The Becker-Posner Blog

Response to Comments on Social Security Privatization-Becker

Plenty of comments and some quite good. Too many to answer in detail, so I pick out a few highlights. I made political aspects of government behavior with regard to pay-as-you-go social security the cornerstone of my argument for privatization. I emphasized the substantial lowering of retirement ages in the U.S. and elsewhere-this is a fact not disputed by anyone who has studied the experiences of Europe, and North and South America-- the political determination of benefits that has little to do with a sensible welfare program for the elderly, and not the least, that governments tend to spend the �surplus� on social security accounts (and other tax revenues) rather than saving them for future generations. Some comments suggest it would be easier to keep the present system and just get rid of these political defects. If only that were possible, governments could operate most industries very well! But I argued these �merely� political defects are intrinsic to a government operated social security system since they occur in pretty much every single pay-as-you-go system, including the European and South American ones. The case for privatization of different industries throughout the world has always ultimately rested on similar �political� defects in government-run enterprises. As an example of government behavior, publicly owned banks in China continue to make capital available to government enterprises, even though these enterprises for the most part have no chance of ever repaying these �loans�. Some government involvement in a private system is necessary, especially if, as in my system, there is a guarantee to all retirees of a minimum standard of living. I suggested the use of index funds for private retirement social security accounts as a way to keep government involvement in the new system to a minimum. The index fund industry is easy to enter, and so is very competitive. Since it involves minimal buying and selling of stocks and bonds, and no research on picking �winners�, this industry has very low administrative charges. So private accounts under this system would most definitely not have high administrative management costs. Moreover, these funds still provide individuals with choice since they can and do vary the fraction of their assets allocated to bonds vs stocks. Those savers who are very risk averse could choose funds with high bond-stock ratios. Moreover, one should not oversell the low cost of the present social security system. The government managed social security system until recently gave individuals little choice and information about their accounts, and still gives little updating of payments and expected benefits. The cost of promoting and selling cars would be greatly reduced too if they were all produced by the government and were all the same kind. That was the Soviet method, not one to be envied. Private funds would be more costly to operate, but they would give more choice and information as they compete for customers. Index funds, in particular, would still have very low administrative expenses, as exemplified by Vanguard. It is inconsistent to complain about the low savings in U.S. and object to a system that is likely to increase savings, much of which would be invested in equities. Higher savings would reduce dependence on foreign investments in the country, and would make more domestic capital available to be invested. Is that bad? Withdrawals upon retirement or reaching a fixed age would be either in a lump sum, or spread out on an actuarially-determined basis. Individuals who are �worried� (not a bad worry!) about living longer than average could save part of their withdrawals. To be sure, some persons might underestimate the amounts needed because they live longer than average, and some of them would fall below the minimum income guaranteed level. If that is considered a problem, the actuarial calculations could err on the conservative side. Some of you were dubious about my argument that governments tend to spend any surplus. But that is not theory alone, it is also fact, as shown in a Journal of Law and Economics Article of October 2003 by my colleague Casey Mulligan and myself. We look at several types of evidence from different countries, and from US history. Not included in our discussion, but valuable additional evidence comes from the 1990�s. Federal spending was first kept down by deficits, but then both Federal and state government spending grew rapidly with the sharp increase in tax revenues during the boom years at the end of the century. This pattern indicates that the large current federal deficits will force a slowdown of government spending growth during the next several years. Privatization would significantly increase the pressure on spending by removing some personal retirement savings from government revenues. I do believe that one of the most important obligations of a decent government is to provide a safety net for individuals and families who, for one reason or other, are in bad economic circumstances. My proposal does take care of the elderly who fall into this category since I support a minimum income floor for all retirees. Contrary to comments that complain my proposal would not have a �progressive� structure, this floor does give the payout system a sizeable progressive element. I conclude by emphasizing that privatizing any government activity, including social security programs as well as many others, does presume some confidence in the ability of the great majority of individuals and families to look out for their own interests, however they define them, in a market-oriented competitive environment. To be sure, people make mistakes, but the crucial issue is whether they can usually promote their own interests better than government officials and bureaucrats can do it for them. I have great confidence in the ability of even the poor and less educated to generally look out for themselves, however limited are the opportunities available to them. Judging from some but far from all the comments, many of you do not have that confidence in people.

