CRS Examines Social Security Benefits for Past, Present, and Future Retirees

                       CRS REPORT FOR CONGRESS

SUMMARY

[1] In recent years considerable public attention has focused on Social Security's treatment of younger versus older workers. Analysts sometimes have addressed this issue by examining the value Social Security provides each generation of workers in relationship to the Social Security taxes they pay. These are referred to as "moneysworth" analyses.

[2] This paper approaches the issue by presenting illustrations of the time it takes retirees to recover the value of the Social Security taxes they paid. The illustrations cover a range of outcomes, depending on variations in assumptions used. The range of the illustrations is wide. The variables with the largest effect are the earnings history of the worker, whether or not the employer share of taxes is included, and the rate of interest imputed on the value of these taxes. Following are examples:

     Under the economic assumptions most commonly used by the Social

                              CONTENTS

[4] In recent years, considerable public attention has been focused on the broad philosophical issue of the relative economic well-being of the young versus the old. Various commentators have stated that today's children may be worse off than their parents. The improvement of economic conditions of the elderly compared to the increase in poverty among children, the increased income and payroll tax burden of current and future workers compared to that levied on retirees while they were working, and the high costs associated with the looming retirement of the "baby boom" that will force cuts in benefits and/or increases in taxes, are among the reasons cited for concern about relative well-being among the generations.

[5] Social Security is at the center of these discussions. Commentators point out that most current recipients receive more, often far more, than the value of the Social Security taxes they have paid, even when the employer's matching share and interest are included, whereas projections show that future recipients will receive a much poorer deal from the program. Such descriptions are usually referred to as "moneysworth" analyses.

THIS PAPER'S APPROACH TO THE ISSUE

[6] In addition to all the technical factors that must be addressed, the nature of the Social Security law complicates such computations. Not only do analysts disagree on the proper techniques to use in making calculations, there are often fundamental disagreements involving subjective factors: what work patterns to use; what part of the Social Security tax to count; whether or not to include dependents' benefits; whether or not to include the employer's share of the tax; what rate of interest to use; and whether or not to include the effect of income taxation of benefits.

[7] This paper's analysis seeks to avoid judgmental conclusions by providing a range of illustrations that vary these subjective factors. It does not evaluate the "moneysworth" of Social Security, nor is it an "actuarial analysis" of how whole age cohorts fare. Rather, it simply presents illustrations of the amount of time it takes, and is projected to take, to recover the value of taxes paid plus interest. ¹ The illustrations represent a range of possible payback times, depending on variations in the assumptions used. In this way, no judgments need be made -- the use of the illustrations is the reader's choice.

METHODOLOGY

[8] The illustrations in this report show the payback times of workers with three representative levels of earnings who retire at age 65 ² after working a full career in jobs covered by Social Security. The lifetime earnings patterns are workers who always earned either: (1) the federal minimum wage; (2) a wage equal to Social Security's "average wage series," or (3) the maximum wage creditable under Social Security. A range of outcomes is provided by varying several factors. These factors are:

     what part of the Social Security tax is counted -- payment of

ILLUSTRATIVE PAYBACK TIMES

[10] Table 1 shows past and projected payback times for workers retiring in various years from 1940 to 2030. The examples end at 2030 because, as discussed in the appendix, taxes and benefits under current law would not be applicable after 2038, as the system is projected to be insolvent unless action is taken to increase revenue and/or reduce benefits.

[11] The interest rates used in Tables 1, 2 and 4 are equivalent to those earned, and projected to be earned, on a portfolio of long-term U.S. government bonds. Table 3 shows payback times for workers retiring in 2030 with various real (i.e., above inflation) interest rate assumptions. Table 4 shows a range of payback times for workers retiring in 2030 with benefits reduced to reflect the potential minimum and maximum effect of the taxation of benefits.

