Empowering People: The Privatization of Social Security

June 26, 1997 • Testimony

Mr. Chairman, distinguished members of the subcommittee:

My name is José Piñera and I am a Chilean citizen. I spent several years of my life at Harvard University, earning a Master in Arts and a Ph.D. in economics, and my eldest son was born during those years in Boston. Today, I am president of the International Center for Pension Reform in Santiago, Chile, and co‐​chairman of the Cato Institute’s Project on Social Security Privatization. As Minister of Labor and Social Security from 1978 to 1980, I was responsible for the privatization of the Chilean pension system.

I want to thank Chairman Gramm for his invitation to me to testify in the U.S. Senate. In keeping with the truth in testimony requirements, let me first note that neither the Cato Institute nor the International Center for Pension Reform receives any government money of any kind.

Social Security in the United States

The Social Security system of the United States is facing a profound crisis. [1] According to the latest report of its Board of Trustees, the system will be insolvent by 2029, back from 2030 in last year’s report. That represents the eighth time in the last 10 years that the insolvency date has been lowered.

But even that fact does not provide the full story of Social Security’s looming crisis. The important date is 2012. Social Security taxes currently bring in more revenue than the system pays out in benefits. The surplus theoretically accumulates in the Social Security Trust Fund. However, in 2012 the situation will reverse. Social Security will begin paying out more in benefits than what it collects in revenues. To continue meeting its obligations, it will have to begin drawing on the Trust Fund. The trouble is that this Fund is really little more than a polite fiction. For years the federal government has used the Trust Fund, borrowing money from it to pay current operating expenses and replacing the money with government bonds.

Even if Congress can find a way to redeem the bonds, the Trust Fund will be completely exhausted by 2029. At that point, Social Security will have to rely solely on revenue from the payroll tax. But such revenues will not be sufficient to pay all promised benefits. Either payroll taxes will have to be increased–to between 28 percent and 40 percent, according to various estimates by the trustees–or benefits will have to be reduced by as much as one‐​third.

Moreover, even if Social Security’s financial difficulties could be fixed, the system remains a bad deal for most Americans, a situation that is growing worse for today’s young workers. Payroll taxes are already so high that even if today’s young workers receive the promised benefits, such benefits will amount to a low, below‐​market return on those taxes. While today’s retirees will generally get back all they paid into Social Security (including the employer’s portion) plus a modest return on their investment, when today’s young workers retire, they will actually receive a negative rate of return–that is, less than what they paid in. A typical young worker today would actually be better off stuffing his Social Security taxes in a mattress than counting on benefits from the program.

Almost 30 years ago, visionary people like Nobel laureate James Buchanan understood that radical reform of the U.S. Social Security system would be needed. [2]

Also economic growth in the United States would be higher if Social Security were to be privatized. Of course, I am not talking about the government investing the Social Security Trust Fund in private capital markets, a proposal fraught with danger, as professor Krzysztof Ostaszewski has clearly shown. [3]

I believe that the benefits of privatization would be substantial. Harvard Professor Martin Feldstein, for example, estimates that “the combination of the improved labor market incentives and the higher real return on savings (of moving to a fully funded Social Security system) has a net present value gain of more than $15 trillion, an amount equivalent to 3 percent of each future year’s GDP forever.” [4]

Some might believe that privatizing social security would not benefit the poor, but the opposite is the case, according to Michael Tanner of the Cato Institute. [5] Moreover, as Boston University Professor Laurence Kotlikoff and others have shown,

Privatizing social security can, intragenerationally speaking, be quite progressive notwithstanding the fact that such a policy would eliminate the ongoing use of highly progressive social security benefit schedule. Part of the reason for this outcome is that being forced to contribute to and receive a less‐​than‐​market rate of return from “pay‐​as‐​you‐​go” social security represents a larger implicit tax for the poor because of the ceiling on social security taxable earnings. The other part is that the growth in the tax base permits the government to cut general revenues in a way that can differentially benefit the poor. [6]

And, most important, today’s young workers would be assured that when they retired, they would be able to do so with dignity and security.

