Hostname: page-component-cd9895bd7-hc48f Total loading time: 0 Render date: 2024-12-26T04:37:41.740Z Has data issue: false hasContentIssue false
  • English
  • Français

Can Old-Age Social Insurance Be Justified?*

Published online by Cambridge University Press:  13 January 2009

Daniel Shapiro
Daniel Shapiro
Affiliation:
Philosophy, West Virginia University
Get access Rights & Permissions [Opens in a new window]

Extract

While in America most people think of “welfare” as means-tested programs such as Aid to Families with Dependent Children, in reality in the United States and other affluent democracies the heart of the welfare state is social insurance programs, such as health insurance, old-age or retirement pensions, and unemployment insurance. They are insurance programs in the sense that they protect against common risks of a loss of income if and/or when certain events come to pass (illness, old-age or retirement, unemployment); they are “social” because unlike market insurance they are not run on a sound actuarial basis, the premiums are not voluntarily incurred but compulsory, and there is very limited choice or flexibility concerning the type of policy one can purchase. Why have social insurance rather than market insurance? In this essay, I take up this question with regard to old-age or retirement pensions, which at present absorb around 9 percent of the gross domestic product (GDP) and 25 percent of government spending of the affluent industrial countries comprising the Organization for Economic Cooperation and Development (OECD). My aim is to show that old-age or retirement social insurance (henceforth “SI”) is worse in virtually every relevant normative respect than its alternative, some form of market or private pensions. By relevant normative respect, I mean those values or principles which are used by contemporary political philosophers in their discussions and justifications of welfare-state policies, and which are applicable to assessments of different systems of old-age or retirement pensions. (Although they are applicable, almost no contemporary political philosophers have in fact applied them—an amazing state of affairs which I hope to remedy here.)

Type
Research Article
Information
Social Philosophy and Policy , Volume 14 , Issue 2 , Summer 1997 , pp. 116 - 144
Copyright
Copyright © Social Philosophy and Policy Foundation 1997

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1 These figures for OECD countries refer to the years 1986 through 1991 and are taken from the World Bank Policy Research Report, Averting the Old Age Crisis: Policies to Protect the Old and Promote Growth (New York: Oxford University Press, 1994), pp. 358–60Google Scholar. Unless significant structural changes are made, expenditures on pensions as a percentage of GDP are expected to double from 1990 to 2050 in the OECD countries (ibid., p. 7). See also Davis, E. Philip, Pension Funds: Retirement-Income Security and Capital Markets (Oxford: Clarendon Press, 1995), p. 46.Google Scholar

2 One cannot find such a discussion by liberal defenders of the welfare state, such as John Rawls, Ronald Dworkin, Thomas Nagel, Will Kymlicka, or David A. J. Richards, or by communitarian advocates, such as Michael Walzer, Charles Taylor, or Daniel Bell. Robert Goodin and Brian Barry do provide a defense of social insurance and income-replacement programs, but they do not discuss old-age insurance in any detail, and their defense of these programs rests crucially on the claim that the market cannot provide certain kinds of insurance—a claim which is clearly false in the case of pensions. See Barry, Brian, “The Welfare State versus the Relief of Poverty,” Ethics, vol. 100, no. 3 (04 1990), pp. 503–29CrossRefGoogle Scholar; and Goodin, Robert, “Stabilizing Expectations: The Role of Earnings-Related Benefits in Social Welfare Policy,” Ethics, vol. 100, no. 3 (04 1990), pp. 530–53.CrossRefGoogle Scholar

3 A fully funded pension is usually, but not always, a private pension. A few countries do have government-managed, fully funded pensions, rather than PAYG systems. See note 25.

4 See World Bank, Averting the Old Age Crisis, pp. 110–12Google Scholar; and Weaver, Carolyn L., “Controlling the Risks Posed by Advance Funding—Prospects for Reform,” in Social Security's Looming Surpluses: Prospects and Implications, ed. Weaver, Carolyn L. (Washington, DC: American Enterprise Institute, 1990), pp. 167–84.Google Scholar

5 On the misleading idea that PAYG systems have genuine trust funds, see Ferrara, Peter J., Social Security: The Inherent Contradiction (Washington, DC: The Cato Institute, 1980), pp. 49–51.Google Scholar

