving the minimum benefit guarantee, increasing income disparity. Do you believe this is correct, and why?

That claim is partially accurate. It is true that women and low-wage workers are likely to accumulate less than the average worker. Women because they tend to earn less than men, have more irregular professional lives, and may stop contributing to their accounts at age 60 (that age is set at 65 for men). (Women also tend to live longer, a factor that also contributes to making the average pension for women lower than the average pension for men, all other things being equal.) All those characteristics are common to women everywhere and not just Chilean women and should not be considered features of the Chilean system. Since the new system gives every worker property rights in his or her contributions, every worker with 20 years of contributions will receive at least the minimum pension. That was not the case in the old pay-as-you-go government system, a system that especially penalized women (and other workers) with irregular professional lives.
Low-wage workers in general accumulate less than average workers because they are low-wage workers. Low-wage workers also tend to start working at an earlier age than other workers, which conceivably can make up for the smaller amount contributed per period, and to have a shorter life expectancy, which conceivably can allow workers to make larger withdrawals per period of time than other workers with a longer life expectancy.
Therefore, it is not correct to say that women and low-wage workers will disproportionally end up receiving the minimum pension. The reform was undertaken under the assumption that if a worker contributes to his account 10 percent of his salary for 35 years, and the real rate of return on his investment is 4 percent on average, he will have enough funds accumulated in his account upon retirement to fund a pension that is equivalent to 70 percent of the average salary over the last 10 years of his working life.
I think that focusing on whether income disparity increases under a private system or not is mistaken. What matters is that poor workers (as well as rich ones) have property rights in their contributions and can invest their savings in productive investments, so that they live their old age with comfortable means, even if other workers are much wealthier. The income disparity between Bill Gates and I, for instance, matters nothing to me. What matters to me is that Bill Gates has developed tools that allow me to become a more productive worker and, consequently, earn a higher salary, which in turn allows me to live more comfortably now and, hopefully, in my old age.

4. You mention that the current commission structure encourages funds to seek out higher-wage workers. How would your suggestions to liberalize commission structure (allow funds to offer discount and different combinations of price and service) affect low-wage workers? Would funds be interested in attracting low-wage workers?

AFPs are not allowed to offer discounts for permanence, for making voluntary contributions, for groups, or for maintaining a specific balance in an account. For instance, if workers were able to negotiate group discounts, then their bargaining power would significantly increase. That would allow them to negotiate lower commissions, which would benefit low-wage workers the most. Funds would continue to seek out low-wage workers so long as the marginal cost of administering the account of a low-wage worker (or a group of low-wage workers) does not exceed the marginal revenue derived from administering those accounts. If the administration companies were allowed to adjust their service to the ability and desire of workers to pay for those services, low-wage workers would have nothing to lose if the commission structure were liberalized. Those concerned that the services provided to low-wage workers would drop to unacceptable low levels need not be, as the government already mandates a minimum of services that AFPs have to provide to their clients.

5. If the worker dies before retirement, what happens to the account balance? What if the worker dies after retirement?

If a worker dies before retirement, the balance in his account belongs to the beneficiaries of his estate, as workers now have property rights in their contributions. If a worker dies after retirement and if he chooses the programmed withdrawal option, then the balance in his account belongs to the beneficiaries of his estate. If he chooses to purchase an annuity from an insurance company, the balance in his account upon retirement is used to purchase the annuity and the account is closed, so money is left to the beneficiaries of his account.

6. The government has started allowing companies to lower their variable fees while raising flat fees. What effect will this have on workers at different wage levels?

Increases in flat fees and reductions in variable fees would eliminate the cross-subsidy from high-wage workers to low-wage workers that is present today.