Let me, for the remainder of my time, bring you an idea, a powerful idea, that can save social security in the United States by privatizing its provision. That is exactly what we did in Chile 16 years ago.

The Chilean PSA System [7]

In 1980, Chile approved a law (D.L. 3500) to fully replace a government‐​run pension system with a privately administered, national system of Pension Savings Accounts.

The new system began to operate on May 1, 1981 (Labor Day in Chile). After 16 years of operation, the results speak for themselves. Pensions in the new private system already are 50 to 100 percent higher–depending on whether they are old‐​age, disability, or survivor pensions–than they were in the pay‐​as‐​you‐​go system. The resources administered by the private pension funds amount to around 42 percent of Chile’s GNP. By improving the functioning of both the capital and the labor markets, pension privatization has been one of the key initiatives that, in conjunction with other free‐​market oriented structural reforms, has pushed the growth rate of the economy upwards from the historical 3 percent a year to 7.0 percent on average during the last 12 years.

In a recent work, UCLA Professor Sebastian Edwards has stated that, “The [Chilean] pension reform has had important effects on the overall functioning of the economy. Perhaps one of the most important effects is that it has contributed to the phenomenal increase in the country’s saving rate, from less than 10% in 1986 to almost 29% in 1996.” [8] He goes on to say that, “The pension reform has also had an important effect on the functioning of the labor market. First, by reducing the total rate of social security contributions it has reduced the cost of labor. Second, by relying on a capitalization system it has eliminated the labor tax component of the retirement system.”

Under Chile’s Pension Savings Account (PSA) system, what determines a worker’s pension level is the amount of money he accumulates during his working years. Neither the worker nor the employer pays a social security tax to the state. Nor does the worker collect a government‐​funded pension. Instead, during his working life, he automatically has 10 percent of his wages deposited by his employer each month in his own, individual PSA. A worker may contribute an additional 10 percent of his wages each month, which is also deductible from taxable income, as a form of voluntary savings.

A worker chooses one of the private Pension Fund Administration companies (“Administradoras de Fondos de Pensiones,” AFPs) to manage his PSA. These companies can engage in no other activities and are subject to government regulation intended to guarantee a diversified and low‐​risk portfolio and to prevent theft or fraud. A separate government entity, a highly technical “AFP Superintendency,” provides oversight. Of course, there is free entry to the AFP industry.

Each AFP operates the equivalent of a mutual fund that invests in stocks and bonds. Investment decisions are made by the AFP. Government regulation sets only maximum percentage limits both for specific types of instruments and for the overall mix of the portfolio; and the spirit of the reform is that those regulations should be reduced constantly with the passage of time and as the AFP companies gain experience. There is no obligation whatsoever to invest in government or any other type of bonds. Legally, the AFP company and the mutual fund that it administers are two separate entities. Thus, should an AFP go under, the assets of the mutual fund–that is, the workers’ investments–are not affected.

Workers are free to change from one AFP company to another. For this reason there is competition among the companies to provide a higher return on investment, better customer service, or a lower commission. Each worker is given a PSA passbook and every three months receives a regular statement informing him how much money has been accumulated in his retirement account and how well his investment fund has performed. The account bears the worker’s name, is his property, and will be used to pay his old age pension (with a provision for survivors’ benefits).

As should be expected, individual preferences about old age differ as much as any other preferences. The old, pay‐​as‐​you‐​go system does not permit the satisfaction of such preferences, except through collective pressure to have, for example, an early retirement age for powerful political constituencies. It is a one‐​size‐​fits‐​all scheme that exacts a price in human happiness.

The PSA system, on the other hand, allows for individual preferences to be translated into individual decisions that will produce the desired outcome. In the branch offices of many AFPs there are user‐​friendly computer terminals that permit the worker to calculate the expected value of his future pension, based on the money in his account, and the year in which he wishes to retire. Alternatively, the worker can specify the pension amount he hopes to receive and ask the computer how much he must deposit each month if he wants to retire at a given age. Once he gets the answer, he simply asks his employer to withdraw that new percentage from his salary. Of course, he can adjust that figure as time goes on, depending on the actual yield of his pension fund or the changes in the life expectancy of his age group. The bottom line is that a worker can determine his desired pension and retirement age in the same way one can order a tailor‐​made suit.

