THE ART OF DECOUPLING: KEEPING SOCIAL SECURITY'S PROMISE UP-TO-DATE*

Peter W. Martin

A public pension system should be predictable and reliable. People need to know what the order of magnitude of their publicly financed old age benefits will be; their private investment decisions, as well as the psychological well-being, to some extent depend on this knowledge.

M. Derthick, Policymaking for Social Security.(1)

The right to alter, amend, or repeal any provision of this Act is hereby reserved to the Congress.

Social Security Act of 1935.(2)

I

Social Security Adjustments to Changing Prices and Wages Prior to 1977

A. The Principle of Actuarial Soundness

From its inception, the Social Security program has incorporated a forward-looking perspective. The original legislation combined two features that required attention to the future operation of the program despite distant shifts in economic conditions: (1) a financial base for present and future periods resting on new taxes levied against covered employees' annual earnings up to a fixed ceiling, and (2) a benefit formula, applicable to both individuals approaching retirement and to those still many years away, that tied the level of monthly payments to the worker's taxed wages over a lengthy period.(3)

Enacted in 1935, the legislation provided that monthly retirement benefits would not be paid until 1942.(4) It specified the monthly benefits the first beneficiaries would receive over the balance of their lives. More ambitious still, the Act specified the taxes to be paid by young workers over a period of approximately forty years and the monthly benefits they would receive upon retirement. The drafters and later revisers sought to frame a law that would operate sensibly over a very long stretch. Perhaps naively, they contemplated charts showing how the 1935 Act would perform in 1980 and beyond.(5)

The insurance concept with which the program was draped produced the phrase by which this notion of long-term reasonableness was most often expressed. From the start, the drafters said that Social Security should be "actuarially sound."(6) Experts advised Congress that the 1935 program met this test.(7) Four years later, in 1939, Congress enacted major Social Security revisions.(8) Again, it sought to enact an actuarially sound system that would function properly over a long period. Congress also added a device to provide warning of developing unsoundness. Drafters of the 1935 Act had avoided linking the program's tax and benefit provisions for constitutional reasons.(9) In 1937, however, the Supreme Court removed those concerns.(10) Congress was free to express a direct connection between taxes and benefits. The 1939 amendments, thus, employed insurance terminology throughout, established the Social Security "trust fund" as an earmarking method, created a Board of Trustees comprised of the Chairman of the Social Security Board, the Secretary of the Treasury and the Secretary of Labor, and called upon that body to report annually to Congress on the operation and status of the fund.(11) Congress intended that the mandatory trustees' report not be simply a retrospective accounting. The Act called for it to cover both the "expected operation and status [of the trust fund] during the next ensuing five fiscal years" and its "actuarial status."(12)

Despite its concern with the fund's future financial health, Congress failed to consider the likely impact of time and economic change on the adequacy of the benefits promised by the statutory formula. Instead, Congress focused almost entirely on the internal relationships of the program-the long-term balance between projected revenues and benefits.

Actuarial soundness did not, particularly in the early years, hold the same meaning for all. Views differed sharply over whether the tax rates written into the law should provide the revenue necessary to pay all future benefits, or only benefits for a shorter future period, leaving the problem of distant projected deficits to later Congresses. The first and more extreme view was, at President Roosevelt's insistence, incorporated in the 1935 legislation.(13)

The 1939 amendments increased benefit payments for those nearing retirement, and delayed a payroll tax increase that the 1935 Act had scheduled for 1940. The changes reflected a looser view of balance and growing misgivings about tax and benefit schedules that would build up large reserve funds.(14) In 1939, the House Ways and Means Committee and the Senate Finance Committee adopted Treasury Secretary Morgenthau's view that the tax rates and benefit schedules should not be set to yield an "eventual reserve amounting to . . . more than three times the highest prospective annual benefits in the ensuing 5 years."(15) The committees were persuaded that the revised system was financially sound, but acknowledged the possibility that unpredicted changes might require additional sources of funding.(16)

In summary, two opposing concerns about the future operation of the new Social Security system marked the Act, as amended in 1939: a concern that the payroll tax might not generate enough revenue to cover benefits promised by the statute, and conversely, a concern that revenues might exceed benefits by such a large measure that an enormous reserve fund would develop. The Act therefore directed the Board of Trustees to:

Report immediately to the Congress whenever [it] is of the opinion that during the ensuing five fiscal years the Trust Fund will exceed three times the highest annual expenditures anticipated during that five-fiscal-year period, and whenever [it] is of the opinion that the amount of the Trust Fund is unduly small.(17)

B. Congress' Reserved Power to Amend

Although the 1939 amendments raised benefits generally, they did not increase payments for all classes of future beneficiaries. The revised schedule provided for smaller monthly payments for workers with many working years before retirement, unless household family members were eligible for the newly-created dependents and survivors benefits.(18) Moreover, the amendments eliminated the lump sum return of contribution which the 1935 Act had furnished workers who paid the social security tax but failed to qualify for monthly benefits.(19) Altogether, the changes enhanced the value of benefits promised by the 1935 Act for most but not all contributors.

This restructuring would have been impossible if the 1935 Act had unconditionally guaranteed benefits. Quite clearly, however, it did not. Section 1104 of the Act provided: "The right to alter, amend, or repeal any provision of this Act is hereby reserved to the Congress."(20) This section ultimately protected the program from an uncertain future. Even the best actuarial projections were understood to be guesses. Therefore, flexibility was essential.(21)

In the years after 1935, Congress has repeatedly exercised this reserved flexibility; indeed, section 1104 is one of the few provisions of the Social Security Act that Congress has not "altered, amended, or repealed." Like the 1939 amendments, most revisions of the formula have increased benefits.(22) But section 1104 has also permitted Congress to reduce or eliminate expected benefits.(23) Some of those revisions have gone so far as to reduce the value of monthly payments to individuals already receiving them.(24)

C. Unprincipled Benefit Revision

The Board of Trustees soon had occasion to report a reserve fund exceeding the "three times annual expenditures" benchmark set in 1939. Social Security's 1% payroll tax, levied on the first $3,000 of annual earnings against both employer and employee, generated far greater revenues during the 1940's than Congress had anticipated. World War II drove both employment and earnings levels substantially above pre-war figures. Growth in the number of covered workers produced an immediate gain in income while the inevitable rise in the number of beneficiaries lay off in the distant future.(25) Climbing wage levels also produced a direct effect on the program's income. Individual earnings subject to the payroll tax rose sharply.(26) Although higher covered wages meant that the system had to pay out higher benefits, those benefits would be paid in a future period. In addition, under the 1939 formula, growth in wages above the lowest levels produced only a minimal effect on benefits.(27)

These developments posed no threat to actuarial soundness; instead, they increased the program's fiscal strength. Indeed, Congress and the President disagreed over whether these wartime economic conditions might lead to the reverse problem-a great imbalance between program revenues and expenditures creating an excessively large reserve fund.(28) As early; as 1941, the Board of Trustees reported to congress that the reserves would exceed three times annual expenditures within five years.(29) Neither the Trustees nor the administration advocated an adjustment in tax or benefits,(30) but Congress did. The 1939 amendments had scheduled a contribution rate increase for 1943. Despite President Roosevelt's opposition, Congress attached a rider to the Revenue Act of 1942 that postponed this increase until January 1, 1944.(31)

Actuarial principles furnished no guide when taxes and benefits were judged to be out of balance because of price, wage and employment trends of the war years. In order to adjust the balance, Congress might have increased benefits, or allowed the reserve to grow. Instead, Congress repeatedly chose to cut taxes against the wishes of Presidents Roosevelt and Truman. After 1942, Congress postponed the payroll tax increase six more times, holding the rate at 1% until 1950.(32) It enacted the 1944 postponement over a presidential veto.(33) In response to the administration's concern with actuarial soundness, however, Congress accompanied the 1944 postponement with a provision authorizing appropriation from general revenue to the trust fund of "such additional sums as may be required to finance the benefits and payments under this title."(34) Despite these delays, the system generated a substantial reserve, totaling 23.4 times annual expenditures in 1945, and 3.4 times annual expenditures in 1950.(35)

Standards or reference points outside the Act were needed to understand the effect these dramatic economic changes were having on the benefit scheme. In the immediate post-war years, President Truman emphasized wage increases and the higher cost of living when he called on Congress to increase Social Security benefits.(36) An Advisory Council on Social Security appointed by the Senate Finance Committee presented similar arguments in its 1948 report calling for extension of Social Security coverage to additional workers and dramatic benefit increases. The council stressed the inadequacy of benefits in relation to the buying power of past benefits and associated earnings levels, the current public assistance standards, and the cost of the basic necessities of life.(37) It pointed out that in 1939, the $3,000 statutory limitation on annual wages taxed and counted for Social Security purposes had covered total earnings for nearly 97 percent of those subject to the tax, but that the subsequent rise in wage levels was excluding growing amounts of the earnings of a significant proportion of covered workers.(38) The Council did not recommend a simple up-dating of the 1939 provisions measured against wage or price increases, but placed the direction and size of its proposed adjustments in the context of such comparisons with 1939.

The 1948 Advisory Council's report also expressed a general view on benefit adequacy during periods of economic change that indicted a need for repeated revisions of tax and benefit schedules: "A social insurance program must be adjusted periodically to basic economic changes. In a dynamic economy, provisions which were appropriate at the time they became effective inevitably become outmoded."(39)

The Democratic Congress elected in 1948 eventually restructured Social Security along the lines recommended by the President and the 1948 Advisory Council.(40) The Social Security Act Amendments of 1950 both expanded the program's coverage and boosted benefits through a total revision of the benefit formula.(41) Both existing beneficiaries and future retirees were affected.(42)

The 1950 amendments reemphasized the principle of actuarial soundness by repealing the 1944 authorization for appropriations from general revenue.(43) The Finance Committee report on the amendments took the position that the program "should [operate] on a completely self-supporting basis."(44) Actuarial balance did not, however, pose an obstacle to benefit improvement. In fact, Congress was able to increase benefits by more than 75% without dramatically raising the tax.(45) The 1950 amendments boosted the tax base-the ceiling on taxable wages-to $3,600 beginning in 1951 and increased tax rates to 1 1/2% for employees and employers for 1950-53, 2% for 1954-59 (prior to the amendments, that increase was scheduled for 1952), 2 1/2% for 1960-64, 3% for 1965-69 and 3 1/4% thereafter (higher rates than previously set in the law).(46) Actuarial projections estimated these tax rates would adequately cover future benefits payable under the revised schedule.(47)

From 1950 through 1972, Congress clung to the view that the Social Security program should be self-supporting. Each year the Board of Trustees of the Trust Fund reported on the actuarial soundness of the program. As payroll tax receipts rose with individual earnings, Congress responded from time to time with benefit increases. Each such change maintained the discipline of actuarial balance.(48) Thus, Congress frequently employed tax rate adjustments and tax base increases to achieve balance. Inevitably, the Act specified substantially larger future tax rates, which were at least partially rolled back as the future drew near.