[12] In the illustrations in Tables 1, 3 and 4, benefits are for the worker alone. However, the value of the benefit could be higher if the worker had dependents who were eligible for benefits. For example, if these workers had spouses who also were the full retirement age and were not entitled to a Social Security benefit on their own account, then the value of the monthly benefit would increase by 50%. ³ Table 2 shows illustrations that assume that a dependent spouse of the same age is eligible for full spousal benefits.

      TABLE 1. NUMBER OF YEARS TO RECOVER TAXES PLUS INTEREST FOR

______________________________________________________________________ YEAR OF MINIMUM AVERAGE MAXIMUM RETIREMENT EARNER EARNER EARNER ______________________________________________________________________ Illustration 5: Years to recover employee's OASI taxes

1940 ** ** 0.1 1960 0.3 0.5 0.6 1980 1.0 1.3 1.4 2001 4.6 6.4 8.5 2010 5.2 7.5 10.3 2020 5.4 8.0 12.3 2030 4.9 8.0 12.9

Illustration 6: Years to recover combined employee-employer OASI taxes

1940 ** 0.2 0.3 1960 0.7 1.0 1.3 1980 2.0 2.6 2.9 2001 10.0 14.4 19.9 2010 11.5 17.1 25.3 2020 11.9 18.5 32.0 2030 10.7 18.4 34.0

Illustration 7: Years to recover retirement portion of employee's OASI taxes

1940 ** ** 0.1 1960 0.3 0.4 0.5 1980 0.7 1.0 1.1 2001 3.4 4.7 6.2 2010 3.9 5.5 7.5 2020 4.1 6.0 9.1 2030 3.9 6.2 9.8

Illustration 8: Years to recover retirement portion of combined employee-employer OASI taxes

1940 ** 0.2 0.3 1960 0.5 0.7 0.9 1980 1.5 1.9 2.1 2001 7.2 10.2 13.9 2010 8.2 12.1 17.3 2020 8.7 13.3 21.7 2030 8.2 13.7 23.8 ______________________________________________________________________

/**/ Less than 0.1 years.

______________________________________________________________________ MINIMUM AVERAGE MAXIMUM EARNER EARNER EARNER ______________________________________________________________________

Illustration 9: 1%real (4.33% nominal)

Employee's OASI taxes 4.6 7.3 11.2 Combined OASI taxes 9.4 15.1 23.9 Retirement portion of OASI taxes 3.7 5.8 8.9 Retirement portion of combined 7.5 11.9 18.7 OASI taxes

Illustration 10: 2% real (5.37% nominal)

Employee's OASI taxes 5.9 9.4 15.1 Combined OASI taxes 12.5 21.1 36.7 Retirement portion of OASI taxes 4.6 7.4 11.7 Retirement portion of combined 9.7 16.1 26.8 OASI taxes

Illustration 11: 3% real (6.40% nominal)

Employee's OASI taxes 7.7 12.8 21.8 Combined OASI taxes 17.7 33.6 102.6 Retirement portion of OASI taxes 6.3 9.5 14.9 Retirement portion of combined 13.2 23.5 47.9 OASI taxes

Illustration 12: 4% real (7.43% nominal)

Employee's OASI taxes 10.5 18.7 39.3 Combined OASI taxes 28.8 infinite infinite Retirement portion of OASI taxes 7.9 13.5 24.9 Retirement portion of combined 19.5 44.4 infinite OASI taxes

Illustration 13: 5% real (8.46% nominal)

Employee's OASI taxes 15.4 33.9 infinite Combined OASI taxes infinite infinite infinite Retirement portion of OASI taxes 11.1 23.3 infinite Retirement portion of combined 36.8 infinite infinite OASI taxes ______________________________________________________________________

______________________________________________________________________ MINIMUM AVERAGE MAXIMUM EARNER EARNER EARNER ______________________________________________________________________

Illustration 14: No effect of taxation of benefits

Employee's OASI taxes 7.7 12.8 21.8 Combined OASI taxes 17.7 33.6 102.6 Retirement portion of OASI taxes 6.3 9.5 14.9 Retirement portion of combined 13.2 23.5 47.9 OASI taxes

