Problems with the private pension system and how they should inform the Social Security reform debate

Leverage, linkage, and leakage: Problems with the private pension system and how they should inform the Social Security reform debate

Stein, Norman P

I. Introduction

The argument for Social Security privatization is, at bottom, simple: we need more, and better, advance funding of the public retirement system.1 In particular, we need to commit a portion of FICA tax to privately managed investment accounts, which will purchase investment instruments that promise higher rates of return than the government debt instruments in which the Social Security surplus is currently invested.2 The privatization debate has centered on the extent to which Social Security faces an impending demographic crisis during the coming decades,3 whether privatization is fundamentally inconsistent with the idea of social insurance,4 whether privatization financial projections are accurate,5 and whether privatization is a more rational means of securing and improving the financial status of retirees than the current Social Security system.6

tributes mightily to the system through tax subsidies,7 and many employmentbased pension plans are maintained by public entities for their employees rather than by private-sector employers.8 We nevertheless use the term “private pension system” here, in large part because we focus on concerns with private sector retirement plans.

We attempt in this Article to advance the debate over whether to privatize Social Security by reflecting on three of the private pension system’s major problems and their relevance in the Social Security reform debate. We refer to these issues as ones of leverage, linkage, and leakage.

provide retirement income, and, therefore, plan assets should not leak out of the plan for non-retirement purposes.12 The private pension system, in our view, has inadequate leverage and linkage and an alarming degree of leakage.

Reflecting on these problems of the private pension system can inform the Social Security reform debate in two significant ways. First, to the extent that Social Security and the private pension system are understood as two components of a unified national retirement policy, they should work in complementary fashion to produce a coherent policy result that addresses the income security needs of all, or almost all, of the nation’s nonworking-aged population. From this perspective, Social Security should backstop the weaknesses of the private system and satisfy objectives it does not adequately address. Second, the private system is a finded system similar in some ways to proposed privatization models and may hold some lessons for how a privatized system should be designed, if one is to be designed at all. Our own view, though, is that no design features can adequately harmonize individual investment accounts with the concept of social insurance.13 We leave the second question to those either less committed than we to the idea of social insurance or less skeptical about the compatibility of that idea with a Social Security system based on private investment accounts.

Parts II, III, and IV discuss the problems of leverage, linkage, and leakage in the private sector pension system. Part V explores the meaning of these problems in the private pension system for the Social Security reform debate.

II. The Problem of Leverage

from those provisions in the Internal Revenue Code that depart from the ordinary structure of an income tax.14 Among the provisions considered to produce tax expenditures are those that govern the tax treatment of “qualified” employer pension plans.15 The Congressional Budget Office estimates the lost tax revenue related to public and private pensions to be more than $90 billion dollars and more than $100 billion if we include individual retirement accounts.16 This is a sizable sum; in fact, it is the largest of all the expenditures in the tax expenditure budget.17

Article, we accept the conventional understanding that the Internal Revenue Code does provide a valuable tax subsidy for qualified pension plans and that the subsidy should be justified by some purpose extrinsic to the goals of an income tax.

The orthodox explanation for the subsidy is that it provides retirement income security for employees who would not otherwise save adequately for retirement.21 The intended primary beneficiaries of the tax expenditure, then, are low- and moderate-income employees, who often find it difficult to save on their own for retirement because of immediate consumption demands. More affluent individuals have greater capacity to save for their retirement without governmental assistance.

This strategy is first to make the tax benefits of qualified plans sufficiently attractive to the tax-sensitive people who own and manage businesses so that they decide to set up plans to capture tax benefits for themselves; and second to require such plans, once established, to provide meaningful benefits not only to the people who set them up but also to their low- and moderateincome employees.24 The Code effects the latter part of the strategy through a series of statutory provisions, most prominently the nondiscrimination rules.25 Professor Dan Halperin has used a (tax) carrot and (regulatory) stick metaphor to describe the strategy.26 Some have called this trickle-down benefits policy.27 This is, in any event, the idea that we refer to here as leverage. As the pension economist Alicia Munnell put it, “The rationale for favorable tax treatment of qualified plans is that retirement benefits for rankand-file employees will exist if Congress provides tax incentives that induce higher paid employees to support the establishment of employer-sponsored pension plans.”28

and moderate-income employees because those employees, at least as a group, do not value deferred compensation at its cost to the firm.30 Accordingly, some employers who participate in the system play a game of statutory limbo, being under the regulatory stick by manipulating the complexities of the nondiscrimination rules to minimize or eliminate benefits for rank-and-file employees, while maximizing benefits for the highly compensated.31 Moreover, many employers simply do not respond to the incentives and fail to sponsor pension plans.32 Thus, the system is both overinclusive in that it provides benefits for those who can save for their own retirement absent governmental incentive, and underinclusive because it fails to cover many low- and moderate-income workers at all and pays only minimum benefits to some of the covered low- and moderate-income employees.

Despite these criticisms, the basic paradigm – tax benefits to encourage plans and nondiscrimination rules to ensure that the plans provide meaningful benefits to regular employees – has endured as the rationale for the favored tax treatment of qualified plans. Furthermore, qualified plan coverage of the private workforce has remained pretty much steady at around 50%.33

Not surprisingly, the coverage rates decline with income: the coverage rate for the top quintile by earnings approaches 76%, drops to approximately 68% for the second quintile, 58% for the third quintile, 40% for the fourth quintile, and 18% for the lowest quintile.36

There are two explanations for low coverage rates: some employees work for firms that do not sponsor a plan, and some employees do not participate in plans sponsored by their firms. There are three reasons for the latter explanation: (1) regulations permitting firms to exclude employees with certain characteristics (for example, part-time employees, employees covered by collective bargaining agreements, employees with less than a year of service, and employees younger than age twenty-one);37 (2) regulations permitting firms to develop additional criteria for plan coverage;38 and (3) regulations permitting firms to establish plans – primarily 401(k) plans – under which employees must elect to receive reduced current wages in order to participate.39

their plans for employees who have not worked at least five years.42 Because average job tenures decline with income level, such forfeiture provisions may disproportionately affect low- and moderate-income workers.43

According to one extrapolation of 1998 Social Security data, private sector pension plan benefits provide 29.8% of the retirement income for the top quintile of income of the population above age 55; 28.1% for the second quintile; 16.1% for the third quintile; 6.8% for the fourth quintile; and 3.3% for the fifth quintile.44

Thus, there are two issues that subvert the rationale for subsidizing qualified plans: low- and moderate-income employees have low rates of coverage and they earn relatively low levels of benefits when they are covered. The two issues might be viewed collectively as one of low effective coverage, i.e., coverage that results in meaningful levels of benefits for low- and moderateincome workers.

added a new provision to the Internal Revenue Code that required every plan to cover at least the lesser of 40% of the workforce or fifty employees.48 To increase benefits for rank-and-file employees, Congress limited the degree to which a plan’s benefit formula could be integrated with Social Security;49 created a category of plan – the top heavy plan, in which benefits are concentrated in the accounts of “key employees” – that requires accelerated vesting standards and at least a minimum benefit for all non-key employees;52 accelerated the statutory vesting standards for plans generally;51 and imposed caps on the compensation that could be considered in benefit formulas.52

