Strengthening Social Security: Can We Learn from Other Nations?
Social Security Reform in Chile


My name is L. Jacobo Rodríguez and I am a financialservices analyst at the Cato Institute. I would like to thankChairman Craig and Ranking Member Breaux for inviting me to testifyon social security reform in Chile and its lessons for the UnitedStates. In the interest of transparency, let me point out thatneither the Cato Institute nor I receive government money of anykind.

The aging of the world's population is the result of twodemographic trends. First, life expectancies at birth and atretirement have increased substantially as a result oftechnological and medical advances. Second, fertility rates havedecreased drastically, the result in part of economic progress andgreater opportunities for women around the world. Those two trendscombined mean that in the future the rates of growth of thepopulation and the labor force will slow down or even decrease, andthe ratio of the elderly to the working-age population willincrease. While the aging of the population per se is not a badthing, especially because the elderly today can have a much higherquality of life than in the past, it does have important effects onthe fiscal situation of countries.

Although the prospects for the United States are not as severeas those for some industrialized nations of Europe and Japan, U.S.policymakers will nonetheless face daunting challenges as they seekto reform and strengthen Social Security in the context of an agingpopulation. In the absence of any reform, Social Security willstart to pay out in benefits a larger amount than what it collectsin payroll taxes in 2018, according to the Social SecurityAdministration's own actuaries. Trust fund assets and payroll taxesare projected to be sufficient to pay out scheduled benefits onlyuntil 2042. My colleague Jagadeesh Gokhale, who testified beforethis Committee in January 2004, estimates that Social Security'sfiscal imbalance -- that is, the total financial shortfall thatSocial Security faces -- is approximately $7 trillion.1

Fortunately for the United States, there are other countries,both industrialized and developing, that have already addressed thechallenge of structurally reforming their retirement system underconditions that were similar or even more drastic than those theUnited States faces today. In my remarks today I will focus on thepioneering reform of Chile, because I think that it still remainsthe standard against which other private pension systems in LatinAmerica should be and are measured. Indeed if there is a mainlesson to be drawn from the collective experiences of LatinAmerican countries is that not all reforms are created equal. SomeLatin American countries -- notably, Argentina, Uruguay, andColombia -- introduced important flaws in the design of theirprivate pension systems that have limited the success andpopularity of those systems.

In 1981 Chile replaced its bankrupt pay-as-you-go retirementsystem with a fully funded system of individual retirement accountsmanaged by the private sector.2 That revolutionary reform defused the fiscal timebomb that is ticking for countries with pay-as-you-go systems underwhich fewer and fewer workers have to pay for the retirementbenefits of more and more retirees. More important, Chile created aretirement system that, by giving workers clearly defined propertyrights in their pension contributions, offers proper work andinvestment incentives; and acts as an engine of, not an impedimentto, economic growth.

Since the Chilean system was implemented, labor forceparticipation, pension fund assets, and benefits have all grown.Today, more than 95 percent of Chilean workers have joined thesystem; the pension funds have accumulated over $50 billion inassets, a sum that is equivalent to about 67 percent of Chileangross domestic product; and the average real rate of return hasbeen over 10 percent per year.3

If imitation is the sincerest form of flattery, the Chileansystem should be blushing from the accolades it has received. Since1993 10 other Latin American nations have implemented pensionreforms modeled after Chile's.4 In March of 1999 Poland became the first countryin Eastern Europe to implement a partial privatization reform basedon the Chilean model. In short, the Chilean system has clearlybecome the point of reference for countries interested in findingan enduring solution to the problem of paying for the retirementbenefits of aging populations.

Although the basic story is well known, it is worth recappingbriefly. Every month workers deposit 10 percent of the first$22,000 of earned income in their own individual pension savingsaccounts, which are managed by the specialized pension fundadministration company of their choice.5 Those companies invest workers' savings in aportfolio of bonds and stocks, subject to government regulations onthe specific types of instruments and the overall mix of theportfolio. Contrary to a common misconception, fund managers areunder no obligation to buy government securities, arequirement that would not be consistent with the notion of pensionprivatization, and can invest up to 30 percent of the portfoliooverseas, a measure that allows workers to hedge against currencyfluctuations and country risk. At retirement, workers use the fundsaccumulated in their accounts to purchase annuities from insurancecompanies. Alternatively, workers make programmed withdrawals fromtheir accounts; the amount of those withdrawals depends on theworker's life expectancy and those of his dependents. Thegovernment provides a safety net for those workers who, atretirement, do not have enough funds in their accounts to provide aminimum pension. But because the new system is much more efficientthan the old government-run system and because, to qualify for theminimum pension under the new system, a worker must have at least20 years of contributions, the cost to the taxpayer of providing aminimum pension funded from general government revenues has so farbeen negligible. (Of course, that cost is not new; the governmentalso provided a safety net under the old program.)