Illustration 15: Maximum effect of taxation of benefits

Employee's OASI taxes 12.4 21.8 42.7 Combined OASI taxes 32.3 100.9 infinite Retirement portion of OASI taxes 9.5 16.0 28.6 Retirement portion of combined 22.6 47.6 infinite OASI taxes ______________________________________________________________________

                  RETIREMENT PORTION OF EMPLOYEE'S

                  Table 5. Life Expectancy at Age 65

[15] Defenders of Social Security tend to discount the phenomenon of lengthening payback times, arguing that the program serves social ends that transcend calculations of which individuals, or generations, obtain some sort of balance-sheet profit or loss. They say that it is incorrect to evaluate the program's worth by simply measuring the speed or rate of return of a worker's contributions because such computations do not adequately reflect the program's social features, such as its anti-poverty value (it is estimated that 39% of the elderly are kept out of poverty by their Social Security benefits). They point out that pay-as-you-go retirement systems such as Social Security by their nature often provide large returns on the contributions of the initial generations. In the early years of such programs, the ratio of workers to recipients is very high, allowing tax or contribution rates to be low. As the program matures, rates rise to reflect the increase in the number of beneficiaries. This is not unique to Social Security. Establishing benefit levels for early recipients in excess of what contributions would dictate is also found in private pension systems.

[16] Furthermore, proponents of Social Security note that providing "adequate" benefits to initial Social Security recipients that were essentially "unearned" in relation to their contributions to the system was deliberate social policy. Providing a minimum level of protection to the first workers to participate in the system was considered more important, in a period of economic depression, than concerns about excessive rates of return on taxes paid. Besides, the social benefits of giving a measure of economic independence for the elderly, and later for orphaned children, surviving spouses, and the disabled, are said to be immense. For example, younger workers are in large part relieved from the financial burden of supporting their parents, and the elderly are afforded an opportunity to live independently and with dignity.

[17] Critics of Social Security point to these social welfare features as a basic flaw in the program. They argue that by combining the goals of social adequacy, which is welfare-related, with individual equity, which loosely ties benefits to taxes paid, the program has become a mishmash that accomplishes neither goal well and creates inequities. One inequity they cite is that future beneficiaries will on the whole receive retirement benefits inferior to those that the equivalence of their taxes could purchase in the private sector. Furthermore, they say when interest is included, some categories of workers will not recoup what they and their employer paid in taxes. Often buttressing these arguments are calculations that show what individuals could receive if their Social Security taxes were invested privately.

[18] As shown in the illustrations, this latter argument is dependent on the interest rate assumed on such investment. For example, workers retiring at age 65 in 2030 who always earned the average wage would recover the retirement portion of their and their employer's OASI taxes in 11.9 years at a 1% real interest rate, in 16.1 years at a 2% real interest rate, in 23.5 years at a 3% real interest rate, and 44.4 years at a 4% real interest rate. At a 5% real interest rate, however, the point is reached where the annual interest earned on the accumulation of taxes plus prior interest income exceeds the annual benefit payments, so that the principal is never depleted -- in fact, it plus interest grows indefinitely.

[19] The "proper" interest rate is completely problematic. Those who project high investment returns often refer to the historical performance of the stock market, showing that a portfolio of broad-based stocks would have earned on average substantial rates of return over the years, and that this performance can be expected to continue in the future. Also, high real interest rates may not seem so unlikely given the relationship of nominal interest rates and inflation over the past decade.

[20] Critics of such analysis point out that such investments have an element of risk that they believe should be unacceptable in providing a national system of retirement income, and that if a safe- as-possible mix of investment vehicles were used instead, projected rates of return would be smaller. They also contend that recent high real interest rates were a historical anomaly that will not be sustained in the future. The key point for the reader is to be aware of the influence exerted by the projected real interest rate in these sorts of calculations and the large degree to which the argument hinges around it.

