The Chilean Social Security System

Anand Shetty
Charles F. O’Donnell
Iona College

I    Introduction
     There has been and continues to be a growing interest in the U.S. and elsewhere in the world in a market-based model of social security delivery system to replace fully or partially the government run pay-as-you-go system.  The shortcomings of the government run systems are well known.  They have been inefficient, less reliable, and in some cases, outright discriminatory.  The main problem in the government-run system is the dependency of the social security on the political process.  Groups interested in changing the system have been tracking the developments in the market-based system introduced in Chile in 1981 and its variants in other Latin-American countries (Peru, Argentina, Columbia, Uruguay, and Mexico) and elsewhere in the world.  International organizations such as IMF have also been paying close attention to the Chilean model and advising third world countries to introduce similar reforms.  The main purpose of this paper is to examine the key features of the Chilean privatized social security system and to focus on the main problems facing it.  The paper will also draw attention to lessons it may have for the social security reformers in the U. S. and elsewhere.
     In Chile, mandatory social insurance began in 1924 and evolved as a defined benefit (pay-as-you-go) system.  By the 1970s, it had developed a structure with a hodgepodge of separate defined systems for different industries and occupations with different benefit levels with multiple bureaucracies.  The lack of uniformity meant large differences in retirement benefits, some retiring early with large benefits and while other, particularly the blue-color workers had to wait until age 65 to qualify for benefits. More over, some, not all, pensions had automatic cost-of-living adjustments. The system also suffered from declining contributors. It had 12 active contributors per retiree in 1955 and by 1979, the number stood at 2.5 contributors per retiree.¹ The political involvement in the determination of benefit levels had resulted in a high rate of contributions for workers over the years.  In 1975, this rate, including the contributions for health benefits, was in the range of 51-59% of the income.²   The declining contributors also meant greater burden on the government budget.  In 1980, the government had to cover 40 per cent of pension expenses from its budgetary sources.³  A group of economists investing the problems of the social security system in 1973 pointed out excessive contributions, discrimination and injustice in benefits, high administrative cost, low quality of insurance, lack of political will to limit the fiscal cost as the main sources of its problems.  The group proposed a reform which would include separation of income distribution and insurance functions, options for lump-sum withdrawal and annuities, replacement of legislated rules by contractual rules, and a full privatization of the selection of provider.  This proposal had great impact on the reform measures that were introduced in 1981.  Based on the findings of a government study group on the cost of implementing the reform plan in 1974, a decision was made at the government level to build budget surplus in anticipation of the implementation of the plan.
     The reforms introduced in 1981 established a privatized mandatory savings system along with a government operated assistance and minimum pension program to deliver social security benefits to the people.  Under the new system, benefits are to be delivered at three levels. At the first level, the benefit delivery is made from a privately-managed defined contribution pension plan; at the second level, the delivery is made from a government guaranteed program to those whose private pension accumulations are inadequate (minimum benefits); and at the third level, benefits were delivered to the elderly and poor who do not qualify for another pension (assistance benefits).  The second and third levels of benefit delivery reflect the redistributive aspects of the new system. 
II     The New Social Security System
Privatized Mandatory Defined Contribution Plan
     Under the new social security system, workers entering the labor force after Dec 31, 1982 were required to join the new system and the existing members of the work force had the option of staying in the old system. The covered workers must deposit 10% of their wages into an individual retirement account at an approved AFP (Administradora de Fondos de Pensiones) of their choice.  The account is held in the worker’s (affiliate’s) name, and the accumulation in the account is used to pay his/her retirement pension.  Self-employed are not required to participate, but they can voluntarily set up accounts with the same features.  The accumulation in the account is delivered to the affiliate in the form of an indexed annuity sold by an insurance company, a programmed withdrawal or a combination of the two on reaching a pensionable age of 65 years for men and 60 years for women.  Since Chile had a long history of issuing indexed debt, commitment to issuing indexed annuity on retirement did not pose any problem. An individual contributor can choose to retire early if his or her account balance when converted into annuity exceeds 50% of the average annual income and 110% of the minimum pension. The contributions of workers and the investments of the AFPs are free of taxes.  Taxes are, however, paid on withdrawals upon retirement at a lower rate.⁴
     The AFPs also provide disability and survivor insurance coverage to its account holders through private insurance companies. This coverage is paid for by additional contribution of 3% of the salary.  Another 7% of the salary is withheld to finance medical insurance under the new system.  The total mandatory contribution to the new system, including coverage for insurance and provision for income to AFPs, amounts to 13 to 14 percent of the wages.  Government also mandated a before-tax wage increase to participants that equaled the employer financed portion of the workers’ contribution under the old system to ensure that workers compensation is not lowered as a result of eliminating employer contribution.