Although the 1950 benefit increases were achieved through a complete restructuring of the benefit formula,(49) Congress made subsequent adjustments within the framework it established in 1950. That framework based benefits on average covered wages between 1950 and the year of entitlement. Essentially the formula had two parts-a formula for computing the Average Monthly Wage (AMW),(50) and a set of multipliers that converted the AMW into a monthly benefit amount payable to a worker retiring at age sixty-five. The statute termed this basic benefit "primary insurance amount" (PIA).(51) The multipliers varied according to the size of the worker's average wage so that individuals with low AMWs received a greater relative benefit than those with high AMWs. The 1950 formula had two broad AMW brackets with associated multipliers and a graduated lower bracket bottomed on a minimum PIA. (See Table 1.)

Table 1

Benefit Formula Established By 1950 Amendments

AMW
FORMULA
PIA
$30 or less   $20 (minimum PIA)
31
Set by statutory formula
21
32   22
33   23
34   24
35-49   25
     
50
PIA = 50% AMW
25
60   30
80   40
100   50
     
140
PIA = $50 plus 15% (AMW - $100)
56
180   62
220   68
260   74
300   80

Source: Social Security Act Amendments of 1950, Pub. L. No. 81-734, ch. 809, 104, 64 Stat. 477.

Later amendments modified the benefit based by eliminating a number of low-income years from the AMW calculation(52) and providing for optional receipt of benefits prior to the program's original age sixty-five threshold, with a consequent reduction in benefit amount.(53) The latter was not accomplished by employing a different PIA formula, but rather by stipulating that those receiving benefits before age sixty-five would receive only a percentage of their full PIA.(54) However, the most dramatic changes in the formula during the years after 1950 took the form of multiplier changes designed to respond to wage and price increases. Prior to 1972, the process was totally ad hoc. Approximately once every two or three years, generally shortly before elections, Congress raised benefits by changing the multiplier, matching and sometimes surpassing increases in living costs.(55) While the principal purpose of the adjustments appeared to be to prevent devaluation of the otherwise fixed dollar incomes of Social Security pensioners, the same adjusted schedule applied in each case to future retirees.(56)

During the same period, Congress also occasionally amended the maximum amount of taxable and countable wages.(57) This produced corresponding increases in the maximum attainable AMW. At first, the amendments dealt with the wider spread of possible AMWs by expanding the two basic brackets of the benefit formula. Until 1965, the formula contained only two brackets. Thereafter, Congress left the lower AMW brackets undistributed and responded to each instance of higher potential AMWs by adding a new bracket with a different and, with one exception, lower multiplier. Under the 1967 amendments, the number of brackets increased to four. By 1972 there were six. (See Table 2.)

At no point during this period did the legislation itself guarantee that these adjustments wuold continue. As it up-dated the 1950 formula from time to time, Congress reiterated the view of the 1948 Advisory Council that the purposes of Social Security would be served only if the benefits being paid are adjusted to changes in economic conditions.(58) Committee reports invariably justified benefit increases by citing wage and price rises since prior adjustments. But Congress did not attempt in any specific way to relate such changes to the proposed formula revision. Wage and price changes merely furnished the justification for the direction and relative size of benefit increases.(59) Occasionally Congress saw itself not only up-dating Social Security benefits in light of inflation, but providing, in an absolute sense, "some improvement in the adequacy of benefits."(60)

Although Congress retained full control over Social Security adjustments, pressure for both up-dating and absolute increases came from the Social Security Administration, organized labor, other less influential interest groups and periodically-appointed advisory councils. The 1956 amendments established advisory councils as a formal part of the program's policy-making structure, providing that such bodies should be appointed in advance of all future tax increases.(61) The 1956 provisions merely charged the councils with the study of financing issues; but a 1960 amendment expanded their purview to include possible extensions of coverage and improvements in benefit adequacy.(62) The statutory councils recommended specific legislative techniques for adjusting the system to changing economic conditions.(63)

This process of up-dating through amendment posed a serious dilemma for those charged with plotting the program's actuarial soundness. Should their calculations reflect present benefits and taxes, or attempt to predict the future Congressional changes that were inevitable given foreseeable increases in wages and prices? The program actuaries and Board of Trustees adopted the former approach during the forties and applied it long after it became evident that wage and price increases would not soon cease and that Congress would up-date benefits periodically. The approach assumed no change in the statutory benefit and tax schedules and matched no change in the statutory benefit and tax schedules and matched that assumption with an equally unrealistically one of legal wages.(64) A 1947 Trustees' report, which employed this method of calculating balance, noted that wages had been increasing and explained that if the trend continued, contributions would rise more steeply than benefits.(65) In sum, program actuaries based their projections on unrealistic assumptions of static wages and benefit rates. They nevertheless began to recognize the impact of wage growth and benefit revision, but believed that the effect of the former would sufficiently finance the latter.(66)

Table 2

Year of Social Act Amendments
Rate of Increase Over Prior Benefits
Amended Formula for Calculating Primary Insurance Amount (PIA) from Worker's Average Monthly Wage (AMW)
1950 Approx. 77% 50% of first $100 plus 15% of next $200
1952 Greater of 12 1/2% or $5 55% of first $100 plus 15% of next $200
1954 Approx. 13% 55% of first $110 plus 20% of next $240
1958 Greater of 7% or $3 58.85% of first $110 plus 21.40% of next $290
1965 Greater of 7% or $4 62.97% of first $110 plus 22.90% of next $290 plus 21.40% of next $150
1967 Approx. 13% 71.16% of first $110 plus 25.88% of next $290 plus 24.18% of next $150 plus 28.43% of next $100
1969 Approx. 15% 81.83% of first $100 plus 29.76% of next $290 plus 27.81% of next $150 plus 32.69% of next $100
1971 10% 90.01% of first $110 plus 32.76% of next $290 plus 30.59% of next $150 plus 35.96% of next $100 plus 20% of next $100
1972 20% 108.01% of first $110 plus 39.29% of next $290 plus 36.71% of next $150 plus 43.15% of next $100 plus 24% of next $100 plus 20% of next $250

Source: Soc. Sec. Bull.-Annual Statistical Supplement, 1975, at 18-19.

In the early sixties, the Board of Trustees added a new actuarial projection to their annual report: "medium-range cost estimates for the next 15 to 20 years . . . [taking] into account possible variations in economic factors, such as earnings and employment levels, as well as in demographic developments."(67) Although these estimates reflected predicted wage growth, initially they did not reflect benefit changes:

The medium-range estimates, like the short-range and long-range estimates, assume no future changes in the law. However, over the period covered by the estimates, many changes will undoubtedly be made. In particular, if earnings levels rise, changes will almost certainly be made in benefit levels. The extent and timing of whatever changes are made in the law are, of course, unpredictable.(68)

Before long, the Board of Trustees began constructing two medium-range projections-one assuming no change in the law and the second assuming "that the law is frequently amended to keep the system fully up to date."(69)

By the late sixties, the pattern of repeated Congressional revisions of the benefit and tax provisions had become fairly clear to program administrators and other experts. By then, it also seemed obvious that the steady growth in wages furnished a reliable source of revenue from which to finance benefit improvements.(70) The 1971 Advisory Council recommended that Congress incorporate this financing scheme into the Social Security Act. It proposed an amendment to provide for automatic benefit increases tied to increases in the Consumer Price Index and for automatic wage base increases based on increases in the average covered wage. In effect, the proposal was an automatic modification mechanism following the pattern of adjustment Congress had practiced in crude form since 1950.(71)

The Council also urged that the method of performing actuarial balance be changed to reflect the availability of additional revenue from future wage increases.(72) The proposed automatic adjustment feature will remove any ground for assuming level benefits or wages and compel the use of projected wages and benefits.(73) In 1972, Congress adopted the Council's automatic adjustment recommendation and, relying on the revised actuarial principles, enacted an immediate 20% across-the-board increase.(74)

D. The Search for Appropriate External Standards

The 1972 amendments codified a particular relationship between the important components of the system, benefits and covered wages, and specific measurements of economic change. Thus, Congress placed the Consumer Price Index and average covered wages alongside actuarial soundness as statutory standards.

Previously there was no discernible policy on how benefits should relate to prior earnings, or what wages should be subject to the payroll tax, or how large that tax should be. The Act simply set various parameters, and from time to time Congress changed them as a consequence of altered economic, social and political circumstances. The early amendments did not implement any coherent policy.