6 In the OECD countries, the implicit public pension debt in 1990 ranged from 90 percent of the GDP (the U.S.) to almost 250 percent of the GDP (Italy). Sweden at that time was the country with the greatest degree of advance funding, with a “trust fund” of about 30 percent of the GDP. Japan's reserve was 18 percent of the GDP, and the U.S. reserve was around 5 percent (it has since gone up to almost 10 percent). On the implicit public pension debt, see World Bank, Averting the Old Age Crisis, pp. 139–40Google Scholar. On Sweden and Japan, see Munnell, Alicia H. and Ernsberger, C. Nicole, “Foreign Experience with Public Pension Surpluses and National Savings,” in Weaver, ed., Social Security's Looming Surpluses, pp. 85–118Google Scholar. I obtained the figures on the U.S. from Statistical Abstracts of the United States (Washington, DC: U.S. Government Printing Office, 1995), pp. 334, 379, 451Google Scholar. But see note 7 on why the U.S. trust fund is probably just an accounting device that does not reduce future SI liabilities.

7 To see why the “trust fund” in the U.S. may very well be an accounting device that does not reduce the implicit public pension debt, one needs to understand how advance funding can actually reduce that debt. This happens when payroll taxes are used to purchase outstanding government debt held by private investors, and then those investors substitute new private securities for government debt they relinquish. That action by private investors allows for increased capital formation and ultimately higher future incomes to meet the cost of retirement benefits in coming decades. Payroll taxes can be lower than otherwise in the meantime, because of substantial interest accruing to the trust funds. And income taxes that finance the payment of the interest need not be any higher, because there is no change in the government's total indebtedness—only a change in the ownership of the debt from the public to the trust funds (in other words, money that would have gone to private investors goes to the trust funds instead). Thus, with increased capital formation, lower payroll taxes, and no increase in income taxes, the burden of SI on present and future workers decreases.

The crucial assumptions in the above scenario are (1) that payroll taxes are used to purchase outstanding government debt held by private investors; that is, there is net government savings; and (2) that this leads to private investors substituting private securities for the government debt they previously held, which in turn increases national savings. The first condition fails to obtain when taxes raised for the purpose of advance funding are used to finance the general operations of the government's budget and/or if these taxes loosen fiscal restraint in the sense that the government's budget deficit increases. The second condition fails to obtain if the net government savings is offset by a decrease in private savings. Condition (1) may have been met in Japan and Sweden, as the taxes raised by partial funding were placed in a separate account which was not allowed to be used to finance the general operations of the budget; and the partial funding did not seem to loosen fiscal restraint in these countries, at least in the sense that budget deficits did not increase (though since we do not know whether or not deficits would have increased in the absence of partial funding, the effects of that funding on budget deficits are hard to determine). In the U.S., however, there is little if any reason to believe that the first condition has been met. Under current law, all tax revenues not needed to meet current benefits are “invested” in new special-issue government bonds. The trust funds are credited with a bond—an IOU from one part of government to another—and the Treasury gets the cash. From the Treasury's viewpoint, this cash is available to finance the general operations of the federal government. We do not know whether it is being used to retire outstanding publicly held debt, that is, whether it has produced any net government savings. To determine that, either we would have to know what the size and composition of the budget would have been in the absence of trust-fund buildup, or the budget excluding Social Security would have to be in balance. For a clear exposition of the problems with advance funding in the U.S., see Weaver, , “Controlling the Risks Posed by Advance Funding,” in Weaver, ed., Social Security's Looming Surpluses, pp. 168–72Google Scholar. On the special conditions in Japan and Sweden that may have helped advance funding to genuinely reduce future SI liabilities, see Munnell, and Ernsberger, , “Foreign Experience with Public Pension Surpluses and National Savings,”Google Scholar in ibid., pp. 90–106, 114–18.

8 Barr, Nicholas, The Economics of the Welfare State (Stanford: Stanford University Press, 1993), p. 220Google Scholar, emphasis in original.

9 I have been aided here by Barry, Norman P., “The State, Pensions, and the Philosophy of Welfare,” Journal of Social Policy, vol. 14 (1985), pp. 479–80.CrossRefGoogle Scholar

10 That modern economies are not subject to repeated depressions is a fact; the explanation for this fact is a matter of controversy, which revolves around the extent to which various government policies are or are not responsible for the depression-free record of modern economies. It might seem that this controversy is relevant for the evaluation of SI versus CP, for if CP were to threaten the ability of modern economies to prevent long-term depressions, then that would be an excellent argument against CP. However, the government policies that are often claimed to be the defense against economic depressions—such as preventing bank failures and the shrinkage of the money supply, keeping world trade reasonably free, the existence of automatic “stabilizers” (programs that transfer income to the unemployed during economic downturns), etc.—have no necessary connection with an SI system, and could perform their function within a CP system, as will become clear from the description of CP that occurs within the text.