7. Why did Chile choose to primarily base administrative fees on contributions and not assets?

When the system began, AFPs were allowed to charge fixed and variable commissions on assets under management, fixed and variable commissions on contributions, or any combination thereof. AFPs were not allowed to offer discounts for permanence, group discounts, discounts for making voluntary contributions, or for maintaining a specific balance in the account. In 1987, the commission structure was changed by eliminating all commissions on assets under management. This change had the effect of providing a cross subsidy to (1) workers who do not contribute to their accounts regularly, because the fund manager is still providing a service (administering the account of those workers) for which he is not receiving compensation; and (2) to low-income workers, because the administrative costs of managing the account of wealthier workers are not proportionally higher than the administrative costs of managing the accounts of low-income workers, although the commissions paid by high-income workers are proportionally higher than those paid by low-income workers. In that sense, it cannot be said that the commission structure is fair, because some workers are paying more than others are for the same type of service.
The rigidity in the commission structure prevents the AFPs from adapting the quality of their service to the ability to pay for that service of each segment of the population and also explains why the AFPs have an incentive to capture the accounts of high-income workers and attempt to do so by offering them better customer service. AFPs will continue to spend money until the marginal cost of trying to capture new accounts is equal to the marginal revenue derived from those accounts. In addition, the AFPs generally do not charge entry fees, even though the law allows them to do that, which means that consumers do not pay a penalty by changing from one AFP to another.

8. How does the government certify the companies that offer individual accounts? How does the government keep politics out of the decision on what companies to certify and what investments they may use?

There is free entry and exit into the industry, even for foreign companies, provided that certain capital requirements, which are specified in advance, are met. The minimum capital required to create an AFP is 5,000 Unidades de Fomento (UF), a Chilean indexed unit of account. If an AFP has 5,000 affiliates, then the minimum increases to 10,000 UF; if it has 7,500 affiliates, then it increases to 15,000 UF; and when an AFP reaches 10,000 affiliates, the minimum capital requirement increases to 20,000 UF. By specifying clear and simple rules in advance, the whole process of creation of management companies is completely depoliticized. The government agency that regulates the industry sets, within the framework of the law, general investment rules in conjunction with the Central Bank of Chile. Both the Central Bank and the regulatory agency are highly technical and independent agencies.

9. Could you explain in more detail how the government’s rate of return guarantee works? For example, doesn’t the government require that investment returns exceeding certain amounts be set aside for buffering returns in case they fall below certain prescribed amounts in the future? Doesn’t the government guarantee funds that go bankrupt? How many funds have gone bankrupt and at what cost to the government?

Each year each AFP must guarantee that the real return of the AFP is not lower than the lesser of (1) the average real return of all AFPs in the last 36 months minus 2 or 4 percentage points, depending on the type of fund, and (2) 50 percent of the average real return of all AFPs in the last 36 months. If the returns are higher than 2 or 4 percentage points above the average return of all AFPs over the last 36 months, or higher than 50 percent of the average return of all AFPs over the preceding 36 months, the “excess returns” are placed in a profitability fluctuation reserve, from which funds are drawn in the event that the returns fall below the minimum return required. For instance, if the industry’s average return for the preceding 36 months is 10 percent and an AFP has a return of 17 percent, then the “excess returns” are 2 percentage points (10 percent plus 50 percent of the average return, which is 5 percent, equals 15 percent, which is the threshold in this case). If, on the other hand, the industry’s average return is 2 percent and an AFP has a return of 4.5 percent, then the “excess returns” are 0.5 percentage points (2 percents plus two percentage points equals 4 percent, which is the threshold in this case, since it is higher than 2 percent plus 50 percentage of the average, 1 percent, which would be equal to 3 percent. Should an AFP not have enough funds in the profitability reserve, funds are drawn from a cash reserve, which is equivalent to 1 percent of total assets under management. If that reserve does not have enough funds, then the government makes up the difference and the AFP is liquidated. To date, no AFP has gone bankrupt, although three have been liquidated for not meeting the minimum capital requirements, so the cost to Chilean taxpayers has been zero. It is also worth noting that the system establishes two different legal entities for the management company and the fund it administers, which is the property of workers. So, it is possible that a management company go bankrupt (that is, its net worth is negative) without it affecting the fund.