By the mid-sixties, experts compared the operation of the benefit formula for average retirees of various vintages to the movement of prices and wages. In 1966, for example, this analysis demonstrated that current payments received by many beneficiaries did not provide the purchasing power of the benefits they received when first going on the rolls. Moreover, for those whose benefits had equivalent purchasing power, there had been substantial periods of lag in correction for price inflation. Finally, these comparisons revealed that average benefits increased at a much slower pace than wage levels.(75)

National concern with poverty in the 1960's also influenced Social Security planners. Experts and advisory councils began to compare the lower part of the Social Security benefit scale to general standards of minimum income.(76) The comparison revealed that Social Security produced benefits near or below the poverty level for workers who had been self-sufficient with low but steady earnings throughout their working years.(77) That observation supported repeated efforts to improve Social Security's performance at the bottom as a minimum income guarantee for retired persons.(78) In addition, experts continued to debate the validity of the connection between program benefits and the payroll tax.(79)

Prior to the early seventies, however, no widely-used analytic tool existed that permitted assessment of the performance of the Social Security benefit formula above the realm of minimal adequacy. The scheme's fairness to similar individuals of different age and years of retirement, its fairness to groups with differing wage patterns retiring at the same time and its performance over time as amended by Congress, were matters that received inadequate attention in the debates over Social Security, both inside and outside Congress. Some of that inattention must be blamed on the difficulty of describing and measuring the relevant characteristics of the program. The terms of the legislation, which provided the framework for most discussion of the benefit schedule, did not (given changing wage and price levels) facilitate consideration of questions of equity and adequacy.(80) The system had so completely rejected return on contribution as the benefit model, that the program's policymakers rejected insights based on calculations employing that model.(81)

In the early seventies, the concept of "replacement" rate or ratio began to fill the analytic void. Within five years, it predominated in discussions of Social Security benefit schedule revision.(82) The phrase had been used in a number of earlier studies, but its pre-1970 meaning was almost invariably linked to the Social Security benefit formula itself. Thus, replacement's former orientation made it a poor device for viewing the benefit formula critically.(83)

As the term came to be used in the seventies, replacement rate referred to the ratio of benefits, for an individual or class of individuals, to immediate pre-retirement earnings. The concept was logically derived from the avowed purpose of social insurance programs like OASDI, namely, the partial replacement of lost earnings. Replacement furnished a useful extrinsic measure of the performance of the benefit formula.(84) The Social Security benefit amendments of 1972, which included the first automatic adjustment provisions, were enacted without reference to replacement rates in this sense. However, a panel of economists and actuaries reporting to a subcommittee of the 1971 Advisory Council, which recommended the 1972 automatic adjustment provisions, warned:

While past and proposed legislative actions have approximately achieved the goal of maintenance of purchasing power [of benefits], the replacement rates have shifted over time and between different levels of average wages. There has been insufficient analyses or public discussion of the role of replacement rate in prescribing the benefit formulas. . . . Careful study and serious consideration should be given before establishing a specific policy about replacement rates. However, the policy should be explicit and not implicit.(85)

By 1975, replacement rates were receiving a great deal of attention. The 1975 Advisory Council's report demonstrated that, with continuing high inflation, the automatic adjustment provisions in the amended act would produce absurdly generous replacement rates, paying more in benefits than a retiree's prior earnings. On the other hand, different wages and price conditions would produce replacement rates far below current levels.(86) "The result is that replacement rates, instead of being controlled by conscious and deliberate policy set by Congress, can fluctuate up or down with the tides of inflation."(87) The Advisory Council characterized this quite accurately as "a flaw [of the system] which was probably never intended."(88) It was a feature of the system difficult to perceive without the new concept of replacement rate.(89) It was also a flaw that had relatively slight consequences under the conservative wage and price estimates underlying the proposals and legislation of 1971 and 1972. A 1977 Congressional Budget Office study on Social Security financing observed: "The automatic provision for indexing future benefits received little attention because combinations of wage and price increases that cause great instability were not seriously considered."(90)

II

The "Flaws" in the 1972 "Coupled" formula

Without considering post-1972 concepts and economic experience, it is difficult to view the 1972 amendments as "flawed" or "mistaken." The 1972 adjustment provisions were attractively simple and had the advantage of treating equally those a few years before retirement and persons already receiving benefits. They also reflected the pattern of ad hoc adjustments legislated in preceding years. When passed, they drew little criticism and considerable favorable comment.(91) The first recommendation of the 1971 Advisory Council on Social Security was that the program's benefit and contribution formulae should be automatically adjusted to changes in prices and earnings levels.(92) The 1972 amendment did just that.

Experts thought the financing demands of benefit adjustments and the principle of actuarial soundness were adequately dealt with by the parallel automatic adjustment of wages subject to the Social Security tax. The Council's report explained that earnings levels tend to increase faster than prices. Planners concluded that the resulting increase in contributions would finance benefit increases attributable both to price increases and higher earnings records.(93)

A report by the Office of the Actuary of the Social Security Administration supported the Advisory Council's estimate, but was more careful in noting its underlying assumptions about economic conditions over the next 75 years. One assumption was that "the rate of increase in wage levels [would remain] approximately twice the rate of increase in price levels." The actuary cautioned: "The estimates [of fiscal balance] are extremely sensitive to the stated assumptions as to wage and price increases."(94)

Congress ignored this warning. Welfare reform was the dominant Social Security Act issue in 1972 and Congress enacted the automatic adjustment and 20% one-time increase without extensive debate or hearings. Sponsors of the provisions in both the House and Senate relied solely on the recommendations of the 1971 Advisory Council. The amendments were tacked on to a bill to extend the debt ceiling which had to be passed in the final days of June, before the end of the fiscal year.(95)

A year later, in June 1973, Congress again adjusted the benefit formula without benefit of hearings or careful deliberation.(96) The rapid climb in the Consumer Price Index during the first six months of 1973 prompted the Senate Finance Committee to amend that year's public debt limit bill to move up the effective date of the first automatic increase from January 1975 to June 1974.(97)

The continuing climb of prices led the Senate Finance Committee to introduce another amendment in late 1973. Signed by President Nixon on the final day of the year, the amendment substituted a two step 11% increase in the benefit formula for the 5.9% increase the automatic adjustment provision would have produced in June 1974.(98) The first increase of 7% took effect immediately.(99) The amendment not only produced a larger and earlier increase than the automatic adjustment provisions to reduce the lag between price increases and benefit schedule response.(100) Both 1973 amendments increased the annual tax base to help meet the additional costs of the larger benefits.(101)

It was not until June 1975 that the automatic adjustment provisions of 1972, as amended in 1973, were allowed to generate their first increase-a benefit rise of 8%. By then it was clear that the program had a long-term deficit due in part to the interplay between the automatic adjustment provisions and revised economic projections.(102) Social Security's automatically adjusting benefit formula thus posed a serious thread to actuarial soundness. The formula was drawn into a growing Social Security debate energized by an ever-increasing projected actuarial deficit.(103) Benefit formula revision was pushed into the policy arena not because it could offer more equitable treatment of groups with differing wage histories and dates of retirement-though it could-but because without dramatic change on the benefit side, a major long-term deficit loomed. Interest in that long-term deficit was fueled by serious short-term financial problems.(104)

The automatic adjustment provisions created another serious problem. As the 1975 Advisory Council pointed out, the 1972 amendments, and the ad hoc adjustments that preceded them, were designed to allow benefits to keep pace with inflation.(105) However, in periods of rapid inflation of the sort experienced in the early seventies, the formula overcompensated future retirees. The new replacement rate analysis revealed this otherwise latent problem. Studies employing plausible assumptions about wage and price growth estimated that replacement rates would climb to absurd levels under the 1972 formula. Since low-wage workers forty or fifty years hence would have replacement rates in excess of 100%, they would receive retirement benefits exceeding their pre-retirement earnings.(106) Clearly, Congress did not intend this result. Congress expected that benefits would be adjusted to keep pace with inflation. It had not intended them to outpace inflation; yet the 1972 formula allowed that to happen by counting wage and price increases twice for future retirees.

Understanding how the old formula counted inflation twice requires a closer look at how inflation was handled within the benefit framework established in 1950. The 1950. The 1950 amendments responded to wage changes with a formula based on post-amendment wages. Those retiring in the fifties received benefits calculated by applying stipulated percentages to their post-1950, pre-retirement covered earnings. Pre-1950 taxed and covered earnings were ignored for most retirees. For later retirees, the earnings starting point for benefit calculation remained essentially the same-average covered earnings computed over a post-1950, pre-retirement base period.(107) Congress occasionally raised the wage base to adjust for increased average wages. The Act set the base at $3,000 for the years 1951-54, $4,200 for 1955-58, $4,800 for 1959-65, and so on.(108) Each increase in the wage base meant that more workers' wage increases were reflected in their average covered wage figures. Nonetheless, the benefit calculation for a retiree collecting Social Security benefits for the first time in the late seventies began with an average earnings figure that included wages paid at fifties' rate and subject to fifties ceilings.(109)

The principal price inflation adjustment was accomplished by increasing the multipliers applied to the average earnings figure. From 1965 on, the pattern of adjustment was roughly this: Benefit increases were achieved by increasing the multipliers, increasing the wage base (which had a long-term effect on possible benefits), and adding on a new wage bracket at the top to cover the larger average earnings figures possible under the new wage base.(110) The automatic adjustment pattern of the 1972 amendments took this same form.(111)

Such a technique of adjustment overcompensated for inflation in the following sense. A worker experiencing typical wage growth between 1975 and 1976, for example, first benefitted from that increase through its effect on his average earnings figure. Assuming price growth during the same period, the worker would also have higher multipliers applied to the higher average wage. One year's inflation was thus counted twice.(112) This double-counting arose because the system used the same method to adjust benefits for present recipients and future retirees. Both received identical adjustment of the multipliers; post-retirement and pre-retirement inflation adjustment were "coupled."(113) Replacement rate studies demonstrated that the cumulative effects of the system's double-counting could be extreme during periods of high wage and price inflation.(114)

The 1977 decoupling amendments were politically feasible because, under the old coupled formula, actuarial projections showed a serious long-term deficit and replacement rate analysis forecast benefits exceeding historic levels and the absolute benchmark of full replacement. Equally critical was Congress' perception that decoupling did not really revise benefits, but rather replaced a faulty method of inflation adjustment with a more accurate mechanism-one that did not count inflation twice.

III

The Response: How the 1977 Amendments Work

The 1977 amendments "decoupled"(115) the Social Security benefit formula by establishing two distinct systems: one for adjusting the wage record on which an individual's benefits are calculated to reflect changing economic conditions, and another for adjusting benefits, once available, to prevent purchasing power erosion.

The formula adopted in 1977 indexes the worker's wage record by increasing his past wages to reflect their value in terms of more recent wage levels. The formula then applies a set of unchanging multipliers to the individual's average indexed earnings.(116) The benefit amount produced by that formula is subsequently adjusted upward to reflect cost-of-living increases.(117) Any year's wage and price movements count only once in the benefit calculation for a particular individual. Those movements, however, have different effects on employed and retired workers. In the case of employed workers, a given year's increase in average wages is ultimately reflected through wage-indexing. For those already entitled to benefits, price increases, measured by the Consumer Price Index, produce a proportionate benefit increase.

The following computations illustrate how a new formula works for high, medium, and low-wage workers.

A. Average Indexed Monthly Earnings (AIME)

The indexing of earnings called for by the decoupled formula uses average earnings figures promulgated by the Department of Health, Education and Welfare.(118) The formula uses average earnings for the year two years before the individual's sixty-second birthday, death or disability as its indexing base.(119) Thus, 1978 is the base year for an individual attaining sixty-two, dying or becoming disabled in 1980. The HEW average wage figure for 1978 is $10,556.03.(120) A worker's indexed earnings for past years are calculated by multiplying his or her covered earnings for each year by an index number equal to the ratio of the average earnings figure for 1978 ($10,556.03) to the average earnings figure for the year being adjusted.(121) For example, of a worker had the maximum possible $9,000 multiplied by $10,556.03/$7,133.80 (the average earnings figure for 1972), or $13,317.48.