Through their pension accounts, Chilean workers have becomeowners of the means of production in Chile and, consequently, havegrown much more attached to the free market and to a free society.This has had the effect of reducing class conflicts, which in turnhas promoted political stability and helped to depoliticize theChilean economy. Pensions today do not depend on the government'sability to tax future generations of workers, nor are they a sourceof election-time demagoguery. To the contrary, pensions depend on aworker's own efforts and thereby afford workers satisfaction anddignity.

Critics of the Chilean system, however, often point to highadministrative costs, lack of portfolio choice, and the high numberof transfers from one fund to another as evidence that the systemis inherently flawed and inappropriate for other countries,including the United States and European countries. Some of thosecriticisms are misinformed. For example, administrative costs areabout 1 percent of assets under management, a figure similar tomanagement costs in the U.S. mutual fund industry. To the extentthe criticisms are valid, they result from a single problem:excessive government regulation.

In Chile pension fund managers compete with each other forworkers' savings by offering lower prices, products of a higherquality, better service or a combination of the three. The pricesor commissions workers pay the managers are heavily regulated bythe government. For example, commissions must be a certainpercentage of contributions regardless of a worker's income. As aresult, fund managers are prevented from adjusting the quality oftheir service to the ability (or willingness) of each segment ofthe population to pay for that service. That rigidity also explainswhy the fund managers have an incentive to capture the accounts ofhigh-income workers, since the profit margins on those accounts aremuch higher than on the accounts of low-income workers.

The product that the managers provide -- that is, return oninvestment -- is subject to a government-mandated minimum returnguarantee (a fund's return cannot be more than 2 or 4 percentagepoints, depending on the type of fund, or 50 percent below theindustry's average real return in the last 36 months).6 That regulation forces the funds to makevery similar investments and, consequently, have very similarportfolios and returns.

Thus, the easiest way for a pension fund company todifferentiate itself from the competition is by offering bettercustomer service, which explains why marketing costs and salesrepresentatives are such an integral part of the fund managers'overall strategy and why workers often switch from one company toanother.

Government restrictions on fees and returns have probablycreated distortions in the optimal mix of price, quality, andservice each fund manager would offer his customers under a moreliberalized regime. As a result of those restrictions, fundmanagers emphasize the one variable over which they have the mostdiscretionary power: quality of the service. (Before the airlineindustry was deregulated in the United States, airlines competed onservice, rather than on price. That service might be thought of asthe equivalent of "wasteful administration costs" in the absence ofprice competition. Similarly, banks in the United States competedon service before deregulation of the banking industry allowed themto engage in other forms of competition, such as offering betterinterest rates or lower fees.)

Although, in the eyes of the Chilean reformers, restrictionsmade sense at the beginning of the system in a country with littleexperience in the private management of long-term savings, it isclear that such regulations have become outdated and may negativelyaffect the future performance of the system. Thus, in addressingthe challenges of the system as it reaches adulthood, Chileanauthorities should act with the same boldness and vision theyexhibited 24 years ago when they drafted the pension reformlaw.

Fortunately, they have taken some important steps, but there areother equally important steps that are yet to be taken. The mostimportant structural reform of the last 3 or 4 years is theintroduction of multiple investment funds. Up until 2000, thepension fund management companies could only manage one fund. Thatyear, the regulatory framework was changed to allow the AFPs tooffer a second fund, invested only in fixed income instruments.That reform proved to be insufficient, as very few workers decidedto switch their savings from the diversified fund to thefixed-income one. Indeed, consumer demand for the fixed-income fundwas negligible. What was needed was to let pension fund managementcompanies manage more than one variable-income fund.7 Chilean authorities finally adoptedthis reform in early 2002 when they instituted a rule that mandatedAFPs to offer 5 different funds that range from very low risk tohigh risk. One advantage of having several funds administered bythe same company is that that could reduce administrative costs ifworkers were allowed to invest in more than one fund within thesame company. This adjustment also allows workers to make prudentchanges to the risk profile of their portfolios as they get older.For instance, they could invest all the mandatory savings in alow-risk fund and any voluntary savings in a riskier fund. Or theycould invest in higher risk funds in their early working years andthen transfer their savings to a more conservative fund as theyapproached retirement. Table 1 shows the maximum percentages ofequity investment allowed in each fund:

Table 1
Maximum Percentage
Mandatory Minimum
Fund A 80% 40%
Fund B 60% 25%
Fund C 40% 15%
Fund D 20% 5%
Fund E Not Allowed Not Allowed