We want to emphasize that during this period Congress did not create important additional incentives to encourage plan formation. Moreover, the addition of new statutory provisions to improve effective coverage came during an era when actual coverage rates dipped somewhat, and it has been suggested that the coverage dip was, in part, a response to the cost of the new regulations.53 It also, however, has been suggested that the coverage dip occurred primarily because of the severe economic downturn in the middle of the 1980s.54 In any event, the coverage rates crept back to the 50% mark in the 1990s, possibly attributable to the strong economy we enjoyed during that decade.55

bottom two earnings quintiles because we have no statistics on what we earlier referred to as the effective coverage rate. We do not know, for example, how many ofthe people who lost coverage during the 1980s were actually accruing more than trivial benefits or whether they were vesting in the benefits they did accrue. It seems likely that the firms most sensitive to the costs of the new regulatory measures, and thus most likely to drop their plans in response to tightened regulation, would have been those firms whose plans provided little more than nominal coverage to their low- and moderate-income employees. It is thus conceivable that the decline in nominal coverage rates was due principally to firms dropping plans that failed to satisfy the purpose of the tax subsidy, at least as we have defined it for purposes of this Article.

Similarly, Treasury proposed in 1989 and promulgated in 1991 nondiscrimination regulations that tolerate if not encourage substantial disparities in benefit accruals between the highly-compensated employee and the non-highly compensated employee, particularly in smaller plans.63 At least with respect to smaller firms, the regulations often permit the construction of plans that maximize benefits for the highly compensated while minimizing benefits for low- and moderate-income workers. The so-called “cross-testing rules,”64 permitting some firms to establish defined contribution plans in which the most highly compensated employees receive annual account allocations that are more than twenty times the allocations provided for lower-paid employees, have been the most visible and controversial part of the regulations enabling firms to effect such goals.65

It is difficult to assess the net effect of the legislative and regulatory modifications of the nondiscrimination rules during the 1980s, but our sense is that these opposing forces probably left the overall effective coverage rate for the lower two quintiles about where it was at the beginning of the 1980s, despite the congressional agenda to improve benefits for low- and moderate– income workers. Had the regulatory agenda been consistent with congressional policy decisions, it is possible that the level of effective coverage would have increased even though the nominal coverage levels may have declined.

crafted to cajole new plan sponsorship, however, will likely have the effect of reducing benefit adequacy for low- and moderate-income workers because the approach is me of reduced regulation.67

Congressman Benjamin Cardin, an architect of this approach, wrote a legislative package, now largely enacted, that he helped design, with the goal that

H.R. 10 will help extend the opportunity for tax-favored retirement savings for workers in small businesses. To date, only a small proportion of small businesses have set up retirement savings plans for their workers.

Among companies with fewer than 100 employees, 80% of the workforce has no pension or retirement plan. Compared to large companies, where 75% of the work force has a retirement plan, this demonstrates the urgent need to make it easier for small businesses to set up retirement savings plans. H.R. 10 will remove burdensome regulations that have made it difficult for small businesses to give their workers the opportunity to save for retirement.68

Burdensome regulations make a convenient windmill for any brave legislative knight. Indeed, everyone can agree that regulations should be simplified when a regulatory burden is generated by complexity alone, i.e., when plan sponsors have to pay consultants to decipher the meaning of a regulation’s requirement or to determine whether they are in regulatory compliance (“first, let’s pay all the lawyers”). But the regulations that Congress recently eliminated or softened did not impose complexity burdens of this variety. The burdens they imposed were simply the financial costs of providing nontrivial levels of benefits for low- and moderate-income employees.

satisfy the testing, impose administrative burdens an the sponsoring firm, but a firm may design its plan and procedures to simplify these processes.71

The 1996 legislation permitted employers to eschew the testing process if the plan design provides that the employer will make matching contributions for the employee equal to 100% of the first 3% of compensation that the employee voluntarily contributes to the plan, and equal to 50% of the next 2% of compensation.72 For small employers, the same legislation created SIMPLE, where the plan satisfies the nondiscrimination rules if it provides a match for only the first 3% of compensation. These types of plans are easier to administer than traditional 401(k) plans, but the result will be smaller benefits for low- and moderate-income employees in many plans.

The rationale for these safe-harbor nondiscrimination requirements is that they reduce regulation without reducing benefits because empirically we know that employees at all income levels respond to employer matches in traditional 401 (k) plans.74 But the matching requirements, which permit the employer to condition ultimate receipt of the match on three years of vesting service, were arbitrarily selected without empirical verification that the matches would be adequate to stimulate employee contributions.

If the effect were only to induce some firms that do not have plans to adopt new plans, the overall effect on retirement savings for moderate-income employees would, of course, be positive, even if small. The issue then would be simply whether the additional coverage was worth the tax cost of these new plans. But when Congress created these new types of plans, it also encouraged firms with traditional 401(k) plans and firms with employer-funded retirement plans to consider substituting a SIMPLE or safe-harbor 401(k) plan for their existing plan. If firms do this, an actual decline in the retirement savings of their moderate-income employees might occur. In order for the aggregate effect on moderate-income employees to be positive, the number of employees who were not covered by any plan prior to their firm’s adoption of a SIMPLE or safe-harbor 401(k) would have to exceed the number of employees who save less as a result of their firms replacing an existing plan with a SIMPLE or safe-harbor 401(k) plan.

There is some reason to believe that negative effects – lost savings – may predominate. First, most firms that do not sponsor plans say they lack sufficient business profit to justify adopting a plan.75 Firms in which business profits are low typically will not have any highly-compensated employees. These firms were within an effective safe harbor before the introduction of the SIMPLE and safe-harbor 401(k) plan because plans with no highly compensated employees automatically pass nondiscrimination testing. The SIMPLE and safe-harbor 401(k) plans add little to the mix of incentives for such firms. Second, the recently enacted Economic Growth and Tax Relief Reconciliation Act of 2001 (EGTRA) increases the amounts that employers and employees can contribute to SIMPLE and safe-harbor 401(k) plans, making them attractive to a wider range of firms already sponsoring these plans.76

EGTRA also increases the deductibility limits for individual retirement plans to $5,000, providing some small business owners with an alternative to adoption of a qualified plan that requires them to provide some benefits to their employees.77

in which “key employees” have accumulated 60% or more of the aggregate benefits under the plan – must provide accelerated vesting and a minimum benefit for each plan participant.78 For defined contribution plans, the minimum benefit is generally equal to 3% of a participant’s compensation.79 In many 401(k) plans and in so-called “new comparability” plans, the sole reason that most non-highly paid employees are accumulating even a marginally meaningful benefit is this minimum contribution rule.80 (In new comparability plans, key employees may be receiving benefits in excess of 20% of their compensation.81) EGTRA makes several definitional changes that will reduce the number of plans that must provide minimum benefits to non-key employees.82 It also permits top-heavy plans to credit employer matching contributions toward the minimum benefit requirement.83