Why I Support a Privatized Individual Account Social Security System-BECKER

Today both Republicans and Democrats are passionately arguing about the future of social security. Although there is merit in each side�s argument, neither side is portraying the situation accurately. In my view, movement toward a privatized individual account social security system offers the best option, where individuals save and accumulate assets to provide for their retirement. It is true, as the critics correctly observe, that there is no magical gain in privatizing since all systems have to provide incomes for retired persons. But there is also no magical gain in privatizing a government steel plant since steel still has to be produced, yet there are good reasons to privatize steel. I also believe that there are excellent reasons to aim for a privatized individual account social security system. Pay as you go social security started first in Europe as a relatively easy way to provide a minimum standard of living for the elderly. It was introduced in the United States during the 1930�s partly also to discourage the elderly from competing for jobs when unemployment of younger workers was staggeringly high. It was a cheap system then because there were more than 10 workers per retired person, so the social security tax could be small relative to the benefits received by retirees. Indeed, the first several generations of retirees earned very high returns in retirement income on their accumulated social security tax payments. But as birth rates fell drastically, and the life expectancy at age 60 expanded enormously, fewer workers are now being forced to support more and more retirees. The result is a huge rise in social security taxes in every nation with a pay as you go system. The combined tax on employees and employers in the United States, excluding contributions to medicare, is now 12.4 per cent and rising, and that percentage is much higher in Japan and most Western European nations. The expectation of continuing growth in this tax rate explains why countries as different as Sweden and Great Britain have partially moved toward a privatized individual account system. It also helps understand why Hong Kong, Poland, and other countries with low birth rates that recently introduced social security have important components of individual accounts in their systems. I do not believe that the main advantage of a private account system is that individuals can get a higher return on their old age savings by investing in stocks. There are no free lunches from such investments since the higher return on stocks is related to their greater risk and other trade offs between stocks and different assets. However, neither is there any special �transition� problem in moving to a fully funded privatized system since future generations in some way or another have to pay for the implicit debt due to commitments toward present and future retirees. But it is better to transit smoothly to fund this debt rather than require a sharp increases in taxes on later generations. Retirees for whom social security income is not a major part of their retired assets will invest much of their own savings in stocks. Studies indicate that this is precisely what they generally do with their IRA�s in order to have a balanced portfolio between stocks and the implicit social security assets guaranteed them. Since lower income men and women accumulate few assets other than their social security assets, a fully funded system through personal savings would enable them to have more balanced portfolios between stocks and bonds. If there is no obvious gain from allowing most individuals to invest in stocks to help cover their retirement, and if there is no fundamental transition problem, what, if any, are the advantages of a funded privatized system? I believe the advantages are mainly political, not �economic�, that privatization helps to separate saving for retirement from interest group politics, taxation, and government spending. Pay as you go systems are in trouble throughout the world in good part because of changes in the number of workers per retiree, but also because of politically determined decisions that changed the system from saving for old age to an inefficient and complicated welfare system for some of the elderly. For example, despite the growing mental and physical health of older persons, political pressures in all nations with pay as you go systems forced a restructuring of social security payouts to encourage retirements at earlier ages than even the originally established age 65. In the United States many retirements occur age 62 or earlier, while Italians retire frequently while in their mid fifties, and very early retirement is not uncommon also in Germany, Belgium, and many other European countries. In addition, the link between contributions and benefits has been separated, so that each additional dollar contributed in taxes pays on the average