As noted above, worker contributions are deductible for income tax purposes. The return on the PSA is tax free. Upon retirement, when funds are withdrawn, taxes are paid according to the income tax bracket at that moment.

The Chilean PSA system includes both private and public sector employees. The only ones excluded are members of the police and armed forces, whose pension systems, as in other countries, are built into their pay and working conditions system. (In my opinion–but not theirs yet–they would also be better off with a PSA). All other employed workers must have a PSA. Self‐​employed workers may enter the system, if they wish, thus creating an incentive for informal workers to join the formal economy.

A worker who has contributed for at least 20 years but whose pension fund, upon reaching retirement age, is below the legally defined “minimum pension” receives that pension from the state once his PSA has been depleted. What should be stressed here is that no one is defined as “poor” a priori. Only a posteriori, after his working life has ended and his PSA has been depleted, does a poor pensioner receive a government subsidy. (Those without 20 years of contributions can apply for a welfare‐​type pension at a lower level).

The PSA system also includes insurance against premature death and disability. Each AFP provides this service to its clients by taking out group life and disability coverage from private life insurance companies. This coverage is paid for by an additional worker contribution of around 2.9 percent of salary, which includes the commission to the AFP.

The mandatory minimum savings level of 10 percent was calculated on the assumption of a 4 percent average net yield during the whole working life, so that the typical worker would have sufficient money in his PSA to fund a pension equal to 70 percent of his final salary.

Upon retiring, a worker may choose from two general payout options. In one case, a retiree may use the capital in his PSA to purchase an annuity from any private life insurance company. The annuity guarantees a constant monthly income for life, indexed to inflation (there are indexed bonds available in the Chilean capital market so that companies can invest accordingly), plus survivors’ benefits for the worker’s dependents. Alternatively, a retiree may leave his funds in the PSA and make programmed withdrawals, subject to limits based on the life expectancy of the retiree and his dependents. In the latter case, if he dies, the remaining funds in his account form a part of his estate. In both cases, he can withdraw as a lump sum the capital in excess of that needed to obtain an annuity or programmed withdrawal equal to 70 percent of his last wages.

The PSA system solves the typical problem of pay‐​as‐​you‐​go systems with respect to labor demographics: in an aging population the number of workers per retiree decreases. Under the PSA system, the working population does not pay for the retired population. Thus, in contrast with the pay‐​as‐​you‐​go system, the potential for inter‐​generational conflict and eventual bankruptcy is avoided. The problem that many countries face–unfunded pension liabilities–does not exist under the PSA system.

In contrast to company‐​based private pension systems that generally impose costs on workers who leave before a given number of years and that sometimes result in bankruptcy of the workers’ pension funds–thus depriving workers of both their jobs and their pension rights–the PSA system is completely independent of the company employing the worker. Since the PSA is tied to the worker, not the company, the account is fully portable. Given that the pension funds must be invested in tradable securities, the PSA has a daily value and therefore is easy to transfer from one AFP to another. The problem of “job lock” is entirely avoided. By not impinging on labor mobility, both inside a country and internationally, the PSA system helps create labor market flexibility and neither subsidizes nor penalizes immigrants.

A PSA system is also much more efficient in promoting a flexible labor market. In fact, people are increasingly deciding to work only a few hours a day or to interrupt their working lives–especially women and young people. In pay‐​as‐​you‐​go systems, those flexible working styles generally create the problem of filling the gaps in contributions. Not so in a PSA scheme where stop‐​and‐​go contributions are no problem whatsoever.

The Transition

One challenge is to define the permanent PSA system. Another, in countries that already have a pay‐​as‐​you‐​go system, is to manage the transition to a PSA system. Of course, the transition has to take into account the particular characteristics of each country, especially constraints posed by the budget situation.