EGTRA weakened another Internal Revenue Code provision – the compensation cap – which is designed to increase benefits for all but the highest paid employees.86 The compensation cap limits the amount of compensation that an employer may consider in a plan’s benefit formula. The effect of the compensation cap generally comes from the intersection between a plan’s benefit formula and the dollar limits that Internal Revenue Code (sec) 415 imposes on contributions and benefits.’ Before EGTRA, the statute set the compensation cap at $150,000 with adjustments for cost of living increases.88 In the year 2000, the inflation-adjusted index reached $170,000.(89)

In 2000, a firm could contribute a maximum amount of $30,000 to a defined contribution plan.90 Assume that a firm’s owner had $200,000 in compensation and that the firm’s owner wanted the maximum $30,000 contribution to his account. Also assume, for purposes of simplicity, that the Internal Revenue Code required the firm to contribute the same percentage of compensation for each plan participant.91 Because the compensation cap forces the plan to treat the firm owner as if he were earning $170,000, the plan would have to use a 17.65% contribution rate for all participants in order to provide the owner $30,000.(92)

this would be to reduce the contributions to the other plan participants to 15%.(94)

EGTRA made other changes that will permit or encourage some existing plans to reduce benefits for low- and moderate-income employees.95 Realistically, none of these changes are likely to result in substantial new plan sponsorship. To the extent that the changes do encourage the formation of new plans by lowering the amount of benefits that plans must provide to low- and moderate-income employees, those plans will provide lower levels of benefits to such employees than the Internal Revenue Code has required previously. Moreover, another likely effect of these EGTRA changes will be the reduction of benefits for low- and moderate-income employees in many already-existing plans.

The only provision aimed directly at helping low- and moderate-income workers is a government matching credit for certain lower-income individuals who elect to contribute to an employer cash or deferral plan or to an individual retirement account.96 The maximum credit is 50% of the amount contributed (up to $2,000 per person or $4,000 for married couples), which applies to individuals with $15,000 or less in adjusted gross income.97 The credit then drops to 20% for individuals with income over $15,000 but less than $16,250, and then drops to 10% for those with up to $25,000 of income.98

The legislative decisions reflected in recent pension legislation mark a retreat from the traditional carrot/stick blueprint for the tax treatment of qualified plans, that of encouraging plan sponsorship through tax incentives for the highly-paid and of forcing plans to provide the social benefit of retirement savings for lower-paid employees through regulation. As Professor Bruce Wolk explained in 1982:

As the discrimination rules require more in the way of contributions for lower paid employees, the employer’s costs increase. For any given employer, the costs may eventually exceed the benefits of covering the highly paid employees. At that point, the employer would decline to establish or continue a retirement plan. Thus, an aggressive congressional stance against discrimination might effectively preclude many lower paid employees from receiving retirement benefits….

Congress could avoid the adverse effect of aggressive discrimination rules by designing rules to ensure a high level of tax subsidy in relation to employer costs. Presumably this would result in a larger number of employers establishing or maintaining plans. Rulesbrngmng about this result, however, would risk wasting the tax subsidy. To the extent that such rules would encourage employers to establish plans by excluding lower paid employees, the subsidy would be applied ineffectively.

From Congress’s perspective, the optimum level of tax subsidy is that which encourages the establishment of a refinement plan only if the social benefit of the plan equals or exceeds its costs.108

Congress has moved in the direction of higher subsidies for the highly paid and of less regulation – sweetening the carrot and softening the stick.109 As Dan Halperin has noted, this will result in more but worse plans.110 If the only result were the addition of new plans, the only problem would be cost: are we getting enough social benefit given the tax costs of such new plans? Because of the softening of regulations, however, a reduction of benefits for lower-paid employees in existing plans also will result.

for retirement. Another possible explanation for the qualified plan tax subsidy is that it introduces a consumption tax element into our tax system, providing a partial balance to the perceived bias of an income tax against savings.111 A third possibility, although one that by its nature cannot be explicitly acknowledged in political discourse, is that the qualified plan subsidy is designed to indirectly provide a reduction in the effective tax rates of relatively highincome taxpayers. This explanation assumes that providing the reduction indirectly through qualified plans accords political cover to what could not be legislated directly. Finally, one might argue that the tax subsidy primarily is intended to increase retirement savings for the upper strata of the middle class. The plan participation rate for the upper-income band of the middle class is high (approximately 80%), and it is reasonable to think that some if not most of the high utilization is the tax subsidy embedded in qualified retirement plans.

We accept that the alternative explanations for the subsidy outlined above already play some role in the political sustainability of the qualified plan subsidy. But one of the explanations – the idea that the subsidy simply provides a tax rate reduction – is seldom acknowledged, and for the past forty years, none of the alternative explanations have been nearly as instrumental to the discussion as what we have sometimes called the qualified plan paradigm. This explanation has been the most important for the past forty years in shaping the intellectual and rhetorical landscape that provides the tax subsidy its public justificatory context.112 Indeed, the qualified plan paradigm is so dominant that the sponsors and supporters of recent pension legislation argue that such proposals are consistent with this paradigm, even though the proposals could be justified more easily under one or more of the other explanations.113

ers and thus are consistent with the traditional qualified plan paradigm, the congressional champions of EGTRA close off serious consideration of other measures that might in fact have expanded coverage and enriched retirement benefits for these workers. If instead they had sought to justify their proposals on the basis of one of the alternatives for the qualified plan subsidy, we might have expected two positive political outcomes: explicit discussion of the merits and costs of their suggested justification (whatever it might be), and consideration of proposals to help low- and moderate-income workers build retirement security apart from the universe of employer-sponsored plans. Inevitably, the latter consideration should include proposals to modify Social Security in this direction.

III. The Problems of Linkage

Certainty and understanding are virtues in a retirement program whether or not it is designed to provide an old age or disability pension, and whether or not the program provides income replacement or in-kind benefits, such as medical benefits. Participants who do not understand what they are promised, or who cannot rely with certainty on the promises made, may reach a time of dependency with inadequate resources – a time when it is too late to make alternative arrangements. This Part of the Article considers this idea of linkage between employee understanding of a plan’s benefit and the employee’s ultimate receipt of the expected benefit.

The problems of linkage occur when participants believe employers are promising something different from that for which they have an enforceable contractual right. Despite the goal of the Employee Retirement Income Security Act of 1974 (ERISA) of ensuring the certainty of the benefit promise, the problems of linkage continue to be a fixture in the retirement benefits landscape. This Part of the Article reviews some linkage problems that haunt participants in private pension plans.

