Social Security: The Chilean Approach to Retirement
The Chilean Approach
May 17, 2007
Analyst in Social Security
Domestic Social Policy Division
Social Security: The Chilean Approach to Retirement
Over the past few years, there has been intense debate about Social Security
reform in the United States. A number of options, ranging from changing the benefit
formula to adding individual accounts, has been discussed. The policy debate takes
place against the backdrop of an aging population, rising longevity, and relatively
low fertility rates, which pose long-range financial challenges to the Social Security
system. According to the 2007 Social Security Trustees Report’s intermediate
assumptions, the Social Security trust funds are projected to experience cash-flow
deficits in 2017 and to become exhausted in 2041.
As policymakers consider how to address Social Security’s financing
challenges, efforts of Social Security reform across the world have gained attention.
One of the most oft-cited international cases of reform is Chile. Chile initiated
sweeping retirement reforms in 1981 that replaced a state-run, pay-as-you-go defined
benefit retirement system with a private, mandatory system of individual retirement
accounts where benefits are dependent on the account balance. As a pioneer of
individual retirement accounts, Chile has become a case study of pension reform
around the world. Although Chile’s experience is not directly comparable to the
situation in the United States because of large differences between the countries,
knowledge of the case may be useful for American policymakers.
This CRS report focuses on the Chilean individual retirement accounts system.
It begins with a description of the U.S. Social Security policy debate, along with a
brief comparison of Chile and the United States. Next, the report explains what
Chile’s individual retirement accounts system is and how it works. The pension
reform bill sent to the Chilean Congress for debate in 2007 is also discussed. The
report does not address other components of Chile’s social security system, such as
maternity, work injury, and unemployment.
The final section provides an assessment of Chile’s now 26-year-old individual
retirement accounts system. Pension reforms have contributed to the rapid growth
in the Chilean economy over the past two decades and returns on pension fund
investments have been greater than expected. Administrative costs, however, have
been high and participation rates have been modest at best. There is concern that the
system does not cover the entire labor force and provides inadequate benefits to low
This report will not be updated.
In troduction ......................................................1
The U.S. Social Security System..................................2
The Social Security Debate..................................3
Chile and the United States in Comparative Perspective................5
Political and Economic Differences............................5
Overview of Chilean Individual Retirement Accounts.....................7
Compensation for Previous Contributions to the Old System............9
Voluntary Savings Mechanisms..................................10
Management of Individual Accounts..............................11
Pension Fund Investments..................................12
Assessing the Chilean Approach to Retirement..........................20
Efficiency and Costs to Participants..............................22
Appendix. Chilean Pension Funds ...................................28
List of Figures
Figure 1. Aggregate Allocation of Chilean Pension Funds, 1983-2003.......13
Figure 2. Value of Chilean Pension Funds as % of GDP, 1981-2005........14
Figure 3. Annual Real Rates of Return in Chilean Pension Funds, 1981-2003..14
List of Tables
Table 1. Basic Demographics, the United States and Chile..................5
Table 2. Features of the Chilean Individual Retirement Accounts System.....8
Table 3. Administrative Costs, Chilean Individual Retirement Accounts,
Table 4. Features of Multi-Funds, Equity Allowance, and Default Age.......16
Table 5. Active Members and Total Contributors, Chilean Individual
Retirement Accounts, 1981-2005................................25
Table A1. Allocation of Chilean Pension Funds, 1981-2003...............28
The Chilean Approach to Retirement
As policymakers contemplate ways to address Social Security’s long-term
financial challenges, pension reforms across the world have gained new attention.1
This report focuses on Chile, one of the most oft-cited cases of pension reform
internationally. In 1981, Chile initiated sweeping reforms that replaced a state-run,
pay-as-you-go defined benefit retirement system with a private, defined contribution
individual retirement accounts system.2 As a pioneer of individual accounts, Chile
has become a case study for many countries seeking to reform their retirement
systems. Although the Chilean experience is not directly comparable to the United
States situation because of large differences between the countries, the case may offer
some valuable insights for policymakers who are interested in individual retirement
The report3 begins with a description of the U.S. Social Security policy debate
and a brief comparison of Chile and the United States. It discusses the backdrop
against which the Chilean pension reforms were implemented. Next, the report
explains what Chile’s individual retirement accounts system is and how it works. A
pension reform bill, which is scheduled for consideration by the Chilean Congress
1 U.S. Government Accountability Office, Social Security Reform: Other Countries’
Experiences Provide Lessons for the United States, United States Government
Accountability Office Report to Congressional Requesters, 2005, GAO-06-126, at
[http://www.gao.gov/new.items/d06126.pdf]; (Hereafter cited as GAO, 2005). U.S.
Congressional Budget Office, Social Security Privatization: Experiences Abroad, January
1999, at [http://www.cbo.gov/ftpdocs/10xx/doc1065/ssabroad.pdf]; (Hereafter cited as
CBO, 1999). U.S. Social Security Administration, Social Security Programs Throughout the
World: The Americas, 2005 (Washington: GPO, 2006), [http://www.socialsecurity.gov/
policy/docs/progdesc/ssptw/2004-2005/americas/ssptw05americas.pdf]; (Hereafter cited as
2 Retirement programs are legally classified as either defined benefit plans or defined
contribution plans. In defined benefit or “DB” plans, the retirement benefit is normally tied
to an employee’s earnings history, years of service, and age of retirement, among other
factors. A defined contribution or “DC” plan operates much like a savings account in which
the retirement benefit is tied to an employee’s history of contributions, as well as
administrative costs, investment returns and payout options, among other factors. See CRS
Report RL30122, Pension Sponsorship and Participation: Summary of Recent Trends, by
3 This report was written by Christopher R. Tamborini while on detail to CRS. Questions
should be addressed to Kathleen Romig at (7-3742).
in 2007 and expected to be passed before May 2008, is detailed. The final section
provides an assessment of the individual retirement accounts system’s performance
relative to some of its initial goals.
The U.S. Social Security System
The U.S. Old-Age, Survivors, and Disability Insurance (OASDI) programs
collectively make up the system referred to as Social Security. The program is a
social insurance system, whereby premiums are paid by workers to obtain coverage
and benefits are intended to replace part of the earnings lost to the worker and the
family when the worker retires, becomes disabled, or dies. Virtually all working men
and women in the United States are covered by Social Security — about 96% of the
labor force pay payroll taxes.4
Social Security is financed primarily on a pay-as-you-go (PAYGO hereafter)
basis, in which today’s workers pay for the benefits of today’s retirees. The primary
revenue source is a payroll tax paid by current workers and their employers.5 When
revenues exceed outgo, as they do now, surpluses are invested in bonds and credited
to the Social Security trust funds managed by the Treasury Department.
Financial Challenges. The Social Security system faces a long-term
financing problem. Under the intermediate assumptions of the Social Security
trustees, the system is projected to begin running cash flow deficits in the year 2017,
at which point the system must begin redeeming any bonds (including interest)
accumulated in previous years. Financial projections of the trustees also show that
the trust funds will be exhausted in 2041, at which point 75% of scheduled annual
benefits would be payable with income revenue.6
A primary factor underlying Social Security’s long-term financial problem is the
program’s PAYGO financing structure in combination with the demographics of an
aging society, specifically the looming retirement of the large baby boom cohort
(persons born between 1946 and 1964), along with rising longevity and a low birth
4 Being covered by Social Security means that a worker is employed in a job or is self-
employed and contributes a portion of his or her earnings to Social Security. Workers not
covered by Social Security are either covered by a similar eligible contributory system
offered by their employers outside of Social Security (such as some local and state
employees), do not have high enough earnings for mandatory participation, or have another
special exemption. See CRS Report 94-27, Social Security Brief Facts and Statistics, by
5 Covered workers and their employers each pay 6.2% of wages to Social Security up to a
taxable maximum ($97,500 in 2007). Self-employed workers pay 12.4% of wages to Social
Security up to the taxable maximum.
6 U.S. Social Security Administration, 2007 Annual Report of the Board of Trustees of the
Federal Old-Age and Survivors Insurance and Disability Insurance Trust Funds, April 23,
rate.7 Demographics are important because PAYGO systems, such as Social
Security, are sensitive to the ratio of workers to beneficiaries, which is declining in
the United States.8 Presently, there are about three covered workers for every
beneficiary, and according to the Social Security trustees, this ratio will eventually
fall to less than two to one. Between 2010 and 2030, the number of individuals age
65 or older is projected to grow by 76%, while the number of workers supporting the
system is projected to grow by 6%.9
The Social Security Debate. The longer it takes to address Social
Security’s financing challenge, the greater the changes will need to be. There are,
however, fundamentally diverging views on reform. One approach would maintain
the current program structure and make relatively modest changes to restore the
system’s long-term solvency, such as increasing the retirement age, reducing the cost-
of-living adjustments, or raising the amount of earnings subject to the payroll tax.10
A second approach would change the program’s underlying structure and create a
partially or fully funded system based on personal savings and investments in11
individual retirement accounts.