     The annuity option guarantees a constant monthly income for life, indexed to inflation, plus survivor’s benefits for the worker’s dependents.  Under this option, any accumulation in excess of what is needed to fund basic benefits can be withdrawn at retirement and can be used for any purpose.  The mandatory savings levels under the system are determined in such a way that the typical worker will have sufficient funds to finance a pension that will provide 70% of average salary over the last 10 years plus disability and survivor’s benefits equal to 50% of the retirement pension. 
     While the new system brought the contribution rate down from 22% close to 14%, those who chose to remain in the old system did not receive any cut in the contribution rate.  This resulted in a massive switch into the new system by those below the age of 45 years when the new system was introduced.
     The subscribers to the new system are not allowed to hold accounts in more than one AFP, but they are free to switch after being with an AFP at least for four months. The freedom to switch is intended to generate competition among the AFPs and hold the administrative cost down.  
     The AFPs are subjected to strict government regulations to protect the interest of the workers.  The AFPs must be specialized institutions, separate from other financial institutions.  Each AFP is allowed to manage a single fund, and the return on the fund is allocated completely to the individual accounts.  They are allowed to charge a commission for their management services, and the structure of this commission is regulated.  The regulation also covers the portfolio composition and organization of the fund.  These regulations require that pension manager’s rate of return fall within +2/-2 percent of the average of all-AFPs. Until 1985, no equity investment was allowed in the AFP portfolios. The maximum amounts now allowed for investment in specific securities are fixed at 30% for domestic equity, 20% for foreign equity, and 45% for government bonds.  The AFPs also have to satisfy a minimum reserve requirement.  No AFP can purchase more that 20% of a bond issue.  Equity holding in a given Chilean company is limited to 7% of the outstanding equity.  They are not allowed to hold equity in other AFPs, mutual funds, insurance companies, and stockbrokerage firms.  Banks were not allowed to provide pension benefits under the new system. 
     The system also provided for individual savings for retirement outside the mandatory savings system through incentive schemes consisting of special tax treatments.  Long-term savings through AFPs are exempted from income tax on the principal as well as interest earned.   Withdrawals from these savings are, however, limited to pensions.  Voluntary short-term savings at the AFP are available for withdrawal, but the tax treatment is different from that of long-term savings.
Minimum and Assistance Pension Guarantees
     The government guarantees minimum pension to those whose accumulated funds through contribution for at least 20 years falls short of the minimum pension (about 75% of the minimum wages).  This pension is protected against inflation.  People who do not qualify for minimum pension and are without any means of support are eligible to receive an assistance pension.  The amount allocated for this is limited by legislation and is not automatically adjusted for inflation. Both payments are funded from the general revenue source.
Financing the Transition
     The pension reform implied that the government must find funds to finance the existing system in the face of declining worker contributions.  The Chilean government had prepared to meet the resulting budget deficit by generating budget surpluses in the years preceding the reform.  The level of surplus stood close at 5.5% of the GDP in the year 1980.⁵  The most of the transition deficit - the deficit in the old pension system - was financed out of a primary surplus.  Table below shows the extent of actual budget deficits (contributions minus pensions plus the value of the recognition bonds that come due each year) up to1988 and their estimates attributable to the reform.

See Table 1

Estimates are based on a constant GDP growth rate of 4%.  The financing of the deficits relies on funds from four sources: accumulated government surpluses, privatization of state-owned businesses, issue of government bonds (the recognition bonds to be given to those who left the old system in lieu of their past contributions), general revenue, and increase in the retirement age.  The recognition bonds valued at 80%-100% of GDP was expected to cover the transition liability over time.  These bonds would mature when the worker reached the retirement age and earn 4% interest.