A. The Reservation of Rights Clause and Retiree Health

Competently designed employee benefits plans generally include a waiver of rights clause under which the sponsoring firm can modify or terminate the plan at any time. Such clauses sometimes begin with strong endorsements of the plan and then disavow any obligation to employees to continue to maintain the plan. For example, General Motors’ retiree health plan included the following language:

General Motors believes wholeheartedly in this Insurance Program for GM men and women, and expects to continue the Program indefinitely. However, GM reserves the right to modify, revoke, suspend, terminate, or change the Program, in whole or in part, at any time.115

A law review note written in 1940 referred to such clauses as weasel clauses.116

coverages will be provided at GM’s expense for your lifetime.”120 Moreover, a large group of the retirees who retired under special early retirement programs also received additional oral and written communications promising them retiree medical care.121 Many of them signed early retirement contracts after receiving a description of their retirement benefits that included the continuation of health care coverage at no cost to them.122 Finally, according to a lawyer for the participants, a long history of uninterrupted free retiree health care, continuous improvements in such health care, and a corporate ethos and employee culture of which retiree benefits were an important part, all reinforced the employee belief that the rights were permanent.123

they believed included health benefits. But the Sixth Circuit reversed in part and instead held that the reservation of rights clause trumped all other representations and expectations because it was part of the written plan document. The Sixth Circuit noted:

ERISA “has an elaborate scheme in place for beneficiaries to learn their rights and obligations at any time, a scheme that is built around reliance on the face of written plan documents.” To implement this scheme, ERISA requires that every plan “shall be established and maintained pursuant to a written instrument.” 29 U.S.C. (sec) 1102(a)(1). ERISA also requires, as we have said, a written summary plan description that will “reasonably apprise … participants and beneficiaries oftheir rights and obligations under the plan.” 29 U.S.C. (sec) 1022(a).

The writing requirement ensures that “every employee may, on examining the plan documents, determine exactly what his rights and obligations are under the plan.” And the requirement lends predictability and certainty to employee benefit plans. This serves the interests of both employers and employees. “Congress intended that plan documents and SPDs exclusively govern an employer’s obligations under ERISA plans.” We recognize that “It]his may not be a foolproof informational scheme, although it is quite thorough. Either way, it is the scheme that Congress devised.”

Our court has consistently refused to recognize oral modifications to written plan documents. “[Ve are quite certain,” we have explained, “that Congress, in passing ERISA, did not intend that participants in employee benefit plans should be left to the uncertainties of oral communications in finding out precisely what rights they were given under their plan.” Therefore, the “clear terms of a written employee benefit plan may not be modified or superseded by oral undertakings on the part of the employer.”125

document over other forms of communication contradicts basic behavioral rules governing communication in the workplace.127

Reservation-of-rights clauses, even when employees are aware of them and understand that a firm might exercise its rights thereunder, create problems of certainty for employees trying to plan for their retirement security. When such clauses exist, employees cannot depend on employer-provided health benefits still being available by the time they retire, and they cannot depend on their indefinite continuation after they retire. Rational employees would make other arrangements for health care before they retire because once they retire their fixed financial resources will constrain their ability to engage in other arrangements.128 Employees who respond in this manner would probably attach little or no value to the employer’s suggestion that it will provide its retirees with health care. Receipt of such benefits, then, might be regarded by lucky retired employees as a windfall. In such an environment, a firm probably would not make such quasi-promises to its employees.

B. The Implicit Bargain in Defined Benefit Plans

an employee ages. There are two reasons for backloading. First, the value of a dollar of promised retirement income is directly related to the length of the discount period – the interval between benefit accrual and retirement.129 Thus, the older the employee, the shorter the discount period and the greater the value of a dollar’s worth of benefit. Second, many defined benefit plans are based on a formula incorporating the employee’s final pay – for example, 1% of final pay times years of service.130 An increase in compensation for one year increases the value not only of that particular year’s nominal benefit accrual but also of the nominal benefit accrual of all prior years.

The firm, however, has the legal right to terminate a defined benefit plan or to modify its benefit formula prospectively.132 The financial gain to the employer who does so can be substantial. While an employee’s accrued benefit at the time of a termination or modification is protected, the employee is deprived of the implicit bargain the employer offered, that is, the substantial benefits the employee would have earned during the last period of employment.133 The loss to a middle-age employee can be substantial.

C. Other Defined Benefit Plan Issues

There are two other significant linkage problems in defined benefit plans. One involves the subsidized early retirement benefit included in some plans for people who retire after a certain age or after a specified number of years of service.134 An employer can amend the plan, however, to eliminate or reduce such subsidies for most employees.135 For example, a plan might provide that an employee with thirty years of service can retire at age fifty-five.136 An employee who is age fifty-five with twenty-five years of service begins planning to retire at age sixty, but in the interim, the employer amends the plan to eliminate the benefit. While the employee still can retire at age sixty with the benefit calculated based on his service and salary as of age fifty-five, the firm has reduced the employee’s age-sixty retirement benefit.

Corporation (PBGC) takes over the plan and pays participants guaranteed benefits.137 That can be far less than the employee’s vested accrued benefits.138

D. Standard of Judicial Review of Benefit Denials

The Supreme Court in 1989 resolved a circuit split over whether a court reviewing a benefit denial should accord deference to the plan administrator’s factual findings and plan interpretations.139 The Court held that judges should defer to a plan administrator if the plan’s language vested the plan administrator with discretionary authority to decide benefit claims.140 This approach to judicial review of benefit denials creates linkage issues because a participant is not entitled to the court’s interpretation of a plan’s provisions or resolution of factual issues; the participant only is entitled to the court’s determination of whether the plan administrator’s interpretation of the plan or findings of fact is unreasonable.

that a plan administrator did not behave arbitrarily in denying disability benefits to a participant whom the Social Seauity Administration has found to be totally disabled.142

E. Obscure Plan Provisions

Employee benefit plans sometimes include provisions that are not understood by plan participants and have the effect of reducing the benefits to which employees believed the plan entitled them. We offer two examples from the files of a pension clinic at the University of Alabama School of Law.

benefits. He ultimately settled his worker’s compensation claim for a lump sum payment of $50,000, 20% of which went to his lawyer. The offset formula in the plan reduced his monthly pension to zero. The participant was unaware that his settlement would have any effect on his pension.

In the second case, a large national employer sponsored a defined benefit plan.144 The plan’s benefit formula used a multiple of years of service and the average compensation during the highest consecutive five years of pay during the most recent ten years of employment. A participant in the plan took parttime status, reducing her pay by approximately 50%. At the time she took part-time status her accrued retirement benefit, a monthly annuity that began at age sixty-five, was approximately $800 per month. In her sixth year of parttime status, her accrued benefit began to drop in value as the ten-year reference period began dropping off her full-time compensation years. Her benefit ultimately declined in value to approximately $500 per month despite her additional years of service for the firm. Her benefit would have thus been greater had she quit rather than continued to work part-time. She was unaware that her decision to work part-time could reduce her already-accrued pension benefits, although she might have been able to determine this before she took part-time status if she had had a pension consultant review the plan document, Such cases, although perhaps not common because they involve a plan provision and an employee adversely affected by it, can result in devastating effects on affected individuals.145

F. Theft and Mismanagement of Plan Assets

Others involve churning by investment managers.148 Others involve firms, often capital-strapped, dipping into plan assets or delaying payment of employee elective deferrals to 401(k) plans.149 Some involve plan managers selling fraudulently valued company stock to the plan or the plan selling it to plan managers.150

The consequences to participants of a decline in plan assets attributable to mismanagement or theft vary depending on the circumstances. If plan fiduciaries or plan participants learn of a resulting decline in plan assets, they can bring a civil action to recoup plan loss.151 They also may bring the loss to the attention of the Department of Labor, which can investigate152 and, if necessary, initiate civil proceedings against the wrongdoers153 or refer the matter for criminal prosecution.154 In some cases, however, the firm and its principals may lack the ability to make the plan whole, or procedural or practical obstacles may prevent correction.155

pant benefits will be reduced by the amount they exceed PBGC guarantees.156 Participants may also suffer indirect reductions to fixture benefits if the plan sponsor reacts to a loss in plan assets by either t the plan or amending the plan to reduce the rate of future benefit accruals.

