FLORIDA TAX REVIEW VOLUME 4 1999 NUMBER 5 Book Review Perspectives on Social Security Reform Framing the Social Security Debate: Values, Politics, and Economics (R. Douglas Arnold, Michael J. Graetz & Alicia H. Munnell eds., 1998) Prospects for Social Security Reform (Olivia S. Mitchell, Robert J. Myers & Howard Young eds., 1999). Reviewed by Karen C. Burke* and Grayson M.P. McCouch.*" These two volumes offer an interdisciplinary perspective on major issues which lie at the heart of the current debate over social security reform. Each volume contains a series of papers from a broad range of contributors, including economists, actuaries, investment managers, political scientists, and policymakers. In analyzing various specific proposals and conceptual models for reform, the contributors identify some areas of emerging consensus as well as numerous conflicts, tensions, and tradeoffs which ultimately must be resolved through the political process. The result is a valuable survey of proposals and trends which will determine the direction of future reform. Options For RefonLn Projected long-term shortfalls in social security financing have prompted all sorts of proposals for modifying or restructuring the existing system. Many of these proposals involve two basic reforms which hold special interest for economists: (1) moving away from a pay-as- you-go system in the direction of advance funding; and (2) adopting a more diversified investment strategy for funds accumulated in the social security system. Economists generally agree that advance funding and investment diversification might bring substantial benefits but would also pose significant risks and costs. Thus, they emphasize that the focus of debate should be on identifying and comparing the tradeoffs required by various approaches rather than on whether there is a "free lunch."' * Professor of Law, University of Minnesota Law School. ** Professor of Law, University of Miami School of Law. 1. See John Geanakoplos, Olivia Mitchell & Stephen P. Zeldes, Would a Privatized Social Security System Really Pay a Higher Rate of Return?, in Framing the Social Security Debate: Values, Politics, and Economics 137, 150 (R. Douglas Arnold, Michael J. Graetz & Alicia H. Munnell eds., 1998) [hereinafter Framing the Social Security Debate]. Florida Tax Review Some proposals would go further and introduce a system of private accounts owned and controlled by individual participants. In effect, these proposals would privatize social security by shifting from a defined-benefit system in the direction of a defined-contribution system. Although the concepts of advance funding, diversification, and privatization are often linked together, they are analytically distinct: in theory, it is possible to introduce any one of them independently of the others.2 All three issues are central to the social security debate, but the implications of privatization remain far more controversial than those of advance funding or investment diversification. Thus, in analyzing and comparing different reform proposals, economists treat the three concepts separately. In particular, they point out that advance funding and investment diversification could be implemented either under a collective defined-benefit system or under a privatized defined- contribution system. Advance Funding and Transition Costs. In the abstract, advance funding is attractive because it would potentially increase national savings and economic growth (to the extent that increased social security fund accumulations would not be offset by dissaving in other sectors of the economy). The resulting higher return on the accumulated funds might ultimately make it easier to provide a given level of benefits without increasing payroll taxes. Under the existing pay-as-you-go system, almost all of the revenue collected each year in payroll taxes is used to pay current benefits. (The social security trust funds are presently accumulating modest annual surpluses, but these accumulations are expected to be exhausted by 2032.) In effect, the burden of financing benefits for each generation of retirees falls on succeeding generations of workers. By contrast, under a system with full advance funding, payroll taxes would be set aside and used, with accumulated earnings, to pay benefits for current workers in retirement. In effect, advance funding would place on each generation of workers the burden of financing their own retirement. Economists point out, however, that the transition to advance funding would not be costless. Even if it were possible to replace the existing system prospectively with full advance funding, it would still be necessary to pay future benefits that have been promised to current retirees and workers under the existing system based on their previously credited earnings. By one estimate, this "unfunded liability" already stands at around $9 trillion.3 Given the magnitude of this transition cost, any proposal to implement advance 2. See id. at 138-41. 