Further Changes to the System
     Many changes to the new system have been made since 1981, leaving, however, the basic structure unaltered. The following is a list of some of the major changes:

1. Authorization of investment in domestic equity in 1985 and foreign investment in 1990.
2. Redesign of disability and survivor insurance in 1988.  Under the original reform, the insurance would pay the real annuity defined by the law to the survivor or disabled.  The 1988 change required the insurance company to make a lump sum to the beneficiary’s account, and then the beneficiary may use the funds to purchase a real annuity.
3. Authorization of early pension in 1988. This reform waived the age requirement for the affiliates who could show that the funds they hold are sufficient to purchase a real annuity that is adequate.
4. Other reforms include indexing the assistance and minimum pensions, improving and designing the existing regulations, establishing an AFP clearing house, helping reduce the level of annuity fee, introducing the group purchase of annuity and allowing AFPs to offer multiple funds.

III     Evaluation of the System
     The mandatory system introduced in Chile and in other Latin American countries provide an effective solution to the problem of intergeneration conflict originating from the labor demographics that the traditional pay-as-you-go social security systems have been faced with.  That is, the working population does not have to subsidize the retiring population under the privatized mandatory system.  By letting the funds to accumulate in the name of the employee, the system has changed the modus operandi of delivering earnings related retirement provision and insulated it from political influence.
     The system became popular and it was instrumental in promoting economic growth and stability in Chile.  The benefit payments were freed from the political process and the uncertainty that goes with it.  It replaced a system that was essentially discriminatory and unfair as some powerful group could retire their members with nearly full benefits after as few as 25 years of work.  The workers acquire a direct stake in the economy when they become the investors. 
     When the reform gave the workers the choice to join the new system, most young workers chose to join the new system.  The number of people who joined the system and had an AFP account increased gradually, while those who remained in the old declined over the years.  Between 1981 and 1996, the number of people holding an account increased from 1.4 million to 5.57 million, representing 99% of the labor force.  Due to lack of continuity in contribution to the account, the number of people who regularly contribute is, however, far less than the account holders. This number stood at 3.12 million in 1996, just 59% of those employed. ⁶
     Besides being successful in transitioning from a traditional pay-as-you-go to a fully funded privatized system, the new system made noteworthy contributions to the Chilean economy in many areas.  It helped to lower the long-term public sector liabilities and contributed to the development of efficient capital markets and financial institutions. Chile started pension reforms without a strong background in financial regulation. Important financial regulations were introduced later in the 1980s and in 1990s in response to the needs of growing pension funds.  Increased specialization and the creation of new financial instruments and markets such as long-term corporate bond market created in 1988-89 are the direct results of the new system.  Additional reforms in the financial markets and growth of privatized pension funds contributed to greater transparency and efficiency in the financial markets. Transparency comes from workers knowing exactly how much they have contributed and from their freedom to choose the fund manager, and the security of having retirement funds outside the influence of changing government policies. Financial disintermediation process that began in the mid-1980s resulted in pension funds holding fewer instruments that were intermediated through the banking system.  The Chilean experimentation demonstrates that pension reforms can be started without having a sound financial regulatory system. 
     One noteworthy change that was essential for the pension reform in Chile to succeed was the development in 1974-78 of a legal framework that permitted the trading in a full range of CPI-indexed debt by all market participants, including banks, firms, and households.  The government had also indexed the tax collection to make the tax revenue independent of inflation and enable the government to issue indexed bonds safely.