The introduction of a family of funds is an important step andthere are indications that consumers are behaving as one wouldexpect -- that is, by diversifying their investments across themenu of funds. Other steps that have been taken in the recent pastinclude:

  • The lengthening of the investment period over which the minimumreturn guarantee is computed to 36 months from 12 months and thewidening of the band from 2 to 4 percentage points for some type offunds;
  • The further liberalization of the investment rules, so thatworkers with different tolerances for risk can choose funds thatare optimal for them; and
  • The expansion of consumer choice with the signature of abilateral accord with Peru that allows workers from those twocountries to choose the pension system with which they want to beaffiliated.8

Other specific steps that Chilean regulators should take toensure the continuing success of the private pension systeminclude:

  • Liberalize the commission structure to allow fund managers tooffer discounts and different combinations of price and quality ofservice, which would introduce greater price competition andpossibly reduce administrative costs to the benefit of allworkers.
  • Let other financial institutions, such as banks or regularmutual funds, enter the industry. If financial institutions wereallowed to establish one-stop financial supermarkets, whereconsumers could obtain all their financial services if they sochose, the duplication of commercial and operational infrastructurecould be eliminated and administrative costs could be reduced.
  • Give workers the option of personally managing their accounts.Thanks to the emergence of the World Wide Web as an investmenttool, individuals could gain greater control over their retirementsavings if they decided to administer their accountsthemselves.
  • Reduce the moral hazard created by the government safety net bylinking the minimum pension to the number of years (or months)workers contribute.
  • Adjust contribution rates in such a way that workers have tocontribute only that percentage of their income that will allowthem to purchase an annuity equal to the minimum pension. In otherwords, if a high-income worker can obtain an annuity equal to theminimum pension by contributing only 1 percent of his income, heshould be able to do so and decide for himself how to allocate therest of his income between present and future consumption.

Those adjustments would be consistent with the spirit of thereform, which has been to adapt the regulatory structure as thesystem has matured and as the fund managers have gained experience.All the ingredients for the system's success -- individual choice,clearly defined property rights in contributions, and privateadministration of accounts -- have been present since 1981. Someshortcomings remain, to be sure, but the Chilean model stillprovides an excellent example to those countries -- industrializedand developing alike -- that are thinking about reforming theirretirement systems. Unlike a pay-as-you-go system, a fully fundedindividual capitalization system such as Chile's can anticipatefewer problems as it matures.

Let me conclude by commending this Committee for its willingnessto learn from the experiences of other countries and how thoseexperiences may be applied to the United States. I believe there'smuch to learn from the experiences of Latin American countries,both from their successes as well as from their mistakes, and Ithank you very much for the opportunity you have given metoday.

Frequently Asked Questions About Chile's Private PensionSystem

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

As of March 2002, the government had supplemented 33,029pensions, including 11,759 old-age pensions, out of over 400,000pensions, in its role as the financial guarantor of last resort inthe new private system. Because the new system has tougherrequirements to qualify for the minimum pension and is far moreefficient than the old one, the cost to the Chilean taxpayer ofproviding a general safety net is lower than under the old system.Indeed the cost to the government of supplementing these pensionshas been about $33 million. Projections about the percentage ofpensions that will receive a government subsidy range from about 10percent to close to 50 percent, but if returns continue to be above4 percent in real terms and workers contribute to their accountsregularly, the government contribution will continue to beminimal.

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

The often-cited figure of 18-20 percent representsadministrative costs as a percentage of current contributions,which is not how administrative costs are usually measured. Thisfigure is usually obtained by dividing the commission fee, which ison average equivalent to 2.3 percent of taxable wages, by the totalcontribution (10 percent plus the commission).9 This calculation fails to take into accountthat the 2.3 percent includes the life and disability insurancepremiums (about 0.7 percent of taxable wages on average) thatworkers pay, which are deducted from the variable commission, andthus overstates administrative costs as a percentage of totalcontributions.10 Also, if,for instance, the mandatory contribution were lowered to 5 percentof total wages instead of 10 percent, then administrative costsmeasured as a percentage of the total contribution would increasefrom 18.69 percent to 31.51 percent (2.3/(2.3 + 5)), even if thosecosts measured in absolute terms or as a percentage of assets undermanagement remained the same.

When administrative costs are compared to the old government-runsystem, the criticism is not accurate. Chilean economistRaúl Bustos Castillo has estimated the costs of the newsystem to be 42 percent lower than the average costs of the oldsystem.11 However,comparing the administrative costs of the old system with those ofthe new one is inappropriate, because the underlying assumptionwhen making that comparison is that the quality of the product (orthe product itself) being provided is similar under both systems,which is certainly not the case in Chile.