11 World Bank, Averting the Old Age Crisis, pp. 102–5Google Scholar; Ferrara, , Social Security: The Inherent Contradiction, pp. 5–7, 53–55.Google Scholar

12 World Bank, Averting the Old Age Crisis, pp. 315–16Google Scholar. The World Bank uses a more fine-grained analysis, dividing PAYG into three stages instead of two, but for my purposes, such detail is not necessary.

13 For example, disability and survivors' benefits were often added later. That SI schemes combine retirement pensions with these benefits is a complication which for the most part I shall ignore in this essay, as it does not affect my central arguments.

14 World Bank, Averting the Old Age Crisis, p. 105.Google Scholar

15 That later retirees cannot get the windfall that early retirees receive is a separate matter from whether the former can get a greater-than-market “rate of return,” an issue I shall discuss shortly.

16 If the payroll-tax increase is sizable, and the taxes are actually invested and not used to cover other government spending, the surpluses created at this point can be considerable, but they still do not come even close to eliminating the public pension debt, as I discussed in note 6.

17 Samuelson, Paul, “An Exact Consumption-Loan Model of Interest with or without the Social Contrivance of Money,” in The Collected Scientific Papers of Paul Samuelson, ed. Stiglitz, Joseph (Cambridge, MA: MIT Press, 1966), vol. 1, pp. 219–34Google Scholar. Samuelson's argument is explained, minus the complex mathematics, in Ferrara, , Social Security: The Inherent Contradiction, pp. 293–94Google Scholar; and Tullock, Gordon, The Economics of Income Redistribution (Boston: Kluwer-Nijhoff Publishing, 1983), pp. 111–22.CrossRefGoogle Scholar

18 From 1971 to 1990, the average annual rate of return for a portfolio containing half stocks and half bonds was around 3 percent above the growth in real wages, and for a portfolio containing only stocks, the difference is considerably greater. This data is from the World Bank, Averting the Old Age Crisis, pp. 299–302, 355Google Scholar. During the same period, population growth was slow—indeed, in many OECD countries the growth rate was (and continues to be) barely above the replacement rate. Since SI's “rate of return” declines as the mature stage progresses, one suspects that the advantage of equities over PAYG would look even better if the period from 1980 to the present were used. Note also that with capital becoming more mobile, an international portfolio is likely to do better than a national one, and is likely to provide advantages to those living in countries with sluggish economies, which should give an additional advantage to equities over PAYG in the future.

The point about unanticipated inflation comes from Tullock, Gordon, Welfare for the Well-to-Do (Dallas: Fisher Institute, 1983), p. 54Google Scholar. He provides no data, but since the point is an obvious one, data is probably unnecessary.

19 World Bank, Averting the Old Age Crisis, pp. 131–34.Google Scholar

20 The World Bank's Averting the Old Age Crisis does not indicate whether “contribute less” means less total taxes or less proportionatelyGoogle Scholar. Either way, however, the result is not a progressive intragenerational redistribution.

21 On the spotty record of SI systems keeping their promises, see World Bank, Averting the Old Age Crisis, pp. 112–13.Google Scholar

22 It was the belief that a significant amount of defined-benefit occupational pensions were not being fully funded by U.S. employers that helped to produce the Employment Retirement Income Security Act (ERISA) in 1974. The act provided both tax incentives and legal penalties for failure to fund pensions adequately and created a set of regulations governing employee participation and the vesting of benefits (i.e., the conditions under which employees could leave the company without losing their pensions). For further information, see Coleman, Barbara J., Primer on ERISA (Washington, DC: Bureau of National Affairs, 1985)Google Scholar. Laws similar to ERISA exist in many of the OECD countries. See World Bank, Averting the Old Age Crisis, pp. 193–97.Google Scholar