G. Some Thoughts About Linkage and Defined Contribution Plans

In the past two decades, defined contribution plans have replaced defined benefit plans as the most common form of retirement vehicle.157 Defined contribution plans generally do not create linkage problems; the employee receives exactly what is in his or her account, and there is no difference between what the employee expects to receive from the plan and what the employee in fact receives from the plan. However, the deeper structural issue in linkage is not the disparity between what employees believe they have been promised and what they actually have been promised, but rather it is the gap between what the employee perceives to be the value of his or her benefits and the actual value of those benefits. From this perspective, defined contribution plans pose an issue similar to leakage because of a tendency for employees to overvalue the benefits they will receive from defined contribution plans.158

There are two ways in which employees apparently overvalue defined contribution accounts. First, many employees overestimate the rate of return they are likely to achieve on their accounts.159 Second, employees tend to overestimate the amount of retirement income their account will provide once they retire.160 Educational efforts, of course, may counter these effects to a certain extent.

IV. The Problem of Leakage

death. In the case of married participants, a plan would provide a periodic and constant annuity until the later of the death of the participant or his spouse, perhaps with an appropriate downward adjustment after the first death to reflect the reduced expenses of a one-pension household.161 Each year, however, retirement plans pay billions of dollars of benefits that are not so applied. Participants either spend these funds before retirement or exhaust these funds too quickly in retirement.

The most cited form of leakage is the lump sum payment of pension benefits when a participant leaves employment or takes an in-service distribution from a 401(k) or other profit-sharing plan.162 Plans may choose to pay participants cash when they leave if the benefit has a present value of less than $5,000 or ifthe employee consents.163 Statistics derived from the 1993 Current Population Survey indicate that 20% of the population who received lump sums rolled over the entire amount into an IRA or other qualified plan, 40% rolled over part of their distribution, and 40% did not roll over any part of their distribution.164 In sum, participants rolled over approximately two-thirds of the total value of the distributions. Hewitt Company, in a survey of its 1996 database, indicated that 40% of participants rolled over their distributions, representing 79% of the total distributed assets.165 One should also note that not all of the money rolled over into individual retirement accounts will stay there because amounts in such accounts are easily accessible to the IRA owner.166

Using the mid-range assumption, the thirty-year old’s failure to preserve a $5,000 distribution causes a loss of almost $75,000 of retirement benefits.

Congress has attempted to control leakage, in part, through the assessment of a 10% excise tax on premature plan distributions that are not rolled over.168 A distribution is generally premature if it is made to a participant prior to the year in which the participant attains age 59.5.(169) Over the last fifteen years, however, Congress has carved out exceptions to the penalty tax for plan and IRA withdrawals to pay for college tuition,170 to pay health care expenses,171 or to help pay the down payment on a first home.172

One can argue that pre-retirement leakage from safe retirement plans is not invariably a bad thing. Purchasing a home can be an important investment for retirement, as can reducing high interest debt. If the home purchase or debt reduction would not otherwise have taken place, perhaps permitting access to retirement finds for such purposes is defensible policy. Moreover, one can defend allowing access to retirement savings to pay for a child’s education on general policy grounds. To the extent it will enhance the child’s lifetime earnings and thus put the child in a better financial position, the child’s education might be considered an asset that enhances the parent’s old age security.

Some commentators argue that employees would be less likely to make elective deferrals to 401(k) plans if they could not access their accounts in times of financial stress.175 Thus, in a voluntary retirement system in which employee willingness to participate is necessary to the system’s viability, it is essential to fashion a policy compromise between locking up benefits until retirement and encouraging voluntary participation by giving employees preretirement access to their benefits in certain circumstances.

Discussion of leakage generally focuses only on participant access to retirement plan assets prior to reaching retirement age. However, if the purpose of qualified retirement plans is to ensure adequate income in retirement, premature exhaustion of benefits (before death) or failure to exhaust assets in retirement (by death) are also forms of leakage. The former is a source of leakage if we conceptualize the idea of retirement security as a method of providing a sufficient and generally steady stream of income after an individual permanently leaves the labor force because of age or disability. Front– loading consumption by drawing down financial resources early in retirement is inconsistent with this goal, and thus, one may characterize it as a form of leakage. Moreover, in an economy in which some level of inflation is a permanent feature and in a world in which expenses, particularly medical expenses, increase with age, some degree of backloading of retirement benefits may be necessary to maintain a stable standard of living. In addition, dying without exhausting retirement resources and leaving the excess assets to nondependent heirs is arguably a form of leakage from a system designed to provide retirement income.

Accepting such limits, however, suggests that the only way to eliminate postretirement leakage is through mandatory, inflation-indexed annuitization of retirement benefits for the lifetime of an individual and, in most cases, the individual’s spouse or domestic partner.

Recent pension legislation and trends suggest that some of the problems of post-retirement leakage are worsening and will continue to worsen. In the last two decades, there has been a shift from defined benefit plans, in which annuitization is common, to a defined contribution world in which it is not.116 Moreover, the creation of cash-balance and similar types of defined benefit plans, which state benefits in the form of a notional account balance rather than a life annuity, has increased the likelihood of a cash-out on separation of service and has decreased the likelihood of annuitization on retirement.177

The Economic Growth and Tax Relief Reconciliation Act of 2001 (EGTRA) includes provisions that will exacerbate leakage problems. First, EGTRA greatly increases the attractiveness of profit-sharing plans over money-purchase plans, making it likely that many firms will abandon the latter plan form.178 Money purchase plans, however, include an important anti-leakage feature: a married participant generally must obtain spousal consent before taking a benefit distribution in any form except a qualified joint-and-survivor annuity.179 Spousal consent to a lump sum payment is not mandatory in a profit-sharing plan.

There is also some encouraging news about leakage. First pre-retirement leakage may diminish as educational efforts stress the value of saving for retirement in tax deferred vehicles.181 Second, EGTRA does include some provisions that make it mildly easier to roll over assets between different forms of tax-qualified plans, although the rules have long permitted roll overs to individual retirement accounts and annuity contracts.182

Overall, however, the problems of leakage are difficult ones for our private pension system. It is difficult to envision a strong and politically viable natural constituency supporting legislative adoption of controls on leakage from the private pension system.

V. Social Security Privatization

The problems of leverage, leakage, and linkage in private sector retirement plans have implications for the debate over suggestions to privatize Social Security. All three issues represent gaps in the private pension system for which the Social Security system, as currently structured, constitutes the public response. Privatization of Social Security would greatly weaken or, in some cases, eliminate the campe features of the current system – mandatory and universal coverage, life annuity payments with no lump sum or early withdrawal possibility, and guaranteed wage and price indexed benefits backed by the federal taxing power. Fully or partially replacing the current defined benefit structure with a system of personal IRA-style accounts would simply import the gaps of the current private pension system into the Social Security system, inevitably resulting in much less retirement security for most workers.