Table 3 sets forth the HEW average earnings figures beginning with the year 1951 (the earliest includable year under the formula), the ratio of each year's average to that for 1978 (the base year), and maximum countable earnings for each year.

Table 3

Year Average Earnings

Earnings Index (1978 Base)

Max. Countable Earnings Indexed Max. Earnings
1951 $2,799.16 3.7711 $3,600 $13,576.11
1952 2,937.32 3.5503 3,600

12,780.90

1953 3,139.44 3.3624 3,600

12,104.61

1954 3,155.64 3.3451 3,600

12,042.47

1955 3,301.44 3.1974 4,200

13,429.09

1956 3,532.36 2.9884 4,200

12,551.19

1957 3,641.72 2.8986 4,200 12,174.28
1958 3,673.80 2.8733 4,200 12,067.97
1960 4,007.12 2.6343 4,800

12,644.73

1961 4,086.76 2.5830 4,800 12,398.32
1962 4,291.40 2.4598 4,800

11,807.09

1963 4,396.64 2.4009 4,800 11,524.47
1964 4,576.32 2.3067 4,800 11,071.98
1965 4,658.72 2.2659 4,800 10,876.15
1966 4,938.36 2.1376 6,600

14,107.88

1967 5,213.44 2.0248 6,600 13,363.50
1968 5,571.76 1.8946 7,800 14,777.56
1969 5,893.76 1.7911 7,800

13,970.20

1970 6,186.24 1.7064 7,800

13,309.71

1971 6,497.08 1.6247 7,800

12,672.93

1972 7,133.80 1.4797 9,000 13,317.48
1973 7,580.16 1.3926 10,800 15,039.94
1974 8,030.76 1.3144 13,200 17,350.73
1975 8,630.92 1.2230 14,100

17,244.98

1976 9,226.48 1.1441 15,300 17,504.75
1977 9,779.44 1.0794 16,500

17,810.27

1978 10,556.03 1.0000 17,700 17,700.00
1979     22,900  
1980     25,900  

Sources: 43 Fed. Reg. 61,016 (1978); 44 Fed Reg. 62,956 (1979)

A person reaching sixty-two in 1980 has a benefit base period of twenty-four years. His or her average indexed earnings figure is calculated over that twenty-four year base, using the twenty-four year after 1950 with highest indexed earnings. The period will be twenty-five years for those reaching sixty-two in 1981, twenty-six for those reaching sixty-three in 1982 and so on until the maximum base period of thirty-five years is reached-for those turning sixty-two in 1991.(122)

1. High-Earnings Worker

For a person with high earnings-earnings at least equal to the maximum countable earnings level each year-the Average Indexed Monthly Earnings (AIME) will be based on indexed earnings in the years 1951 through 1979, omitting 1962-65 and 1954. The base year, 1978, and any subsequent years (in this case 1979) are counted, but not indexed.(123) This averaging process yields an annual indexed earnings figure of $14,497 for the high-earnings individual and an AIME of $1,208.00.

2. Middle- and Low-Earnings Worker

The averaging process is the same for all workers reaching sixty-two in 1980. Although the five lowest indexed years for our hypothetical high-earnings worker were defined by the system (those five years when maximum countable earnings stood lowest in relation to average earnings), the low five years for middle- and low-earnings workers will depend on the individual's earnings history. If the worker's earnings increased in relation to average earnings over the post-1950 period, the five lowest indexed years will be 1951 through 1955. (A worker reaching sixty-two in 1980 was already thirty-three in 1951--past the years when most have their lowest earnings relative to the average.) Some workers will have lower later years. For those who have dropped out of covered employment in order to care for children, or because of involuntary unemployment, the lowest five years will include years with zero indexed earnings. Table 4 shows the annual earnings and indexed earnings for a worker reaching age sixty-two in 1980 who received earnings equal to the median for workers of the same age each year since 1950. This worker's lowest five indexed years are 1951, 1958, 1959, 1963 and 1973; Average Indexed Monthly Earnings (AIME) are $864.00.

Table 4

 

Middle-Earnings Worker

  Low-Earnings Worker  
  Annual Earnings Indexed Earnings Annual Earnings Indexed Earnings
1951 $2,689.00 $10,140.60 $1,344.50 $5,070.30
1952 2,893.00 10,270.87 1,446.50 5,135.44
1953 3,112.00 10,464.70 1,556.00 5,232.35
1954 3,117.00 10,426.77 1,558.50 5,213.39
1955 3,225.00 10,311.62 1,612.50 5,155.81
1956 3,412.00 10,196.35 1,706.00 5,098.17
1957 3,562.00 10,324.95 1,781.00 5,162.48
1958 3,499.00 10,053.77 1,749.50 5,026.89
1959 3,706.00 10,145.92 1,853.00 5,072.96
1960 3,891.00 10,250.13 1,945.50 5,125.07
1961 3.990.00 10,306.10 1,995.00 5,153.05
1962 4,177.00 10,247.63 2,088.50 5,137.31
1963 4,209.00 10,105.52 2,104.50 5,052.76
1964 4,447.00 10,257.73 2,223.50 5,128.87
1965 4,665.00 10,570.26 2,332.50 5,285.13
1966 4,959.00 10,600.15 2,479.50 5,300.07
1967 5,217.00 10,563.24 2,608.50 5,281.62
1968 5,409.00 10,247.67 2,704.50 5,123.84
1969 5,811.00 10,407.80 2,905.50 5,203.90
1970 6,105.00 10,417.40 3,052.50 5,208.70
1971 6,498.00 10,557.52 3,249.00 5,278.76
1972 7,042.00 10,420.19 3,521.00 5,210.10
1973 7,255.00 10,103.22 3,627.50 5,051.61
1974 7,810.00 10,265.85 3,905.00 5,132.93
1975 8,305.00 10,157.41 4,152.50 5,078.71
1976 8,900.00 10,182.50 4,450.00 5,091.25
1977 9,500.00 10,254.40 4,750.00 5,127.20
1978 10,250.00 10,250.00 5,125.00 5,125.00
1979 10,850.00 10,850.00 5,425.00 5,425.00

Sources Soc. Bull-annual statistical supplement, 1965, Table 33, at 36; Soc. Sec. Bull-Annual Statistical Supplement, 1975, Table 42, at 75.

Table 4 also shows earnings for a hypothetical low-earnings worker. This worker received one-half the median annual earnings for workers of the same age since 1950 -- a figure falling slightly below earnings at the minimum wage with steady employment.(124) The AIME for such a worker is $432.00.

B. Converting AIME to Primary Insurance Amount (PIA)

The 1977 Amendments substituted wage indexing for annual adjustments in the multipliers previously used to convert average covered earnings to benefit amounts. The new formula applies constant multipliers to three different AIME brackets. Although the multipliers do not change, the boundaries between brackets shift annually in response to movement in average wages.(125) The formula for converting AIME to Primary Insurance Amount (PIA) for workers reaching age sixty-two in 1980 is as follows:

PIA = 90% of the first $194 of AIME, plus

32% of the next $977 of AIME, plus

15% of any balance (exceeding $1,171).(126)

Table 5 shows the PIA's this produces for the high-, middle- and low-earnings worker for whom AIMEs were calculated in the prior section.

Table 5


High-Earnings Worker
Middle-Earnings Worker
Low-Earnings Worker
AIME
$1,208.00
$864.00
$432.00
PIA
90% of first $194 of AIME
$ 174.60
$174.60
$174.60
32% of next $977 of AIME
312.64
214.40
76.16
15% of any balance
5.55

Total

$ 492.49

$389.00

$250.76

Transitional Guarantee Old Formula, Dec. 1978 Schedule
AME $ 729.50 $498.20 $249.10
PIA $ 503.40 $388.20 $248.70

Sources: 44 Fed. Reg. 62,956 (1979); 20 C.F.R. 404.223 (1979).

C. Price-Indexing for Those Eligible to Receive Benefits

The PIA is adjusted as prices increase, using the Consumer Price Index (CPI). The PIA of a worker turning sixty-two in 1980 will be increased as of June 1980 and thereafter in proportion to CPI increases.(127)

D. Working Past Sixty-Two

The PIA formula outline in the preceding subsections applies whether or not an individual actually claims retired-worker benefits at age sixty-two. Years of covered employment after the wage-indexing base year (after 1978 in the case of those reaching sixty-two in 1980) can be substituted for lower, earlier years in the AIME calculation, but they are not indexed.(128) If our hypothetical high-earnings worker continued in covered employment until age sixty-five (1983), he would have earnings figures of $25,900 for 1980, $29,700 for 1981, and $31,800 for 1982 -- the maximum covered earnings figures for those years established by the 1977 amendments.(129) The worker's AIME would still be based on the best twenty-four years after 1950. These three most recent unindexed years would be over twice the indexed earnings for 1958, 1953 and 1957, otherwise included in the average. Consequently, they will produce a better AIME.

For a high-earnings worker, the additional dollars of AIME fall entirely in the 15% bracket and thereby yield only modest improvement in PIA.(130) However, delaying benefits has a second effect o the monthly amount. A retired worker who begins collecting benefits at age sixty-two receives only eighty percent of his PIA.(131) For each month benefits are delayed, the monthly amount increases by 5/9 of 1% of the PIA, until age sixty-five when 100% of the PIA is available.(123) Further delay adds to the monthly amount at a lower rate-3% per year up to age seventy-two.(133) Whether or not benefits are received at age sixty-two, PIA adjustments based on price increases begin with the sixty-second year.(134)

E. Disability and Survivors Benefits

The benefit formula treats death and disability as equivalents of attaining age sixty-two. A worker dying or becoming disabled in 1980 has past earnings indexed and converted into a PIA as outlined in subsections A and B.(135)

In cases of death and disability, the worker may be much younger when his earnings are indexed. Thus, the AIME may be calculated over a substantially shorter averaging period. The statute calculated benefits based on earnings in the years after 1950 or the year of the worker's twenty-first birthday, if later.(136) The lowest five earnings years are omitted from the calculation.(137) A worker dying at age thirty in 1980 would, therefore, have an AIME based on only three years' earnings.(138) Table 6 shows the PIA calculation for such a worker. Eligible surviving spouses, children, and parents receive benefits based on the worker's PIA.(139) If the same worker were disabled in 1980, he would receive a disabled worker benefit equal to 100% of the PIA.(140) Subject to the family maximum, the worker's spouse and children would also each receive individual benefits at a rate of 50% of the worker's PIA.(141)

Table 6

Benefit Calculation for High-Earnings Worker

Dying in 1980 at Age 30

 
Indexed Earnings
1972
$13,317.48*
1973
15,039.94*
1974
17,350.73*
1975
17,244.98*
1976
17,504.75*
1977
17,810.27
1978
17,700.00
1979
22,900.00
   
AIME

$ 1,622.00

   
PIA  
90% of first $194 of AIME
$ 174.60
32% of next $977 of AIME
312.64
15% of any balance
67.65
   
Total
$ 554.89

* Five "dropout years."