Furthermore, the Congressional Budget Office reported in 1999that, "In Chile, the country with the longest experience withprivate retirement accounts, [administrative costs] can beequivalently expressed as 1 percent of assets, which is similar tocosts of mutual funds in the United States."12 The CBO report goes on to say that, "It isdifficult to convert a charge on contributions to a charge onassets (typical for a U.S. mutual fund). The calculation depends onthe rate of return and the length of the investment horizon andtherefore does not yield a single figure."13 Chilean economist Salvador Valdés hasestimated the average annual cost of the AFP system to beequivalent to 0.84 percent of total assets under management overthe life cycle of the worker, which is lower than the average costof the mutual fund industry in Chile but higher than other savingsalternatives.14

To the extent that such administrative costs are stillconsidered too high, that is the result of government regulationson the commissions the AFPs can charge and on the investments thesecompanies can make. The existence of a "return band" preventsinvestment product differentiation among the different AFPs. As aresult, the way an individual AFP tries to differentiate itselffrom the competition is by offering better service to itscustomers. One way to provide better service would be to offer adiscount on the commission fee to workers who fit a certain profile-- e.g., workers who have maintained their account for an extendedperiod of time or who contribute a certain amount of money to theiraccounts; however, government regulations do not allow that. Thoseregulations state that the AFPs may only charge a commission basedon the worker's taxable income and expressed as a percentage ofthat income.

Another reason administrative costs are not as low as they couldbe is that AFPs have a monopoly in the administration of pensionsavings accounts. Mutual funds, banks, insurance companies, andindividuals themselves are not allowed to manage those accounts.The existence of this monopoly (which is part of the fragmentationof 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 sochoose.15 Such supermarketswould substantially reduce administrative costs by eliminating theduplication of commercial and operational infrastructure.

The average rate of return net of administrative costs for theaverage retirement savings account has ranged from 7.18 percent to7.50 percent, depending on the type of account, from 1981 to April2001, according to the Chilean government agency that regulates theindustry.

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

That claim is partially accurate. It is true that women andlow-wage workers are likely to accumulate less than the averageworker. Women because they tend to earn less than men, have moreirregular professional lives, and may stop contributing to theiraccounts at age 60 (that age is set at 65 for men). (Women alsotend to live longer, a factor that also contributes to making theaverage pension for women lower than the average pension for men,all other things being equal.) All those characteristics are commonto women everywhere and not just Chilean women and should not beconsidered features of the Chilean system. Since the new systemgives every worker property rights in his or her contributions,every worker with 20 years of contributions will receive at leastthe minimum pension. That was not the case in the old pay-as-you-gogovernment system, a system that especially penalized women (andother workers) with irregular professional lives.

Low-wage workers in general accumulate less than average workersbecause they are low-wage workers. Low-wage workers also tend tostart working at an earlier age than other workers, whichconceivably can make up for the smaller amount contributed perperiod, and to have a shorter life expectancy, which conceivablycan allow workers to make larger withdrawals per period of timethan other workers with a longer life expectancy.

Therefore, it is not correct to say that women and low-wageworkers will disproportionally end up receiving the minimumpension. The reform was undertaken under the assumption that if aworker contributes to his account 10 percent of his salary for 35years, and the real rate of return on his investment is 4 percenton average, he will have enough funds accumulated in his accountupon retirement to fund a pension that is equivalent to 70 percentof the average salary over the last 10 years of his workinglife.

I think that focusing on whether income disparity increasesunder a private system or not is mistaken. What matters is thatpoor workers (as well as rich ones) have property rights in theircontributions and can invest their savings in productiveinvestments, so that they live their old age with comfortablemeans, even if other workers are much wealthier. The incomedisparity between Bill Gates and I, for instance, matters nothingto me. What matters to me is that Bill Gates has developed toolsthat allow me to become a more productive worker and, consequently,earn a higher salary, which in turn allows me to live morecomfortably now and, hopefully, in my old age.

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

AFPs are not allowed to offer discounts for permanence, formaking voluntary contributions, for groups, or for maintaining aspecific balance in an account. For instance, if workers were ableto negotiate group discounts, then their bargaining power wouldsignificantly increase. That would allow them to negotiate lowercommissions, which would benefit low-wage workers the most. Fundswould continue to seek out low-wage workers so long as the marginalcost of administering the account of a low-wage worker (or a groupof low-wage workers) does not exceed the marginal revenue derivedfrom administering those accounts. If the administration companieswere allowed to adjust their service to the ability and desire ofworkers to pay for those services, low-wage workers would havenothing to lose if the commission structure were liberalized. Thoseconcerned that the services provided to low-wage workers would dropto unacceptable low levels need not be, as the government alreadymandates a minimum of services that AFPs have to provide to theirclients.