A. Loss of Leverage in the Private Pension System

make up for the lack of private pension income, and the situation is likely to be worse under any privatized version of Social Security.

Social Security is currently the sole or primary source of retirement income for half of all Americans over age sixty-five, as well as for many young families relying on survivor benefits in the wake of the loss of the family wage earner.184 The low level of benefits available under Social Security, particularly for elderly retirees and even more particularly for elderly women, generally relegates those relying primarily or solely on Social Security to living at or below the poverty line in old age.185 It is very unlikely that privatizing Social Security would improve the situation of the aged poor, who, as this Article suggests, are almost sure to receive even less as a group from private sector retirement plans in the future than they have in the past.186

working lifetimes, are probably equally unlikely to see their retirement incomes improved or even maintained at current levels in any private account retirement system. Many in this second cohort of retirees may have enough private pension and savings resources that, when combined with Social Security, allow them to live comfortably above the poverty line in the context of the current system.187 However, the features of the current system that most strongly underwrite the maintenance during retirement of a middle-class preretirement standard of living and the survival at the poverty level of poorer workers in retirement are generally absent from a privatized retirement system model. Furthermore, additional changes suggested to finance the transition to a privatized system would only further decrease the adequacy of Social Security benefits that even now are barely adequate to compensate for the gaps in the private retirement system.

First, both low- and moderate-income workers and retirees benefit from the redistributive features of the Social Security benefit structure.188 Over the years, there has been substantial criticism of this redistributive feature, both from those who would like to see greater redistribution of income within a broader social welfare system189 and from those who largely base their advocacy of privatization on their opposition to income redistribution in principle.190 Nonetheless, the weighting in the benefit formula, which provides a higher than proportional benefit for low-wage workers and a lower than proportional benefit for high-wage workers, remains one of the few mechanisms in American social welfare programs for direct income redistribution in a largely politically successful context.

private pension benefits.191 This pattern is inextricably linked to the principle of leverage – without the possibility of, as it were, reverse income distribution, highly paid employees would have little incentive to participate in or to pressure employers to establish qualified pension plans. However, without the redistributive features of Social Security that provide at least some retirement income protection for low-wage workers, the degree of coverage for such workers that a privatized system might as a consequence require of private pensions would be difficult to achieve under the current tax structure.192

Second, the indexing of wage records to increases in average wages during the worker’s lifetime assures the relationship of the Social Security benefit to the worker’s pre-retirement income level.193 Similarly, yearly automatic costof-living benefit increases preserve the purchasing power of the benefit in retirement necessary for maintenance of the standard of living established at retirement.194 Neither feature exists in private defined benefit pension systems, and there is no real equivalent to insure adequate old age income in the private account systems suggested as replacements for Social Security.

Moreover, any tax deferral inherent in privatized investment accounts will work to the relative advantage of high-income workers whose Social Security accounts would receive a larger tax subsidy than workers who pay tax at low marginal tax rates.196 Thus, there are reasons to suspect that the benefit structure under a privatized account system will be regressive in practice, thus importing into Social Security one of the main deficiencies of the private system’s reliance on leverage to increase coverage of low-wage workers.

It might be argued that although this issue is genuine, preserving an adequate minimum benefit can mitigate it. However, preserving an adequate minimum benefit for the poor may face political obstacles because the larger the minimum benefit is, the smaller the size of guaranteed benefits for other beneficiaries is likely to be. Moreover, a minimum benefit approach to preserving adequacy completely ignores the essential, broad-based middle class focus of Social Security, which the government designed to prevent poverty in old age by linking benefits to lifetime wage patterns rather than to alleviate it for those who are so poor in old age as to qualify for any minimum benefit that Congress might enact.197

It is perhaps more troubling that proponents of privatization have suggested paying for privatization with Social Security benefit decreases. Two of these decreases would greatly affect the elderly poor even if the current minimum benefit remained intact and also would substantially decrease benefits for middle-class retirees, resulting in a much larger group of poor elderly than currently exists.

terms.199 Finally, even those who are not “poor” when they initially retire (because they have other sources of income such as a private pension and savings) generally come to depend more and more on Social Security as a source of income precisely because Social Security benefits are indexed for inflation while other sources are not.200 One should regard changing the indexation formula as an undesirable benefit cut, not simply a “technical correction,” that could have dramatic effects in generally increasing poverty rates for the elderly.

The second suggested decrease would raise the Social Security normal retirement age.201 Of course, calling this change “raising the retirement age” is really a misnomer because mandatory retirement is no longer legal. The only change Congress really can make is to reduce benefits for those who continue to retire at the current early or normal retirement ages, which Congress accomplished in the 1983 Social Security Amendments.202 Workers are free to continue to retire at age sixty-two or at age sixty-five, but the benefits for those retiring in 2000 and later will be lower than under prior law.203

the normal retirement age will result in a reduction of monthly income support. In addition, even for middle-class workers, the possibility of illness, injury, or layoffs forcing retirement at or before age sixty-two is very real, so that any decrease in benefits designed to induce workers to stay in the labor force longer is just as likely to push middle-class workers into poverty as it is to “raise the retirement age.”

The final characteristic of Social Security that most directly contrast the private pension system’s need for “leverage” is the extent of coverage – the Social Security system is virtually universal and mandatory.206 One cannot overstate the importance of universal mandatory coverage and complete portability from job to job, particularly in contrast to the voluntary nature of the private pension system. The leverage feature encourages pension formation by employers with stable, highly-paid work forces and has inevitably produced only partial coverage, largely leaving out those groups most in need of supplemental old age income – the working poor, the underemployed, and those who work no longer than two or three years at any one job.207

It may well be possible to construct a mandatory and universal privatized Social Security system if the current payroll tax withholding structure could bear the administrative burden of collecting contributions. However, the lower overall level of retirement income that would likely result from a private account system would vitiate the effect of universal coverage.

B. Leakage

assets by death. In addition, the period of retirement should extend to the second death for a married couple, in recognition of the reality of household planning and saving for retirement.

In the private pension system, it is likely that the problem of pre-retirement leakage will at least continue and probably increase, although educational efforts focused on the importance of preserving savings for retirement and the increased proclivity toward retirement savings of an aging workforce might mitigate the problem somewhat. Nonetheless, the trend toward cash balance defined benefit plans, defined contribution plans, and particularly 401(k) plans, in which the concept of voluntary salary deferral fosters a sense of immediate ownership and access, as well as recent legislative changes encouraging some firms to abandon their money purchase pension plans, will create additional leakage pressures on the system.208

The increasing prevalence of cash balance and defined contribution plans will also contribute to other leakage issues, particularly the failure to annuitize. Furthermore, indexation of benefits, such as post-retirement cost-ofliving increases, in defined benefit plans, which commonly took place on an ad hoc basis through the mid-1980s, is now a rarity.