The AIME is lowered from $1,622.50 to $1,622.00 because of 20 C.F.R. 404.212a(b)(v) (1979), which states: "If the AIME as computed is not a multiple of $1, reduce it to the next lower multiple of $1." Id.

IV

Some Policy Consequences of the New Formula

A. Greater Equity for Those with Atypical Earnings Histories

The pre-1977 benefit formula effectively weighed later earnings more heavily than earlier earnings. This occurred because the formula averaged different years' earnings together without adjusting for general wage or price level differences among those years. In 1956, a $4,000 annual earnings figure represented above average earnings, falling just short of the $4,200 maximum countable amount.(142) By 1966, $4,000 was below average earnings.(143) Ten years later, $4,000 was below the poverty-level for a family of three and less than one-half the average earnings figure.(144) Prior to the 1977 amendments, however, all $4,000 years counted equally in the Social Security benefit calculation.

If a worker had a combination of average earnings and below average or no earnings years, the chronological placement of those years made a large difference. If the average earnings years were recent, they would court for $8,631 (1975) or $9,226 (1976); if they fell instead during the mid- to late-fifties they would be averaged in at less than $4,000.(145) Workers with earnings histories that equaled or exceeded the increase in average wages were adequately handled by the periodic adjustment of the multipliers used to convert average earnings to benefit amounts. Those who earned more in earlier years in relation to the average suffered. The system also disadvantaged workers who dropped out of the covered work force later in their careers. Conversely, the worker who had little covered employment until shortly before retirement received disproportionately generous benefits. In many instances, a worker's late spurt of covered work followed a full career in uncovered public employment with ample compensation and pension. This relative generosity seemed especially inequitable.

Since recent years' earnings counted for more under the pre-1977 formula, those whose benefits rested on a brief, recent wage record fared better than those with similar or higher recent earnings, but averaged over a longer period which included earlier-and therefore substantially lower-earnings. In this way the formula provided more favorable treatment to workers who wee disabled or died at a relatively young age than to workers with similar earnings who survived and contributed to a much later age. If a worker became disabled or died in his early twenties, his or his survivors' basic benefit might rest on average earnings over only his last two years of employment.(146) By contrast, the worker whose first claim for benefits came at age sixty-two had recent earnings averaged in with those paid some twenty years earlier.

The greater equity of a system of earnings-record indexing should be obvious. Earnings of $4,000 in 1956, 1966, and 1976 represent drastically different positions in the work force and amounts of purchasing power. Equitable treatment of individuals with disparate earnings profiles calls for some reflection of these differences, especially when benefit calculation rests on a lengthy period of past earnings. The two most obvious approaches involve weighting past earnings according to their purchasing power or relative position in the work force. Either approach provides more equitable treatment of workers with non-average wage profiles than does a system without indexing.

B. Further Growth of Benefits

Although it is clear that an indexed system is more desirable than the 1972 formula, the choice between price and wage indexing is far less obvious. Finding the appropriate measure for either approach is itself a challenge since no single price or wage benchmark seems appropriate for adjusting the past earnings of all covered workers.(147)

Beyond the significant problem of implementation -- which alone might influence the ultimate choice of method -- lies a more fundamental difference. Since 1950, wages have grown at a faster pace than prices due to increased economic productivity.(148) Indexing 1956 earnings as equal to their 1976 earnings equivalent, as distinct from what they would purchase in the latter year, builds that productivity gain into the Social Security benefit structure. The worker's retirement benefit bears a steady relationship (determined by the 1977 amendment multipliers) to earnings levels shortly before retirement.

By contrast, a system of price-indexing counts 1956 earnings based on their purchasing power in the later base year. Subsequent productivity gains are not factored in. Future benefits are not adjusted for wage increases that surpass price inflation. Over a lengthy period of productivity gains, a price-indexing system yields benefits that seem less adequate when measured against immediate pre-retirement earnings. A wage-indexing system ensures that future benefits will keep up with wage growth.

If future benefits are compared to present payments or to the gross national product, wage-indexing appears to commit the system to growth whereas price-indexing yields stability. On the other hand, when the criterion of replacement rate is used, wage-indexing appears to maintain a stable benefit level while price-indexing leads to gradual shrinkage-higher benefits but lower replacement rates.(149) Supporters of the 1977 Congressional wage-indexing scheme characterized it as the path of stability; opponents viewed it as a commitment to benefit growth. Both were right.(150)


Endnotes

* First appeared in 65 Cornell Law Review 748 (1980). Copyright © 1980, 2000 Peter W. Martin.

1. M. Derthick, Policymaking for Social Security 423 (1979).

2. Ch. 531, § 1104, 49 Stat. 620 (1935).

3. See Social Security Act of 1935, ch. 531, 201-210, 49 Stat. 620.

4. See Social Security Board, Social Security in America 222-24 (1937).

5. See A. Altmeyer, The Formative Years of Social Security 26 (1966).

6. See Social Security Board, supra note 4; E. Witte, The Development of the Social Security Act 147-51 (1962). For an account of the enduring tension between actuarial principles appropriate to private insurance and the demands of Social Security, see M. Derthick, supra note 1, 170-82.

7. The 1935 Act provided that appropriations to the Old-Age Reserve Account should be "determined on a reserve basis in accordance with accepted actuarial principles, and based upon such tables of mortality as the Secretary of the Treasury shall from time to time adopt. . . . " It also required the Secretary to report annually on the "actuarial status" of the account. Social Security Act of 1935, ch. 531, 201(a), 49 Stat. 620.

8. Social Security Act Amendments of 1939, ch. 666, 201, 53 Stat. 1360. See Martin, Public Assurance of an Adequate Minimum Income in Old Age: The Erratic Partnership Between Social Insurance and Public Assistance, 64 Cornell L. Rev. 437, 460-61 (1979).

9. At the time, the drafters believed that they could avoid tenth amendment problems by isolating the tax and benefit provisions of the Act. They hoped that the law would be justified as a valid exercise of Congress' taxing power. See E. Witte, supra note 6, at 146-47.

10. See Helvering v. Davis, 301 U.S. 619, 645 (1937) (validating the 1935 Act in spite of possible connection between tax and benefits). After the Court's decision, the Social Security Board began to refer to the program as "old-age insurance," terminology it had previously avoided. See Wollenberg, Vested Rights in Social-Security Benefits, 37 Oregon L. Rev. 299, 307 n. 25 (1958).

11. Social Security Act Amendments of 1939, ch. 666, 201, 53 Stat. 1360.

12. Id. 201(b).

13. The original proposal by the Committee on Economic Security suggested lower payroll taxes and contemplated substantial general revenue financing beginning in 1965. President Roosevelt insisted that the proposed system should balance without use of general revenues. The 1935 Act reflected this view. It contained a schedule of tax rates projected to yield a very large reserve. See E. Witte, supra note 6, at 74, 1470-51.

This potential reserve was one of the most controversial features of the legislation. See Crowley, Financing the Social Security Program -- Then and Now, in Subcomm. on Fiscal Policy of Joint Economic Comm., 93d Cong., 2d Sess., Studies in Public Welfare: Paper No. 18, Issues in Financing Retirement Income 24-26 (Comm. Print 1974) [hereinafter cited as Retirement Income Study].

14. See Retirement Income Study, supra note 13, at 26-29. To assure that heavy future benefit demands could be met, the 1935 legislation had set tax rates for the 1940's estimated to yield a large reserve. See note 13 supra.

15. Retirement Income Study, supra note 13, at 28. See Hearings Relative to the Social Security Amendments of 1939 Before the House Comm. On Ways and Means, 76th Cong., 1st Sess. 2113 (1939).

16. The Senate Finance Committee stated:

Unforeseen contingencies may, however, change the entire operation of the plan. It is important, therefore, that Congress be kept fully informed of the probable future obligations being incurred under the insurance plan. . . . Each generation may then meet the situation before it in such manner that it deems best.

If future annual pay-roll tax collections plus available interest are insufficient to meet future annual benefits it will be necessary, in order to pay the promised benefits, to increase the pay-roll tax or provide for ;the deficiency out of other general taxes, or do both.

S. Rep. No. 734, 76th Cong., 1st Sess. 18 (1939) (accompanying H.R. 6635).

17. Social Security Act Amendments of 1939, ch. 666, 201(b)(3), 53 Stat. 1360.

18. See Martin, supra note 8, at 460-61.

19. See Social Security Act Amendments of 1939, ch. 666, 902(g), 53 Stat. 1360.

20. Social Security Act of 1935, ch. 531, 1104, 49 Stat. 620.

21. The Supreme Court wrote in Flemming v. Nestor, 363 U.S. 603 (1960):

[The Social Security] program was designed to function into the indefinite future, and its specific provisions rest on predictions as to expected economic conditions which must inevitably prove less than wholly accurate, and on judgments and preferences as to the proper allocation of the Nation's resources which evolving economic and social conditions will of necessity in some degree modify.

To engraft upon the Social Security system a concept of "accrued property rights" would deprive it of the flexibility and boldness in adjustment to ever-changing conditions which it demands.

Id. at 610.

22. See text accompanying notes 40-56 infra.

23. See generally Wollenberg, supra note 10.

24. See Bernstein v. Ribicoff, 299 F.2d 248 (3d Cir.), cert. denied, 369 U.S. 887 (1962).

25. See Board of Trustees of the Federal Old-Age and Survivors Insurance Trust Fund, Second Annual Report, H.R. Doc. No. 694, 77th Cong., 2d Sess. 1-2 (1942).