23 World Bank, Averting the Old Age Crisis, pp. 167–69, 182–83.Google Scholar

24 Partly for these reasons, and partly because some employers consider the ERISA regulations burdensome, occupational pension plans in the U.S. have been shifting from defined-benefit to defined-contribution. That trend also exists in some other OECD countries, such as Australia and Switzerland. For the U.S., see Ferguson, Karen and Blackwell, Kate, Pensions in Crisis (New York: Arcade Publishing Company, 1995), pp. 168–69, 173Google Scholar; for the OECD countries, see World Bank, Averting the Old Age Crisis, pp. 198–200.Google Scholar

25 My information about the Chilean system was obtained from Davis, , Pension Funds (supra note 1), pp. 250–53Google Scholar; World Bank, Averting the Old Age Crisis, ch. 6Google Scholar; and Ferrara, Peter J., Goodman, John C., and Matthews, Merrill Jr., Private Alternatives to Social Security in Other Countries, NCPA Study No. 200 (Dallas: National Center for Policy Analysis, 1995), pp. 13–23.Google Scholar

A system that is intermediate between SI and CP is a system called “provident savings,” found in a number of countries, and most effectively run in Singapore. Here employees and/or employers are required to save, but the savings must be deposited in a government-managed system which invests primarily in government securities. The system is fully funded—one gets back one's contribution plus interest—and contributors do have a property right in their accounts, but since the government manages the system, it would be too large a stretch to place this in the category of private pensions. On Singapore, see Davis, , Pension Funds, pp. 253–55Google Scholar; World Bank, Averting the Old Age Crisis, ch. 6Google Scholar; and Ferrara, et al. , Private Alternatives to Social Security in Other Countries, pp. 7–13.Google Scholar

26 How much larger the rate of return will be in a CP system as compared with what later generations get in an SI system will depend upon the particular country as well as the degree of maturity of the SI system (for the longer an SI system goes on, the worse its “rate of return”). It is worth stressing that all that my arguments in Section III will require is that CP has a nontrivial advantage here, an advantage which grows as SI matures (see note 18 for the advantage that investing in equities has over PAYG). It is also worth stressing that my arguments will not rely upon the extremely impressive 13 percent average annual rate of return that the Chilean system has had since its inception in 1981. See Davis, , Pension Funds, p. 252Google Scholar; and Ferrara, et al. , Private Alternatives to Social Security in Other Countries, p. 21.Google Scholar

27 Note that if one views these different stages as different selves, in the manner of Derek Parfit's theory of personal identity, then the distinction between (1) transfers between different parts of a person's life and (2) transfers between different persons is not metaphysically significant. For purposes of this essay, however, Parfit's views are not relevant, since whether or not they are true, they are not politically significant, as people generally view “redistribution” across different stages of a life as belonging in a vastly different category than redistribution across different persons. For Parfit's views, see his Reasons and Persons (New York: Oxford University Press, 1985), pp. 199–347.Google Scholar

28 On this conceptual or terminological matter, I have been influenced by Temkin, Larry S., Inequality (New York: Oxford University Press), pp. 7–8.Google Scholar

29 Rawls, John, A Theory of Justice (Cambridge: Harvard University Press, 1971), p. 302Google Scholar; Rawls, , Political Liberalism (New York: Columbia University Press, 1993), p. 291.Google Scholar

30 Two helpful summaries of the “Equality of what?” literature are Cohen, G. A., “On the Currency of Egalitarian Justice,” Ethics, vol. 99, no. 4 (07 1989), pp. 906–44CrossRefGoogle Scholar; and Sen, Amartya, Inequality Reexamined (Cambridge: Harvard University Press, 1992).Google Scholar

31 John Rawls's difference principle is the great exception here: his conception of the worst off does not incorporate a responsibility or choice condition. But post-Rawlsian egalitari-anism invariably incorporates such a condition, for two reasons. First, it is quite implausible that persons or institutions have stringent obligations to rectify an inequality or a burden if it is voluntarily incurred or if one can be held responsible for incurring it. Second, many egalitarians also defend some basic individual rights, and such a defense almost inevitably involves accepting the principle that one has a right to act in accordance with (certain of) one's choices. That principle would be difficult to reconcile with the view that voluntarily acquired inequalities or burdens require rectification, or so I argue in “Liberal Egalitarianism, Basic Rights, and Free Market Capitalism,” Reason Papers, vol. 18 (Fall 1993), pp. 171–73.Google Scholar

Ronald Dworkin's writings are probably most responsible for the presence of a responsibility or choice condition in contemporary egalitarianism. For Dworkin's theory, which he calls “equality of resources,” see his “What is Equality? Part I: Equality of Welfare,” Philosophy and Public Affairs, vol. 10, no. 3 (Summer 1981), pp. 185–262Google Scholar, and “What is Equality? Part II: Equality of Resources,” Philosophy and Public Affairs, vol. 10, no. 4 (Fall 1981), pp. 283–345.Google Scholar

32 Another way to put this is that those who reject obligations to future generations, i.e., who relativize justice to time, are primarily rejecting such obligations for strangers.