In Chile we set three basic rules for the transition:

1. The government guaranteed those already receiving a pension that their pensions would be unaffected by the reform. This rule was important because the social security authority would obviously cease to receive the contributions from the workers who moved to the new system. Therefore the authority would be unable to continue paying pensioners with its own resources. Moreover, it would be unfair to the elderly to change their benefits or expectations at this point in their lives.

2. Every worker already contributing to the pay‐​as‐​you‐ go system was given the choice of staying in that system or moving to the new PSA system. Those who left the old system were given a “recognition bond” that was deposited in their new PSAs. (The bond was indexed and carried a 4 percent real interest rate). The government pays the bond only when the worker reaches the legal retirement age. The bonds are traded in secondary markets, so as to allow them to be used for early retirement. This bond reflected the rights the worker had already acquired in the pay‐​as‐​you‐​go system. Thus, a worker who had made pension contributions for years did not have to start at zero when he entered the new system.

3. All new entrants to the labor force were required to enter the PSA system. The door was closed to the pay‐​as‐​you‐ go system because it was unsustainable. This requirement assured the complete end of the old system once the last worker who remained in it reaches retirement age (from then on, and during a limited period of time, the government has only to pay pensions to retirees of the old system).

After several months of national debate on the proposed reforms, and a communication and education effort to explain the reform to the people, the pension reform law was approved on November 4, 1980.

Together with the creation of the new system, all gross wages were redefined to include most of the employer’s contribution to the old pension system. (The rest of the employer’s contribution was turned into a transitory tax on the use of labor to help the financing of the transition; once that tax was completely phased out, as established in the pension reform law, the cost to the employer of hiring workers decreased). The worker’s contribution was deducted from the increased gross wage. Because the total contribution was lower in the new system than in the old, net salaries for those who moved to the new system increased by around 5 percent.

In that way, we ended the illusion that both the employer and the worker contribute to social security, a device that allows political manipulation of those rates. From an economic standpoint, all the contributions are ultimately paid from the worker’s marginal productivity, because employers take into account all labor costs–whether termed salary or social security contributions–in making their hiring and pay decisions. By renaming the employer’s contribution, the system makes it evident that workers make all contributions. In this scenario, of course, the final wage level is determined by the interplay of market forces.

The financing of the transition is a complex technical issue and each country must address this problem according to its own circumstances. The implicit pay‐​as‐​you‐​go debt of the Chilean system in 1980 has been estimated by a recent World Bank study [9] at around 80 percent of GDP. (The value of that debt had been reduced by a reform of the old system in 1978, especially by the rationalization of indexing, the elimination of special regimes, and the raising of the retirement age.) That study stated that “Chile shows that a country with a reasonably competitive banking system, a well‐​functioning debt market, and a fair degree of macroeconomic stability can finance large transition deficits without large interest rate repercussions.”

Chile used five methods to finance the transition to a PSA system:

1. Since the contribution needed in a capitalization system to finance adequate pension levels is generally lower than the current payroll taxes, a fraction of the difference between them was used as a temporary transition payroll tax without reducing net wages or increasing the cost of labor to the employer (the gradual elimination of that tax was considered in the original law and, in fact, that happened, so that today it does not exist).

2. Using debt, the transition cost was shared by future generations. In Chile roughly 40 percent of the cost has been financed issuing government bonds at market rates of interest. These bonds have been bought mainly by the AFPs as part of their investment portfolios and that “bridge debt” should be completely redeemed when the pensioners of the old system are no longer with us.

3. The need to finance the transition was a powerful incentive to reduce wasteful government spending. For years, the budget director has been able to use this argument to kill unjustified new spending or to reduce wasteful government programs, thereby making a crucial contribution to the increase in the national savings rate.

4. The increased economic growth that the PSA system promoted substantially increased tax revenues. Only 15 years after the pension reform, Chile is running fiscal budget surpluses of around 2 percent of GNP.

5. In a theoretical state’s balance sheet (where each government should show its assets and liabilities), state pension obligations may be offset to some extent by the value of state‐​owned enterprises and other types of assets. Privatizations in Chile were not only one way to contribute, although marginally, to finance the transition, but had several additional benefits such as increasing efficiency, spreading ownership, and depoliticizing the economy.