26. See Soc. Sec. Bull.-Annual Statistical Supplement, 1975, Table 36, at 69.

27. See Martin, supra note 8, at 461-62.

28. See Retirement Income Study, supra note 13, at 29.

29. See Board of Trustees of the Federal Old-Age and Survivors Insurance Trust Fund, supra note 36, at 9.

The board of trustees, in its first annual report to the Congress, indicated that in the 5 fiscal years, 1941 to 1945, the trust fund would apparently exceed three times the highest annual expenditures anticipated during the 5-fiscal-year period. A similar relationship is now anticipated for the ensuing 5 fiscal years (1942 through 1946). This condition is again reported to the Congress in accordance with section 201 (b) (3) of the Social Security Act.

30. See id. The administration argued that the "three times rule" was not intended to apply to the early years of the system. It also asserted that the higher tax rates written into the law should be allowed to take effect to help finance the war effort and hold down inflation. See Retirement Income Study, supra note 13, at 29.

31. See A. Altmeyer, supra note 5, at 137-38.

32. See id. At 149-60.

33. See id. At 149.

34. See Revenue Act of 1943, ch. 63, 902, 58 Stat. 21 (1944).

35. See Korns, The Future of Social Security, in Retirement Income Study, supra note 13, at 1, 4. Estimates of the reserve fund in 1939 projected that the fund would reach $7-8 billion by the end of 1955. Notwithstanding the tax rate freeze during the forties, the fund was nearly $12.9 billion by the end of June 1950. See Retirement Income Study, supra note 13, at 29.

36. President Truman's May 24, 1948 Special Message to Congress on Social Security urged more adequate benefits on the following grounds:

The benefits being paid under old-age and survivors insurance are seriously inadequate. They were adjusted last in 1939. Even then, the benefits in most cases replaced only a small part of the income that the worker or his survivors had lost because of his retirement or death. Earnings and the cost of living have risen sharply since that time and cannot be expected to return to prewar levels. Consequently, further adjustments in benefit rates are imperative.

People whose sole income is from social security payments have just about reached the breaking point. . . .

The present average payment for a retired worker is only about $25 a month, and is substantially less for dependents and survivors. If the insurance system is to prevent dependency upon public and private charity this amount is obviously far too low. I recommend that the Congress increase benefits by at least 50 per cent.

1948 Pub. Papers 272-73 (1964).

37. See Senate Committee on Finance, Recommendations for Social Security Legislation: The Reports of the Advisory Council on Social Security, S. Doc. No. 208, 80th Cong., 2d Sess. 9-10, 35 (1949) [hereinafter cited as 1948 Advisory Council Report]. The Advisory council was appointed September 17, 1947, as a concession by the Senate Finance Committee responding to the administration's displeasure with the repeated tax increase postponements. See Altmeyer, supra note 16, at 160; M. Derthick, supra note 1, at 91-92. Its final report was dated December 1948, but the OASI portions were issued separately on April 20, 1948. See Advisory Council on Social Security, Old-Age and Survivors Insurance: A Report to the Senate Committee on Finance, S. Doc. No. 149, 80th Cong., 2d Sess. (1948).

38. 1948 Advisory Council Report, supra note 37, at 32. "In 1945 about 14 percent of all covered workers had wages exceeding $3,000. . . . " Id.

39. Id. At 31-32. The report stated:

For millions of persons the social security system represents a guaranty of future security. If that guaranty is to be valid and meaningful, the purchasing power of benefits must not be destroyed by large increases in price levels. A special obligation rests on the Government and all groups in the community with an interest in the social-insurance system and the security it offers to make sure that the monetary policies, price policies, and wage policies contribute to the objective of preventing such a large rise in the price level. If the people of the United States are unable to prevent steep increases in price levels, benefits will have to be readjusted to preserve their purchasing power for unless the purchasing power of th3ee benefits is preserved, the security guaranteed by the social-insurance plan will be illusory.

Id. at 13-14.

40. See Social Security Act Amendments of 1950, ch. 809, 64 Stat. 477; A. Alteyer, supra note 5, at 163, 169-87.

41. See Martin, supra note 8, at 464.

42. See id.

43. Social Security Act Amendments of 1950, ch. 809, 109(a)(3), 64 Stat. 477.

44. S. Rep. No. 1669, 81st Cong., 2d Sess. 33, reprinted in [1950] U.S. Code Cong. & Ad. News 3287, 3322:

Your committee has very carefully considered the problems of cost in determining the benefit provisions recommended. Also your committee is of the belief that the old-age and survivors insurance program should be on a completely self-supporting basis. Accordingly, the committee-approved bill, just as the House-approved bill, eliminates the provision added in 1943 authorizing appropriations to the program from general revenues. At the same time, your committee has recommended a tax schedule which it believes will make the system self-supporting as nearly as can be foreseen under present circumstances.

Id.

45. See notes 49-51 and accompanying text infra.

46. Social Security Act Amendments of 1950, ch. 809, 201(a), 203(a), 64 Stat. 477.

47. See S. Rep. No. 1669, 81st Cong., 2d Sess. 33-52, reprinted in [1950] U.S. Code Cong. & Ad. News 3287, 3322-43.

48. Actuarial balance calculations accompanied all important Social Security amendments during this period. Congress consistently declared fiscal balance to be a constraining principle. See, e.g., S. Rep. No. 1856, 86th Cong., 2d Sess. 37, reprinted in [1960] U.S. Code Cong. & Ad. News 3608, 3644:

The Congress has always carefully considered the cost aspects of the old-age, survivors, and disability insurance system when amendments to the program have been made. In connection with the 1950 amendments, the Congress was of the belief that the program should be completely self-supporting from the contributions of covered individuals and employers. Accordingly, in that legislation, the provision permitting appropriations to the system from general revenues of the Treasury was repealed. This policy has been continued in subsequent amendments. Thus, the Congress has always very strongly believed that the tax schedule in the law should make the system self-supporting as nearly as can be foreseen and, therefore, actuarially sound.

The reserve fund bench mark of three times annual expenditures established in 1939 was formally abandoned in 1960. See 1960 Social Security Amendments, Pub. L. No. 86-778, 701, 74 Stat. 924. By then the reserve no longer exceeded that standard and the 1959 Advisory Council recommended the deletion: "We do not believe that the trust funds should be held to any arbitrary relationship to expected annual expenditures, and we recommend that the provision be repealed." Board of Trustees of the Federal Old-Age and Survivors Insurance Trust Fund and the Federal Disability Insurance Trust Fund, 19th Annual Report, H.R. Doc. No. 181, 86th Cong., 1st Sess. 68, app. Iv (1959).

49. See Social Security Act Amendments of 1950, ch. 809, 104, 64 Stat. 477.

50. The statutory term was always "Average Monthly Wage," but Social Security literature, including agency publications, has also used the term "Average Monthly Earnings." See, e.g., J. Pechman, H. Aaron, & M. Taussig, Social Security: Perspectives for Reform 79 (1968); Soc. Sec. Bull.-Annual Statistical Supplement, 1975, at 7.

51. Social Security Act Amendments of 1950, ch. 809, 104, 64 Stat. 477.

52. See Soc. Sec. Bull.--Annual Statistical Supplement, 1975, at 18.

53. Id. At 20-22.

54. See 42 U.S.C. 402(q)(1) (1976).

55. See Soc Sec. Bull--Annual Statistical Supplement, 1975, at 18-19.

56. Typically the amendment would set forth a new benefit table with one column converting AIAs established under prior law into the new higher amounts, and a second column applying to subsequently qualifying beneficiaries, that contained the new Average Monthly Wage-PIA relationship. See, e.g., Social Security Amendments of 1965, Pub. L. No. 89-97, 301, 79 Stat. 286.

57. See Soc Sec. Bull--Annual Statistical Supplement, 1975, at 32.

58. See, e.g., S. Rep. No. 1987, 83d Cong., 2d Sess. 1-3, reprinted in [1954] U.S. Code Cong. & Ad. News 3710, 3710-12.

59. See, e.g., S. Rep. No. 2388, 85d Cong., 2d Sess., reprinted in [1958] U.S. Code Cong. & Ad. News 4218, 4220.

60. S. Rep. No. 1987, 90d Cong., 1st Sess. 1-3, reprinted in [1967] U.S. Code Cong. & Ad. News 2834, 2873.

61. See Social Security Amendments of 1956, ch. 836, 116(e), 70 Stat. 807.

62. See Social Security Amendments of 1960, Pub. L. No. 86-778, 704(b), 74 Stat. 924.

63. See, e.g., Advisory Council on Social Security, The Status of the Social Security Program and Recommendations for Its Improvement 4, 54-60 (1965). On the important role of the advisory councils, see M. Derthick, supra note 1, at 89-109.

64. The assumption of a static law was often justified by Social Security Administration actuaries on the ground that it would be presumptuous for them to assume specific Congressional changes. See Retirement Income Study, supra note 13, at 95.

65. See S. Doc. No. 18, 80th Cong., 1st Sess. 29 (1947). The Council continued:

With such assumptions, the rise in wages would seem to offer significant financial help in the financing of benefits because pay-roll contributions at a fixed percentage rate would increase steadily relative to benefit disbursements; but benefits paid to beneficiaries would steadily diminish in relation to current wage levels. In such a case, it is likely that the present formula would not be maintained, but rather revisions would be adopted which would make average benefits at least as adequate relative to the then existing average wage level as average benefits under the present formula were in relation to the 1939 wage level.

Id. at 30.

66. See, e.g., Advisory Council on Social Security, The Status of the Social Security Program and Recommendations for Its Improvement 16 (1965):

The Council believes that making the estimates on a level-wage assumption allows for a desirable margin of safety and recommends that the practice be continued in making the long-range estimates.

67. Board of Trustees of the Federal Old-Age and Survivors Insurance Trust Fund and the Federal Disability Insurance Trust Fund, 25th Annual Report. H.R. Doc. No. 80, 88th Cong., 1st Sess. 32-33 (1963).

68. Id. at 33.

69. Board of Trustees of the Federal Old-Age and Survivors Insurance Trust Fund and the Federal Disability Insurance Trust Fund, 25th Annual Report. H.R. Doc. No. 80, 89th Cong., 1st Sess. 32 (1965).

70. In 1967, Social Security Commissioner Robert M. Ball wrote:

There is some leeway for improvement in [benefits in] the future without a Government contribution [from general revenues] and without increasing the contribution rate. . . . [I]f the maximum earnings base is increased somewhat from time to time (it does not have to be increased proportionately to increases in earnings levels for this purpose), contribution rates in present law will produce sufficient income to considerably more than keep benefits adjusted to future increases in prices.