3. See Stephen C. Goss, Measuring Solvency in the Social Security System, in Prospects for Social Security Reform 16, 34 (Olivia S. Mitchell, Robert J. Myers & Howard Young eds., 1999) [hereinafter Prospects for Social Security Reform). [Vol 4:5 Perspectives on Social Security Reform funding should specify how the cost is to be financed (e.g., from additional government debt, increased taxes, or revenue diverted from other uses). Obviously, the choice of financing method would determine how the transition costs would be spread among current and future generations. Furthermore, it might be imperative as a political matter to impose costs as gradually and imperceptibly as possible, especially in the absence of immediate and obvious benefits.' Proponents of privatization sometimes argue that the projected surpluses in the unified budget should be used to finance a "painless" transition to an advance-funded system of private accounts. The projected surpluses, attributable primarily to the temporary accumulation of social security funds, may seem to offer a ready source of funds to defray transition costs. However, diverting the projected surpluses to fund private accounts would merely exacerbate the unfunded liability of the existing system. This approach, described by one commentator as "stealth privatization," appears to be a political tactic designed to jump-start the process of privatization and preempt full consideration of the direction of future reform.5 The current budgetary skirmishing highlights the need to devise a mechanism to ensure that accumulated social security funds cannot be raided to finance other government expenditures. Investment Returns. The prospect of accumulating substantial funds through partial or full advance funding raises the issue of how those funds should be invested to achieve optimal returns. Many of the current reform proposals would abandon the practice of investing funds exclusively in government bonds and permit investment in a more diversified portfolio of stocks and bonds. There seems to be little reason, on purely economic grounds, to deny access to higher returns from private capital markets; indeed, modem financial theory treats diversification as an important aspect of prudent investment. At the same time, it is important to understand the risks and limitations associated with a more flexible investment strategy. Proponents of privatization often complain that the existing system produces unreasonably low returns, and argue that all current and future workers would be better off under a privatized system that allowed individual workers to invest funds directly in private capital markets. Careful economic analysis reveals two major flaws in this standard "moneys-worth" claim. First, the claim ignores the massive transition costs associated with the shift to advance funding. If those costs were taken into account (e.g., by converting 4. See RL Douglas Arnold, The Political Feasibility of Social Security Reform, in Framing the Social Security Debate, supra note 1, at 389, 397-408. 5. See id. at 414. 19991 Florida Tax Review the existing unfunded liability into explicit debt financed with new taxes along a similar time path), the return on private accounts invested in government bonds would be no higher than the low projected return on current contributions under the existing system. Second, the money's-worth claim fails to make a proper adjustment for risk. Investors who already hold properly diversified portfolios will not necessarily choose to invest their marginal dollars entirely in stocks rather than in government bonds, even though the stocks have higher expected returns. Indeed, these investors may rationally choose to offset any change in social security investments by making a countervailing change in the rest of their portfolios. The real beneficiaries of diversifying social security investments would be "constrained" investors who would otherwise lack access to private capital markets. After taking account of transition costs and risk, the improvement in investment returns would be neither as dramatic nor as widespread as they might appear at first glance.6 Of course, higher investment returns could be also be achieved in a centralized system without private accounts. Here again, the correlation between risk and return is relevant in determining how the accumulated funds should be invested. Moving social security funds from government bonds into private capital markets would imply a corresponding shift of other funds in the opposite direction, leaving a larger portion of non-social security funds invested in government bonds. Such an "asset swap" might well result in higher investment returns for the social security funds, which would offer the political advantages of improving the financial condition of social security and reducing the need for payroll tax increases (or benefit cuts). At the same time, higher returns would be accompanied by increased risk and would presumably trigger countervailing adjustments in the rates of risk and return for non-social security funds. Control of Investments. The prospect of investing accumulated funds in private capital markets also raises the issue of control over investment strategy. If the funds are held by the government in a centralized system, there are grounds for concern that government control might lead to "social investing" or to unwarranted interference in corporate decision-making.7 One response to these concerns would be to introduce institutional safeguards to insulate investment strategy from political pressure. The leading example of this approach is the federal Thrift Savings Plan (TSP), which has adopted a combination of safeguards: the investment function is delegated to an inde- 6. See Geanakoplos et al., supra note 1, at 148-57. 