The Debate and the Chilean Case. As policymakers consider how to
address Social Security’s challenges in the United States, other countries’
experiences with retirement reforms have gained attention.12 Much of this attention
has been directed toward countries that have adopted some sort of individual
accounts program as part of their retirement system.13 The case of Chile, the first
7 A PAYGO system of financing would be sustainable if payroll taxes collected on behalf
of current workers exceed benefits paid to current beneficiaries.
8 This can be referred to as the age dependency ratio or support ratio. See CRS Report
RL32981, Age Dependency Ratios and Social Security Solvency and CRS Report RL32701,
The Changing Demographic Profile of the United States, both by Laura B. Shrestha.
9 Another factor influencing the long-term financial health of Social Security is the projected
increase in the real value of Social Security benefits due in large part to wage-indexing rules
put into effect in 1979. Wages are projected to exceed price growth in the future. This
results in greater Social Security benefits for future retirees since initial benefits are indexed
to wages. See Congressional Budget Office, The Future Growth of Social Security: It’s Not
Just Society’s Aging, An Issue Summary from CBO, no. 9, July 2003, at
[ h t t p : / / www.cbo.gov] .
10 See CRS Report RL33840, Options to Address Social Security Solvency and Their Impact
on Beneficiaries: Results from the Dynasim Microsimulation Model, by Laura Haltzel,
Dawn Nuschler, Kathleen Romig, Gary Sidor, Scott Szymendera, Mikki Waid, and Debra
11 See CRS Report RL33544, Social Security Reform: Current Issues and Legislation, by
12 See CBO, 1999; GAO, 2005; SSA, 2006.
13 For a useful review of pension reforms in Latin America and Eastern Europe see Barbara
Kritzer, “Individual Accounts in Other Countries,” Social Security Bulletin, vol. 66, no. 1
(2005), pp. 31-37; and Stephen Kay and Barbara Kritzer, “Social Security Reform in Latin
America: Policy Challenges,”Journal of Aging & Social Policy, vol. 14, no. 1 (2002), pp.
country to introduce a fully funded individual retirement accounts system in 1981,
is often cited by both proponents and opponents of individual accounts.
Policymakers who advocate introducing individual accounts in the United States
have tended to point to Chile as a successful case. Adopting a Chilean-type system,
some argue, would put the Social Security system on a path of sustainable solvency
beyond the traditional 75-year projection period, since a fully funded system is not
sensitive to changes in the number of workers per beneficiary. These proponents
believe government-run, PAYGO systems are unsustainable in aging societies. They
argue that a fully funded system, such as Chile’s, would reduce future demands on
the government for financing the growing costs associated with an increasingly
In addition, advocates of individual accounts maintain that a Chilean-style
model would change the way Americans save for retirement, providing workers with
a sense of ownership over their retirement savings. Individual retirement accounts,
proponents maintain, would strengthen the link between contributions and benefits
and thus would provide more incentives for workers to save for retirement.
Moreover, American workers may earn significant returns on their contributions
under an individual retirement accounts system, because their capital would be
invested in stocks and bonds.
By contrast, policymakers who advocate a more traditional approach to reform
have tended to highlight the risks of a Chilean-type system. They argue that
individual accounts would expose workers to the risk of investment market volatility.
Implementing individual retirement accounts, others maintain, would also erode the
social insurance nature of the U.S. system, which is designed to pool risk and protect
workers and their families against loss of earnings due to retirement, disability, or
death. A Chilean-type model, it is pointed out, would eliminate the system’s
progressive benefit formula, which replaces a higher share of earnings for lower
earners than for higher earners. Under an individual accounts system, benefits would
become strictly a function of workers’ earnings and the returns achieved by their
plan’s investments. As a result, some workers could be worse off for reasons
including poor investment decisions or downturns in financial markets. Some also
argue that individual accounts do not necessarily provide the annuity features of the
current U.S. system. Furthermore, critics may cite problems facing Chile’s
individual retirement accounts system, such as coverage gaps and high administrative
costs, as evidence that implementing individual accounts in the United States would
Another concern raised by opponents relates to the transition costs from
switching from a PAYGO to a fully funded system — that is, the cost of paying
accrued obligations while funding individual accounts.14 Adopting a Chilean-type
funded system would require today’s younger workers to save for their own
retirements while continuing to pay taxes to cover current retirees’ benefits. Others
14 CRS Report RS22010, Social Security: “Transition Costs,” by Laura Haltzel.
contend that, depending on the funding mechanism, an individual retirement accounts
system could worsen the U.S. program’s financial outlook and exacerbate current
budget deficits. For example, diverting revenues from payroll taxes into individual
accounts, such as carve-out accounts, would only reduce the system’s long-term
financing problem to the extent that the benefit offset (from lower benefits) is greater
than the diverted revenues.15
Chile and the United States in Comparative Perspective
Before examining how Chile’s individual retirement accounts system works, it
is important to note some of the major demographic, political, and economic
differences between Chile and the United States.
Basic Demographics. Chile has a relatively small population compared to
the United States (roughly 16 million compared to 300 million respectively), and a
smaller share of persons 65 or older (see Table 1). Both Chile and the United States
are projected to face population aging, although Chile is expected to remain younger
than the United States. In 2025, 18% of Americans and 14% of Chileans are
projected to be age 65 years or older. Also note that the share of workers (aged 20-
Table 1. Basic Demographics, the United States and Chile
PopulationExpectancy Median AgePercent AgeWorking
(in thousands)(at birth)(both sexes)65 or olderAge (20-64)
Country 2005202520052025200520252005 20252005 2025
United295,734349,666 77.7 80.5 36.3 38.512.4 18.260.0 55.4
Chile15,981 18,52176.6 79.9 30.1 36.8 8.014.258.0 60.2
Source: U.S. Census Bureau, International Data Base, August 2006 version.
Political and Economic Differences. The political situation in Chile when
the individual retirement accounts system was first adopted is very different than the
United States today. Individual accounts were implemented in Chile under the
military dictatorship of Augusto Pinochet (1973-1990). According to many analysts,
this allowed the Chilean government to implement far-reaching pension reforms
quickly and without a great amount of political consensus building.16 A democratic
15 CRS Report RL31498, Social Security Reform: Economic Issues, by Jane Gravelle and
Marc Labonte; CRS Report RL33544, Social Security Reform: Current Issues and
Legislation, by Dawn Nuschler.
16 Studies show that democratic and authoritarian regimes have been equally as likely to
implement private pension schemes, albeit in varying scopes. See Carmelo Mesa-Lago and
Katharina Mueller, “The Politics of Pension Reform in Latin America,” Journal of Latin
government was reinstated in Chile in 1990 and has continued the individual
retirement accounts system.
The fiscal conditions of Chile prior to the 1981 privatization reforms and the
United States today are very different. A large budget surplus (e.g., 5% of Gross
Domestic Product (GDP) in 1980) helped the Chilean government cover the
transition costs to the new system. By contrast, in 2006 the U.S. federal budget
recorded a deficit of $248 billion (representing 1.9% of GDP).17 Moreover, Chile
used individual accounts to help develop its financial markets. Such markets are
already well established in the United States.
Another difference between the countries is the condition of the PAYGO
retirement system. Chile’s PAYGO system was insolvent at the time that their
retirement system was overhauled. In 1980, the program had a deficit equivalent to
2.7% of gross domestic product, and general revenues financed roughly 28% of
outgoing payments.18 The U.S. Social Security system, by contrast, currently has a
surplus (1.4% of GDP in 2006)19 and is projected to remain solvent until 2041.
Moreover, in Chile, factors such as government mismanagement, high rates of
contribution evasion, and public distrust, among other reasons, helped set the stage
for replacing their PAYGO system.20 By contrast, the U.S. Social Security program
is well-managed, covers almost the entire labor force, and enjoys broad public
The two countries differ in the size and scope of their economies. Chile is a
relatively small developing county recognized for its robust economic performance
in the Latin American region since the 1980s. With a per capita GDP of $5,747 in
whereas the United States, with a per capita GDP of $37,574, is classified by the
World Bank as an OECD developed high income country.22 Chile has a large
number of self-employed workers (around 27% of the labor force), many of whom
American Studies, vol. 34, no. 3 (August 2002), pp. 687-715; and Raúl L. Madrid, Retiring
the State: The Politics of Pension Privatization in Latin America and Beyond (Stanford:
Stanford University Press, 2003).