Subcommittee on Fiscal Policy of the Joint Economic Comm. 90th Cong., 2d Sess., Old Age Income Assurance 56 (Comm. Print 1967).

Annual reports of the Board of Trustees during the late sixties and early seventies contained the following warning:

It is important to note that these estimates are made on the assumption that earnings will remain at about the level prevailing in [the previous year]. If earnings levels rise, as they have in the past, the benefits and the taxable earnings base under the program will undoubtedly be modified. In fact, if all other assumed cost factors are closely followed by the experience, then increasing wage levels will automatically generate positive actuarial balances that can be used to increase benefit levels without changing the financing provisions.

See, e.g., Board of Trustees of the Federal Old-Age and Survivors Insurance Trust Fund and the Federal Disability Insurance Trust Fund, 1971 Annual Report. H.R. Doc. No. 88, 92nd Cong., 1st Sess. 39 (1971).

71. The precise terms of the recommendation were:

The law should provide a guarantee that, in the absence of congressional action, social security benefits will be at least kept up to date, automatically, with increases in prices, and that the contribution and benefit base and the retirement test exempt amount will be automatically adjusted to increases in earnings levels.

Advisory Council on Social Security, Reports on the Old-Age, Survivors, and Disability Insurance and Medicare Programs17 (1971).

The Council's recommendation did not reflect a belief that Congress should abstain from future ad hoc increases:

[I]f social security benefits are increased to keep up only with increases in the cost of living, the levels of living of people on the benefit rolls will gradually decline relative to families of people still working. To assure that this situation will not continue over a long period of time, the Council believes that the Congress in its future deliberations should take into account increases in the general standard of living and provide benefit increases from time to time that are higher than increases in the cost of living.

Id. at 20-21.

72.

[T]he actuarial cost estimates for the cash benefits program [should] be based-as the estimates for the hospital insurance program now are-on the assumptions that earnings levels will rise, that the contribution and benefit base will be increased as earnings levels rise, and that benefit payments will be increased as prices rise.

Id. at 85.

73. The actuarial recommendation was not, however, conditioned on adoption of the automatic adjustment recommendation. See id. at 86.

74. See Act of July 1, 1972, Pub. L. No. 92-336, 201, 202, 86 Stat. 406.

75. See Haldman, OASDI Benefits, Prices, and Wages: 1966 Experience, Soc. Sec. Bull., June 1967, at 9.

76. See, e.g., L. Epstein, Income Security Standards in Old-Age 12-13 (Social Security Administration Research Rep. No. 3, 1963); 1971 Advisory Council on Social Security, Reports on the Old-Age, Survivors, and Disability Insurance and Medicare Programs 19-20 (1971); Reno & Zuckert, Benefit Levels of Newly Retired Workers: Findings From the Survey of New Beneficiaries, Soc. Sec. Bull., July 1971, at 3, 22-24.

77. See, e.g., Advisory Council on Social Security, Reports on the Old-Age, Survivors, and Disability Insurance and Medicare Programs 19 (1971) ("very close to the poverty threshold").

78. See Martin, supra note 8, at 479-88.

79. See Retirement Income Study, supra note 13, at 39-47, 65-82; J. Brittain, The Payroll Tax for Social Security 1-20 (1972); J. Pechman, H. Aaron, & M. Taussig, supra note 50, at 173-213.

80. Some experts writing on Social Security during the 1960s expressed the need for an appropriate external standard for evaluating system performance. See, e.g., L. Epstein, supra note 76; Bok, Emerging Issues in Social Legislation: Social Security, 80 Harv. L. Rev. 717, 752-59 (1967).

81. During the early years of the system, nearly all age cohorts, family groups, and income segments received benefits far in excess of contribution. Attention to relative return on contribution has grown in recent years as certain sizable groups (e.g., working spouses) have attacked the system's failure to give them any appreciable benefit increment for their contribution and as the prospect of an inadequate return for young worker-contributors becomes a reality. See Advisory Council on Social Security, Social Security Financing and Benefits 61-63 (1979); HEW, Social Security and the Changing Roles of Men and Women 28-29 (1979); R. Kaplan, Indexing Social Security: An Analysis of the Issues 66-67 (1977).

82. A Social Security Administration study of new beneficiaries published in 1971 made no use of the replacement rate concept. See Reno & Zuckert, supra note 76. By the mid-seventies, however, some were suggesting that replacement rates held too great a dominance over Social Security analysis. See R. Kaplan, supra note 81, at 41-43. See also Consultant Panel on Social Security to the Congressional Research Service, 94th Cong., 2d Sess. Report 21 (Joint Comm. Print 1976).

83. For example, the important book by Pechman, Aaron and Taussig, Social Security: Perspectives for Reform, published in 1968, analyzed the "replacement rates" under the 1967 amendments to the Social Security Act for various beneficiary groups, but calculated its replacement rate by comparing "monthly benefit awards" to "average monthly earnings." See J. Pechman, H. Aaron, & M. Taussig, supra note 50, at 80. The book does, however, sketch the advantages of indexing wage records particularly during periods of inflation. See id. at 97-98.

During the late fifties and early sixties there was far less reason to construct a distinct replacement rate concept since the Social Security benefit formula came close to resting on immediate pre-retirement earnings. The retired worker's average monthly earnings figure represented only a few years, and for many workers, the highest years were those immediately prior to retirement.

Social Security Administration research reports on replacement rates across time were similarly based on the legislation's benefit formula "average monthly earnings" figure. See Ceccarelli & Skolnik, Wage-Replacement Rates Under the OASDHI Retirement Provisions (Social Security Administration, Research & Statistics Note No. 18, 1973).

On those rare occasions when Congressional commmittees used the term, benefits were compared to average earnings. See H.R. Rep. No. 544, 90th Cong., 1st Sess. 22 (1967):

[Y]our committee has carefully studied the matter of wage-replacement upon retirement, disability or death of the wage-earner and has sought to establish a reasonable relationship between former wages and benefits. Thus, the bill embodies the principle that the retirement benefit for a man age 65 and his wife should represent at least 50 percent of his average wages under the social security system. (Emphasis added).

84. See Merriam, Income Maintenance, in Social Security in International Perspective 55-82 (S. Jenkins ed. 1969); J. Schulz, G. Carrin, H. Krupp, M. Peschke, F. Sclar, & J. Van Steenberge, Providing Adequate Retirement Income: Pension Reform in the United States and Abroad 18-20 (1974); Henley, Recent Trends in Retirement Benefits Related to Earnings, Monthly Lab. Rev., June 1972, at 12; Horlick, The Earnings Replacement Rate of Old-Age Benefits: An International Comparison, Soc. Sec. Bul., Mar. 1970, at 3.

85. Advisory Council on Social Security, supra note 71, at 125.

86. Quadrennial Advisory Council on Social Security, Report, H.R. Doc. No. 75, 94th Cong., 1st Sess. 15, 103-06.

87. Id. at 15.

88. Id. at 103.

89. This potential instability of the "coupled" formula was recognized by the experts, but discounted Robert Myers, Chief Actuary for the Social Security Administration, testified in 1969 that the automatic adjustment technique (enacted eventually in 1972) would work well "unless we came into a time when there was a run-away inflation," but he concluded that such a possibility could "be ignored." Social Security and Welfare Proposals, Hearings Before the House Comm. On Ways and Means, 91st Cong., 1st Sess. 204 (1970).

Decoupled methods of indexing had been discussed prior to 1972, but compared to the 1972 approach they seemed more complicated and a sharper break with the past. See Cohen, Federalism and Social Insurance, in The Princeton Symposium on the American System of Social Insurance 12-14 (1968).

90. Congressional Budget Office, Financing Social Security: Issues for the Short and Long Term 42 (1977). Early conservative wage and price predictions produced stable projected replacement rates. See id.; Board of Trustees of the Federal Old-Age and Survivors Insurance and Disability Insurance Trust Funds, 1973 Annual Report, 1973 Annual Report H.R. Doc. No. No. 130, 93d Cong., 1st Sess. 38 (1973).

91. Indeed, an automatic adjustment provision has previously passed the House in 1970 with the support of the Republicans, who favored automatic increases over the ad hoc increases used by a Democratic Congress for political gain. See M. Derthick, supra note 1, at 345-50.

The 20% benefit increase, passed at the same time as the automatic adjustment provision, was the controversial feature of the 1972 legislation. See 28 Congressional Quarterly Almanic 399 (1972).

92. The Council urged:

The law should provide a guarantee that, in the absence of congressional action, social security benefits will be at least kept up to date, automatically, with increases in prices, and that the contribution and benefit base and the retirement test exempt amount will be automatically adjusted to increases in earnings levels.

Advisory Council on Social Security, supra note 71, at 7.

93. The automatic adjustment of benefits to changes in prices can be fully financed over the long run without any increases in the contribution rates if the contribution and benefit base is increased as wage levels increase. Earnings levels tend to increase faster than prices, since they reflect productivity increases as well as price increases, and only about one half of the increases in income to the social security program that result from rising wages will be needed to meet the cost of higher benefit payments resulting from the higher wages credited to the worker's social security earnings record. If the contribution and benefit base is kept up to date with rising earnings, the additional income that becomes available to the program and that is not needed to meet additional benefit liabilities based on the higher covered earnings would at least be sufficient over the long run to finance automatic increases in benefits.

Id. at 18.

94. Id. at 113.

The calculations attesting to the actuarial soundness of the 1972 amendments assumed average annual price increases of 2 3/4% and 5% average annual wage increases. The 1973 trustees' report estimated that a difference of 1/2% in the average rate of price increases would produce a significant long-term deficit. See Board of Trustees of the Federal Old-Age and Survivors Insurance and Disability Insurance Trust Funds, 1973 Annual Report, H.R. Doc. No. 130, 93d Cong., 1st Sess. 31, 41-42 (1973).

95. See Act of July 1, 1972, Pub. L. No. 92-336, 86 Stat. 406; M. Derthick, supra note 24, at 58-62. John Byrnes, senior Republican member of the House Ways and Means Committee, protested strenuously on the floor of the House:

[T]o appear here as a member of the Committee on Ways and Means, to discuss a matter of this importance with so little information for the members of . . . the Congress, is something that I thought I would never have to experience in my service in this House.

This is no way to treat vitally important legislation.

. . .