The Results

The PSAs have already accumulated an investment fund of $30 billion, an unusually large pool of internally generated capital for a developing country of 14 million people and a GDP of $70 billion.

This long‐​term investment capital has not only helped fund economic growth but has spurred the development of efficient financial markets and institutions. The decision to create the PSA system first, and then privatize the large state‐​owned companies second, resulted in a “virtuous sequence.” It gave workers the possibility of benefiting handsomely from the enormous increase in productivity of the privatized companies by allowing workers, through higher stock prices that increased the yield of their PSAs, to capture a large share of the wealth created by the privatization process.

One of the key results of the new system has been to increase the productivity of capital and thus the rate of economic growth in the Chilean economy. The PSA system has made the capital market more efficient and influenced its growth over the last 15 years. The vast resources administered by the AFPs have encouraged the creation of new kinds of financial instruments while enhancing others already in existence, but not fully developed. Another of Chile’s pension reform contributions to the sound operation and transparency of the capital market has been the creation of a domestic risk‐​rating industry and the improvement of corporate governance. (The AFPs appoint outside directors in the companies in which they own shares, thus shattering complacency at board meetings.)

Since the system began to operate on May 1, 1981, the average real return on investment has been 12 percent per year (more than three times higher than the anticipated yield of 4 percent). Of course, the annual yield has shown the oscillations that are intrinsic to the free market–ranging from minus 3 percent to plus 30 percent in real terms–but the important yield is the average one over the long term.

Pensions under the new system have been significantly higher than under the old, state‐​administered system, which required a total payroll tax of around 25 percent. According to a recent study, the average AFP retiree is receiving a pension equal to 78 percent of his mean annual income over the previous 10 years of his working life. As mentioned, upon retirement workers may withdraw in a lump sum their “excess savings” (above the 70 percent of salary threshold). If that money were included in calculating the value of the pension, the total value would come close to 84 percent of working income. Recipients of disability pensions also receive, on average, 70 percent of their working income.

The new pension system, therefore, has made a significant contribution to the reduction of poverty by increasing the size and certainty of old‐​age, survivors, and disability pensions, and by the indirect but very powerful effect of promoting economic growth and employment.

When the PSA was inaugurated in Chile in 1981, workers were given the choice of entering the new system or remaining in the old one. One fourth of the eligible workforce chose the new system by joining in the first month of operation alone. Today, more than 93 percent of Chilean workers are in the new system.

As the state pension system disappears, politicians will no longer decide whether pension checks need to be increased and in what amount or for which groups. Thus, pensions are no longer a key source of political conflict. A person’s retirement income will depend on his own work and on the success of the economy, not on the government or on the pressures brought by special interest groups.

For Chileans, pension savings accounts now represent real and visible property rights–they are the primary sources of security for retirement. After 16 years of operation of the new system, in fact, the typical Chilean worker’s main asset is not his used car or even his small house (probably still mortgaged), but the capital in his PSA.

Finally, the private pension system has had a very important political and cultural consequence. Indeed, the new pension system gives Chileans a personal stake in the economy. A typical Chilean worker is not indifferent to the behavior of the stock market or interest rates. Intuitively he knows that his old age security depends on the wellbeing of the companies that represent the backbone of the economy.

The Global Pension Crisis

The real specter haunting the world these days is the specter of bankrupt state‐​run pension systems. The pay‐​as‐​you‐​go pension system that has reigned supreme through most of this century has a fundamental flaw, one rooted in a false conception of how human beings behave: it destroys, at the individual level, the essential link between effort and reward–in other words, between personal responsibilities and personal rights. Whenever that happens on a massive scale and for a long period of time, the result is disaster.

Two exogenous factors aggravate the results of that flaw: (1) the global demographic trend toward decreasing fertility rates; and, (2) medical advances that are lengthening life. As a result, fewer and fewer workers are supporting more and more retirees. Since the raising of both the retirement age and payroll taxes has an upper limit, sooner or later the system has to reduce the promised benefits, a telltale sign of a bankrupt system.