17 CRS Report RS22550, The Federal Budget: Sources of the Movement from Surplus to
Deficit, by Marc Labonte; and CRS Report RL31235, The Economics of the Federal Budget
Deficit, by Brian W. Cashell.
18 CBO, 1999, pp. 11-12.
19 CRS Report RL31498.
20 Rodrigo Acuña and Augusto Iglesias, “Chile’s Pension Reform after 20 Years,” Social
Protection Discussion Paper No. 0129 (2001), The World Bank.
21 CRS Report RL33544.
22 Per capita GDP is GDP divided by midyear population. It is the sum of gross value added
by all resident producers in the economy plus any product taxes and minus any subsidies not
included in the value of the products. For more information see World Bank, World
Development Indicators, 2006, at [http://devdata.worldbank.org/data-query/].
are part of the informal sector — a segment that did not participate widely in the
PAYGO system or in the current individual retirement accounts system. In contrast,
the United States has a very small informal sector and the self-employed must
contribute to Social Security.
Finally, the definition of social security in Chile, as in many countries, includes
sickness and maternity insurance, work injury insurance, unemployment and family
allowances in addition to old-age, survivors and disability insurance. This CRS
report focuses on the retirement component of Chile’s social security system.
Overview of Chilean Individual Retirement Accounts
In 1924, Chile became the first Latin American country to establish a national
social insurance system, with the goal of insuring against elderly poverty, disability
and death. It was a state-run, pay-as-you-go system with defined benefits, primarily
financed by payroll contributions from employees and employers. While the program
grew in size and complexity between 1924 and 1980, its basic structure remained
Several major problems associated with Chile’s PAYGO retirement system set
the stage for pension reforms in 1981. It was highly fragmented, with more than
three dozen different retirement schemes, each with different eligibility requirements
and contribution rates.23 In 1973, for example, the combined employer/employee
payroll contributions ranged from 19.5% to 26.0% of wages, depending on
occupational type.24 The system was also vulnerable to political manipulation and
was widely seen as inequitable. White-collar workers tended to fare better than those
on the lower end of the economic spectrum. The multitude of plans and contribution
rates resulted in another difficulty: high administrative inefficiencies and expenses.
In May 1981, Chile replaced its state-run, PAYGO system with a private, fully
funded individual retirement accounts system.25 The switch had a number of goals,
including to restore the long-term financial balance of the system; to provide
efficiency gains in the system; to reduce inequities of the old system and cover more
23 For example, government employees qualified for a full pension after 30 years of service,
bank employees 24 years, and legislators 15 years. For more information see R. Myers,
“Chile’s Social Security Reform, After Ten Years,” Benefits Quarterly, vol. 8, Third Quarter
(1992), pp. 41-55.
24 Wage workers contributed 6.5% of wages (7% in arduous occupations); self-employed
contributed 10% of earnings; and salaried employees, 8% of salary. Employers contributed
13% of wages for wage earners (15% to 17% in arduous occupations), and 17.8% of salary
for salaried employees. See U.S. Department of Health, Education, and Welfare (Social
Security Administration), Social Security Programs Throughout the World, 1973, 1973, p.
25 The implementation of individual retirement accounts (Decree Law 3500) was part of a
broader set of free-market reforms.
workers; to give workers “ownership” over their retirement resources; to increase
national savings; and to stimulate the national economy.
Mandatory individual retirement accounts comprise the centerpiece of the new
Chilean retirement system. Individual accounts are supplemented by a minimum
guaranteed pension program, a social assistance pension program, and voluntary
private savings accounts system. The main features of the system are summarized
in Table 2 and in the sections that follow.26
Table 2. Features of the Chilean Individual Retirement
ElementIndividual Retirement Accounts System
Year Started— 1981
Participation— Mandatory for new workers
— Voluntary for self-employed
— Optional for workers under the old system
Compensation from prior— Yes, recognition bonds
Mandatory contribution rate — Employee (10%) — capped at 60 UFs
(% of taxable wages)— Employer (none)
— Additional contribution for administrative fees and
survivors/disability insurance (See text box)
Additional savings— Additional contribution on top of the mandatory 10% of
— Separate savings account
Management of individual— Private pension fund management companies
accounts(Administradoras de Fondos de Pensiones, or AFPs)
Fees — Charged to participant by AFP
Pension fund investments — Five funds varying by risk (beginning in 2002)
Default fund on investments— Depends on age, with older workers defaulted in fixed-
income securities (since 2002).
Legal retirement age— 65 (men), 60 (women)
Payout options— Annuity (joint if married)
— Programmed withdrawals
— Programmed withdrawals with immediate annuity/deferred
Government benefit guarantee— Minimum pension guarantee (with 20 years ofa
contributions to individual account)
Source: Congressional Research Service.
a. An “assistance pension” (PASIS) may be available to indigent elderly with less than 20 years of
contributions. Not all who qualify, however, receive an assistance pension. To control costs,
the government sets a limit on the number of persons who can receive social assistance benefits.
A recent reform bill under consideration in the Chilean Congress attempts to extend a PASIS-
type benefit to all who qualify.
26 For a detailed overview of the system, see the Superintendent of Pension Fund Managers
(SAFP), The Chilean Pension System, 4th edition (Santiago de Chile, 2003), online at
[http://www.safp.cl/sischilpen/english.html]. (SAFP, 2003 hereafter.) See also Robertst
Holzmann and Richard Hinz, Old Age Income Support in the 21 Century (Washington, DC:
The World Bank, 2005).
Workers who entered the Chilean labor force after January 1, 1983 were no
longer covered by the old system. Instead, they were required to pay a proportion of
their earnings into a private pension fund; that is, the individual retirement accounts
system. Participation of self-employed workers in the individual retirement accounts
system was made voluntary. The police and members of the armed forces remain in
their own separate system to date. Those already in the workforce when the reforms
were implemented were permitted to join the new system or remain in the old one,
and persons already receiving a pension continued under the old law.
Compensation for Previous Contributions to the Old System
Workers who switched to the new system received government-financed
“recognition bonds” (bonos de reconocimiento) to compensate them for accrued
benefits under the previous system. The recognition bond is paid out of general
revenues into a worker’s individual account at retirement. Its value takes into
account, among other things, the life expectancy of workers and the number of years
they contributed to the old system.27
Workers must contribute 10% of their monthly earnings to an individual
retirement account, plus an additional amount (variable percentage) for
administrative fees and survivors and disability insurance. There is a monthly
maximum earnings limit on contributions of 60 UFs (unidad de fomento): US$2,04328
as of January 2007. Contributions and interest are tax-deferred until retirement.
Employers are responsible for sending the monthly contribution to workers’ pension
fund management companies (Administradoras de Fondos de Pensiones, AFPs).
Employers are not required to contribute but may do so. There is and additional
contribution for fees and survivors and disability insurance. (See text box below.)
27 CBO, 1999, pg. 12.
28 Chilean pensions are expressed in a currency called Unidad de Fomento (UF), which is
adjusted monthly to prices. As of January 2007, one UF is equal to 18,327.50 Chilean pesos
or about $34.
What about Survivor and Disability Benefits?
Under the current Chilean system, survivors and disability pensions are provided through
the private market and not the central government. The insurance is financed as a
fraction of workers’ additional required monthly contribution which varies by AFP —
averaging around 0.75%-0.76% of the worker’s gross earnings per month. These
resources are used by the pension fund management companies (AFPs) to take out
disability and group life (survivor) insurance from private insurance companies.
Workers who lose earnings capacity due to an injury or illness may receive a disability
benefit. A total disability benefit is provided to members who have lost at least 66% of
earning capacity (a smaller benefit is provided for partial disability). A survivor benefit
is provided to the surviving widow, disabled widower, the mother of children of the
insured born out of wedlock, children younger than age 18 (age 24 if a student, no age
limit if disabled), and in some cases, to the parents of the deceased worker.
Qualifying widows and disabled widowers are required to have married the insured
person at least six months before his or her death or three years before if the marriage
took place when the member was already receiving an old-age or disability pension. As
of the time of writing, men can qualify for a survivors pension only if they are totally or
partially disabled. A reform bill under consideration in Chile would extend survivors29
benefits to widowers.
Source: This information has been adapted from the Superintendent of Pension Fund Managementth30
Companies (SAFP), The Chilean Pension System, 4 edition (Santiago de Chile, 2003).