7. See Theodore J. Angelis, Investing Public Money in Private Markets: What Are the Right Questions?, in Framing the Social Security Debate, supra note 1, at 287, 290-93, 304-10. [Vol 4:5 Perspectives on Social Security Reform pendent board, funds are invested exclusively in broad index funds, and stock voting rights are exercised by an independent investment manager.' It is not clear whether this approach would prove as effective for the social security system as it has for the TSP; the sheer size of the funds involved in an advance-funded social security system might pose special risks of government interference. Furthermore, even the most effective safeguards would remain subject to the political risk of revision or repeal by subsequent legislation. Proponents of privatization often argue that putting funds directly into private accounts under the control of individual workers would be a more effective approach to prevent unwarranted government interference in private capital markets. A decentralized system of private accounts, however, would give rise to a fresh range of problems: high administrative costs; excessively risky investment decisions; fraud; and breaches of fiduciary obligations. To protect individual workers and preserve the integrity of the system, the government would almost certainly be called on to regulate financial service providers and perhaps to restrict individual investment choices as well.9 Furthermore, it would be necessary to specify whether the responsibility for establishing and maintaining private accounts should fall on individual workers or on their employers in the first instance. A worker-based system (along the lines of individual retirement accounts) would raise special concerns about administrative costs, funding errors, and imprudent investment choices, while an employer-based system (along the lines of 401(k) plans) would rely heavily on employers who in many cases might be ill-equipped to meet their new obligations.'0 These increased regulatory and administrative burdens should be taken into account in evaluating any proposal to move toward a decentralized system of private accounts. Retirement Security. The debate over privatizing social security reflects a fundamental tension between competing goals of "social adequacy" and "individual equity." In general, reformers who place a high value on social adequacy support retaining the existing collective defined-benefit system with various modifications, while those who emphasize individual equity recommend moving toward a defined-contribution system with private accounts. Even staunch proponents of privatization, however, usually acknowledge the need to maintain some minimally adequate level of retirement benefits. 8. See id. at 293-303; Francis X. Cavanaugh, discussant, in Framing the Social Security Debate, supra note I, at 319, 319-23. 9. See Howell E. Jackson, discussant, in Framing the Social Security Debate, supra note 1, at 329, 335-45. 10. See Janice M. Gregory, Possible Employer Responses to Social Security Reform, in Prospects for Social Security Reform, supra note 3, at 313, 321-23. 19991 Florida Tax Review A defined-contribution system with private accounts automatically serves the goal of individual equity by linking the level of retirement benefits directly to the accumulated value of the participant's prior contributions. A major concern under such a system is that retirement benefits may prove inadequate due to premature account withdrawals, market fluctuations, or low lifetime earnings. To guard against these risks, the system might include one or more of the following: (1) restrictions on the time and form of account withdrawals; (2) guarantees to ensure a minimum account balance or investment return; and (3) government matching contributions for low earners. A separate issue is whether such measures, even if adopted, would prove politically sustainable over time. In the absence of any restrictions on account withdrawals, there is an obvious risk that participants might spend down their entire account balances promptly upon retirement, leaving inadequate resources for later years. (Indeed, the experience with private pension funds suggests that there might be substantial political pressure to allow withdrawals in certain cases even before retirement.) To prevent premature exhaustion of private accounts, participants might be required to use all or part of their account balances at retirement to purchase annuities. Full mandatory annuitization would provide valuable protection for long-lived participants (albeit at the expense of short- lived participants), and would mitigate the effects of adverse selection on annuity prices. Even partial annuitization would provide some of the same protection while allowing individual participants to spend down part of their account balances. Just as some form of mandatory savings is considered necessary to ensure an adequate level of resources at retirement, some level of mandatory annuitization may be desirable to ensure an adequate level of income throughout the retirement period. Restrictions on the time and form of account withdrawals also have important consequences for surviving spouses. Elderly widows run a disproportionately high risk of poverty, despite the mandatory survivor benefits provided by the existing social security system." Under a system of private accounts, special safeguards would be necessary to prevent an even higher poverty rate among elderly widows. Assuming some level of mandatory annuitization, one approach would be to require that benefits be paid out in the form of joint-and-survivor annuities, by analogy to the existing treatment of private pension plans.'2 In the absence of mandatory survivor benefits, it would be necessary to consider whether some form of "earnings sharing" should be required for private accounts. A pure earnings- 11. See Karen C. Holden, Women as Widows Under a Reformed Social Security System, in Prospects for Social Security Reform, supra note 3, at 356, 358-61. 