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

If a worker dies before retirement, the balance in his accountbelongs to the beneficiaries of his estate, as workers now haveproperty rights in their contributions. If a worker dies afterretirement and if he chooses the programmed withdrawaloption, then the balance in his account belongs to thebeneficiaries of his estate. If he chooses to purchase an annuityfrom an insurance company, the balance in his account uponretirement is used to purchase the annuity and the account isclosed, so money is left to the beneficiaries of his account.

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

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

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

When the system began, AFPs were allowed to charge fixed andvariable commissions on assets under management, fixed and variablecommissions on contributions, or any combination thereof. AFPs werenot allowed to offer discounts for permanence, group discounts,discounts for making voluntary contributions, or for maintaining aspecific balance in the account. In 1987, the commission structurewas changed by eliminating all commissions on assets undermanagement.16 This changehad the effect of providing a cross subsidy to (1) workers who donot contribute to their accounts regularly, because the fundmanager is still providing a service (administering the account ofthose workers) for which he is not receiving compensation; and (2)to low-income workers, because the administrative costs of managingthe account of wealthier workers are not proportionally higher thanthe administrative costs of managing the accounts of low-incomeworkers, although the commissions paid by high-income workers areproportionally higher than those paid by low-income workers. Inthat sense, it cannot be said that the commission structure isfair, because some workers are paying more than others are for thesame type of service.17

The rigidity in the commission structure prevents the AFPs fromadapting the quality of their service to the ability to pay forthat service of each segment of the population and also explainswhy the AFPs have an incentive to capture the accounts ofhigh-income workers and attempt to do so by offering them bettercustomer service. .18 AFPswill continue to spend money until the marginal cost of trying tocapture new accounts is equal to the marginal revenue derived fromthose accounts. In addition, the AFPs generally do not charge entryfees, even though the law allows them to do that, which means thatconsumers do not pay a penalty by changing from one AFP to another..19

8. How does the government certify the companies thatoffer individual accounts? How does the government keep politicsout of the decision on what companies to certify and whatinvestments they may use?

There is free entry and exit into the industry, even for foreigncompanies, provided that certain capital requirements, which arespecified in advance, are met. The minimum capital required tocreate an AFP is 5,000 Unidades de Fomento (UF), a Chilean indexedunit of account. If an AFP has 5,000 affiliates, then the minimumincreases to 10,000 UF; if it has 7,500 affiliates, then itincreases to 15,000 UF; and when an AFP reaches 10,000 affiliates,the minimum capital requirement increases to 20,000 UF. Byspecifying clear and simple rules in advance, the whole process ofcreation of management companies is completely depoliticized. Thegovernment agency that regulates the industry sets, within theframework of the law, general investment rules in conjunction withthe Central Bank of Chile. Both the Central Bank and the regulatoryagency are highly technical and independent agencies.

9. Could you explain in more detail how the government'srate of return guarantee works? For example, doesn't the governmentrequire that investment returns exceeding certain amounts be setaside for buffering returns in case they fall below certainprescribed amounts in the future? Doesn't the government guaranteefunds that go bankrupt? How many funds have gone bankrupt and atwhat cost to the government?

Each year each AFP must guarantee that the real return of theAFP is not lower than the lesser of (1) the average real return ofall AFPs in the last 36 months minus 2 or 4 percentage points,depending on the type of fund, and (2) 50 percent of the averagereal return of all AFPs in the last 36 months. If the returns arehigher than 2 or 4 percentage points above the average return ofall AFPs over the last 36 months, or higher than 50 percent of theaverage 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 belowthe minimum return required. For instance, if the industry'saverage return for the preceding 36 months is 10 percent and an AFPhas a return of 17 percent, then the "excess returns" are 2percentage points (10 percent plus 50 percent of the averagereturn, which is 5 percent, equals 15 percent, which is thethreshold in this case). If, on the other hand, the industry'saverage return is 2 percent and an AFP has a return of 4.5 percent,then the "excess returns" are 0.5 percentage points (2 percentsplus two percentage points equals 4 percent, which is the thresholdin this case, since it is higher than 2 percent plus 50 percentageof 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 1percent of total assets under management. If that reserve does nothave enough funds, then the government makes up the difference andthe AFP is liquidated. To date, no AFP has gone bankrupt, althoughthree have been liquidated for not meeting the minimum capitalrequirements, so the cost to Chilean taxpayers has been zero. It isalso worth noting that the system establishes two different legalentities for the management company and the fund it administers,which is the property of workers. So, it is possible that amanagement company go bankrupt (that is, its net worth is negative)without it affecting the fund.

