Mr. Chairman and Members of the Subcommittee:

I want to thank the Subcommittee for inviting me to testify on the experiences of other countries who have privatized their Social Security systems. At this point there are a large number of countries, mostly in the developing world, who have partially or completely privatized their Social Security systems. There are four basic generalizations that can be made based on the experiences of these countries:

1) Privatization invariably increases risks for workers. These risks take three forms: the timing of the worker’s retirement, the risk associated with the worker’s choice of assets, and the risk of having a low income during a working lifetime. The latter point refers to the fact that most traditional Social Security systems, including the system in the United States, are designed to be redistributive to low-wage earners. A system based strictly on individual accounts is not redistributive, although redistributive features can be added to the system.

2) Privatized systems vastly increase the administrative costs of operating a Social Security system. The most efficient privatized systems have annual administrative costs for the retirement program that are 30 to 40 times as high as the current system in the United States. When the costs of annuities are included, financial intermediaries can take as much as 30 cents of every dollar placed in the system.

3) Privatized systems have not proved very popular with the workers they are supposed to benefit. One of the main reasons for introducing defined contribution systems in developing countries is that a privatized system was supposed to extend coverage to the large segment of the workforce employed in the informal sector, most of whom were not covered by the traditional Social Security system. There has been little change in participation rates in the countries with privatized Social Security system. In most cases, the vast majority of workers have voted with their feet against the privatized systems by opting not to participate.

4) Financial markets view borrowing to cover transition costs as real borrowing. The Bush administration has argued that the transition costs associated with switching from the current Social Security system to a system of private accounts should not be viewed as real debt, since the borrowing would be associated with lower Social Security benefits in the future. The evidence from other countries is that financial markets focus on current balance sheets, not speculation about benefit levels in the distant future. Every country that has privatized their Social Security system has attempted to at least partially fund the transition with a combination of tax increases and spending cuts. Argentina, which defaulted on its debt in 2001, is the most prominent example of a country that failed to take adequate steps to offset the cost of the cost of its transition.

I would also add the additional observation that private accounts do not by themselves increase national wealth. This is important in the context of promoting private accounts as a way to increase returns to retirees. Since private accounts do not actually increase wealth, at best they can be a mechanism for redistributing money from the working population to retirees. If Congress intends to redistribute money from workers to retirees, then there are arguably more efficient mechanisms to accomplish this goal.

I will elaborate on each of these points in turn.


Countries that have privatized their Social Security systems have subjected retirees to all three forms of risk noted above, the risk of market timing, the risk associated with asset choice, and the risk of low income during a working lifetime. In the first case, it is a basic fact about financial markets that they are volatile. Even if the average return on equities exceeds the return on riskless assets, there is considerable variation in this return. In Chile, the longest standing experiment with a privatized system, a government minister recommended that workers delay their retirement for a few years after a downturn in the national stock market.

There is no way to avoid this market timing risk. The Bush administration’s suggestion that workers be forced to switch out of stocks approximately 10 years before retirement does little to change the story. If the market plunges just before a worker reaches this switch date, then he or she is almost as bad off as if the plunge occurred just before his or her retirement.

A second sort of risk is associated with the choice of asset. This risk can result from a worker either being too conservative or taking too much risk with their accounts. Many workers are ill-informed about financial markets and may only feel comfortable holding very safe assets with low returns. In the case of the proposal President Bush outlined in his State of the Union Address, this could result in workers losing money on their individual accounts, since they would lose more from their Social Security benefit than they would gain from the investments in their accounts.

Workers can also engage in speculative investments that end up losing money. This happened to some extent in England where there was a “mis-selling scandal” in the mid-nineties. Many financial firms had sold accounts to workers by promising returns on the accounts that workers would not actually realize. The British government eventually forced these firms to make good on these promises. In some cases, where firms had gone out of business, the government was forced to pick up itself the cost of fulfilling these promises.

These risks can be minimized by restricting choice. If the government gives workers a very narrow range of options, then the risk of bad asset selection is reduced. (Of course, this assumes that the government knows better than individual workers how best to invest their money.)