33 It might seem that there is not just a “significant connection,” but that necessarily or by definition the worst off are the poorest. But that is not so. First, as I noted earlier, some egalitarians are concerned with inequalities in welfare, and there is no necessary connection between being unhappy or lacking satisfaction and being poor (though extreme poverty, over the long run, considerably reduces happiness and satisfaction). Second, measuring poverty simply in terms of low income, as is often done, omits some important considerations that affect the quality of people's lives. For example, two people can have equal incomes but not be equally well off because of different capabilities of converting that income into resources: e.g., someone with a high metabolic rate, a large body size, or a parasitic disease that wastes nutrients may have a much harder time meeting minimal nutritional requirements with her income than someone who does not have these characteristics. See Sen, , Inequality Reexamined, ch. 7Google Scholar, for more discussion of this last point.

Thus, although there are obvious connections between the poor and the worst off, egalitarians cannot and do not equate them. The best one can say here is that when concerned with policy or institutional choices, poverty is a plausible but not completely reliable marker for being badly off.

34 Temkin, Larry S. makes this argument about the justifiability of transfers from the young to the old in “Justice and Equality: Some Questions about Scope,” Social Philosophy and Policy, vol. 12, no. 2 (Summer 1995), pp. 96–97CrossRefGoogle Scholar, although he makes no mention of SI.

35 Loren E. Lomasky raised this objection (which he does not, incidentally, endorse). After Lomasky raised this objection, I discovered that Davis, , Pension Funds, pp. 38–39Google Scholar, makes a similar point, as does Temkin, in “Justice and Equality,” pp. 93–94Google Scholar, although Temkin does not do so in the context of a discussion of SI.

36 I add this because it is inconceivable to me that any egalitarian would object to later generations being better off than earlier ones in situations where grinding poverty is the norm.

37 World Bank, Averting the Old Age Crisis, pp. 78–80.Google Scholar

38 Since there is a paucity of discussion of SI by egalitarian philosophers, I should mention one other argument that might seem germane to this topic. Political theorist Albert Weale argues that although a flat-rate form of SI, which provides the same modest pension to everyone, may appear to be more egalitarian than one with an earnings-related component (since the latter increases benefits with increased earnings), in fact the opposite is true. The reason resides in the different incentives the two systems give to the nonpoor or middle class. In the former, the incentive is to supplement the modest flat-rate pension with a private pension which is expected to produce returns which are as lucrative as possible for oneself and one's family; one would try to avoid any pension scheme which redistributes income or includes poor risks (as occurs in some occupational pensions). Indeed, to the extent to which your taxes for the flat-rate scheme reduce your ability to purchase an advantageous private pension, you will resist any attempts to expand the level of coverage. In an SI system with an earnings-related component, on the other hand, the incentive is to expand the level of coverage. Doing so will not be a pure loss, since while benefits are not proportional to earnings, they will rise with earnings to some extent. And there will be little incentive to resist the redistribution that is often a part of earnings-related schemes, since such collective action would be extremely costly. The result of these incentives is that more money from the pension funds reaches the worse-off members of society in an earnings-related scheme than in a pure flat-rate scheme, even though the latter is more egalitarian per unit of resources. In effect, spillover from the larger amount of tax funds expended on earnings-related systems means that there is more total redistribution in such a scheme than in one which is more egalitarian per unit of resources.