Voluntary Savings Mechanisms
Chilean workers can supplement their individual retirement accounts in several
ways. They can contribute on top of their required 10% contribution on earnings, up
to a monthly ceiling (60 UF) to an individual account. Since August 1987, workers
may also put money aside, regularly or sporadically, in a separate voluntary savings
account.31 The tax code provides a variety of incentives for voluntary contributions
and accounts, which were extended to the self-employed not participating in the new
29 For information on Chile’s new civil marriage law as it relates to survivor benefits, see
the Chilean Pension Fund Administrators’ Association, Research Series AFP Association,
Number 55, March 2006, at [http://www.afp-ag.cl/ingles/estudios/Estudio55.pdf].
30 For more information on disability benefits in individual accounts systems, see Patrick
Wiese, “Financing Disability Benefits in a System of Individual Accounts: Lessons from
International Experience,” WP2006-4, Center for Retirement Research at Boston College,
31 These separate savings accounts are limited to four withdrawals per year. Voluntary
contributions and accounts may be transferred to workers’ mandatory individual accounts
at retirement. For more details see Chilean Pension Fund Administrators’ Association,
“Characteristics of APV (Voluntary Social Security Saving Scheme)”, available at
[http://www.afp-ag.cl/ingles/pdf/apv_web.pdf]. See also B. Kritzer, “Recent Changes to the
Chilean System of Individual Accounts,” Social Security Bulletin, vol. 64, no. 4
(2001/2002), pp. 66-71.
system and to members of the old system beginning in 2002. As of December of
2005, there were roughly 1.5 million voluntary savings accounts, compared to 7.4
million mandatory individual retirement accounts.32
Management of Individual Accounts
In Chile, mandatory individual retirement accounts are administered by private
pension fund management companies known as Administradoras de Fondos de
Pensiones (AFPs).33 The design emphasizes competition between AFPs in an
attempt to lower administrative costs, promote higher returns on investments, and
encourage better customer service.
Workers may select one AFP to manage their mandatory retirement accounts,
which are invested in a mix of stocks, bonds, and other financial instruments.
Workers may switch from one AFP to another at any time. When the system began,
there were 12 AFPs operating. The number of AFPs peaked to 21 in 1994, and since
then, a number have merged and some have been liquidated. As of January 2007, six
pension fund management companies were in operation.34
AFP Fees. AFPs may levy an array of different fees for managing workers’
mandatory individual retirement accounts.35 These typically include a fixed flat fee36
when contributions are made (varies by AFP), a proportional fee on contributions,
and a fee to open a new account. All fees are levied when contributions are made.
Since 1987, a management fee on a worker’s account has not been permitted. At
retirement, there are up-front fees for the purchase of an annuity or programmed
withdrawals, but no exit fee or fee to transfer pension fund companies may be levied
In addition to administrative fees, workers pay AFPs for survivors and disability
insurance, and each AFP takes out group life and disability insurance from separate
private insurance companies. While typically included in the calculation of total
administrative costs, survivors and disability insurance premiums are distinct from
fees related to administering the individual accounts and provide protection against
disability and death for the worker and his family.
Although total administrative costs in the Chilean system have declined from
their peak in 1984, when they represented 3.6% of taxable earnings, they remain too
32 Superintendent of Pension Fund Management Companies (SAFP), El Sistema Chileno de
Pensiones, 6th edition (Santiago de Chile, 2007), p. 256.
33 A government agency known as the Superintendent of Pension Fund Managers issues
extensive guidelines for private pension fund companies, among other responsibilities.
Their website is available at [http://www.safp.cl/].
34 Superintendent of Pension Fund Companies (SAFP), Preguntas Frecuentes ¿Que AFP’s
Operan en el Mercado Chileno?, January 2007, at [http://www.safp.cl/].
35 The Chilean government places limits on the structure of fees.
36 The amount and the application of a fixed fee varies by AFPs. Fixed fee means an amount
that does not vary by the level of contribution or quantity of capital in the account.
high, according to observers.37 Table 3 breaks down administrative costs (fees plus
survivors and disability insurance) for the system in 2003. In that year, total
administrative charges averaged 2.26% of taxable earnings, representing 22.6% of
workers’ 10% deposit or around 18% of the total deduction on workers’ wages
(12.26%). Note that, premiums paid for survivors and disability insurance typically
make up around 0.75%-0.76% of the worker’s taxable earnings per month.38
Table 3. Administrative Costs, Chilean Individual
Retirement Accounts, 2003
(a) (b)(c)Administrative Costs as % of
Deposit inAdministrative Total(d) Total (e) Deposit in
Countryindividualaccount as %costs (fees +premium) asdeductionas % ofdeduction individual
of wages% of wages wages account (b÷c) (b÷a)
Source: Carmelo Mesa-Lago, “Evaluation of a Quarter Century of Structural Pension Reforms in
Latin America,” in Carolin A. Crabbe, ed., A Quarter Century of Pension Reform in Latin America
and the Caribbean: Lessons Learned and the Next Steps, Inter-American Development Bank, 2005,
Table 2.6, pp. 43-82. (Hereafter cited as Mesa-Lago, 2005.)
Notes: Administrative costs are deducted on top of the 10% deposit in the individual account. Unlike
the 10% mandatory contribution, however, administrative costs represent a variable percentage (over
time and across AFPs). The 2.26% figure referenced in the table (b) includes roughly 0.75% for
survivors and disability insurance premiums.
Pension Fund Investments. Pension fund investments are subject to a
number of rules set by the Chilean government. To reduce risk, investments were
initially limited to bonds of financial institutions, bank deposits, mortgage bonds,
government securities, and a limited amount of corporate bonds. Investments in
domestic or foreign equities were not permitted. As the system has matured,
investment rules and restrictions have been relaxed. Investment in domestic equities
has been allowed since the mid-1980s, and investment in foreign assets (both fixed
and variable return investments) has been allowed since the mid-1990s. The limit
on asset allocation in foreign instruments was raised from 20% to 30% in 2004;
recent proposals seek to raise the limit to 80%.
37 For general information on administrative costs see, Whitehouse, E., “Administrative
Charges for Fund Pensions: Comparison and Assessment of 13Countries” in Private
Pensions Systems: Administrative Costs and Reforms, Private Pensions Series No. 2, (Paris:
Organization of Economic Co-operation and Development [OECD], 2001).
38 Recent figures from the Superintendent of Pension Fund Management Companies (SAFP)
indicate that administrative charges have remained steady since 2003. In January 2007,
administrative fees and premiums for survivors and disability insurance ranged from 2.23%-
Figure 1. Aggregate Allocation of Chilean Pension Funds, 1983-2003
Source: Based on Arenas de Mesa, 2005, Table 3.5, p. 92.
Notes: Totals may not equal the sums of rounded components. Categories less than 1% were dropped
(i.e., disposable assets). Foreign assets include foreign mutual funds, foreign issued bonds, and
equity of foreign corporations publicly traded on the New York Stock Exchange, NASDAQ, the
London Stock Exchange, etc. Mutual Funds include investment funds of business firms plus others
from the external sector (domestic issuer). Equities include stocks of financial institutions (domestic
issuer). Non-financial institutions include the corporate bonds of business firms. Financial
institutions include banking deposits and mortgage securities.
The asset mix of Chilean pension funds has shifted dramatically since the
system’s inception, moving toward greater diversification (see Figure 1). Investment
in domestic equities began in the mid-1980s, peaked in the early 1990s at 32% of
total assets, then declined to 15% of total assets in 2003 as the portfolio moved
toward international diversification. In recent years, investments in foreign
instruments have grown dramatically, rising from 0.6% of total portfolio assets in
1993 to 24% in 2003 (in 2005, this figure was at 30%).39 A detailed summary of the
asset allocation from 1981 to 2003 is provided in Appendix Table A1.
The growth of the funds’ size between 1981 and 2005 in relation to the Chilean
economy has been considerable (Figure 2). By 2005, pension funds represented
almost 60% of Chile’s GDP. The returns on investments have also been sizeable
(Figure 3). From the system’s inception in 1981 to 2003, the pension funds have
returned an average annual real rate of 10.4%; the average real return between 1991
and 2003 was 8.7% (before administrative costs). Negative returns were recorded
39 SAFP, “Evolucion de la Inversion de los Fondos de Pensiones Por Sector Institucional e
Instrumentos Financieros,” 2006, at [http://www.safp.cl/inf_afiliados/index.html].
in two years — 1995 (-2.5%) and 1998 (-1.1%)40 — highlighting how capital markets
vary during boom or bust times.