10. Could you describe the pay out requirements for personal accounts?

The new private system provides workers with three different types of retirement benefits:
a) Old-Age Pensions. Male workers must reach the age of 65 and female workers the age of 60 to qualify for this pension. However, it is not necessary for men and women who reach these respective ages to retire, nor do they get penalized if they choose to remain in the labor force. No other requirements are necessary.
b) Early-Retirement Pensions. To qualify for this option, a worker must have enough capital accumulated in his account to purchase an annuity that is (1) equal to at least 50 percent of his average salary during the last 10 years of his working life; and (2) at least 110 percent of the minimum pension guaranteed by the state.
c) Disability and Survivor’s Benefits. To qualify for a full disability pension, a worker must have lost at least two thirds of his working ability; to qualify for a partial disability pension a worker must have lost between 50 percent and two thirds of his working ability. Survivor benefits are awarded to a worker’s dependents after the death of said worker. If he did not have any dependent individuals, whatever funds remain in his pension savings account belong to the beneficiaries of his estate.

Types of Pensions. There are three retirement options:

a) Lifetime Annuity. Workers may use the money accumulated in their accounts to purchase a lifetime annuity from an insurance company. This annuity provides a constant income in real terms.

b) Programmed Withdrawals. A second option is to leave the money in the account and make programmed withdrawals, the amount of which depends on the worker’s life expectancy and those of his dependents. If a worker choosing this option dies before the funds in his account are depleted, the remaining balance belongs to the beneficiaries of his estate, since workers now have property rights over their contributions.

c) Temporary Programmed Withdrawals with a Deferred Lifetime Annuity. This pension option is basically a combination of the first two. A worker who chooses this option contracts with an insurance company a lifetime annuity scheduled to begin at a future date. Between the start of retirement and the day when the worker starts receiving the annuity payments, the worker makes programmed withdrawals from his account.
In all three cases a worker may withdraw in a lump-sum (and use for any purpose) those funds accumulated in his account over and above the money necessary to obtain a pension equal to at least 120 percent of the minimum pension and to 70 percent of his average salary over the last 10 years of his working life.

11. If a worker takes programmed withdrawals, but outlives his account balance, what happens? Is there a safety net to insure he still has a source of income?

If a worker outlives the balance in his account, then the government provides the minimum pension, as defined by the Chilean Congress, if that worker has contributed to his account for a minimum of 20 years. If a worker does not have at least 20 years of contributions, he may apply for a welfare-type pension that is lower than the minimum pension. So, yes, there is a safety net under the Chilean private pension system, as there was one under the old government-run system. However, since the new system is far more efficient than the old one, the cost to the Chilean taxpayer is considerably lower.

12. Chile has been criticized in the past for having high rates of transfers between funds. What actions has the government taken to help reduce transfer rates?

Because of investment regulations and rules on fees and commissions, product differentiation is low. Thus companies compete by offering gifts or other incentives for workers to switch to their companies. Switchovers increased dramatically from 1988, the year when the requirement to request in person the change from one AFP to another was eliminated, until 1997, when the government reintroduced some restrictions to make it more difficult for workers to transfer from one AFP to another. The number of transfers in 1998-2000 decreased to less than 700,000, less than 500,000, and slightly more than 250,000, respectively, from an all-time high of almost 1.6 million in 1997. Transfers have stabilized around 250,000 annually since 2000.

13. Are workers aware of the options they have? Do they know through government or industry efforts how the system works? Has there been an educational campaign about the features of the system? Is it not true that this is a system that handicaps low-income workers because they are less likely to not be familiar with investment strategies?

There are three points that I would like to make. First, in Chile there was a roughly six-month period between the day on which the reform was approved (4 November 1980) and the day on which the new system started (1 May 1981). In that time, the architect of the reform, Dr. José Piñera, who was then the Chilean Minister of Labor and Social Security, would appear once a week on national television for three minutes each time to explain different features of the system. Second, the Pension Fund Administration companies also perform an educational campaign, explaining the main features of the system in flyers that are available at the branch offices of those companies. During a trip to Chile, I walked into a branch office of a Pension Fund Administration company in downtown Santiago and I asked the saleswoman some basic questions about the Chilean system. I found her to be very polite, helpful, and knowledgeable of the system. Third, the Pension Fund Administration companies are supervised by a highly technical and very transparent government agency that imposes stiff penalties to those companies that commit fraud or provide misleading information to their clients. Furthermore, that regulatory agency provides very clear and concise information about the private pension system.