Whether this reduction of benefits is done through inflation, as in most developing countries, or through legislation, the final result for the retired worker is the same: anguish in old age created, paradoxically, by the inherent insecurity of the “social security” system.

The success of the Chilean private pension system has led three other South American countries to follow suit. In recent years, Argentina (1994), Peru (1993), and Colombia (1994) undertook a similar reform. Mexico, Bolivia, and El Salvador have already approved laws that will reform their state‐​run pension systems following the Chilean model. The new system in those countries will begin operation in the next few months (for example, July 1 in Mexico).

The Chilean experience can also be instructive to countries around the world, including the United States. In a report to Congress released earlier this year, five members of the Advisory Council on Social Security, led by Ms. Carolyn Weaver, an early advocate of this line of reform, recommended the transformation of the current pay‐​as‐​you‐​go system into a two‐​tier system that would include privately‐​held individual accounts. Already, economists have begun to write alternative plans to finance the transition in the United States. [10] (Without endorsing any specific proposal, the Cato Project on Social Security Privatization will present a number of possible transition scenarios.)

As an indication of the power of ideas, even officials from the People’s Republic of China have come to Chile to study the private pension system. Indeed, I have just returned from a conference in Shanghai, where I met with top government officials who demonstrated a clear interest in Chilean‐​style pension reform. It is possible that before entering the new millennium, several other countries, including all those in the Americas, will have privatized their pension system along the Chilean model. This would mean a massive redistribution of power from the state to individuals, thus enhancing personal freedom, promoting faster economic growth, and alleviating poverty, especially in old age.

Mr. Chairman, let me conclude with a warning about the damaging moral effects of social security and other social insurance programs issued at the dawn of the New Deal:

The lessons of history, confirmed by evidence immediately before me, show conclusively that continued dependence on relief induces a spiritual and moral disintegration fundamentally destructive to the national fiber. To dole out relief in this way is to administer a narcotic, a subtle destroyer of the human spirit. It is inimical to the dictates of sound policy. It is a violation of the traditions of America. [11]

That warning was issued by President Franklin Delano Roosevelt in his 1935 State of the Union address.

Thank you very much.


[1] For an overview of the social security crisis, see Peter G. Peterson, Will America Grow Up Before It Grows Old?, Random House, 1996.

[2] James M. Buchanan, “Social Insurance in a Growing Economy: A Proposal for Radical Reform.” National Tax Journal, Vol. 21 (December 1968): 386–95.

[3] Krzysztof Ostaszewski, “Privatizing the Social Security Trust Fund? Don’t Let the Government Invest.” Cato Institute Social Security Paper No. 6 (January 14, 1997).

[4] Martin Feldstein, “The Missing Piece in the Policy Analysis: Social Security Reform.” American Economic Review, Vol. 86, No. 2 (May 1996).

[5] Michael Tanner, “Privatizing Social Security: A Big Boost for the Poor.” Cato Institute Social Security Paper No. 4 (July 26, 1996).

[6] Laurence Kotlikoff et al., “Privitazing U.S. Social Security‐​A Simulation Study.” Paper presented at the World Bank Conference on “Pension Systems: From Crisis to Reform,” November 1996.

[7] This section follows José Piñera, “Empowering Workers: The Privatization of Social Security in Chile.” Cato’s Letter No. 10, Cato Institute (1996).

[8] Sebastian Edwards, “The Chilean Pension Reform: A Pioneering Program.” Paper presented at the NBER Conference, “Privatizing Social Security,” August 1–2, 1996.

[9] World Bank, Averting the Old Age Crisis (1994).

[10] See, for instance, Peter Ferrara, “A Plan for Privatizing Social Security.” Cato Institute Social Security Paper No. 8 (April 30, 1997).

[11] Franklin D. Roosevelt, The Public Papers and Addresses of Franklin D. Roosevelt, Vol. 4, The Court Disapproves: 1935, Random House, 1938.

About the Author