Social Security, in contrast, is a largely leak-proof system, with no opportunity for pre-retirement consumption of benefits (for which there is little pressure because the program provides survivor and disability benefits independently of the retirement benefit, as well as mandatory annuitized benefits on retirement).209 Moreover, Social Security provides mandatory spousal benefits and indexes all benefits to the cost of living.210 By design, Social Security’s benefit structure and method of payment result in the provision of retirement income, not estate building opportunities, and thus the life annuity form, with no remainder to go to heirs, normally guarantees consumption of benefits during the retiree’s lifetime.

Many commentators have raised the issue of the administrative costs of maintaining and investing private accounts as the hidden cost of privatization, and one should view the issue as a possible major source of pre-retirement, as well as post-retirement, leakage in any privatized version of Social Security. The contrast between low likelihood of leakage in the current Social Security program and the high probability of it in a privatized program should not be a surprise for the following two reasons: the current program’s earnings– based entitlement structure and its lack of need for generation of profits for the managers.

First, the costs of administering the Social Security system are quite low – about one-half of one percent of the total outlay in benefits each year, a figure which covers salaries of Social Security Administration employees and maintenance of hundreds of district offices all over the country (computers, etc.).211 Social Security is mostly an automatic program with few individual entitlement decisions to make (apart from the disability program). Because past earnings recorded through the payroll tax withholding system throughout workers’ careers determine benefits, there is little need for personal meetings between the agency and beneficiaries and little need or scope for judgments on individual cases. Moreover, the only costs associated with running the program are government salaries; nobody gets rich from running a Social Security district office or managing the trust funds.

private account returns was sufficient to induce the chairman of the last Social Security Advisory Board to tell Congress that he could no longer support a private account plan.214

One response to these concerns might be simply to insist that each individual worker can and should manage her own retirement savings in an individual account system.215 This perspective, we suggest, is cynical in the extreme. Most Americans do not have the resources, the expertise, or even the time to make informed judgments on the constant, everyday level that would be necessary to ensure that their individual accounts would increase in value enough to support their retirement. Even during the heyday of the bull market, untrained investors, such as day traders who used computer trading accounts as a new form of casino gambling, frequently lost family savings and went into debt, essentially eliminating their own retirement incomes.216

rumors or that elusive measure “consumer confidence” without any corresponding changes in company earnings or future prospects.217 According to Robert Shiller of Yale:

The high recent valuations in the stock market have come about for no good reasons. The market level does not, as so many imagine, represent the consensus judgment of experts who have carefully weighed the longterm evidence. The market is high because of the combined effect of indifferent thinking by millions of people, very few of whom feel the need to perform careful research onthe long-term investment value ofthe aggregate stock market, and who are motivated substantially by their own emotions, random attentions, and perceptions of conventional wisdom. Their all-too-human behavior is heavily influenced by news media that are interested in attracting viewers or readers, with limited incentive to discipline their readers with the type of quantitative analysis that might give them a correct impression of the aggregate stock market level.218

Of course, Shiller wrote these words before the decline of the market beginning in early 2000, which has resulted in a 15-20% drop in overall market share prices. In a bear market, the shortcomings of a retirement system relying solely on individual savings and investments are more glaringly apparent.

A second “leakage” problem in a newly privatized social security system would be the possibility of access to the funds in an account either before retirement or during retirement in a way that would endanger the life-time income stream necessary to support the individual until death. Most, but not all, privatization advocates eschew the concept of mandatory annuitization.220 Moreover, there would be practical problems if the system used private insurers to underwrite the annuity contracts, as some have proposed: annuitization of small accounts would be costly, insurance companies are not likely to issue indexed annuities, and participants would be at risk of insurer insolvency, which opponents would use as an argument against privatization.221

If we were to implement a private account regime with safeguards against leakage, political pressures still ultimately might push the system to a different design. In the private system, employees may withdraw funds from 401(k) plans in some circumstances, and there is no excise tax on pre-retirement withdrawals from plans and individual retirement accounts when employees make withdrawals for certain approved purposes. The pressures that produced these leaks in the private pension system may well result in similar leaks from a privatized social security system, in which people have accounts to which they contribute and thus may have a sense that they should be able to access that money, at least in emergencies.

Similar pressures would push for modification of any initial rule requiring annuitization. Moreover, societal and political sympathy might be particularly high in cases when people have strong need for access to their accounts, such as sickness or purchasing a home. Any carveout for special purposes would also impose an administrative cost on the system to determine eligibility for an exception.

receiving larger annuities are likely to feel cheated by the system. Insurance companies issuing annuities have much higher administrative costs than the Social Security system, and those costs will be a drag on benefits. Even if a captive market results in a lowering of annuity load factors among private insurers, the necessity of having two systems will impose additional costs, which will lower benefits. It is also possible that insurers will not want to annuitize small account balances, leading to political pressure to limit annuitization options for such individuals. If the government moved in as insurer of last resort for those with small account balances, the annuitization factors may be different, and less favorable, for those with small accounts.

A purely governmental annuitization program might also engender political pressures to opt out of the system. If the government had to annuitize with a single conversion rate, individuals with large account balances, long life expectancies, and consumer savvy might argue that the rules should permit them to use a private insurer if they can secure more favorable annuitization rates. Such pressure might lead to insurers underwriting the best risks and increasing costs for the governmental program. Participants left in the governmental program, faced with higher annuitization rates, might argue that the rules should permit them to opt out of annuitization. One might expect similar results if the initial design of the system contained governmental and private components. Thus, there is a substantial risk that even if the initial system design required annuitization, the concept of mandatory annuitization might erode over time.

Governmental participation in an annuitization program would also create possible public finance issues. If interest rates fall, the government – unless it purchased secure long-term debt instruments, which might include purchasing such instruments from itself – would assume an insurance risk that it may ultimately have to finance out of public revenues. If interest rates rise and the system shows a surplus, there might be pressure from participants to provide upward adjustments to benefits and pressure from other political actors to allow government to dip into the “surplus” for other programs.

Thus, developing a program of private accounts within the Social Security system carries with it the possibility of introducing leakage into the Social Security retirement program. Indeed, one of the strongest attractions of privatization appears to be the accumulation and inheritance of wealth possible in a personal account system, in contrast to the income maintenance without estate building that is an essential feature of social insurance systems generally. The risk of either losing the amounts in such accounts or expending the savings on non-retirement income purposes may to many Americans be worth the possibility of disproportionate gain through investment of an individual account. The problem is that the risk is much higher for those who have the fewest additional resources to support them in old age.

C. Linkage or Failures of Expectation

The final gap in the private pension system that the current structure of Social Security most effectively fills is the problem of expectation of income security in old age. Most people will not risk stopping work while they are still physically and mentally able if they are not sure they have enough income to support them until death. This idea of linkage, the link between an employee’s expectations about private sector benefits and the ultimate realization of those benefits, is critical to the employee’s ability to formulate reliable financial plans for retirement. Yet the private sector retirement system is one in which employee expectations and reality often lack linkage. Moreover, because private systems depend on voluntary employer participation and employer flexibility, the problems of insufficient guarantees have no easy solution short of fundamental redesign of the systems.

An intermediate solution may be to require more candid disclosure about rights retained by employers to alter benefit programs and about how that may affect employees. This solution, however, may alter, but would not increase, the degree of certainty in the system because rational employees would assume that employers would modify benefits to the detriment of employees at a point when it was too late to make alternative arrangements. Ultimate payment of benefits would then become a windfall to employees who have made alternate arrangements.