                         ADDITIONAL READING

[21] Many things complicate any determination of the relationship of benefits to taxes for future retirees. For example, although Social Security tax rates and benefit formulae are set by law, they are not immutable. Indeed, Congress has modified taxes and benefits many times since the beginning of the program. Thus, it is clearly inconsistent with the program's history to calculate taxes and benefits into the future on the assumption that these key elements will not change. There is little doubt they eventually will be altered, as demographic phenomena will cause the program's projected outgo to outstrip its resources significantly in 37 years. ⁴ Higher taxes or benefit cuts would be necessary, at that point or before, if the self-supporting character of the program is to be continued. These changes obviously would affect payback times. However, the nature of future changes is unknown, whereas current law is a given. Therefore, in order to assess the relationship of future taxes and benefits, this paper uses calculations that are useful in presenting possible outcomes of policies currently incorporated in the law. ⁵

[22] Calculations of the relationship of benefits to taxes for future retirees involve many key factors. The rate of Social Security taxation is set by law. The portion of the tax that provides cash benefits (Old-Age, Survivors and Disability Insurance, or OASDI) is 6.2% on employees and their employers, each. The Old-Age and Survivors Insurance portion of the tax, from which retirement benefits are paid, is 5.3%. The tax rate applies to earnings up to a maximum amount. The "maximum taxable earnings" is $80,400 in 2001, but will rise in the future, as prescribed by law, at the same rate as average wages in the economy. Thus, the amount of Social Security taxes an employee will pay under current law is a direct function of his or her earnings. If one knows the amount of the individual's earnings, and what the maximum taxable earnings are each year, the amount of tax paid is easily calculated.

[23] Future initial benefit amounts are also in part a function of one's earnings. They are computed at first eligibility (age 62 for retirement) by a method that indexes both earnings over the worker's career and the benefit formula to changes in average wages in the economy. After age 62, benefits rise in tandem with the cost of living. As these factors are unknown, future benefit amounts cannot be predicted with certainty.

[24] Further complicating the issue is the nature of the program. As a "social insurance" program, Social Security has both social and insurance goals. The social-goal features provide a design that deliberately gives a better return on taxes to some workers than to others. For example, the basic formula for calculating Social Security benefits is tilted to replace a higher proportion of earnings for low-paid workers. Also, a complex array of dependents' benefits is available at no additional cost for workers with families.

[25] As with insurance, the exact relationship of Social Security benefits received to total taxes paid cannot be predicted for each and every worker. For example, workers who die before or shortly after retirement and leave no survivors may collect only a few dollars in benefits or perhaps none at all. Other workers may collect Social Security benefits for many years after retirement and receive benefits that are greater than the value of their Social Security taxes. Workers who become disabled or die at an early age might have paid relatively little in Social Security taxes, but they or their families may receive benefits for many years, recovering the value of the worker's taxes many times.

[26] Also, there really is no "typical" Social Security beneficiary with a "typical" work history. An "average" benefit can be the result of many different work histories and thus be based on different amounts of taxes paid. For example, because the benefit formula does not require that all earnings be used in the benefit computation, workers with gaps in their earnings history may receive the same benefits as other workers but pay less in total taxes.

[27] Nevertheless, models can produce projections of future benefits, based on assumptions about wage and price growth, for workers with designated work histories and characteristics. This paper makes such projections using several common assumptions about illustrative workers. It assumes that each worker retires at age 65 in January of the designated year after having worked full time in employment covered by Social Security beginning at age 22. ⁶ Similarly, all the illustrations reflect three lifetime earnings patterns -- workers who always earned either (1) the federal minimum wage; ⁷ (2) a wage equal to Social Security's "average wage series"; and (3) a wage equal to the maximum amount creditable under Social Security. To show the effect of a dependent spouse, the illustrations in Table 2 assume that the spouse is also age 65 and is not entitled to a Social Security benefit on his or her own account.

[28] These work histories and characteristics are necessarily arbitrary. Many variations could be constructed that would alter the payback times. However, by comparing similar examples of workers in what may be considered illustrative situations one may make a number of observations without having to resolve all the judgmental questions concerning what constitutes a typical worker or having to provide a voluminous array of illustrations.