     As pointed out before, the steadily growing pension funds accompanied by financial market legislation and changes in the policies governing AFPs investment were instrumental in the growth of capital market in Chile.  Pension funds are the largest investors in the Chilean capital market. The assets held by AFPs reached 56% GDP in 2002 and they included a significant share of Chile’s equity and public debt.⁷  The availability of AFP funds supported the privatization of state-owned enterprises.  The capitalization of the Chilean stock market reached the level of its GDP, a feat very few developing countries can be proud of.  Some 40% of the AFP assets are held in government bonds and another one-third in domestic stocks.  They hold 70 percent of the total public debt. The high degree of concentration in the domestic securities in the AFP portfolios is the reflection of high average rates of return during the 1980s.  The new system also had positive impact on the insurance industry because disability, survivors’ benefits and annuities are paid by insurance companies.  Insurance companies and pension funds also helped to develop the real estate market by supplying long-term mortgage loans. About 18% of the pension funds portfolio was invested in mortgage bonds in 1996.⁸  One of the visible areas of benefits of social security privatization in Chile is the high rates return in the 1980s and 1990s.  In spite of periods of low and negative returns, AFP investments earned an annual average return of 10.59% from 1982 to 2005.⁹
IV     Challenges
     The system, however, faces many challenges. Two important challenges today are (1) much lower income-replacement rates than expected and (2) low and declining participation rate. The two are somewhat interconnected. The main factors contributing to these problems are the high management cost, decreasing government-bond yields and lapses in contributions. The workers retiring in the formal sector have observed income-replacement rate as low as 40 percent.¹⁰
     The participation rate has been and continuous to be a problem and it is far from universal. Measured as the coverage ratio (the ratio of individuals contributing to pension fund to labor force), it remains low at about 55%.  This problem is further accentuated by the low density of contribution (actual number of months of contribution over the total potential months of contribution) which is below 60%.¹¹  The contributory factors to low participation include fragmented nature of the Chilean labor market (divided into formal and informal segments), non-participation of self-employed who account for 25% of total employment and for whom participation in the private pension system is voluntary, moral hazard problem created by the minimum pension guarantee as low-income individuals have incentive to contribute as little as possible or not contribute at all. It is estimated that in any given month, only about half of Chileans with personal accounts make contributions.¹²
     In spite of lofty investment returns over the years, the system has failed to keep the income-replacement rate at a desirable level because of the high cost of benefit delivery.  The Chilean system has been frequently criticized for this high cost.  These costs include those of the AFPs, who are allowed to charge fees and commissions freely, and those of the insurance companies that sell disability insurance, life insurance, and annuities.   Valdes-Prieto (1994) estimated that the average administrative cost per participant while active is US$89.10 per year in 1991, which is 2.94 percent of average taxable income and 20 percent of the roughly 13.5% paid for the program. Because the system allows frequent transfers of funds, the competition among AFPs with identical portfolios has increased causing the sales, marketing and administrative costs to rise.  It has been estimated the marketing and sales cost in the early part of 1990s was in excess of one third of total cost.  In a study published in McKinsey Quarterly, Chaia et al. (2007) place the cost of benefit delivery at one-fifth of the potential capital accumulation of savings during 1982 to 2005 period.     
     Though the new system was designed to limit the role of the government, the budgetary adjustments will be required for (1) payment of pensions to those who remained in the old system, (ii) financing the recognition bonds¹³ issued to those who had made contribution to the old before joining the new one (iii) covering the cost of guaranteed minimum and assistance pensions.  Payments of benefits to all the pensioners under the old system are not expected to cease until 2045.¹⁴  Though the budget deficit due to these liabilities is likely to decrease over time as the number of workers receiving benefits under the old system falls, the government will continue to face the pressure to raise tax to cover its obligations to make minimum and assistance pension payments to increasing number of people due to low participation rate and resulting low accumulation under the new system. It is estimated that half of the workers in Chile are not savings enough to qualify for even a minimum pensions.¹⁵  If the current rate of participation continues, it is estimated that the increase in minimum pension and assistance benefits will increase to 2.4 percent of GDP in 2030 and continue to rise thereafter.¹⁶
     Another aspect of the Chilean system is that the workers bear investment risk.  This aspect does not appear to be a problem when seen from the point of high average return earned on the investment overtime.  But there may be periods when the returns are low and even negative causing a major concern for those planning to retire.  In 1994, more than half of the AFPs incurred losses. Between 1995 and 1998, returns were -2.5%, 3.5%, 4.7% and -1.1% respectively. Workers actually lost much when management fees are factored into the return. Chilean government asked workers to defer retirement until the situation improves.¹⁷  The system has built in exposure to both the market and interest rate risks as conversion of accumulated wealth to annuities can occur only at retirement.  There is always a possibility that their funds will have low value and low interest rate when they have to make that decision on conversion. Given that the annuity market in Chile suffers from incomplete information about life expectancy, the retirees may face selection bias as insurance companies try to attract better risk.