no more than about 40 cents in additional benefits. Hence, the social security system has evolved into two largely independent systems: a sizeable tax on wages, starting with the first dollar earned, and retirement benefits that are �guaranteed� by the government. There is only a modest link from an individual�s accumulated tax payments on his earnings to these �guarantees�. Just as important are the political implications of Federal fiscal behavior. Tax revenue from social security taxes at present exceed payments to retirees. This excess is counted as part of the growing Social Security Trust Fund, but in fact also enters into the consolidated Federal budget account, and helps reduce the reported spending deficit. Reported deficits during the past decade would have been much larger if social security was not running a surplus during this whole time period. Social security tax revenues are expected to fall below spending on retirees in about 20 years. If we simply raised social security taxes now-say by two percentage points- consolidated federal deficits would appear much smaller, and the federal government would be under less constraint to reduce spending. Both theory and evidence indicates that a good fraction of the additional revenue would indeed be spent. �Putting aside� assets for the future is very difficult for all governments, subject as they are to immense demands for spending now from various interest groups. A good individual funded savings system would require people to save 4-6 per cent of income (President Bush suggests 4 per cent), and invest these savings in private individual accounts that would meet certain government established criterion. At the same time, social security taxes should be cut by a couple of percentage points from its present level to ease the burden on workers. These taxes could be cut since younger workers would then be contributing to their own retirement. Moreover, a tax cut would reduce the social security surplus, so the government would be less tempted by rapidly growing social security �reserves�. These private accounts would accumulate tax-free until individuals decide to retire. The age of retirement within broad limits would be left to individual choices, but funds would continue to grow with savings for persons who retire at later ages because they like their work and are in good health-this basically is how IRAs work. At retirement, individuals would get access to their assets, either in a lump sum or as an annualized income, and they would then pay taxes on their withdrawals. As in Chile and other countries with private retirement accounts, the government would guarantee every retiree a minimum income-similar to but larger than the minimum social security guaranteed income under the present United States system. Unfortunately, such guarantees create a �moral hazard�; that is, savers may want to make risky investments that give high payoffs if they succeed since the government partly bails them out if they fail. Or they may not save at all. The minimum required savings rate overcomes the latter incentive to �game� the system, and regulation of which types of investment accounts are approved takes care of the incentive to be overly risk-taking. We should follow the President�s proposal, and only allow index funds for social security accounts- that is, funds that do not try to beat the market and invest in a balanced market portfolio of stocks and bonds. Individuals who are contributing more than the 4 per cent minimum could take greater risks if they want to since they do not pose any moral hazard from bad investments on these larger accounts. Index funds both reduce the overall risk of an account, and have very low management fees since it is quite cheap to run these funds, as shown, among others, by Vanguard and Barclays. High management fees is a common complaint about the Chilean system, although this system has yielded high returns to investors even net of these fees, and the fees have come down a lot in recent years. There is no guarantee that government interference would not increase further in such a privatized system since the retired would continue to press for additional benefits. But experience shows that governments interfere less when an industry is privatized than when it is a public enterprise, especially in access to capital and financing of industry budget deficits. So the really strong arguments for privatization are that they reduce the role of government in determining retirement ages and incomes, and improve government accounting of revenues and spending obligations. All the other issues are really diversions because neither advocates or opponents of privatization are asking the most meaningful question about privatizing social security: Is there as strong a political economy case for eliminating government management of the retirement industry as there is for eliminating their management of most other industries? My answer is �yes�.