[29] The model uses the alternative II assumptions of the 2001 Social Security Trustees' report to forecast wage and price growth. Under these assumptions, wages grow for most of the projection period by 4.3% a year, prices by 3.3%.

[30] Although using common assumptions and focusing on certain examples allows comparisons across generations, there are other factors that can be varied depending on one's view of the Social Security system. Among these is whether to count the employer's share of the payroll tax. There is some disagreement concerning who really bears the burden of the Social Security tax paid by employers. Economists generally say that employees pay for it in the form of foregone wages. However, others have maintained that employers are actually paying for income maintenance protection that they would have to pay for anyway in one form or another in the absence of the Social Security program, and that they absorb part of it and pass the rest along to the general public in the form of higher prices. This paper does not attempt to resolve this debate, but rather presents examples using both assumptions.

[31] Another variable subject to the reader's choice is the proportion of the Social Security tax to apply to retirement benefits. The payroll tax consists of three elements -- Old-Age and Survivors Insurance (OASI), Disability Insurance (DI), and Hospital Insurance (HI). Because the DI and HI programs have earmarked taxes, their own trust funds, and designated tax rates specified in the law, they are clearly and easily excludable from computations of taxes that pay for retirement benefits. OASI taxes pay for survivor as well as retirement benefits, and it would be inconsistent to include taxes that pay for survivor benefits on the tax side, but not include the value of survivor benefits on the benefit side, in computing payback times. However, there is no separate allocation of taxes in the law for survivor or old-age benefits. It is possible to derive hypothetical year-by-year tax allocations for old-age benefits by assuming that such taxes would be in the same proportion to OASI tax rates as old-age benefits are to OASI benefits for each year. The Social Security Administration's actuaries have year-by-year projections of these benefits and this paper uses them to compute taxes attributable solely to old-age benefits.

[32] A problem with this approach is that the survivor portion of the tax cannot so easily be assigned to a benefit. While the DI and HI taxes protect against risks that really do not involve an element of choice -- all workers possibly can become too disabled to work or suffer illness in old age -- there is an element of choice in whether a worker has dependents. Nevertheless, the worker still must pay the full OASI tax. An unmarried childless worker can maintain that it is inaccurate to say that only the old-age portion of the OASI tax should be used to compute the payback times of his retirement benefit when he is forced to pay a tax (the survivor portion of the OASI tax) for which he can derive no benefit. Again, this paper does not resolve this argument of whether or not to count the survivor portion of the OASI tax. It simply shows both ways of computing the relationship of benefits to taxes.

[33] Also, any calculation of such a relationship is heavily dependent upon the interest rate assumptions used. The value of taxes over time is equivalent to their worth if invested. However, the amount of interest is not easily determinable. Were the value of taxes paid placed in investments that produce high yields, its total real worth could be many times its nominal value. On the other hand, it is possible that the principal could be virtually wiped out by poor investment choices. To obtain a middle ground, consisting of a reasonable and safe investment history, one could assume that the Social Security contributions were always placed in U.S. Government obligations. Excess Social Security taxes have always been invested in U.S. Government securities, so, to provide illustrations, this paper uses the effective interest rates earned by the Social Security trust funds over the years and those projected for the future. ⁸ Under the alternative II assumptions, average annual interest rates are projected ultimately to be 6.4%, a "real" interest rate of 3.0% (i.e., 3.0% above inflation). The interest is assumed to be tax free.

[34] This assumption about the rate of long-term government securities also is subject to debate. Some analysts maintain that, given the recent insistence of investors to be paid a premium against the threat of inflation, a 3.0% real interest rate is too low. Others may say that in the long run this rate is too high because eventually the differential between interest rates and inflation will return to their historical lower levels. To provide a range of different rates, Table 2 shows payback times using projected real interest rates of 1, 2, 3, 4, and 5%. What is readily apparent is that interest rate assumptions make large differences. This table also may be useful to those who, in evaluating future payback times, would prefer to use the historically higher rates of return provided by long-term investments in the stock market.