V     Problems Continue to Plague the System
     The much-ballyhooed privatized social security system of Chile has not been able to face its challenges effectively. During the 2005 Chile’s presidential election, the problems facing the system were hotly debated and they became big issues in the campaign. Candidates from both the left and the right agree that the system needs fixing.  The program is not delivering what has been promised by the original scheme:  a minimum of 70% of salaries from contributions made over a 20 year period to the system managed by six competing administration companies.  Many of those who started work when the system was first adopted are realizing that they have not contributed enough to generate a significant amount of pension.  They find that they are entitled to a pension that is between 30 to 50 percent of their wages.¹⁸  Main sources of the problems are high administrative cost which soaks up as much as a third of their original investment, too few staying long enough in their jobs to make the contributions needed for the system to work, many workers, mostly women, moving in and out of the jobs too frequently, only half of the working population actually contributing regularly, and opting out of the system by most self-employed workers¹⁹.  This low level of coverage means a heavy burden on the government as it has the obligation to provide a minimum pension for contributors.  Even after 25 years of its existence, the government continues to pour money into the system to make up for the shortcoming.
     Commenting on Chile’s problem with its social security system, Larry Rohter writes in NY times on January 27, 2005 that people who contributed regularly are finding they are not making in pensions as much as they would have made if they had stayed in the old system.  Rohter gives an example of Dagoberto Saez, a 66-year old laboratory technician who plans to retire in March.  Dagoberto Saez earns just under $950 a month. His pension fund has told him that his nearly 24 years of contributions will finance a 20-year annuity paying only $315 a month.  He says, “Colleagues and friend with the same pay grade who stayed in the old system, are retiring with pensions of almost $700 a month – good until they die.”  Thousands of other Chileans approaching retirement age are also finding that they would have been better off if they had remained in the old system. 
     On might argue that the true performance of the system cannot be judged until after 2020 when the first generation of workers who would made contribution during the entire career begins to retire.  If current data on participation and accumulation rates is any guide, it is very difficult to argue that the program is a success.  In a survey conducted in 2004, more than half of Chileans polled showed no confidence in the privatized system.²⁰ 
     Staunch advocates of free market such as Mr. Pinera, brother of Jose Pinera who was instrumental in introducing the personal account system during the dictatorship of Gen. Pinochet and the current co-chairman of the Project on Social Security Choice at the conservative Cato Institute, have also jumped in with criticism.  Added to these is the public anger on high profits pension fund companies are making on the back of pension participants.  It is reported that at least one peso in every five goes to meet the administrative costs and it lacks transparency.  According to one estimate, between 1999 and 2003 the companies made an average annual return on capital of 53 percent.  It needs transparency as most Chilean are unaware of how much they are paying to funds because of lengthy quarterly financial balance sheet they receive which is incomprehensible according to Guillermo Larrain, director of the Superintendency of Pension Funds.
     Many suggestions have been made during the political debate surrounding the presidential election for reforming the system.  The suggested reforms include a subsidy on contributions to encourage participation, allowing contributors to join block schemes that would be auctioned off to lower commission costs, allowing banks and insurers to enter the market to increase competition and reduce cost.  It also included a guaranteed minimum pension for the poorest members of the society, even if they never contributed to the private system.  When the plan was announced in December 2006, Labor Minister Osvaldo Andrade told the reporters, “This is a radical reform, because it moves us from a system based solely on individual savings to one that includes a pillar of solidarity based on one’s rights as a citizen, and not contributions.  We are integrating systems that are fundamentally different.” ²¹
VI     Concluding Remarks
     When considering the relevance of the Chilean system to other countries, one must consider issues such as whether the benefits of a privatized system outweigh its cost, the timing of the change with respect to fiscal situation, the possibility of an alternative system which may have lower administrative cost, an annuity market that functions better than it did in Chile, and financing of the transition cost.
     Chile’s social security system was highly inefficient before the privatization and continues to suffer from many problems including inefficiency after the privatization.  The sources of these problems are, however, different.  The cost of delivery has remained high. This may not be the case for many countries with developed and efficient capital markets such as the U.S.  There is a danger of raising expectation based on early successes of Chilean system in terms of high rates of return on investment made possible by deepening capital market.  The same situation may not prevail in other countries looking into privatization as the point in time Chile introduced the system is different from what it is today.