Figure 2. Value of Chilean Pension Funds as % of GDP, 1981-2005
Source: Data for 1981-2005 are as of December, SAFP, El Sistema Chileno de Pensiones, 6th edition
(Santiago de Chile, 2007), p. 258.
Figure 3. Annual Real Rates of Return in Chilean Pension Funds,
Source: Data reported in Arenas de Mesa, 2005, Table 3.4, p. 91.
40 Arenas de Mesa, 2005, Table 3.4, p. 91.
AFPs must follow a number of other regulations. They have a minimum capital
start up requirement. They have a minimum and maximum rate of return tied to the
average real rate of return of all AFPs over a three-year period. Under the new multi-
funds system established in 2002 (described in more detail below), the minimum and
maximum rates of return are calculated separately for each fund type.41 When an
AFP has “excess returns” based on a defined percentage, they must put the funds into
a reserve fund. Regulations also require that AFPs hold a margin account equal to
1% of the fund’s value. If an AFP does not achieve a minimum return, it makes up
the difference from the aforementioned 1% and excess returns funds. It should also
be mentioned that from the beginning of the program to March 2000, each AFP was
permitted to offer only one fund. Although the intent was to limit investor risk, the
one fund per AFP rule, along with the maximum and minimum rates of return,
resulted in almost identically held portfolios among AFPs.
To provide workers with a greater array of investment choices, legislation
passed in January 2002 established multiple funds, which permitted each AFP to
offer up to five Funds (A, B, C, D, and E), each with differing degrees of risk (see
Table 4).42 Fund A has the highest proportion of its portfolio invested in equities,
therefore the highest risk and potentially the greatest return. Under the multi-funds
system, plan members are also permitted to allocate their contributions between two
funds within the same AFP. Contributors who do not elect a specific fund are
assigned one according to their age, with older workers automatically shifted into
lower risk funds. As of September 2006, the majority of active contributors who
have chosen a fund have selected A- and B-type funds. However, the majority of
pension fund assets are in the C-type fund, which is the default fund for workers aged
41 To avoid excessive divergence from the average yield, the defined minimum rate of return
has been widened for the two funds with the highest share invested in equities within the
multi-funds system. For more information see, SAFP, 2003, pp. 173-211.
42 For more information on multi-funds see Guillermo Larraín Ríos, “Enhancing the Success
of the Chilean Pension System: The Addition of Multiple Funds and Annuities,” in Carolin
A. Crabbe, ed., A Quarter Century of Pension Reform in Latin America and the Caribbean:
Lessons Learned and the Next Steps, Inter-American Development Bank, 2005), pp. 219-
239; and B. Kritzer, “Recent Changes to the Chilean System of Individual Accounts,” Social
Security Bulletin, vol. 64, no. 4 (2001/2002), pp. 66-71.
43 The Chilean Pension Fund Administrators’ Association, “Multi-Funds Results and
Trends,” Bulletin No. 14, 2006, at [http://www.afp-ag.cl/ingles/estudios/multifondos14.pdf].
Table 4. Features of Multi-Funds, Equity Allowance,
and Default Age
Percent of Equities by FundDefault Age Assignment
Minimum Maximum MenWomen
A 40% 80% n/a n/a
B25%60%Up to 35 Up to 35
C15% 40% 36 to 55 36 to 50
D5% 20% 56 or older 51 or older
ENone (fixed-income securities) n/a
Source: SAFP, 2003.
Notes: In March 2000, regulations began permitting AFPs to offer a fixed-income fund (Fund 2) to
workers already receiving a pension or to those within 10 years or less of retirement age. Participants
previously in Fund 2, when there were two funds, who did not subsequently elect a fund after the
multi-funds system was implemented, were assigned to the E-type Fund.
The legal retirement age under Chile’s individual retirement accounts system
is 65 years for men and 60 years for women. An individual need not stop working
to receive his or her pension at legal retirement age. Early retirement is permitted but
only if capital in the individual account would fund an “adequate” annuity, in terms
of a monetary amount and a replacement rate of former earnings.44 The amount of
the pension benefit depends on the amount accumulated in the mandatory individual
account and the type of payout option a worker chooses.
Chilean workers have three options to withdraw income from their retirement
accounts: (1) purchase an annuity, (2) withdraw a predetermined amount each month,
or (3) a combination of the two. A retiree can use the accumulated capital in the
individual account to purchase an annuity from a private life insurance company.
The annuity must be inflation-protected and provide survivors benefits.45 The second
option is a programmed withdrawal, which is actuarially determined using official
life tables, the account balance, and a variable interest rate. If individuals who
choose the monthly programmed withdrawal outlive their resources, they could
qualify for a minimum pension. If the retiree dies before all withdrawals are made,
44 Until 2004, an “adequate” annuity was defined as providing at least 50% of the insured’s
average wage during the past 10 years and at least 110% of the guaranteed minimum
pension. In 2004, this threshold was raised to 55% of the insured’s average wage and at
least 130% of the minimum pension. The percentages will increase gradually until 2010.
45 For analysis on Chile’s annuity market see Roberto Rocha, Marco Morales, and Craig
Thorburn, “An Empirical Analysis of the Annuity Rate in Chile,” World Bank Policy
Research Working Paper 3929, 2006.
the remaining balance forms part of his or her estate. A third option available since
2004 is to combine an annuity and programmed withdrawal, with the advantage that
the retiree is guaranteed an income stream for life, and if the individual dies early, the
remaining account balance can be passed on to dependents.46 A lump-sum
withdrawal is allowed, but only under designated conditions that few workers meet.47
The Chilean government maintains a number of important roles in the individual
retirement accounts system. Regulatory oversight is provided by the governmental
agency known as the Superintendent of Pension Fund Management Companies
(SAFP), which licenses and oversees the pension fund management companies.
The government acts as the guarantor of the AFP system. If an AFP goes
bankrupt, the account holder (who is not retired or a retiree taking programmed
withdrawals) does not lose any money in his or her individual account. The
guarantee is effective only after the AFP uses its reserve funds. Also, the government
insures 75% of a worker’s annuity over the minimum pension level if an insurance
company goes bankrupt. To date the government has yet to need to do so.48 The
government also continues to administer the old PAYGO pension system (current
pensioners under the old system and those who switched from the old system to the
new system) through the National Pension Fund, INP.49
Two government programs form the safety-net under the system of individual
retirement accounts. A minimum pension guarantee (MPG) is available to all
workers who have at least 20 years of contributions but who have not accumulated
enough in their individual account to finance a minimum lifetime pension. The MPG
is a top-up benefit — if the individual account is not sufficient to fund a minimum
pension, the government makes up the difference. If a person who has programmed
withdrawals runs out of money in their individual account, or the amount they receive
is below the minimum benefit, the government also tops this up. The minimum
benefit level is set by the government and equals roughly 75% of Chile’s minimum
wage or 25% of the average wage.50
46 See U.S. Social Security Administration, “International Update,” March 2004, available
47 As of 2006, an account surplus can be withdrawn if the account balance exceeds a
specified threshold — at least 70% of the worker’s average wage over the past ten years and
at least 150% of the minimum pension.
48 Estelle James, Guillermo Martínez, and Augusto Iglesias, “The Payout Stage in Chile:
Who Annuitizes and Why?,” Working Paper no. 14, SAFPs, December 2005, at
49 Created in 1980, the INP (Instituto de Normalización Previsional) unified the diverse
“Cajas de Previsión” (Social Security institutions) under the old system. The INP also
issues the recognition bonds — the credit given for the accrued contributions to the old
pension system from those who switched to the new system.
50 As of 2006, the set minimum monthly level corresponds to about 88,000 Chilean pesos
A second non-contributory program targets the very poor aged. The means-
tested social assistance pension (known as PASIS) is available to the indigent elderly
with fewer than 20 years of contributions. The benefit equals approximately half the
value of the minimum pension. Not all who would qualify for a PASIS pension,
however, receive one. To control costs, the government has set a limit on the number
of persons who can receive the social assistance pension. A large share of qualified
candidates has been placed on a waiting list. A reform bill currently under
consideration by the Chilean Congress aims to extend PASIS-type benefits to all of
the aged population who qualify, as discussed below.51
In recent years, there have been growing calls to modify Chile’s individual
retirement accounts system. Pension reform was a major theme during Chile’s
presidential campaign of 2005 and early 2006. The two main candidates, Sebastian
Piñera of the center right National Renewal Party and Michelle Bachelet of the center
left Socialist Party, agreed on the need to improve the existing individual retirement
In January 2006, Michelle Bachelet won the presidency in a runoff election, and
soon after taking office in March, she created an Advisory Commission on Pension
Reform. In July, the Commission presented the government with 70 proposals for
improving the individual retirement accounts system.53 By December of the same
year, President Bachelet sent a pension reform bill based on the Commission’s
recommendations to the Chilean Congress for debate and approval.