[35] Finally, one feature of the Social Security law makes such comparisons particularly difficult to measure. Since 1984, retirees whose adjusted gross incomes plus one-half of their Social Security benefits exceeds $25,000 (single) or $32,000 (married filing jointly) must pay income tax on a portion -- up to one-half -- of their Social Security benefits. Effective in 1994, those whose combination of other income plus one-half of benefits exceed $34,000 (single) or $44,000 (couple) may have up to 85% of their benefits taxed. Although the Social Security benefit is still paid in full, many people regard this provision as a reduction in benefits. A reduction in benefits, of course, increases the period over which it takes to recover Social Security taxes. About 32% of current beneficiaries are affected (i.e., pay tax on their benefits), but this proportion will grow in the future because the law does not index the income thresholds to take account of expected increases in the income of the future retirees.

[36] Arguments can be made on both sides of the issue on whether to take account of the tax status of Social Security benefits in payback calculations. Because the earnings from interest also could have been taxed, and annuities and pensions which are often compared to Social Security are taxed, a case can be made that including taxation's effects is perhaps inaccurate, and needlessly confusing in any event. However, two factors weigh against this argument. One is that, as the tax status of Social Security changed recently and will be more heavily felt in the future, its effect on equity among the generations should be accounted for. The other is that, because part of the income tax on benefits is credited to the Social Security trust funds, it reduces the payroll tax workers otherwise would pay.

[37] Stating that the taxation of benefits should be considered is easier than calculating its effects. Because the taxation of benefits heavily depends on how much other income an individual has, there is no exact way to estimate the effect of this provision from one Social Security recipient to another. ⁹ Also, what marginal income tax rates at certain income levels will be far into the future is unknown.

[38] Nonetheless, one can produce a range of possibilities that may be informative. At one end would be an individual who would have none of his benefits taxed and at the other end would be one who has the maximum of 85% of his or her benefits taxed at the highest possible tax rate. That rate under the recently-passed tax cut legislation (P.L. 107-16) would be 35% by the end of 2010. However, under the legislation, after 2010 the highest tax rate would revert to 39.6% unless the law was changed further. Thus under "current law," in 2030 the maximum amount of tax paid would be 33.66% (39.6% times 85%) of the Social Security benefit in each year. Thus, the upper limit of the range would be payback times computed on the basis of benefits reduced by 33.66%, and this is illustrated in Table 4. If the 2010 tax rate were extended, in 2030 the maximum amount of tax paid would be 29.75% (35% times 85%) of the Social Security benefit in each year. This would lower the payback times somewhat.

[39] Another way to consider the issue would be to reflect the effect of taxation of benefits on the tax side of the equation rather than on the benefit side. Without the taxation of benefits provision, the long-range deficit under the alternative II assumptions would be higher by 0.62% of taxable payroll. ¹⁰ In other words, to achieve the same actuarial balance as currently projected under alternative II, the payroll tax would have to be raised by 0.62 percentage points (or about 0.3 percentage points on employers and employees, each) over the next 75 years. On an aggregate basis, therefore, one could portray the effect of the taxation of benefits by calculating payback times based on these increased tax rates.

[40] Whether viewed as a surrogate for increased payroll taxes or as a reduction in benefits, the effect would be to lengthen payback times. The difference is that viewed as a reduction in benefits, payback times vary from individual to individual, and have more of an effect on succeeding generations, whereas viewed as a surrogate for taxes, payback times would lengthen somewhat for everybody. In either event, the choice of whether to consider the effect of taxation of benefits is left to the reader.

[41] The self-employed are not portrayed in this analysis. From 1951 to 1983, they paid approximately three-quarters of the combined employee-employer tax rate. Since 1984, they have paid a RATE of tax equivalent to the combined employee-employer rate, but from 1984 to 1990, they received income tax credits that offset a portion of their Social Security tax. Since 1990, they have received an income tax deduction for one-half of their Social Security tax liability. Thus, the true NET AMOUNT of their taxes is variable, because the amount of each individual's tax savings depends on his or her marginal income tax rate during each year. However, if one wishes to approximate the payback times of the self-employed, eventually they will become more or less similar to those for retirees at the combined employee- employer tax rates.