14. Are there any restrictions on how the funds can be used? Can workers use the funds accumulated in their retirement savings accounts for purposes other than retirement?

In Chile, workers are only allowed to use the savings accumulated in their pension savings accounts for retirement purposes. If a worker has enough funds accumulated in his account to obtain an annuity that is equivalent to at least 120 percent of the minimum pension, as defined by the Chilean congress, and to 70 percent of his average salary over the last 10 years of his working life, that worker may withdraw in a lump sum those excess savings and use them for any purpose.
Other countries, such as Mexico, for instance, allow workers who have been unemployed for at least 45 days to withdraw the lesser of 10 percent of the cumulative balance in their account or the equivalent of 75 times their daily taxable base salary if they have contributed to the account for at least 250 weeks and have made no withdrawals in the previous 5 years. Workers with 150 weeks of contributions may withdraw from their account the equivalent of their monthly salary if they are getting married. Although it would probably be best that the savings be used for retirement purposes only—especially in the presence of a government guarantee of some kind, which creates a moral hazard—workers should be the ones deciding what to do with their money.

15. How was the transition financed?

The true net economic costs of moving from an unfunded pay-as-you-go system to a fully funded system are zero. That is to say, the total funded and unfunded debt of a country does not change by moving from an unfunded system to a funded one. There is, however, a cash flow problem when moving toward a fully funded retirement system. In the case of Chile, transition costs can be broken down into three different parts. First, there is the cost of paying for the retirement benefits of those workers who were already retired when the reform was implemented and of those workers who chose to remain in the old system. That makes up by far the largest share of the transition costs at present. These costs, of course, will decline as time goes by. Second, there is the cost of paying for the recognition bonds given to those workers who moved from the old system to the new in acknowledgement of the contributions they had already made to the old system. Since these bonds will be redeemed when the recipients retire, this cost to the government will gradually increase as transition workers retire (but will eventually disappear). It is worth stressing that these are new expenditures only if we assume that the government would renege on its past promises. The third cost to the government is that of providing a safety net to the system, a cost that is not new in the sense that the government also provided a safety net under the old pay-as-you-go system. Because the new private system is much more efficient than the old government-run program and because, as stated above, to qualify for the minimum pension under the new system, a worker must have at least 20 years of contributions, this cost has so far been very close to zero. The size of this expenditure will, of course, depend on the success of the private system.
To finance the transition, Chile used five methods. First, it issued new government bonds to acknowledge part of the unfunded liability of the old pay-as-you-go system. Second, it sold state-owned enterprises. Third, a fraction of the old payroll tax was maintained as a temporary transition tax. That tax had a sunset clause and is zero now. Fourth, it cut government expenditures. And, fifth, pension privatization and other market reforms have contributed to the extraordinary growth of the Chilean economy in the last 13 years, which in turn has increased government revenues, especially those coming from the value added tax.
In sum, the transition to the new system has not been an added burden on Chile because the country was already committed to paying retirement benefits. On the contrary, the transition--the fiscal requirements of which have varied between 1.4 and 4.4 percent of GDP per year--has actually reduced the economic and fiscal burden of maintaining an unsustainable system.

--------------------------------------------------------------------------------

Notes:

[1] 2.3/(10+2.3) = 0.1869, or 18.69 percent.

[2] Commissions are also overstated in the case of workers who receive gifts or outright lump sums from sales agents as an enticement to transfer from one AFP to another.

[3] See Raúl Bustos Castillo, “Reforma a los Sistemas de Pensiones: Peligros de los Programas Opcionales en América Latina.” In Sergio Baeza and Francisco Margozzini, eds., Quince Años Después: Una Mirada al Sistema Privado de Pensiones (Santiago, Chile: Centro de Estudios Públicos, 1995), pp. 230-1.