Suppose Weale's argument is sound. Does his argument about the egalitarian merits of different forms of SI have any application to the subject at hand, the egalitarian merits of SI versus CP? Could CP be a stand-in for a flat-rate form of SI? No, because CP is a fully funded pension, rather than being PAYG, and it is the PAYG nature of SI, plus its regressive financing, that produces its anti-egalitarian effects, as compared to CP. The logic of Weale's argument is to favor forms of SI which absorb a greater share of GDP on the grounds that more of that money will spill over to the poor than in more modest forms of SI; but even if this is true, it ignores the intergenerational effects of SI. An SI system which absorbs a lot of the GDP means a greater burden for later generations—particularly the poor of those generations, for reasons already discussed. See Weale, Albert, “Equality, Social Solidarity, and the Welfare State,” Ethics, vol. 100, no. 3 (04 1990), pp. 480–81CrossRefGoogle Scholar. Since Sweden is Weale's model for the type of SI system he favors, it is worth noting that even if Sweden's system is more egalitarian than a pure flat-rate system would be, it does not have progressive intragenerational transfers. See World Bank, Averting the Old Age Crisis, p. 133.Google Scholar

39 A basic right is a moral right that has a considerable degree of weight, so that it typically defeats perfectionist considerations (claims about individual well-being or virtue) and claims of societal or aggregate well-being.

40 Both arguments have been given by a variety of philosophers. For a helpful summary and critique, see Jacobs, Lesley, Rights and Deprivation (Oxford: Oxford University Press, 1993), ch. 6.Google Scholar

41 See Moon, J. Donald, “The Moral Basis of the Democratic Welfare State,” in Democracy and the Welfare State, ed. Gutmann, Amy (Princeton: Princeton University Press, 1988), pp. 30–36, 41–46Google Scholar; and Moon, , “Introduction: Responsibility, Rights, and Welfare,” in Responsibility, Rights, and Welfare: The Theory of the Welfare State, ed. Moon, J. Donald (Boulder, CO: West-view Press, 1988), pp. 4–8Google Scholar. See also Jacobs, , Rights and Deprivation, pp. 198, 200–202.Google Scholar

42 In “The Moral Basis of the Democratic Welfare State,” p. 46Google Scholar, Moon says that recipients of social insurance benefits have in fact contributed their fair share, but in “Introduction: Responsibility, Rights, and Welfare,” p. 7Google Scholar, he merely says that “at least in theory” the recipients have contributed the resources that make the benefits possible. Hence, it is unclear to what extent Moon really believes that in an SI system one's pension is funded by one's own contributions.

43 Feinberg, Joel, Harm to Self (New York: Oxford University Press, 1986), pp. 16–17.Google Scholar

44 One might argue that this cannot be a paternalist argument, since the persons who cannot be trusted to save are not the persons being required to save. However, as Feinberg suggests, paternalism is a matter of the justification of a law or policy, and if the justification for forcing A to save for B is that otherwise B will harm B (rather than, say, that B will harm A), then the rationale for the law is indeed paternalistic. While this is a rather indirect way of meeting a paternalistic objective, it is paternalistic nonetheless so long as the aim is to prevent B from harming B. Furthermore, as I shall argue later, PAYG systems are not infrequently misunderstood, and workers come to think that their taxes are being invested for their own retirement, in which case they may view the justification of SI as paternalistic in the more obvious sense.

45 Recent polls show that a large majority of those under thirty-five and those over sixty support a proposal to allow people to direct part of their Social Security taxes to a financial institution of their choice and to receive less in Social Security benefits. See Ferrara, Peter J., A Plan for Privatizing Social Security (Washington, DC: Cato Institute, 1997)Google Scholar. The bipartisan Social Security advisory panel, appointed by President Clinton in 1994, has recommended three different plans for reforming Social Security, all of which involve investing at least part of Social Security taxes in the stock market, and two of which endorse taxpayers doing their own investing. See “A Consensus Emerges: Social Security Faces Substantial Makeover,” Wall Street Journal, 07 8, 1996, pp. A1, A4Google Scholar. While these results do not prove the collapse of the belief that Social Security embodies some norm of reciprocity or contribution, they certainly suggest it, for it would be hard to explain these results if Social Security were viewed as embodying that norm.

46 The strength of this reason is unclear. It is difficult to say to what extent the value of self-respect is diminished if it is based on illusions. Furthermore, the sources of self-respect are multiple, and it is unclear to what extent the self-respect of recipients of SI is linked with a view that they are entitled to their pensions or are not harming later generations.

47 It might be objected that this is false, because in the beginning of a CP system those who had not saved much on their own voluntarily, or who invested imprudently, would have little security when they retired. But all this means is that in the beginning of a CP system, the safety net may have to be larger than in the later stages. It does not negate the point that a CP system does not redistribute security through time.