     The twin problems of low income-replacement rate and low participation rate that Chile continued to face is something that others must notice.  In some ways, these problems may be unique to Chile because of the way the new system was designed and the structure of its labor market.  Some of the issues behind these problems such as high administrative cost can be contained and may not even a factor in a country such as the U.S. with highly competitive financial markets.
     It is, however, difficult to assume that other countries can duplicate lofty returns on privatized pension funds along the lines Chile did.  Chilean investors benefited from uncanny market timing and financial market reforms.  Privatization in Chile helped to develop its fledgling capital markets early on and benefited from that growth over time.  For countries with well –developed capital market such as the U.S., the beneficial effects of privatization as seen in the capital market development and yield rates in the 1980s and the 1990s in Chile is less relevant.

Reference
Asociacion Gremial de Administradoras de Fondos Pensiones, The Private Pension Funds System in Chile, December
        1994.
Barreto Flavio Ataliba, and Olivia S. Mitchell, “Privatizing Latin-American Retirement Systems,” Benefits Quarterly, Third
        Quarter 1997.
Chaia, Alberto, Antonio Martinez, Luis Enrique Rodriguez, “Reforming Latin American Pension Systems,” McKinsey
        Quarterly, 2007 Special Edition, pp126-131.
Congressional Budget Office, “Social Security Privatization, Experiences Abroad,” January 1999.
Diamond Peter A., “Proposals to Restructure Social Security,” Journal of Economic Prospectus, Vol. 10, No.3, Summer
        1996.
Diamond, Peter and Salvador Valdes-Prieto, “Social Security Reforms,” in The Chilean Economy by Barry P. Bosworth.,
        Rudigar Dornbusch, and Raul Laban, editors, The Brookings Institution, Washington D.C., 1994.
Dickerson, Maria, “A Personal Burden: Chile Switched to a Privatized Pension System Nearly 25 Years Ago, and Millions
        of Workers Still Fall Through the Cracks,” Los Angeles Times, Feb 13, 2005, pp C.1.
Gill, Indermit S, Truman G. Packard, and Juan Yarmo, “Keeping the Promise of Social Security in Latin American,” World
        Bank Publication, October 2004.
Idemoto, Steve, Social Security Privatization in Chile: A Case for Caution, Economic Opportunity Institute, Sept 29, 2000.
Lapper, Richard and Adam Thompson, Comments & Analysis, FT 11/25/2005.
Organization for Economic Cooperation and Development, The Chilean Pension System, Working paper AWP5.6, 1998.
Organization for Economic Cooperation and Development, Chile, Economic Surveys, 2003.
Rohter, Larry, www.nytimes.com/2005/01/27/business/worldbusiness.
Rohter, Larry, www.nytimes.com/2006/12/26/world/americas/26chile.html
The International Center for Pension Reform, The Chilean Private Pension System, Santiago, Chile.
Valdes-Prieto, Salvador, “Administrative Charges in Pensions in Chile, Malaysia, Zambia, and the United States. “ Policy
        Research Working Paper 1372, World Bank, 1994.

1. Congressional Budget Office, 1999
2. See Peter Diamond and S. Valdes-Prieto, pp.258.
3. OECD, 2003, pp.178
4. Congressional Budget Office, 1999
5. Peter Diamond and S. Valdes-Prieto, pp.19
6. OECD, 1998.
7. OECD, 2003
8. OECD, 1998
9. Chaia et al., 2007, pp.127
10. Chaia et al., 2007, pp.126.
11. OECD, 2005, pp.60
12. Dickerson, 2005, pp.C1
13. Individuals who contributed for at least twelve months in the old system during the previous 5 years to November 1980 is eligible to receive a Recognition Bond. The amount is determined by what is needed to get an annuity equivalent to 80% of the pensionable salary that the individual received between June 1978 and June 1980, weighted by the number of years contributed over thirty-five years. (OECD, 1998, pp.11)
14. OECD, 1998.
15. Dickerson, 2005
16. OECD, 2003, page 43
17. Idemoto, 2000
18. Larry Rohter, 2007
19. Lapper, 2005
20. Maria Dickerson, 2005
21. See Larry Rohter, New York Times, 12/26/06