The pension reform bill would strengthen the role of the state in the existing
system and has several areas of focus. A main part of the reform is the addition of
a new pillar (Sistema de Pensiones Solidarias, SPS) to the individual retirement
accounts system. The SPS, which would take effect July 1, 2008, would increase
retirement benefits to low-income persons aged 65 or older who have lived in Chile
for at least 20 years. It would replace the means-tested PASIS pension and the
guaranteed minimum pension such that (a) individuals not eligible for any other
pension would receive a basic solidarity pension of 60,000 pesos a month ($111);
($166) for 69 and under; 96,000 pesos ($181) for 70 to 74; and 102,000 pesos ($194) for 75
or older, source: SAFP.
51 Both the minimum and assistance pensions are financed through general revenues and are
indexed to prices, as well as ad hoc increases by the government. In April 2006, the Chilean
Congress increased the minimum guaranteed pension and means-tested assistance pension
by 10% (U.S. Social Security Administration, “International Update,” May 2006 at
[ ht t p: / / www.soci al secur i t y.gov/ pol i c y/ docs/ pr ogdesc/ i nt l _updat e / ] .
52 Larry Rohter, “Chile’s Candidates Agree to Agree on Pension Woes,” New York Times,
January 10, 2006.
53 Consejo Asesor Presidencial para la Reforma Previsional. Informe Final, July 2006, at
[http://www.consejoreformaprevisional.cl/view/informe.asp]. See also Reuters News,
“Chile Commission Recommends Pension Overhaul,” July 6, 2006 and “Chile Preparing
Pension reform Bill for December,” November 14, 2006.
and (b) individuals whose pension from the individual accounts equals less than
200,000 pesos ($369) a month would receive a monthly top up benefit of up to
Another goal of the proposed reforms is to increase the participation of self-
employed and informal sector workers and younger low-income workers in
individual retirement accounts. To this end, one provision calls for a government
cash subsidy to low-income workers (defined as earnings of less than 1.5 times the
minimum wage) between the ages 18 and 35 for the first 24 months they are formally
employed.55 Another measure would require self-employed workers, roughly 27%
of the Chilean labor force, to enroll with an AFP seven years after the bill’s
implementation. Their participation is currently voluntary.
Gender equity is also addressed in the pension reform bill. The legal retirement
age would remain 60 for women and 65 for men, but starting July 1, 2009, women
retiring at age 65 would receive a bonus credit for each child’s birth — equivalent
to 12 monthly contributions at minimum wage (at the time the child was born), plus
Survivor pensions would be extended to widowers (currently, men qualify only if
they are disabled). In the case of divorce or annulement, capital in an individual
account would be divided evenly between spouses. Social security law in Chile does
not yet address the new marital status of divorced as provided by the civil marriage
law of 2004 (legalizing divorce).56
The proposed legislation also aims to increase competition among AFPs and to
reduce administrative costs. Among other modifications, the bill would
!allow banks to set up an AFP and offer pension funds;
!permit the Superintendent of Pension Fund Managers (SAFPs) to
direct new entrants in the workforce to the management company
with the lowest commissions;
!eliminate the flat fee charged to plan members by most AFPs, which
is regressive; and
!raise the limit on foreign investment from 30% to 80%, three years
after the reform bill’s enactment.
54 This section draws heavily from the U.S. Social Security Administration, “International
Update,” July 200 and January 2007, at [http://www.socialsecurity.gov/policy/
docs/progdesc/ intl_update/]. See also Larry Rohter, “Chile Proposes to Reform Pension
System,” New York Times, December 26, 2006.
55 The subsidy consists of a direct payment to the worker of up to 5% of minimum wage and
a deposit of 5% of minimum wage into the worker’s individual retirement account.
56 In 2005, the SAFP ruled that ex-wives (through divorce or annulment) have no right to a
widow’s pension. The main issue lies in the interpretation of the original pension law,
which do not envision divorce as a marital status. Divorce was legalized in Chile in 2004.
For more information see the Chilean Pension Fund Administrators’ Association, “Women
who are Divorced or whose Marriages have been Annulled are not Entitled to a Widow’s
Pension,” Research Series AFP Association, Number 55, March 2006, at
Other measures seek to improve financial education and boost retirement
savings. The reform bill calls for the creation of a Pension Education Fund as a way
to help improve workers’ knowledge of the individual retirement accounts system.
This measure comes on the heels of a growing body of research showing a lack of
knowledge of the retirement system, especially among workers with low incomes,
with less education, and among women.57 The bill would also create an employer-
sponsored voluntary retirement savings plan, known as Ahorro Previsional
Voluntario Colectivo (APVC), in which employers, as well as other persons, could
contribute to an individual’s voluntary savings account.
Assessing the Chilean Approach to Retirement
This section provides a brief assessment of some of the major impacts of
Chile’s individual retirement accounts system. The section does not analyze
potential effects of the reform bill under debate in the Chilean Congress as of May
2007. Note that a complete analysis of the Chilean individual retirement accounts
system cannot be made until the first cohort of workers entirely under the new system
reaches retirement age — around 2020.
Many economists argue that individual accounts have significantly contributed
to Chile’s rapid economic growth since the mid-1980s. One study estimates that
Chilean pension reforms contributed roughly 0.5 percentage points per year (in a
range from 0.2 to 0.9 percentage points) to economic growth from 1981 to 2001, a
period when Chile’s economy grew at an average annual rate of 4.6%.58 Individual
accounts are also viewed as having helped the Chilean labor market; for example, by
lowering the cost of labor (i.e., reducing total rate of payroll taxes), which
encouraged employment creation.59
The implementation of individual accounts has helped develop Chile’s domestic
capital markets. Chilean capital markets were very underdeveloped at the time
individual accounts were adopted. AFPs are now the largest institutional investors
in Chile’s capital markets, and pension fund assets have grown to represent a huge
57 Alberto Arenas de Mesa et al., “The Chilean Pension Reform Turns 25: Lessons from the
Social Protection Survey,” National Bureau of Economic Research Working Paper 12401,
2006, at [http://www.nber.org/papers/w12401]; see also Jeremy Skog, “Who Knows What
About Their Pensions? Financial Literacy in the Chilean Individual Account System,”
University of Pennsylvania Working Paper, September 2006, available at
[http://bpp.wharton.upenn.edu/waldfogj /900/ papers0607/Skog-K nowledge-9-11-06.pdf].
58 See Vittorio Corbo and Klaus Schmidt-Hebbel, “Macroeconomic Effects of Pension
Reform in Chile,” in International Federation of Pension Funds Administrators, ed., Pension
Reform: Results and Challenges, 2003.
59 Sebastían Edwards, “The Chilean Pension Reform: A Pioneering Program,” in Martin
Feldstein, ed., Privatizing Social Security , 1998, pp. 33-62. (Hereafter cited as Sebastían
portion of the Chilean economy, from around 1% of GDP in 1981 to 63% of GDP
as of June 2006.60
Among the oft-cited benefits of moving from a PAYGO to a fully funded
system is the increase in national savings.61 Empirical studies, however, have not
reached a consensus on whether pension reforms have directly boosted Chile’s
national savings rate.62 Measuring changes in the savings rate is difficult as estimates
depend on assumptions of the savings levels that would have existed if no reform had
A primary rationale for moving to a fully funded system in Chile was that it
would help the retirement system achieve solvency. The transition to an individual
retirement accounts system from a PAYGO system, however, has proved fiscally
expensive for Chile in the short term. During the first five years of the reforms in
Chile, total transition costs — expenditures for current pensioners and for those who
remained in the old system, plus the costs of redeeming recognition bonds — ranged
from 4.2% to 4.7% of GDP per year, according to the Congressional Budget Office
(CBO).63 Transition costs peaked in the late 1980s, and during the 1990s averaged
roughly 4.1% of GDP annually.64 The entire transition period in Chile is not
expected to end until about 2050 — the year that benefits to those who stayed in the
old system are projected to cease completely.65
Another area of concern relates to new fiscal burdens on the government
resulting from the growth of minimum pension guarantees (not including the
proposed expansion of the program). As Chile’s individual retirement accounts
system matures, forecasts project that an increasing number of future retirees will
60 La Asociación Internacional de Organismos de Supervisión de Fondos de Pensiones
(AIOS), “Los Regímenes de Capitalización Individual en América Latina,” Boletín
Estadístico AIOS, no. 15, June 2006, p. 13.