[4] See Congressional Budget Office, Social Security Privatization: Experiences Abroad, sec. 2, p. 7 (January 1999).

[5] Ibid., sec. 3, p. 11.

[6] See Salvador Valdés, “Las Comisiones de las AFPs ¿Caras o Baratas?” Estudios Públicos, Vol. 73 (Verano 1999): 255-91.

[7] Allowing banks and other financial institutions to enter the AFP industry might present potential conflicts of interest. In principle, so long as those institutions compete under the same rules as other market participants, they should be allowed to administer the pension savings accounts of Chilean workers. It is likely that in a market environment banks would have to develop effective separations between the banking department and the administration of pension accounts to attract and protect workers’ investments. Furthermore, the banks may invest in instruments of a higher quality to allay any fears that the public might have about the safety of the investments.

[8] The issue of the commission structure has generated a vast literature in Chile. See, for instance, Salvador Valdés, “Comisiones de AFPs: Más libertad y menos regulaciones.” Economía y Sociedad (January/March 1997), pp. 24-26; Salvador Valdés, “Libertad de Precios para las AFP: Aún Insuficiente.” Estudios Públicos 68 (Spring 1997), pp. 127-47; José de Gregorio, “Propuesta de Flexibilización de las Comisiones de las AFP: Un Avance para Corregir las Ineficiencias.” Estudios Públicos 68 (Spring 1997), pp. 97-110; and Alvaro Donoso, “Los Riesgos para la Economía Chilena del Proyecto que Modifica la Estructura de las Comisiones de las AFP.” Estudios Públicos 68 (Spring 1997), pp. 111-126.

[9] The unfairness does not come from the fact that some workers are paying more than others for the same type of service. In a free-market economy sellers should be able to price discriminate if they wish to in order to capture the consumer's surplus. The problem here is that the government is mandating this price discrimination.

[10] Critics of privatization often point to the giving of toasters and other consumer goods as incentives to switch from one AFP to another as proof of the excesses of the Chilean system. Retail banks in the United States engage in similar practices on college campuses without any negative effects to the banking system or consumers. Of course, these practices have decreased as the banking industry has been deregulated and banks in the United States have found other ways of competing with each other, such as offering better interest rates or lower fees.

[11] Entry fees are usually given back (or a part thereof) by sales agents as a rebate to their customers as an enticement to switch from one AFP to another. Exit fees are not allowed by law in an effort to promote competition.

[12] There is now a bill before the Chilean congress that would increase the percentage from 110 percent of the minimum pension to 150 percent.

[13] This option is ideal for workers who are about to retire at a time when the value of their accounts is down.

[14] See José Piñera (1991) El Cascabel al Gato. Santiago: Editorial Zig-Zag.

[15] The official website of the Superintendencia de AFPs, as the regulatory body is known, can be found at www.safp.cl.

[16] See L. Jacobo Rodríguez “In Praise and Criticism of Mexico’s Pension Reform.” Cato Institute Policy Analysis no. 340, April 14, 1999.

[17] See Milton Friedman, “Social Security Chimeras.” The New York Times (January 9, 1999). See, also, Milton and Rose Friedman, Free to Choose, (New York: Harcourt Brace Jovanovich, 1990), p. 124.

[18] The value of these recognition bonds was computed by taking 80 percent of the worker’s average salary in the 12 months leading to mid-1979 (indexed for inflation), multiplied by the number of years the worker had contributed to the system (up to a maximum of 35 years), and multiplied by an annuity factor of 10.35 for men 11.36 for women.

[19] This cost is projected to reach a peak of 1.06 of GDP in 2005.

[20] Of course, since the system is only 23 years old, the only workers who would be eligible for the government safety net would be those who contributed to the old system as well because those workers are the only ones who could have today more than 20 years of contributions.

[21] That tax was still lower than the payroll tax of the old system. In fact the total contribution to the new system plus the tax was also lower than payroll taxes under the old system.

[22] The financing of a transition from a pay-as-you-go system to a fully funded individual capitalization one is a complex issue that has to take into account the fiscal resources of each country.