In principle, insofar as workers are too conservative with their investment choices because they are ill-informed about financial markets, the problem can be addressed with better education. However, this raises two additional problems. First economists usually believe that time has an opportunity cost. Time that workers spend learning about financial markets is time that they could have spent with their children or on other activities. If we are designing a system that requires that tens of millions of workers get additional education on financial markets, then we have decided that this is the best use of their time. (In Chile, the schools now have sessions that teach people about the retirement system. This means that time that could have been spent developing math, science, or language skills is instead being used to teach people how to manage their Social Security accounts.)

The other problem with trying to educate workers on their retirement investments is that it is not clear who should be doing the educating. Many of the country’s top financial advisors were recommending that people invest in stock even at the peak of the nineties bubble. It is not clear that advice from such experts would be beneficial to most workers.

Finally, a system of private accounts, by itself, is not redistributive to low wage workers. This means that if a mechanism is not put in place to ensure that workers who put little into these accounts because of low earnings, still have an adequate retirement income, then many low wage earners could end up as losers. While most countries with privatized systems have put some sort of minimum benefit in place, this is not universally the case. For example, Peru does not have a minimum benefit in its system.

It is also important to realize that putting a redistributive mechanism in place today, does not guarantee that it will be there twenty or thirty years in the future. Any redistributive mechanism attached to private accounts will always be subject to political risk. The intention of the designers of the system will matter little if political support does not exist to retain redistributive mechanisms in the future.


There is now a large body of research that shows that the administrative costs of a privatized system of individual accounts vastly exceeds the costs of a centralized defined benefit system like the one in the United States. The administrative costs of the Chilean system have averaged close to 15 percent of the money placed in the accounts each year, while the cost of the British system have averaged close to 20 percent. Administrative costs are much greater in these systems because of the costs associated with servicing an individual account, the costs associated with marketing to individuals, and the profits of the firms who administer these accounts.

In addition to the annual costs associated with operating these accounts, there are also costs associated with turning the accounts into annuities at retirement (which is not generally required). Research indicates that insurance companies charge between 10 and 20 percent of the value of a sum to convert it to an annuity. Roughly half of this fee is associated with the adverse selection that results when annuitization is not mandatory. (Only relatively long-lived individuals are likely to buy annuities.) The other half is due to the administrative costs and profits of the financial firms that issue annuities.

A single centralized system of accounts (which does not exist in any of the countries that have opted for privatization) could in principle lower costs, especially if it minimized workers’ choices in selecting investments and switching between investments. President Bush’s Social Security commission estimated that a bare-bones centralized system would cost roughly ten times as much as the current system. (There has been considerable confusion about this point because of how the commission framed its cost estimate. The commission estimated that the administrative cost would be 0.3 percent of the stock of money in an account. This means that the fee on a dollar placed in an account would be 0.3 percent for each year that dollar is in the account. Some dollars will be in an account for forty years, while some dollars placed, in the account just before a worker retirees, will be there for just a short time. If a dollar is a worker’s account for an average of twenty years, then this 0.3 percent fee will be paid twenty times, making a total administrative cost of 6.0 percent, compared to a cost of just 0.5 percent on the dollar placed in the Social Security system.)

President Bush’s commission also argued that a centralized government run system can radically reduce the cost of issuing annuities. While a centralized system may in principle be vastly more efficient than the current market system, there would still be a problem of adverse selection in any system where buying annuities is optional, as President Bush has proposed. This means that a worker with an average life-span could expect to lose between 5 and 10 percent of their money under such a system, compared to what they would receive with an actuarially fair annuity.


The World Bank recently completed a study of the privatized Social Security systems in Latin America.1 One of the main criticisms of these systems is that they have not substantially increased participation over the rates achieved under the traditional defined benefit systems. The argument that these systems would increase participation claimed that workers view their current Social Security contribution as a tax, whereas they would see their contribution to a private account in a different light. The fact that participation has changed little after privatization, in some cases not even growing more rapidly than what would have been expected if past trends had continued, indicates that workers do not view contributions to these accounts very differently than they do contributions to the traditional defined benefit system.