² The "full" retirement age (the age at which one receives benefits unreduced for "early" retirement). Under current law, the full retirement age will gradually rise from age 65 to age 66 for those born in 1938-1943 over the period 2003 to 2008, and will rise again to age 67 from 2022 to 2027 for those bom in 1955-1960. Benefits will still be available as early as age 62, but the actuarial reduction will increase (benefits taken at age 62 eventually will be 70% of those taken at age 67, whereas today they are 80% of those taken at age 65). Benefits taken at age 65 after 2024 will be 86.67% of those taken at age 67 and this is reflected in the tables.

³ Under Social Security's "dual entitlement" rule, a benefit payable as a dependent is offset by any benefit earned as a worker (the difference, if any, is then paid as a dependent benefit). About 2.8 million (out of 5.3 million) aged wives and husbands receive a full spousal benefit (i.e., have no Social Security benefit of their own earned as a worker).

⁴ As forecast by the Board of Trustees of the Old-Age, Survivors and Disability Insurance (OASDI) Trust Funds in their annual report to Congress. The 2001 report shows that under the alternative II projections (the set of economic and demographic assumptions most often used in forecasting the actuarial status of Social Security) the trust funds will be depleted in 2038. Over the 75 years of the long-range projection period the Social Security program would have an average annual deficit of 1.86% of payroll. This deficit is equivalent to expenditures exceeding income on average by 14.0%.

⁵ A different picture would emerge if Social Security were placed on a true pay-as-you-go basis. Because OASDI tax rates are fixed from 1990 on, income to the system expressed as a percentage of taxable payroll is fairly constant over the years. (It rises gradually, from 12.72% in 2001 to 13.08% in 2030, because the rate attributable to the taxation of benefits increases, a phenomenon discussed later.) However, outgo expressed as a percentage of taxable payroll varies considerably, mainly because the relative size of the retired population changes over the years. The system is said to be solvent until 2038 under the alternative II assumptions because the trust funds accumulate surpluses when the baby boom is in its peak working years, which are then used to finance the baby boom's retirement. If instead, as has been proposed, such surpluses were not accumulated but rather each year's income were made approximately equal to outgo, OASDI tax rates on employees and their employers would be about 5.2% today, but rise to about 8.6% by 2030. Such variations in taxes would produce substantial differences in the relationship of taxes versus benefits, depending on one's age cohort (baby boomers would have a particular advantage over succeeding cohorts).

⁶ Previous versions of this report began the work career at age 21, thus producing a work history with 44 years of earnings. However, a typical work history tends to be shorter than 44 years. Although many workers begin to work before age 22, a sizable number have periods of unemployment thereafter, and most workers retire before age 65.

⁷ The federal minimum wage is set by law, and has changed many times over the years. This paper assumes that in the future it will increase in proportion to the growth in average wages. Historically, the minimum wage has varied in relative value considerably, which has an effect on payback times. For example, the freezing of the minimum wage from 1981 to 1990 is the primary reason why in Tables 1 and 2 a minimum wage worker retiring in 2030 recovers the value of his or her taxes more quickly than a minimum wage worker retiring in 2020.

⁸ Historically, interest rates have fluctuated widely. Swings in these rates can have a significant effect on payback times. This is the primary reason some average wage workers retiring in 2020 appear to have longer payback times than workers retiring earlier and later, as the SSA actuaries have modified their assumptions about real interest rates in the short and long term.

⁹ It is not unreasonable to posit that there is a strong correlation between pre- and post-retirement income, and therefore higher paid workers are more likely to have their benefits taxed. However, this correlation is not exact; also, adjusted gross income includes all income -- including spouse's income, highly variable income from assets, and other unpredictable sources.

¹⁰ 2001 Trustees' report. p. 64.