At no point has there been a study by the Ways and Means Committee of the new method of [calculating actuarial soundness] that has produced the "windfall" that now is going to be used for the 20 percent benefit increase. Not one word of testimony in public or executive session has been received on this subject. This fundamental change in the criteria by which the soundness of the social security trust fund has been measured for one-third of a century is being adopted willy-nilly by the Congress without even a cursory review.

118 Cong. Rev. 23,732-33 (1972).

96. See Act of July 9, 1973, Pub. L. No. 93-66, 201, 87 Stat. 152 (1973).

97. See 119 Cong. Rec. 21,582, 21,587, 21,622-28, 22,394-404, 22,605-10, 22,642-46 (1973). See also Retirement Income Study, supra note 13, at 62-63.

Parliamentary objections in the House caused the amendments to be deleted from the public debt bill and added to a bill extending the Renegotiation Act. Act of July 9, 1973, Pub. L. No. 93-66, 87 Stat. 152. See 119 Cong. Rec. 22,382, 22,398-404 (1973); Retirement Income Study, supra note 13, at 63.

98. See Act of December 31, 1973, Pub. L. No. 93-233, 1, 2, 87 Stat. 947 (1973); Retirement Income Study, supra note 13, at 63-65.

99. See Act of December 31, 1973, Pub. L. No. 93-233, 1, 87 Stat. 947 (1973).

100. Id. 3.

Under the original adjustment provisions, the cost-of-living figures for the second quarter of the year produced benefit increases effective the following January. The amended provisions used first quarter figures and adjusted benefits beginning in June of the same year. See H.R. Rep. No 627, 93d Cong., 1st Sess. 6-7 (1973).

101. See Act of July 9, 1973, Pub. L. No. 93-66, 202, 87 Stat. 152 (1973); Act of December 31, 1973, Pub. L. No. 93-233, 5, 87 Stat. 947 (1973).

102. Notice of the 8% increase was issued on May 22, 1975. See 40 Fed. Reg. 22,289 (1975).

Nearly a year before, the annual report of the Board of Trustees forecast a significant deficit-approximately 3% of taxable payroll. While the report attributed most of the deficit to new demographic assumptions, it also drew attention to the dependence of the financial condition of the system on unpredictable long term economic factors. See Board of Trustees of the Federal Old-Age Survivors Insurance and Disability Insurance Trust Funds, 1974 Annual Report, H.R. Doc. No. 313, 93d Cong., 2d Sess. 27, 34-38 (1974).

103. See M. Derthick, supra note 1, at 381-84.

104. See A. Munnell, The Future of Social Security at 93-99 (1977).

105. See Quadrennial Advisory Council on Social Security, supra note 86, at 13.

106. See, e.g., A. Munnell, supra note 104, at 31; Quadrennial Advisory Council on Social Security, supra note 86, at 131-32; L. Thompson, An Analysis of the Factors Currently Determining Benefit Level Adjustments in the Social Security Retirement Program 16-17 (HEW, Office of Income Security Policy, Technical Analysis Paper No. 1, 1974):

If allowed to operate without any offsetting influences, these adjustment provisions can, over time, have the result of increasing replacement ratios to 1.00 or more, implying retirement benefits that equal or exceed the driver's previous wages. While it is possible for reasonable men to disagree over what constitutes a reasonable replacement ratio, it ought to be possible for them to agree that there is something amiss with the present arrangement. For the behavior of replacement ratios over time ought to be controlled by conscious policy decision and to the point where the average worker will get a retirement benefit that is greater than his wage.

107. See Soc. Sec. Bull.-Annual Statistical Supplement, 1975, at 18.

108. See id. at 32.

109. Assuming that the worker's five lowest earnings years were 1951-55, they were not averaged in, but the formula counted earnings for 1956-58, when the wage base was only $4,200 and the median earnings figures for a covered worker were even less, 1959 when the wage base was $4,800, and so on. See id. at 32, 72.

110. See Table 2 supra.

111. See Act of July 1, 1972, Pub. L. No. 92-336, 215)I)(2)(D)(v), 86 Stat. 406.

112. The effects of this double counting were softened by two additional features of the benefit formula. First, higher average wages moved a worker into different brackets with lower multipliers. While the multipliers in each bracket were adjusted with price increases, the boundaries between brackets remained the same. Second, through 1994 each successive class of retirees had an additional year's earnings averaged in. The additional year was a year of early and therefore relatively low earnings. See A. Munnell, supra note 104, at 33-36; L. Thompson, supra note 106, at 17-24.

113. See Thompson, Indexing Social Security: The Options, in Financing Social Security 41 n.1 (C. Campbell ed. 1979). In addition, the formula might overadjust benefits for inflation because it ignored the economic connection between wage and price inflation, making separate adjustments for both.

114. The House Ways and Means Committee report on the 1977 amendments stated that "[c]urrent projections show that benefit levels will rise by about 50 percent more than wages over the next 75 years, with most of this increase occurring after the 1990's. Replacement rates can fluctuate widely in the future, either up or down, depending on future changes in wages and prices." H.R. Rep. No. 702, 95th Cong., 1st Sess. 22, reprinted in [1977] U.S. Code Cong. & Ad. News 4155, 4179.

A table in the report showed the replacement rate for a low-earnings worker crossing 100% about the year 2040. Id. at 24, reprinted in [1977] U.S. Code Cong. & Ad. News at 4181.

115. Experts adopted the term "decoupling" to describe the policy alternatives devised to correct the problem of "coupling" in the 1972 automatic adjustment provisions. See Thompson, supra note 113, at 41 n.1.

116. See 42 U.S.C. 415(a)-(b) (Supp. I 1977); 20 C.F.R. 404.212 (1979).

117. See 42 U.S.C. 415(i)(2)(a)(ii) (Supp. I 1977); 20 C.F.R. 404.221 (1979).

118. See 42 U.S.C. 415(b)(3)(A) (Supp. I 1977); 43 Fed. Reg. 60,877-78, 61,016 (1978).

119. See 42 U.S.C. 415(b)(3)(A) (Supp. I 1977); 20 C.F.R. 404.212a(b) (1979).

120. See 44 Fed. Reg. 62,956-57 (1979).

121. See 42 U.S.C. 415(b)(3)(A) (Supp. I 1977); 20 C.F.R. 404.212a(b) (1979).

122. See 42 U.S.C. 415(b)(2) (Supp. I 1977).

123. See 42 U.S.C. 415(b)(3)(B) (Supp. I 1977).

A high-earnings worker's indexed earnings will reach the indexed maximum earnings figure each year. Thus, the omitted years will be the five years during the base period with the lowest maximum figure after indexing, in this case, 1965, 1964, 1963, 1962 and 1954. See Table 3 supra.

124. See Martin, supra note 8, at 495.

125. See 42 U.S.C. 415(a)(1)(A)-(1)(B) (Supp. I 1977).

126. See id., 44 Fed. Reg. 62,956 (1979).

127. See 42 U.S.C. 415(i)(2)(A)(ii) (Supp. I 1977). See 45 Fed. Reg. 31,781 (1980).

128. See 42 U.S.C. 415(b)(3)(B) (Supp. I 1977); 20 C.F.R. 404.212a(b)(v) (1979).

129. See 42 U.S.C. 430(c) (Supp. I 1977).

130. See Table 5 supra. During 1980, the 15% rate applies to all amounts of AIME above $1,171. See text at note 126 supra. That is the equivalent of indexed annual earnings slightly above $14,000. Only a worker with maximum or near maximum taxable earnings in recent years will be in that range. Compare the AIMEs calculated for the high-earnings and middle-earnings workers at notes 123-24 and accompanying text supra.

131. See 42 U.S.C. 402(q)(1)(A) (1976).

132. Id.

133. The 3% figure applies to workers reaching sixty-two after December, 1978. See 42 U.S.C. 402(2) (1976).

134. See 42 U.S.C. 415(i)(2)(A)(iii) (Supp. I (1977).

135. See notes 118-26 and accompanying text supra.

136. See 42 U.S.C. 415(b)(2)(B)(iii) (Supp. I 1977). A minimum of two earnings years is required. See 42 U.S.C. 415(b((2)(A) (Supp. I 1977).

137. Id.

138. The worker turned twenty-one in 1971. There are eight years between 1972 and 1980, of which the five lowest indexed earnings years are omitted.

The Social Security Disability Amendments of 1980 altered the benefit calculation for some disability insurance claimants. This occurred as the result of two changes in the formula. First, the legislation substituted a variable dropout year schedule for the flat five years in the present law. A disability benefit recipient under twenty-seven will have his or her AIME calculated with only one low earnings year rather than five excluded from the average. For workers between their twenty-seventh and thirty-second birthdays, the amendments provided two dropout years, and so on, adding one dropout year for each five of additional age. Claimants forty-seven and older will continue to have the full five years. Social Security Disability Amendments of 1980, Pub. L. No. 962-65, 96th Cong.2d Sess. 3102, 94 Stat. 441.

The second change had no effect on the worker's benefit per se, but cut back on the total family benefits payable on account of disability for some middle and low-income workers. The reduction took the form of a new family maximum set in relation to the worker's AIME an PIA. Id. 101.

139. See 42 U.S.C. 402(b)-(d), 403(a) (Supp. I 1977).

140. See 42 U.S.C. 423(a)(1) (1976).

141. See 42 U.S.C. 402(b), (c), (d) (1976 & Supp. I 1977).

142. See Table 3 supra; Soc. Sec. Bull.-Annual Statistical Supplement, 1975, at 32.

143. See Table 3 supra.

144. See id.; Soc. Sec. Bull.-Annual Statistical Supplement, 1975, Table 9, at 46.

145. See Table 3 supra.

146. See note 136 supra.

147. See R. Kaplan, Indexing Social Security: Analysis of the Issues 27-32 (1977).

148. See A. Munnell, supra note 104, at 52-55.

149. See M. Derthick, supra note 1, at 403-04. Derthick sees the replacement rate concept as critical to the selection of wage-indexing.

The success of the [wage-indexing] program's leadership, then, must be attributed in some significant measure to their capacity to define the terms of debate. They supplied more than a proposal for action; they also supplied the language with which perceptions were shaped. With their emphasis on replacement rates, they defined the base from which changes in direction were to be guaged, and in a system where backward steps are inconceivable and the tendency of policymaking is either toward more of the same (inertia) or more of the same plus a bit (incremental change), the capacity to define where one is starting form is of crucial importance.

Id. at 405.