Social Security Reform--Posner

One of the commonest objections to President Bush�s proposal for reform of social security is that there is no need to act now because there is no �crisis.� Yet the same people who say this are wont to say that a weakness of government is its failure to take the long view. Politicians have a short time horizon because they have limited terms of office and are therefore reluctant to address a problem that lies in the future even if the problem could be solved more easily before it assumes crisis proportions. The increasing burden of social security, owing to the declining ratio of workers to retirees, is a problem easier to solve now than when a continued decline in that ratio (it was 3.3 in 1996 and is expected to be only 2 by 2030) brings on a fiscal crisis necessitating either steep tax increases or drastic benefits cuts (one form of which would be raising the retirement age), or both. In 1983 Congress raised the retirement age for full social security benefits from 65 to 67, thus reducing liftetime benefits. But the legislation provided for phasing in the increase over a 22�year period beginning in 2003, which made the discounted present cost of the reduction in benefits negligible for prospective retirees. With the social security �crisis� still in the future, it is possible to begin now to change the system in the direction of a genuine retirement program, that is, one in which people pay for their own retirement rather than having it paid for by current workers. With that shift, it would no longer matter what the ratio of workers to retirees was. The shift is independently desirable. Having people pay through the tax system for other people�s retirement produces a capricious redistribution of income. The elderly have disproportionate voting power: a higher percentage of elderly than of other age groups vote, and they tend to focus laser-like on issues that affect their pocketbook. The shift can be made painless to present and imminent retirees (anyone over 55, say) by keeping benefits at their current level. However, if younger workers are required to place a percentage of their income in a retirement fund, those who do not want to reduce their disposable income by that amount will perceive the requirement as an increase in their taxes, and will squawk. If to quiet them taxes are reduced, the federal deficit will grow because benefits are not being cut. Eventually, fiscal equilibrium will be restored, but until then, the political price paid for the reform may be high. Deficit spending, financed by government borrowing, may be economically superior to higher taxes or inflation, but to the extent that the lenders are foreign, the payment of interest on the loans will transfer wealth from Americans to foreigners. The issue of compulsory pensions should be decoupled from that of welfare. Some people cannot save enough to be able to retire on; the government should guarantee them a decent retirement income. Those are the only people whose retirement should be financed by other people, that is, by taxpayers. All others should be required to finance their own retirement. This applies to health costs as well as to other expenses of retirement. Just as social security should be for people who can�t save enough to retire on, so medicare should be for people who can�t save enough to buy health insurance for their old age. It is ridiculous that people with Becker�s income or my income should be receiving social security and medicare benefits, except insofar as those benefits have actually been financed by our payments of social security taxes rather than by taxes on workers. Extreme libertarians will challenge the assumption that people should be compelled to save for their retirement. They will say that people should be allowed to allocate their income over their life span as they choose; if they choose to allocate nothing to their old age, then let them starve. I do not agree with this position, and for two reasons. First, I believe it is plausible to model the individual as a succession of selves with different preferences. A young person may dislike the idea of growing old and may be inclined therefore to refuse to make provision for his old self. The old self�the self that will not emerge and �take over� the individual for many years�has no control over the decisions of the young self. Compulsory retirement saving gives the old self a �voice� in the decisions of the young self. Second�and a point actually consistent with libertarian thinking�these improvident oldsters will in fact free ride on their children and grandchildren. This, however, suggests an argument (made in my book Aging and Old Age [1995]) for the existing pay-as-you-go system, as distinct from one in which people finance their own retirement. If by lifting the burden of elder care (or much of it, at least) from the shoulders of one�s children, social security confers a benefit on the children, they should be expected to pay for it. A distinct but related point is made by Becker in his book A Treatise on the Family (enlarged ed. 1991). He points out that parents� payment of school taxes finances their children�s education (and hence earning power), and social security can be regarded as repayment to the parents in their old age. The larger the family, the greater the aggregate benefits to the children of a �free� public education, but also the greater social security taxes they pay (because there are more of them). The benefit to the children of shifting the elder-care responsibility is probably independent of family size, since parents� need for such care is independent of the number of children they have. But the matching of benefits and costs under the pay-as-you-go system is both very crude and highly politicized. Compulsory retirement saving would protect children from parental demands for old-age care; and education, like retirement, could be made self-financing, though with appropriate subsidies for children whose parents cannot afford to educate them.