It is worth noting that in Chile, the most developed system, a large percentage of the workers target the minimum benefit. This minimum benefit allows any worker who has been in the system for twenty years to turn over their account to the government, and then get a guaranteed benefit that is tied to the value of the minimum wage. In effect, these workers are voting with their feet for a defined benefit system.

Transition Costs

In the short-term, the switch from a traditional pay-as-you-go Social Security system to a defined contribution system implies a large increase in the government deficit, since the same benefits must still be paid to current retirees, even though the government is collected much less in Social Security contributions. Every country that has opted to privatize its Social Security system has attempted to at least partially cover these transition costs by reducing its deficit, or building up a surplus, with some combination of tax increases and spending cuts. For example, the Chilean government increased the size of its annual surplus to 4 percent of GDP (the equivalent of a surplus of $500 billion in the United States in 2005) at the point where it implemented its privatization plan.

They felt the need to cut their deficits or increase their surpluses precisely because these governments did not believe that the financial markets viewed their implicit commitments to pay Social Security benefits in the distant future as being the same as actual government debt. The one important example of a government that did not take sufficient steps to offset the borrowing needed to finance its Social Security privatization was Argentina. In 2001, it was paying real interest rates of more than 20 percent on its debt, because lenders did not have faith in the government’s ability to pay off its debt.

By contrast, in 1994, the year Argentina put its privatization plan in place, the country was generally regarded as one of the most creditworthy countries in the developing world. Had it not been for the privatization of its Social Security system, Argentina would have been running balanced budgets between 1994 and 2001.

The United States is approaching the question of Social Security privatization at a time when it faces much larger deficits than any of the other countries that have gone this route. The experience of Argentina suggests that it is likely to face a very high price in financial markets if it does not couple privatization with large tax increases and/or spending cuts.

National Wealth and Privatization

No economist believes that the United States would be increasing national wealth if it borrowed $200 billion a year and invested this money in the stock market. It is possible that this will reallocate income, as the government can benefit from the gap between the return on equities and the interest paid on government bonds, but this is not creating additional wealth for the country as a whole. Similarly, it cannot increase national wealth if it borrows $200 billion a year and hands $2,000 a year to 100 million families and tells them to invest it in the stock market. This would simply be changing the allocation of national income.

This is important to recognize because one of the goals often claimed by proponents of privatization is increasing the rate of return on Social Security contributions. Insofar as the money is simply borrowed, as President Bush has proposed, then any increase in the rate of return due to privatization is simply coming at the expense of the rest of the population. This could be seen fairly directly in the case of Chile where accounts earned double-digit real rates of returns through the eighties. The main asset of Chile’s private accounts in the eighties was Chilean government bonds, which paid double-digit real interest rates. In effect, Chile’s workers received high returns on their accounts because Chile’s taxpayers paid high interest rates on the money their government borrowed to finance the accounts. It may have been desirable to transfer money from Chile taxpayers to Chile’s retirees, but this could have been done without going the route of privatization.

In short, it is important that policy makers recognize the distinction between using private accounts as a way to redistribute income – which they may be to some extent – and a mechanism to increase national wealth, which they surely are not.

(There is a separate issue of whether private accounts in the United States will be able to earn the rate of return claimed by proponents of privatization. Stock returns come from either capital gains or dividend payouts. No analyst has yet passed the “No Economist Left Behind Test,” which asks for a set of dividend payouts and capital gains, consistent with the Social Security trustees profit growth projections, that add to the 6.5-7.0 percent returns assumed in analysis of Social Security privatization. Given current price to earnings ratios and low projected profit growth, there is no plausible set of dividend yields and capital gains that will produce 6.5-7.0 percent real stock returns.)

Mr. Chairman, this concludes my testimony and I would be happy to answer any questions from you or other Members of the Subcommittee.

1 Gill, I, T. Packard, and J. Yermo, 2005, Keeping the Promise of Social Security in Latin America, Stanford, CA: Stanford University Press.

Dean Baker is the co-director of the Center for Economic and Policy Research (CEPR). He is the author of Plunder and Blunder: The Rise and Fall of the Bubble Economy. He also has a blog on the American Prospect, "Beat the Press," where he discusses the media's coverage of economic issues.