48 For an argument that public justification is central to liberalism, see Macedo, Stephen, Liberal Virtues: Citizenship, Virtue, and Community in Liberal Constitutionalism (Oxford: Clarendon Press, 1990), pp. 40–48Google Scholar. Liberal theorists differ concerning the subject matter of public justification: sometimes it is principles of justice; at other times it is the social order or basic social and political institutions. I shall assume that a state program or institution like SI, in virtue of its far-reaching effects and the relevance of questions about its justice and fairness, is an institution that falls within the subject matter of public justification.

49 Gerald Gaus makes this point, citing much empirical evidence to support it, in Gaus, , Justificatory Liberalism (New York: Oxford University Press, 1996), pp. 130–36.Google Scholar

50 It is worth noting that in the first thirty years or so of Social Security in the U.S., most citizens had an extremely hazy and poor idea of how it worked. See Derthick, Martha, Policymaking for Social Security (Washington, DC: Brookings Institution, 1979), pp. 188–90.Google Scholar

51 This is allowed, e.g., in Britain and Japan. See Davis, , Pension Funds, pp. 64–65Google Scholar; and Ferrara, et al. , Private Alternatives to Social Security in Other Countries (supra note 25), pp. 1–2, 25–28.Google Scholar

52 Philosopher Michael Robins has objected to this statement, maintaining that the existence of pension plans for state employees in the U.S. (or for certain state employees, such as teachers) shows that a government-run pension system can be close to being fully funded, give quarterly or annual reports to its recipients, and not be subject to political maneuvering. However, these pensions systems are not nationwide systems, while my concern in this essay is with systems that cover (virtually) all citizens in a country; thus, the operation of pension plans for these state employees does not negate my point that no existing SI system has adopted all the reforms mentioned in the text that would make it operate more like market insurance. More importantly, pension plans for state employees are defined-benefit plans. Defined-benefit plans that rely upon taxpayers for funds, even if they are not subject to political maneuvering, have a stronger incentive than private occupational defined-benefit plans not to fund adequately. This is because the former use the power to tax to make up for any mistaken actuarial projections, while private occupational plans that make mistakes are at a competitive disadvantage (since wages for younger workers will have to be lowered to meet the unfunded pensions promised to retirees). Indeed, many pension plans of states and municipalities in the U.S. are in trouble because in the late 1980s and early 1990s, overly optimistic assumptions about future rates of return made them cut their required contribution rates, and as a result, it is estimated that almost one in three of these plans have less than 75 percent of the assets required to meet their liabilities. See World Bank, Averting the Old Age Crisis, pp. 188–89.Google Scholar

53 See Weaver, Carolyn L., The Crisis in Social Security (Durham, NC: Duke University Press, 1982), pp. 80–86, 123–24Google Scholar; and Derthick, , Policymaking for Social Security, pp. 199–201, 204.Google Scholar

54 Daniels, Norman, Am I My Parents' Keeper? (New York: Oxford University Press, 1988), pp. 134–35Google Scholar, defends SI, but since his arguments do not fit neatly within any of the values or principles I have discussed so far, I have put them in this note. Daniels denies that later generations in an SI system have grounds for complaints despite their poor “rate of return” because (1) returns are indexed in an SI system and not in private pensions, (2) private pensions will give a variable return and thus subject the elderly to more significant risks than one gets with SI, and (3) SI reduces the financial burden of supporting one's parents. These are weak arguments. The first argument is weak because SI's record of indexing benefits is spotty, and a CP system could incorporate indexed annuities into one's post-retirement choices, as Chile's system does. The second argument is weak because CP's safety net provides a minimum guarantee, thus limiting the degree of risk, while in an SI system all citizens are subject to the risk of redefined benefits, changing tax rates, etc. The third argument clearly fails, because CP also reduces the burden of children caring for their parents, and compared to the later stages of SI, CP does a better job of reducing that burden in virtue of its superior rate of return and safety net. On the question of indexing, see Davis, , Pension Funds, p. 114Google Scholar; and World Bank, Averting the Old Age Crisis, p. 156.Google Scholar

55 For an argument that the transition to CP would not be terribly difficult or costly, see Ferrara, Peter J., “Privatization of Social Security: The Transition Issue,”Google Scholar elsewhere in this volume. Note that even if all the values or principles discussed in this essay imply support for a transition to CP, it does not follow that they all imply support for the same way of making that transition.