This Article also suggests that the shift to an increasingly defined contribution world has created firther problems of uncertainty that, at least compared to the world of defined benefit plans, compromise the ability of employees to accurately assess their income in retirement, which, in turn, depends on assumptions about future rates of return, life expectancy, and interest rates (or annuity purchase rates). The Social Security system, which historically has provided a strong measure of certainty, provides a counterweight to this instability of employee expectations in the private pension system. In particular, the dual system of indexing provides certainty that benefits themselves are likely at least to maintain an immediate pre-retirement standard of living.

Perhaps more important, the fact that the political will of the American people underwrites the Social Security system and that the taxing and spending power of the federal government supports it makes this public retirement system an almost absolute source of certainty in retirement income for the elderly. Converting the system to one of private accounts will necessarily undermine that certainty, particularly if the system does not require mandatory annuitization of benefits.

First, as discussed above, the Social Security system bases its benefit structure on wage indexing of the earnings records that determine initial retirement benefits as well as on price indexing of benefits in pay status (the cost of living increases or “COLAs”). Thus, the initial benefit more accurately reflects increases in productivity over the worker’s lifetime and guarantees retirees a share of the national increase in standards of living through their Social Security benefits.222 Once benefits begin, the system indexes them to price increases, thus ensuring that retirees’ incomes, at least to some extent, keep pace with their consumption needs.

In contrast to these virtual certainties, a private account system may or may not adequately reflect productivity increases before retirement – returns on investments depend on individual stock performance or on interest rates, which a number of random factors having little to do with overall economic productivity may affect.223 Similarly, after retirement, unless the participant used the account proceeds to purchase an inflation-adjusted annuity, it would be very difficult for an individual retiree to know for certain whether an accumulated amount would be sufficient to support her consumption needs until death. Moreover, retirees must liquidate account balances invested in most types of equity, such as stocks or real estate, to provide cash for direct consumption. As a result, retirees would face a series of timing decisions (if, that is, there were no requirement that they buy life annuities immediately at retirement) concerning total or partial liquidation of their portfolios, decisions which could be beneficial or disastrous depending on circumstances largely out of the individual’s control or predictive capacity.224

However, the contrast between the certainty granted by private ownership and the uncertainty of the political process in this case may be more illusory than real.

Since its enactment in 1935, the Social Security system has undergone many changes, primarily expansions, but also some limited examples of retrenchment. Most notably, in 1977, Congress made a correction in the automatic indexing formula and then, in the 1981-1983 budget and refinancing bills, addressed the short-term financial difficulty the program was then in and achieved some savings in the federal unified budget.226

The most notable characteristic of the changes made to decrease Social Security expenditures, however, is that only one – a six month delay in the COLA in 1984 enacted as part of the 1983 refinancing bill – affected current beneficiaries. All of the other types of cutbacks affected future beneficiaries that is, those not yet eligible to receive benefits.227 Thus, one immediate conclusion might be that unlike the stock market, which can deplete a retiree’s portfolio in a week or even overnight, the political system is extremely loath to touch current benefits in any way except to provide an increase.

Second, since the changes made in 1983, there have been no real changes in benefits promised to future recipients, let alone to current beneficiaries.228 This record is not unexpected given the growing numbers of voters at or nearing retirement age and the looming baby boom generation rapidly closing in on age sixty-five. It is somewhat surprising, though, that so many advocates of privatization would be willing to trade a system over which the public has substantial policy control, through its vote, for one based solely on investment performance, thus placing retirement in the lap of an institution, the stock market, over which no one has any control. This skepticism about the power of the public to work its will to maintain Social Security through the political process may speak more to a generalized hostility to, and doubt about, the efficacy of government programs generally than about specific weakness in the public retirement system.

VI. Conclusion

tem, however, is a voluntary one and employers design the particulars subject to statutory minimum standards. Those minimum standards, as noted in this Article, leave much flexibility to employers as to which employees will accrue and ultimately earn benefits, what those benefits will be, whether those benefits will change or terminate in the future, and whether employees can tap into their benefits prior to retirement. There are advocates for reducing this flexibility, for heightening benefit certainty, for increasing equity in the distribution of the tax subsidy for low- and moderate-income workers, and for reducing the amount of leakage from the system. Nonetheless, the business community, and to some extent organized labor, view flexibility to shape and modify plan design as an essential component of the private pension system, which may mean that political obstacles will interfere with attempts to increase regulatory standards for retirement plans. Indeed, Congress has recently reduced regulatory influence on the design of retirement plans.

One of the likely outcomes of this new congressional direction is a diminution of private sector retirement benefits for low- and moderate-income workers. This Article has argued that a privatized Social Security system involving individual investment accounts has potential to reduce Social Security benefits for this same group of employees.

The Social Security system also provides a type of feature diversification to our nation’s overall retirement portfolio. The private system’s robustness may well require shielding employer flexibility from comprehensive govern ment regulation to protect employee expectation and to ensure preservation of pension assets for retirement. The Social Security system, with its fixed and certain defined benefits payable over the course of a wage earner’s retirement, is a significant counterweight to the weaknesses of the private pension system. Adopting a system of private accounts would introduce many of the private pension system’s most glaring weaknesses into the parallel public systern. This suggests that caution should mark political discourse on the desirability of private Social Security accounts.

If private accounts are to become part of the system, the design of those accounts should take into consideration the need for predictability in benefits and should safeguard against leakage. This would virtually require annuitization of account balances, and the system should index annuities to increases in the cost of living. Moreover, the design of the annuitization mechanism should anticipate potential problems that might create future political pressure to relax mandatory annuitization.

Similarly, it would not be adequate for legislation creating private accounts to bar pre-retirement access to assets. It should also recognize the possibility of evolving pressures to permit such access in the future. This may involve erecting super-majority requirements to amend a private account system to permit pre-retirement access. Of course, if all of these features were part of a private account system, it would look remarkably like the current Social Security program in many ways, thus raising the important issue of whether we should bother to make such a change to start with.

In 1987, Professor Michael Graetz argued that our polity should conceptualize a coherent and unified retirement policy, which begins with the recognition that Social Security provides inadequate benefits for all income classes and that the private pension system heavily favors the wealthy.229 At the time that Professor Graetz was writing, he observed that Congress in 1986 was “willing to go quite far in an effort to ensure some distribution of benefits to low- and moderate-income earners.”230 This Article suggests that agency– initiated regulatory changes have pushed in the opposite direction and that in the last decade Congress itself has reversed direction, increasing benefits for the affluent and reducing regulatory requirements designed to ensure a meaningful level of benefit distribution to others.

Norman P. Stein*

Patricia E. Dilley**

* Douglas Arant Professor, University of Alabama School of Law; J.D., Duke University School of Law; B.A, New College.

** Professor of Law, University of Florida Frederic G. Levin College of Law, J.D., Georgetown University Law Center, B.A., Swarthmore College.

Copyright Washington & Lee University, School of Law Fall 2001

Provided by ProQuest Information and Learning Company. All rights Reserved