61 CRS Report RL30708, Social Security, Saving, and the Economy, by Brian W. Cashell.
62 See Mauricio Soto, “Chilean Pension Reform: the Good, the Bad, and the In Between,”
Issues in Brief Number 31, Center for Retirement Research, 2005, at
[http://www.bc.edu/centers/crr/issues/ib31.pdf], (Hereafter cited as Soto, 2005);and Peter
R. Orszag and Joseph E. Stiglitz, “Rethinking Pension Reform: Ten Myths About Social
Security Systems,” in Robert Holzmann and Joseph Stiglitz, eds., New Ideas About Old Age
Security (Washington, DC: The World Bank, 2001), pp. 17 - 56.
63 The transition cost range from 1982 to 1986 was from 4.2% to 4.7% of GDP (CBO, 1999,
Table 3, p. 18). As a point of comparison, the total cost of the U.S. Social Security system
in 2007 will account for 4.3% of GDP. Social Security’s cost as a percentage of GDP is
projected to rise to 6.2% of GDP in 2030 (2007 Social Security Trustees’ Report, pp. 11-
64 Milko Matijascic and Stephen J. Kay, “Social security at the crossroads: Toward effective
pension reform in Latin America,” International Social Security Review, vol. 59, no. 1,
65 Arenas de Mesa, 2005, p. 89.
qualify for a minimum pension guarantee as many current workers are not
accumulating enough in their mandatory individual accounts to fund at least a
minimum pension at retirement. The share of individual account holders requiring
a “top up” benefit to provide a minimum pension level is projected to increase to
more than 30 percent as the system matures, costing the government up to 1% of
GDP . 66
It is worthwhile to point out several strategies that helped the Chilean
government finance the transition costs, especially during the early and most
expensive years of implementation. These included increasing taxes on
consumption,67 selling a large array of state-owned enterprises, borrowing from the
public, and tightening spending. These methods helped Chile build a substantial
budget surplus prior to the reform (5.5% of GDP in 1980). According to the CBO,
“in general, the Chilean privatization has been quite successful in managing the
transition from a pay-as-you-go to a fully funded privatized retirement system.”68
Efficiency and Costs to Participants
Chile’s switch to individual retirement accounts was also intended to improve
the efficiencies of the old system. However, among the current system’s greatest
problems are high administrative costs. The extent of administrative costs has
become a focal point of controversy in Chile because management fees reduce a
workers’ net investment over their working lives and hence their final pension from
the individual account. If administrative costs are reduced, more of workers’
contributions can be invested (in their IAs) and therefore raise the accumulated
capital in their individual accounts.
Studies have shown that the cumulative impact of administrative charges on a
workers’ final capital accumulation and pension in Chile may be substantial.69 Over
time it has been estimated that administrative fees have consumed a quarter (25%)
of the accumulations of an average Chilean worker who began contributing in 1982
and retired in 2002.70 Administrative charges also reduce workers’ rate of return.
Estimates from the Superintendent of Pension Fund Management Companies (SAFP)
have indicated that when commission fees are considered, average annual returns on
Chilean pension funds between July 1981 and August 2001 decline from 10.83% to
66 Soto, 2005.
67 A value-added tax (VAT) was established in 1975.
68 CBO, 1999, p. 24.
69 Measuring the precise impacts of administrative costs on workers’ accumulated capital
in their individual accounts is difficult. Some fees are proportional to contributions and
some are fixed. Fees also may vary by year, AFP, and investment returns. See, Whitehouse,
E., “Administrative Charges for Funded Pensions: Comparison and Assessment of 13
Countries” in Private Pensions Systems: Administrative Costs and Reforms, Private
Pensions Series No. 2, (Paris: Organization for Economic Co-operation and Development
70 Soto, 2005, Figure 4, pp. 4-5.
7.33% for low earners and 7.59% for high earners.71 The disparity between high and
low earners stems from a flat-fee on monthly contributions that exists in most AFPs.
Unlike proportional fees, fixed charges tend to be regressive.
There are a number of reasons for high administrative costs in the Chilean
system. Given pension fund managers’ incentive to entice plan members to their
AFP, the system, especially in the 1990s, experienced high marketing costs and a
dramatic growth in sales personnel. Increased advertising and sales representatives
also led individual participants to change AFPs excessively, thereby increasing total
operating costs.72 A lack of competition and market transparency may also contribute
to high administrative costs. From the inception of the program until present, there
has been a concentration of assets in a few AFPs. In July 2006, two out of the six
total AFPs controlled roughly 66% of all fund assets. High administrative costs and
industry concentration may have resulted from government regulation in the AFP
system (e.g., regulation of fee structure, legal barriers to entry such as start-up
requirements or minimum reserve fund).73
Furthermore, many Chilean workers do not to appear to be well-informed about
the fees associated with their individual retirement accounts. A high level of
knowledge about administrative charges, some argue, may boost competition in the
AFP system by encouraging more plan members to shop around for the AFP with the
lowest fee structure. A recent analysis conducted by two Chilean pension experts
shows that only 3.7% of plan members were aware of variable commissions charged
by AFPs (and hence unable to compare fees and performance), and 52% did not
know what percentage of their income went toward contributions into an individual
The administrative costs reported in Chile would not necessarily be replicated
in other countries given the diverse factors that drive such costs.75 For example,
whether administrative functions are managed by the government or private entities
and whether administration tasks are covered by a single entity (centralized) or
diverse entities (decentralized) may play a role in administrative costs. A centralized
management system of investment funds could build on the existing tax collection
71 SAFP, Boletín Estadístico, 2001.
72 In October 1997, the government tightened the rules for transferring pension companies.
See Solange Berstein and Alejandro Micco, “Turnover and Regulation in the Chilean
Pension Fund Industry,” Central Bank of Chile Working Paper No. 180, September 2002,
73 Some estimates, however, point out that administrative costs under Chile’s individual
retirement accounts system are on average 42% lower than the old PAYGO system,
(Sebastían Edwards, 1998, p. 45).
74 Solange Berstein and Rubén Castro, “Costos y Rentabilidad de los Fondos de Pensions:
¿Qué Informar a Los Afiliados?,” Working Paper no. 1, Superintendent of Pension Fund
Management Companies, April 2005, at [http://www.safp.cl/files/doctrab/DT00001.pdf].
75 For a general discussion of administrative costs see Olivia S. Mitchell, “Administrative
Costs in Public and Private Pension Systems.” National Bureau of Economic Research
Working Paper No. 5734, 1996, at [http://www.nber.org/papers/w12401].
and record-keeping systems of the government, generate economies of scale, and
spread administrative costs over more workers, thereby lowering total costs.76 The
relatively low administrative costs in the U.S. government employees’ Thrift Savings
Plan (TSP) is often cited as evidence of the cost advantages of a centralized single,
private entity.77 By contrast, Chile has a decentralized structure of private pension
fund management companies.
The individual retirement accounts system was expected to improve
participation rates in Chile, in part by linking benefits more tightly to contributions.
However, active participation in the new system remains lower than expected.78
Whereas almost the entire Chilean workforce is enrolled in the AFP system,79 the
share of persons actively contributing to their accounts is much smaller (see Table
employed, are estimated to have contributed to their individual retirement accounts.
This figure is roughly similar to the level of coverage provided under Chile’s old82
PAYGO retirement system in the mid-1970s.
76 CRS Report RL32756, Social Security Individual Accounts and Employer-Sponsored
Pensions, by Patrick Purcell.
77 However, the TSP may enjoy economies of scale that would be unavailable to private
sector firms if individual accounts were administered competitively (see CRS Report
78 Dow Jones Industrial News. “Chilean Pensions’ Poor Coverage Sparks Reform Debate,”
April 19, 2004.
79 A person is “affiliated” with the system as long as he or she has contributed one month
over a period of 20 years. In 2005, around 7.3 million Chileans were affiliated. See
Superintendent of AFP,”Afliados activos por AFP,” Información Estadistíca y Financiera,
March 17, 2006, at [http://www.safp.cl/inf_estadistica/series_excel/anuales.html].