The Summers Controversy--Posner's Response to Comments

As always, there were many illuminating comments. I shall try to respond to a few, but I want first to direct readers to the article in this morning's New York Times about the controversy. In it we learn that Summers has apologized not once, but "repeatedly," for having raised questions about women's innate scientific abilities, and that he has now appointed two task forces (comprising a total of 22 women and 5 men) to present proposals for increasing the number of women on the Harvard faculty. He has, in short, capitulated (and rather abjectly, it seems to me--why two task forces?). The response by his critics has been ungracious. They remain guarded and suspicious--he is on probation--and they will no doubt press him for further measures. He is now effectively committed to affirmative action for women scientists. Yet of all the nation's problems, and all the claims for affirmative action, the underrepresentation of women on the science faculties at Harvard must be among the least important. Some of the comments suggest that the real significance of Summers's January 14 remarks was to demonstrate that universities are no longer citadels of free speech, though that cannot have been his intention. One comment compares his apology to the confessions of Stalin's purge victims: "Everyone should oppose a 'signal of discouragement to talented girls and women.' [That is a quotation from Summers's first apology.] But the truth is that such a signal, to the extent it occurred, resulted from deliberate, intense, and misleading responses to his remarks. That's classic totalitarian suppression of an unpopular view, with forced public acknowledgment of guilt and forced repudiation of the 'wrong.'" Another comment quoted: George Wills: "Forgive Larry Summers. He did not know where he was...He thought he was speaking in a place that encourages uncircumscribed intellectual explorations. He was not. He was on a university campus." But no one who has spent much time around universities thinks they've ever "encourage[d] uncircumscribed intellectual explorations." The degree of self-censorship in universities, as in all institutions, is considerable. Today in the United States, most of the leading research universities are dominated by persons well to the left of Larry Summers, and they don't take kindly to having their ideology challenged, as Summers has now learned to his grief. There is nothing to be done about this, and thoughtful conservatives should actually be pleased. As John Stuart Mill pointed out in On Liberty, when one's ideas are not challenged, one's ability to defend them weakens. Not being pressed to come up with arguments or evidence to support them, one forgets the arguments and fails to obtain the evidence. One's position becomes increasingly flaccid, producing the paradox of thought that is at once rigid and flabby. And thus the academic left today. One comment makes the interesting point that Summers's post-apology position implies indefinitely deferring any inquiry into possible innate differences in science abilities. The reason is the insistence that before those differences should be studied, every other possible cause--social, economic, psychological, political--for the differences in male and female career choices be studied and eliminated. But many will not be eliminated, not soon, at any rate; some, indeed, may be secondary to innate differences. A serious scholar does not order his or her research priorities by their political inoffensiveness. Such an ordering reflects a methodological fallacy as well as an ideological commitment. It is the fallacy of mechanical extrapolation. In the past, differences between races, the sexes, groups defined by ethnicity, nationality, or even sexual orientation, were exaggerated, and as the exaggerations are identified and unmasked, discrimination declines; it has declined greatly. And now we have an animal rights movement that claims (with some justice, I might add) that differences in cognitive capacity between human beings and other animals have been exaggerated; and there are even proposals afoot to give some of our fellow primates human rights. So it is natural, but it is also fallacious, to assume that eventually all differences (well, almost all), certainly including different career patterns, that are correlated with immutable (or near-immutable) characteristics, such as sex, will disappear; and so let us jump on the bandwagon of history and speed the process. But not only are the costs of trying to speed up history ignored; there is no reason to expect the trend toward less and less differentiation of careers by race, sex, etc. to continue. It could stop at any time. In 1950, looking back at the enormous strides made by labor unions to organize the American workforce, one might have expected that by now 100 percent of the workforce would be unionized. In fact the percentage of unionized workers has fallen precipitately and is now about where it was in 1900. One further note: several comments point out that genetic factors can influence preferences as well as capabilities. Even if women are just as capable of doing science at the Harvard faculty level as men, it is possible that fewer women than men find a career in science congenial. This might or might not reflect innate differences. One notes the enormously greater percentage of men than of women who are in prison. Could not this "discrimination" reflect innate differences in attitudes toward violence, risk, and transgression? And a final, slightly technical note: yes, another way to express increased variance and therefore longer tails in one distribution than in another is to say that the standard deviation is greater.