12. See id. at 361-67. [Vol 4:5 Perspectives on Social Security Reform sharing approach would allocate to each spouse an equal one-half share of the couple's combined earnings during marriage, resulting in a corresponding allocation of balances in their private accounts. In comparison to the existing system of spousal benefits, a system of private accounts with earnings sharing might be viewed as considerably more equitable for many kinds of families. It remains unclear, however, whether earnings sharing would prove any more politically feasible under a privatized system than under the existing system. Another concern raised by private accounts relates to the issue of market risk. To the extent that individual participants had a wide range of investment choices, many of them would likely end up with inadequate balances in their private accounts at retirement, due to poor investment decisions or simple bad luck. It seems likely, therefore, that the government would be called on to guarantee some minimum account balance or rate of return on investments.' 3 To control the cost of such guarantees and address the related problem of moral hazard, the government would presumably insist on restricting the range of investment choices open to individual participants. For example, participants might be allowed to allocate the funds in their private accounts among a limited number of index funds classified according to type and level of investment risk. Even under a closely regulated system of private accounts, the volatility of private capital markets would probably intensify problems of equity among different groups of participants. Those who happened to retire at a moment when the market had soared to unprecedented heights (or dived to unexpected depths) would find themselves locked into abnormally high (or low) levels of retirement income during retirement. 14 Finally, a system of private accounts raises the issue of providing adequate retirement benefits for participants with low lifetime earnings. The existing social security system includes a significant redistributive component, due to a progressive benefit formula which produces disproportionately high benefit levels for low earners. Even in a system of private accounts, it is possible to preserve some degree of progressivity. To the extent that only a portion of the existing system is privatized, the remaining defined-benefit portion could continue to provide higher benefits for low earners. Alternatively, a two-tier system could provide a flat minimum benefit for all participants as well as a supplementary tier of private accounts. If private accounts represent most or all of the overall system, it might be necessary to provide some sort of matching contributions for low earners. It is also 13. See Stephen G. Kellison & Marilyn Moon, New Opportunities for the Social Security System, in Prospects for Social Security Reform, supra note 3, at 60, 72-74. 14. See Lawrence H. Thompson, Individual Uncertainty in Retirement Income Planning under Different Public Pension Regimes, in Framing the Social Security Debate, supra note 1, at 113, 121-29. 1999] Florida Tax Review important to consider the allocation of administrative costs among private accounts. Since charges based on a flat amount per account would disproportionately erode the value of small accounts, it might be fairer to allocate charges based on accumulated account balances. Private Pension Plans. Any fundamental reform of the existing social security system will inevitably affect the structure of private employer- sponsored pension plans. Employer-sponsored plans have expanded in a context of targeted federal tax incentives and complex nondiscrimination rules designed to ensure access for nonhighly compensated employees. Since plan benefit formulas typically take social security benefits into account, any changes in the level of social security benefits or in the payroll tax base will have an immediate impact on the administration and design of "integrated" employer-sponsored plans. In the long run, employer-sponsored plans will respond to altered employee expectations, business needs, and regulatory requirements. 5 Introducing mandatory private accounts into the social security system might induce workers to reduce other savings. For example, nonhighly compensated employees might choose to reduce voluntary contributions to employer-sponsored 401(k) plans, thereby threatening the ability of some employer-sponsored plans to comply with the nondiscrimination rules. Furthermore, allowing individual workers to make additional, voluntary contributions to their private social security accounts might make employer- sponsored plans less attractive for highly-compensated employees. Some observers fear that proposed social security reforms could accelerate a trend toward a "two-tier" system in which employers would offer unfunded, nonqualified plans for highly-compensated employees and funded plans that would be "adequate for the lower-paid and inadequate for employees in the middle."' 16 From the broader perspective of retirement income policy, social security and employer-sponsored pension plans, along with private savings, have long been viewed as essential complementary components of a "three- legged stool." Taken together, social security and employer-sponsored plans in their existing forms combine defined-benefit and defined-contribution approaches in a "mixed" system of retirement savings. Within the context of employer-sponsored plans, voluntary defined-contribution arrangements such as 401(k) plans have substantially encroached on more traditional defined- benefit arrangements, with the result that individual workers bear an 15. See Gregory, supra note 10, at 318-30. 