80 Contributions are mandatory for wage and salary workers, but remain voluntary for self-
employed workers (about 27% of the Chilean labor force). In Chile, the self-employed are
generally located on the lower end of the income distribution, and many are part of the
country’s large informal or “underground” economy. The vast majority of self-employed
do not contribute to the individual retirement accounts system (around 93%). There is also
a segment of wage workers who should be making individual accounts contributions, but
in practice do not comply with the law. Such workers are also likely to be part of the
81 Soto, 2005, pp. 4-5.
82 See SAFP, El Sistema Chileno de Pensiones, 6th edition (Santiago de Chile, 2007), figure
VI.3, p. 149. See also, Rodrigo Acuña and Augusto Iglesias, “Chile’s Pension Reform after
Table 5. Active Members and Total Contributors, Chilean
Individual Retirement Accounts, 1981-2005
YearsNumber of Members aNumber of Contributors b
1981 1,400,000 n/a
1983 1,620,000 1,229,877
1985 2,283,830 1,558,194
1987 2,890,680 2,023,739
1989 3,470,845 2,267,622
1991 4,109,184 2,486,813
1993 4,708,840 2,792,118
1995 5,320,913 2,961,928
1997 5,780,400 3,296,361
1999 6,105,731 3,262,269
2001 6,427,656 3,450,080
2003 6,979,351 3,618,995
2005 7,394,506 3,784,141
Source: SAFP, El Sistema Chileno de Pensiones, 6th edition (Santiago de Chile, 2007), p. 247.
Notes: (a) Members are those who are still alive and who are not receiving a pension and enrolled in
an AFP, December; (b) The number of members who contributed in December in each year.
Furthermore, not all of those that actively contribute to their accounts at one
point in time will contribute to them regularly over their working life.83 One study
estimates that an average Chilean worker entering the labor force at 20 years old and
retiring at 60 years old will have 21 years of contributions.84 Another study estimates
that an average plan member makes contributions for about 54% of his/her potential
working life.85 While around one fifth of plan members make contributions nearly
100% of the time over their careers, a substantial portion of the population does not
make regular contributions. Individuals may not contribute regularly for various
83 Pension analysts call this contribution density — the frequency with which workers
contribute to their individual account over their working lives.
84 Alberto Arenas de Mesa, 2005 (estimates were calculated by taking the number of months
of reported contributions divided by the number of months between the worker’s entry into
the data set — typically age 15 — and their age in 2002). Other Chilean pension observers,
such as Alejandra Cox Edwards, have estimated higher contributions densities, partly due
to using an older set age in which a worker enters the labor force.
85 Estelle James, Guillermo Martínez, and Augusto Iglesias, “The Payout Stage in Chile:
Who Annuitizes and Why?,” Working Paper no. 14, Superintendent of Pension Fund
Companies, 2005, at [http://www.safp.cl/files/doctrab/DT00014.pdf]; see also, Alberto
Arenas de Mesa, Jere Behrman, and David Bravo, “Characteristics of and Determinants of
the Density of Contributions in a Private Social Security System,” WP 2004-077, University
of Michigan Retirement Research Center, 2004, at [http://www.mrrc.isr.umich.edu/
reasons, such as interruptions in employment or job seasonality, non-employment,
and low wages in relation to daily expenses over a lifetime.
These trends reflect a growing concern in Chile that the current system does not
cover the entire labor force and provides inadequate benefits to an important segment
of workers. According to a Chilean government’s study baseline projection, nearly
40% of workers affiliated with an AFP will accumulate enough capital to fund a
benefit above the minimum level, 10% will qualify for a minimum benefit “top up,”
and nearly half will reach retirement with less than the minimum pension and fewer
than 20 years of contributions.86
Several factors help explain why participation in the Chilean individual
retirement accounts system has not been greater. The labor market in Chile has a
high share of self-employed workers, roughly 27% of the labor force, and a large
informal sector. The participation rate for self-employed workers (which is
voluntary) dropped from 12% to 7% between 1986 and 2003, while salaried workers
rose from 63% to 76%.87 Chilean men contribute to their accounts almost 40% more
months than women, who work fewer years, experience interrupted periods of
employment due to child rearing and have lower lifetime earnings.88 Another factor
may be the minimum pension guarantee, which may create incentives for lower
income workers to contribute just long enough to qualify for the minimum pension
(20 years), and soon thereafter stop making contributions.89
Obtaining knowledge of other countries’ experience with reforms in their social
insurance programs has gained importance in recent years as policymakers
contemplate how to address Social Security’s long-term financial challenges. It is
difficult, however, to draw general lessons from one country’s experience, as there
is no universal solution to reform. Social security systems operate differently in
different countries; each faces its own unique set of political and socioeconomic
conditions and has a different set of income support programs for the elderly.
Moreover, what is viewed as successful or desirable in one country may not be in
another. Nevertheless, information of reformed retirement systems around the globe
can provide valuable insight to policymakers.
86 Cited in Matijascic and Kay, 2006, p. 6.
87 Arenas de Mesa, 2005.
88 Arenas de Mesa, Behrman, and Bravo, 2004.
89 Salvador Valdés-Prieto, “Social Security Coverage in Chile, 1990-2001,” (Office of the
Chief Economist, Latin America and Caribbean Region, The World Bank Background Paper
for Regional Study on Social Security Reform, 2004).
The performance of Chile’s individual retirement accounts system is mixed.90
Individual retirement accounts have contributed to the Chilean economy in a number
of ways and the returns on pension fund investments have been greater than expected.
The system seems to work well for workers with stable jobs who contribute regularly
to their accounts over their working lives. However, the transition to an individual
retirement accounts system has proved fiscally expensive for Chile. Concerns remain
about low participation rates, especially among women and low-income workers,
including members of the informal sector. Analysts also agree that administrative
costs have been too high.
In an attempt to address these problems and improve the system, a pension
reform bill was sent to the Chilean Congress for consideration in 2007. The bill aims
to reduce administrative costs, increase benefits, and expand coverage to supplement
the shortfalls in the existing system’s safety net. This development suggests that the
Chilean individual retirement accounts model continues to evolve after 26 years of
90 A number of comprehensive studies have gauged the system’s performance to date. See,
for example, Alberto Arenas de Mesa, et al., “The Chilean Pension Reform Turns 25:
Lessons from the Social Protection Survey,” National Bureau of Economic Research
Working Paper 12401, 2006, at [http://www.nber.org/papers/w12401]; see also Mesa-Lago,
Appendix. Chilean Pension Funds
Table A1. Allocation of Chilean Pension Funds, 1981-2003
(as a percent of total investments)
Financial Mutual F oreign
YearGovernmentSecuritiesFinancialInstitutions InstitutionsEquities Funds and Assets
1981 28.1 71.3 0.6 0.0 0.0 0.0
1982 26.0 73.4 0.6 0.0 0.0 0.0
1983 44.5 53.4 2.2 0.0 0.0 0.0
1984 42.1 55.6 1.8 0.0 0.0 0.0
1985 42.4 56.0 1.1 0.0 0.0 0.0
1987 41.4 49.4 2.6 6.2 0.0 0.0
1988 35.4 50.1 6.4 8.1 0.0 0.0
1989 41.6 39.2 9.1 10.1 0.0 0.0
1990 44.1 33.4 11.1 11.3 0.0 0.0
1991 38.3 26.7 11.1 23.8 0.0 0.0
1992 40.9 25.2 9.6 24.0 0.2 0.0
1993 39.3 20.6 7.3 31.9 0.3 0.6
1994 39.7 20.0 6.3 32.2 0.9 0.9
1995 39.4 22.4 5.2 30.1 2.6 0.2
1996 42.1 23.6 4.7 26.0 3.0 0.5
1998 41.0 31.7 3.8 14.9 3.0 5.6
1999 34.6 33.2 3.8 12.4 2.7 13.3
2000 35.7 35.1 4.0 11.6 2.5 10.8
2001 35.0 32.4 6.2 10.6 2.5 13.2
2003 24.7 26.3 7.7 14.5 2.9 23.8
Source: Based on Arenas de Mesa, A. “Fiscal and Institutional Considerations and the Chilean
Prescription,” Table 3.5, p. 92 (using data from the Superintendent of Pension Fund Companies).
Notes: Totals may not equal the sums of rounded components. Categories less than 1% were dropped
(i.e., disposable assets). Financial institutions include banking deposits and mortgage securities, but
not equities of financial institutions. Non-financial institutions include corporate bonds of business
firms. Equities are stocks of financial institutions plus those of the business sector (domestic issuer).
Mutual Funds and others include investment funds of the business firms plus others from the external
sector (domestic issuer). Foreign assets includes foreign issuers less others from the external sector
(includes foreign mutual funds, foreign issue bonds, and equity of foreign corporations publicly traded
on the New York Stock Exchange, NASDAQ, the London Stock Exchange, etc.)