10. Could you describe the pay out requirements forpersonal accounts?

The new private system provides workers with three differenttypes of retirement benefits:

  1. Old-Age Pensions. Male workers must reach the age of65 and female workers the age of 60 to qualify for this pension.However, it is not necessary for men and women who reach theserespective ages to retire, nor do they get penalized if they chooseto remain in the labor force. No other requirements arenecessary.
  2. Early-Retirement Pensions. To qualify for this option,a worker must have enough capital accumulated in his account topurchase an annuity that is (1) equal to at least 50 percent of hisaverage salary during the last 10 years of his working life; and(2) at least 110 percent of the minimum pension guaranteed by thestate.20
  3. Disability and Survivor's Benefits. To qualify for afull disability pension, a worker must have lost at least twothirds of his working ability; to qualify for a partial disabilitypension a worker must have lost between 50 percent and two thirdsof his working ability. Survivor benefits are awarded to a worker'sdependents after the death of said worker. If he did not have anydependent individuals, whatever funds remain in his pension savingsaccount belong to the beneficiaries of his estate.

Types of Pensions. There are three retirementoptions:

  1. Lifetime Annuity. Workers may use the moneyaccumulated in their accounts to purchase a lifetime annuity froman insurance company. This annuity provides a constant income inreal terms.
  2. Programmed Withdrawals. A second option is to leavethe money in the account and make programmed withdrawals, theamount of which depends on the worker's life expectancy and thoseof his dependents. If a worker choosing this option dies before thefunds in his account are depleted, the remaining balance belongs tothe beneficiaries of his estate, since workers now have propertyrights over their contributions.
  3. Temporary Programmed Withdrawals with a Deferred LifetimeAnnuity. This pension option is basically a combination of thefirst two. A worker who chooses this option contracts with aninsurance company a lifetime annuity scheduled to begin at a futuredate. Between the start of retirement and the day when the workerstarts receiving the annuity payments, the worker makes programmedwithdrawals from his account.21

In all three cases a worker may withdraw in a lump-sum (and usefor any purpose) those funds accumulated in his account over andabove the money necessary to obtain a pension equal to at least 120percent of the minimum pension and to 70 percent of his averagesalary over the last 10 years of his working life.

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

If a worker outlives the balance in his account, then thegovernment provides the minimum pension, as defined by the ChileanCongress, if that worker has contributed to his account for aminimum of 20 years. If a worker does not have at least 20 years ofcontributions, he may apply for a welfare-type pension that islower than the minimum pension. So, yes, there is a safety netunder the Chilean private pension system, as there was one underthe old government-run system. However, since the new system is farmore efficient than the old one, the cost to the Chilean taxpayeris considerably lower.

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

Because of investment regulations and rules on fees andcommissions, product differentiation is low. Thus companies competeby offering gifts or other incentives for workers to switch totheir companies. Switchovers increased dramatically from 1988, theyear when the requirement to request in person the change from oneAFP to another was eliminated, until 1997, when the governmentreintroduced some restrictions to make it more difficult forworkers to transfer from one AFP to another. The number oftransfers in 1998-2000 decreased to less than 700,000, less than500,000, and slightly more than 250,000, respectively, from anall-time high of almost 1.6 million in 1997. Transfers havestabilized around 250,000 annually since 2000.

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

There are three points that I would like to make. First, inChile there was a roughly six-month period between the day on whichthe reform was approved (4 November 1980) and the day on which thenew system started (1 May 1981). In that time, the architect of thereform, Dr. José Piñera, who was then the ChileanMinister of Labor and Social Security, would appear once a week onnational television for three minutes each time to explaindifferent features of the system.22 Second, the Pension Fund Administrationcompanies also perform an educational campaign, explaining the mainfeatures of the system in flyers that are available at the branchoffices of those companies. During a trip to Chile, I walked into abranch office of a Pension Fund Administration company in downtownSantiago and I asked the saleswoman some basic questions about theChilean system. I found her to be very polite, helpful, andknowledgeable of the system. Third, the Pension Fund Administrationcompanies are supervised by a highly technical and very transparentgovernment agency that imposes stiff penalties to those companiesthat commit fraud or provide misleading information to theirclients. Furthermore, that regulatory agency provides very clearand concise information about the private pensionsystem.23

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

In Chile, workers are only allowed to use the savingsaccumulated in their pension savings accounts for retirementpurposes. If a worker has enough funds accumulated in his accountto obtain an annuity that is equivalent to at least 120 percent ofthe minimum pension, as defined by the Chilean congress, and to 70percent of his average salary over the last 10 years of his workinglife, that worker may withdraw in a lump sum those excess savingsand use them for any purpose.