16. See Christopher Bone, An Actuarial Perspective on How Social Security Reform Could Influence Employer-Sponsored Pensions, in Prospects for Social Security Reform, supra note 3, at 333, 346. [Vol 4:5 Perspectives on Social SecuritY Reform increasing share of investment risks compared to employers. Introducing a system of private social security accounts would shift the balance in the combined public and private systems even more decisively toward a defined- contribution approach. Such a shift might raise questions both about the desirability of creating a new type of savings vehicle (in addition to existing 401(k) plans and IRAs) and about the long-term stability of employer- sponsored plans. Political Risk. All retirement programs are subject to political risks in the sense that their original design, operation, and purposes may prove unstable or become unsustainable over time due to evolving political conditions. For example, if the social security system is bifurcated into a defined-contribution component and a defined-benefit component, the main political risk is that the goal of social adequacy will ultimately be undermined. Initially, the proposed reform might retain some version of the existing defined-benefit system with a progressive benefit formula, coupled with a supplementary tier of private accounts. Such a bifurcated system would concentrate the redistributive function in the defined-benefit component, which would be more isolated and vulnerable than under the existing system. Indeed, the reformed system itself might generate a new political dynamic in which middle- and high-income participants sought to expand the defined- contribution component at the expense of the defined-benefit component. Although long-term predictions are hazardous, there is little reason to expect that introducing private accounts would ultimately strengthen the collective retirement security system or mitigate existing disparities of income among retired workers. 17 By contrast, the main political risk for the existing system is that the public will be unwilling to accept the tax increases (or benefit reductions) necessary to restore actuarial balance. With this risk in mind, many commentators assert that promised benefits cannot be paid without raising payroll taxes to politically unacceptable levels and conclude that social security in its existing form is unsustainable.' Despite repeated predictions that the existing system will inevitably lose political support, the situation appears to be far more complex. Careful analysis of polling data confims low public confidence in the existing system but also indicates strong and 17. See Hugh Heclo, A Political Science Perspective on Social Security Reform. in Framing the Social Security Debate, supra note 1, at 65, 83-89. 18. See Gordon P. Goodfellow & Sylvester J. Schieber, Simulating Benefit Levels Under Alternative Social Security Reforms, in Prospects for Social Security Reform, supra note 3, at 152, 180-81. 19991 Florida Tax Review sustained public support for social security. 9 Moreover, in analyzing attitudes toward reform proposals, there appear to be significant differences along income and gender lines but, somewhat surprisingly, age does not appear to be an especially salient factor.2" Value judgments as well as economic factors may play an important role in shaping public opinion about the existing system and proposals for its reform. Polling data also give grounds for concern about the level of public understanding of basic issues in the social security debate. Many people express general support for social security yet oppose raising taxes (or cutting benefits) because they do not believe such measures are needed to preserve the existing system. Furthermore, people who appear worried about their own retirement security nevertheless tend to underestimate the need for increased individual saving. These gaps and inconsistencies suggest that political leaders, policy analysts, and the news media all have important roles to play in framing the debate and shaping public opinion. Developing a political consensus will not be easy. Nevertheless, delay will only increase the cost of implementing reform. At present, the changes needed to restore actuarial balance to the existing system are still relatively modest, but the required tax increases (or benefit cuts) will rise steeply with the passage of time. Proponents of privatization also cannot afford delay, since the costs of maintaining promised benefits under the existing system while setting aside additional funds for a new system of private accounts will eventually become prohibitive. Even in the absence of an immediate liquidity "crisis," there is general agreement that social security reform should be undertaken sooner rather than later. Conclusion. These volumes provide insightful and balanced analysis of a broad range of issues in the social security debate. Despite the diverse approaches taken by the contributors, there appears to be remarkable agreement about the economic tradeoffs inherent in various reform proposals. Indeed, according to one of the editors, "the economics, while interesting, complicated, and often misreported, is not controversial."'" At the same time, the various reform proposals elicit differing assessments which inevitably reflect highly controversial judgments about values and politics. This is hardly surprising, for values and politics lie at the heart of the social 19. See Lawrence R. Jacobs & Robert Y. Shapiro, Myths and Misunderstandings about Public Opinion toward Social Security, in Framing the Social Security Debate, supra note 1, at 355, 364-76. 20. See John Rother & William E. Wright, Americans' Views of Social Security and Social Security Reforms, in Prospects for Social Security Reform, supra note 3, at 380, 392-93. 21. Alicia H. Munnell, Introduction, in Framing the Social Security Debate, supra note 1, at 1, 28. [VoL 4:5 19991 Perspectives on Social Security Reform 427 security debate. It would be unrealistic to expect unanimous agreement on the criteria for evaluating competing reform proposals, much less on their substantive merits. For all of the differences in their premises and viewpoints, the contributors bring a welcome note of clarity and candor to the debate.