Other countries, such as Mexico, for instance, allow workers whohave been unemployed for at least 45 days to withdraw the lesser of10 percent of the cumulative balance in their account or theequivalent of 75 times their daily taxable base salary if they havecontributed to the account for at least 250 weeks and have made nowithdrawals in the previous 5 years. Workers with 150 weeks ofcontributions may withdraw from their account the equivalent oftheir monthly salary if they are getting married. Although it wouldprobably be best that the savings be used for retirement purposesonly -- especially in the presence of a government guarantee ofsome kind, which creates a moral hazard --workers should be theones deciding what to do with their money.24

15. How was the transition financed?

The true net economic costs of moving from an unfundedpay-as-you-go system to a fully funded system are zero. That is tosay, the total funded and unfunded debt of a country does notchange by moving from an unfunded system to a fundedone.25 There is, however, acash flow problem when moving toward a fully funded retirementsystem. In the case of Chile, transition costs can be broken downinto three different parts. First, there is the cost of paying forthe retirement benefits of those workers who were already retiredwhen the reform was implemented and of those workers who chose toremain in the old system. That makes up by far the largest share ofthe transition costs at present. These costs, of course, willdecline as time goes by. Second, there is the cost of paying forthe recognition bonds given to those workers who moved from the oldsystem to the new in acknowledgement of the contributions they hadalready made to the old system.26 Since these bonds will be redeemed when therecipients retire, this cost to the government will graduallyincrease as transition workers retire (but will eventuallydisappear).27 It is worthstressing that these are new expenditures only if weassume that the government would renege on its past promises. Thethird cost to the government is that of providing a safety net tothe system, a cost that is not new in the sense that the governmentalso provided a safety net under the old pay-as-you-go system.Because the new private system is much more efficient than the oldgovernment-run program and because, as stated above, to qualify forthe minimum pension under the new system, a worker must have atleast 20 years of contributions, this cost has so far been veryclose to zero.28 The sizeof this expenditure will, of course, depend on the success of theprivate system.

To finance the transition, Chile used five methods. First, itissued new government bonds to acknowledge part of the unfundedliability of the old pay-as-you-go system. Second, it soldstate-owned enterprises. Third, a fraction of the old payroll taxwas maintained as a temporary transition tax. That tax had a sunsetclause and is zero now.29Fourth, it cut government expenditures. And, fifth, pensionprivatization and other market reforms have contributed to theextraordinary growth of the Chilean economy in the last 13 years,which in turn has increased government revenues, especially thosecoming from the value added tax.30

In sum, the transition to the new system has not been an addedburden on Chile because the country was already committed to payingretirement benefits. On the contrary, the transition -- the fiscalrequirements of which have varied between 1.4 and 4.4 percent ofGDP per year -- has actually reduced the economic and fiscal burdenof maintaining an unsustainable system.


1. See Jagadeesh Gokhale, "The Futureof Retirement in the United States." Statement of Jagadeesh Gokhalebefore the Special Committee on Aging of the United States Senate,January 22, 2004.

2. A lengthier treatment of theChilean reform can be found in L. Jacobo Rodríguez "Chile'sPrivate Pension System at 18: Its Current State and FutureChallenges," Cato Institute Social Security Paper no.17, July 30, 1999. An updated summary of the Chilean system isSuperintendencia de Administradoras de Fondos de Pensiones, ElSistema Chileno de Pensiones (5th ed.), The fourthedition of that publication is available in English at

3. For more statistical informationon the Chilean system, see the official website of theSuperintendencia de AFPs, the Chilean government regulator of theprivate pension system, at

4. These countries (and the year ofimplementation of the new system) are: Peru (1993), Argentina(1994), Colombia (1994), Uruguay (1996), Bolivia (1997), Mexico(1997), El Salvador (1998), Dominican Republic (2003), Nicaragua(2004), and Ecuador (2004). A good summary of all these systems canbe found in Asociación Internacional de Organismos deSupervisión de Fondos de Pensiones, LaCapitalización Individual en los Sistemas Previsionales deAmérica Latina (December 2003),

5. At present there are 6 AFPs. Thesystem began with 12 AFPs, reached a high of 23, and has graduallyconsolidated to the present number. Most of the consolidation hasoccurred through mergers. There have been, however, three AFPs thatwere closed down by the government for not meeting the minimumcapital requirements.

6. Until recently, the period forcomputing the minimum return was 12 months, which increased the"herd effect" of having a minimum return guarantee.

7. I first made this proposal inRodríguez (1999). The reform adopted by the Chileanregulators closely resembles the proposal that I made, although Icannot determine whether it was influenced by my research ornot.

8. Again, this measure resembles aproposal that I made in Rodríguez (1999). In that paper Irecommended that, "As Latin American markets become moreintegrated, [pension regulators should] expand consumer sovereigntyby allowing workers to choose among the systems in Latin Americathat have been privatized, which would put an immediate (and veryeffective) check on excessive regulations."

9. 2.3/(10+2.3) = 0.1869, or 18.69percent.

10. Commissions are also overstatedin the case of workers who receive gifts or outright lump sums fromsales agents as an enticement to transfer from one AFP toanother.

11. See Raúl BustosCastillo, "Reforma a los Sistemas de Pensiones: Peligros de losProgramas Opcionales en América Latina." In Sergio Baeza andFrancisco Margozzini, eds., Quince Años Después:Una Mirada al Sistema Privado de Pensiones (Santiago, Chile:Centro de Estudios Públicos, 1995), pp. 230-1.

12. See Congressional BudgetOffice, Social Security Privatization: Experiences Abroad,sec. 2, p. 7 (January 1999).

13. Ibid., sec. 3, p. 11.

14. See Salvador Valdés,"Las Comisiones de las AFPs ¿Caras o Baratas?" EstudiosPúblicos, Vol. 73 (Verano 1999): 255-91.

15. Allowing banks and otherfinancial institutions to enter the AFP industry might presentpotential conflicts of interest. In principle, so long as thoseinstitutions compete under the same rules as other marketparticipants, they should be allowed to administer the pensionsavings accounts of Chilean workers. It is likely that in a marketenvironment banks would have to develop effective separationsbetween the banking department and the administration of pensionaccounts to attract and protect workers' investments. Furthermore,the banks may invest in instruments of a higher quality to allayany fears that the public might have about the safety of theinvestments.

16. The issue of the commissionstructure has generated a vast literature in Chile. See, forinstance, Salvador Valdés, "Comisiones de AFPs: Máslibertad y menos regulaciones." Economía y Sociedad(January/March 1997), pp. 24-26; Salvador Valdés, "Libertadde Precios para las AFP: Aún Insuficiente." EstudiosPúblicos 68 (Spring 1997), pp. 127-47; José deGregorio, "Propuesta de Flexibilización de las Comisiones delas AFP: Un Avance para Corregir las Ineficiencias." EstudiosPúblicos 68 (Spring 1997), pp. 97-110; and AlvaroDonoso, "Los Riesgos para la Economía Chilena del Proyectoque Modifica la Estructura de las Comisiones de las AFP."Estudios Públicos 68 (Spring 1997), pp.111-126.

17. The unfairness does not comefrom the fact that some workers are paying more than others for thesame type of service. In a free-market economy sellers should beable to price discriminate if they wish to in order to capture theconsumer's surplus. The problem here is that the government ismandating this price discrimination.

18. Critics of privatization oftenpoint to the giving of toasters and other consumer goods asincentives to switch from one AFP to another as proof of theexcesses of the Chilean system. Retail banks in the United Statesengage in similar practices on college campuses without anynegative effects to the banking system or consumers. Of course,these practices have decreased as the banking industry has beenderegulated and banks in the United States have found other ways ofcompeting with each other, such as offering better interest ratesor lower fees.

19. Entry fees are usually givenback (or a part thereof) by sales agents as a rebate to theircustomers as an enticement to switch from one AFP to another. Exitfees are not allowed by law in an effort to promotecompetition.

20. There is now a bill before theChilean congress that would increase the percentage from 110percent of the minimum pension to 150 percent.

21. This option is ideal forworkers who are about to retire at a time when the value of theiraccounts is down.

22. See José Piñera(1991) El Cascabel al Gato. Santiago: EditorialZig-Zag.

23. The official website of theSuperintendencia de AFPs, as the regulatory body is known, can befound at

24. See L. Jacobo Rodríguez"In Praise and Criticism of Mexico's Pension Reform." CatoInstitute Policy Analysis no. 340, April 14, 1999.

25. See Milton Friedman, "SocialSecurity Chimeras." The New York Times (January 9, 1999).See, also, Milton and Rose Friedman, Free to Choose, (NewYork: Harcourt Brace Jovanovich, 1990), p. 124.

26. The value of these recognitionbonds was computed by taking 80 percent of the worker's averagesalary in the 12 months leading to mid-1979 (indexed forinflation), multiplied by the number of years the worker hadcontributed to the system (up to a maximum of 35 years), andmultiplied by an annuity factor of 10.35 for men 11.36 forwomen.

27. This cost is projected to reacha peak of 1.06 of GDP in 2005.

28. Of course, since the system isonly 23 years old, the only workers who would be eligible for thegovernment safety net would be those who contributed to the oldsystem as well because those workers are the only ones who couldhave today more than 20 years of contributions.

29. That tax was still lower thanthe payroll tax of the old system. In fact the total contributionto the new system plus the tax was also lower than payroll taxesunder the old system.

30. The financing of a transitionfrom a pay-as-you-go system to a fully funded individualcapitalization one is a complex issue that has to take into accountthe fiscal resources of each country.