How Generous Are Social Security and Medicare?

Studies | Federal Spending | Health | Social Security

No. 290
Friday, October 27, 2006
by Andrew J. Rettenmaier & Thomas R. Saving


  1. The estimates assume scheduled and projected benefits will actually be paid to beneficiaries and will be financed with the current statutory payroll taxes, taxes on benefits and general federal revenues. The projections assume that growth in the size of government will not affect gross domestic product (GDP) growth. Also, the term consumption as used here refers to spending and spending opportunities including health care.
  2. See Table VI.F10 in the 2006 Social Security Trustees Report; available at
  3. Changes to the benefit formula in the 1983 reforms produce Social Security benefits that replace relatively stable shares of wage-indexed preretirement earnings for future retirees. The reforms also gradually raised the retirement age and increased taxes paid by workers and retirees.
  4. The particular CPI used to adjust postretirement Social Security benefits is the CPI-W. The CPI-W is used in the examples reported here.
  5. The two age-earnings profiles depicted in the figure represent workers with adjusted average earnings born in 1941. The Social Security Administration's Actuarial Note Number 2005.3 describes how the Social Security Average Wage is adjusted to produce Table VI. F10 in the 2006 Social Security Trustees Report. See Michael Clingman and Orlo Nichols, "Scaled Factors for Hypothetical Earnings Examples under the 2005 Trustees Report Assumptions," Social Security Administration, Actuarial Note Number 2005.3, September 2005.
  6. This wage index is the one used in the Social Security Trustees Report to calculate replacement rates. In the Trustees Report, hypothetical earnings for each income group are wage-indexed to the year prior to retirement. Using this method, calculations show that Social Security benefits replace 45.4 percent of the average worker's preretirement earnings. Note that this method is not the one used by the Social Security Administration to determine actual benefits under current law. The Social Security benefit formula averages the 35 highest-earning years, wage-indexed to age 60, and includes nominal earnings after age 60. Using the benefit formula method of calculating average earnings produces a replacement rate of 48.3 percent. Using the same method, Olivia Mitchell and John W.R. Phillips, "Social Security Replacement Rates for Alternative Earnings Benchmarks," Pension Research Council, Working Paper 2006-6, report a replacement rate of 48 percent for workers born in 1936. The replacement rate used in the benefit formula is 6.4 percent higher than the replacement rate comparable to the one reported in the Trustees Report. The comparison is highlighted here because the term "replacement rate" commonly refers to the replacement of the earnings used in a pension's benefit formula.
  7. The estimate of a 45.4 percent replacement rate for average workers born in 1941 is based on the methodology described for the estimates reported in Table VI. F10 in the 2006 Social Security Trustees Report. However, the Trustees Report estimate for the same group of workers is 42.9 percent. The use of 2005 scaling factors here is the likely cause of the difference in estimates. By the time individuals born in 1960 retire the replacement rates for each earnings category will stabilize given that the 1983 adjustments to the benefit formula will be fully phased in and the relationship between price and wage growth are projected to be stable in the years thereafter. The low, medium, high and taxable maximum replacement rates are projected to be 55.4 percent, 41.0 percent, 34.0 percent and 27.3 percent, respectively, in the long run.
  8. The value of fringe benefits is determined using historical compensation-to-wage ratios. In addition, past payroll taxes and nonpayroll tax federal revenues are based on their historical levels and are used to estimate compensation net of federal taxes. Historical compensation-to-earnings ratios are from Eugene B. Yang and Stephen C. Goss, "Economic Projections for OASDHI Cost and Income Estimates: 1992," Social Security Administration, Actuarial Study No. 108, December 1992. Historical federal nonpayroll taxes collected from labor are used to calculate an average tax on wages earnings. Given progressive tax rates this average overstates taxes collected from low and average earners.
  9. Medicare benefits are identified as the annual lifetime premium that is sufficient to fund each group of new beneficiaries' lifetime Medicare spending net of Part B and Part D premiums.
  10. The comparison is limited to average workers with scaled medium lifetime earnings born in 1941. Scaled earnings reflect the earnings profile for average workers. That is, earnings are less than the average earnings of all workers in the economy when workers are young, rise above average earnings when workers are in their mid-30s to mid-50s and then decline below the average at later ages. For more information, see the Social Security Administration's Actuarial Note Number 2005.3; available at
  11. Social Security limits the amount of earnings subject to taxation in a given year. This limit increases each year with increases in the national average wage index. For earnings in 2006, the limit is $94,200. For the purpose of calculating the average wage reported in the 2006 Social Security Trustees Report, Table VI. F6, earnings above the taxable maximum were included.
  12. Income taxes on Social Security benefits have been netted out of total Old Age and Survivors Insurance (OASI) benefits before calculating the average benefit. These taxes include those credited to both the OASI Trust Fund and Medicare's Hospital Insurance Trust Fund. The share of aged to non-aged OASI beneficiaries is estimated from Table V.C4 from the 2006 Social Security Trustees Report, and the relative size of the average benefit for the aged and non-aged comes from the 2004 Annual Statistical Supplement. Given that income taxes on Social Security benefits affect higher-income retirees, the average benefit received is understated. Also, the calculation of average Social Security benefit per retiree in this section includes benefits awarded to spouses and the effects of benefit reductions due to early retirement.
  13. In calculating net Medicare benefits, Hospital Insurance (Part A) premiums paid by voluntary enrollees, as well as premium payments for Supplementary Medical Insurance (Parts B and D) are deducted from total Medicare expenditures. State transfers for Part D are also omitted. Thus, the average Medicare benefit only includes the net federal transfer. This average is adjusted by the historical relationship between the benefits received by the aged and the disabled to produce the average benefit for aged beneficiaries.
  14. The ratio of compensation to GDP is assumed to be constant in the Trustees Reports (see page 86 of the 2006 Trustees Report). The ratio of compensation to GDP calculated here is 0.60 for most years using the 2006 Trustees Report.
  15. Average compensation is reduced by federal payroll and income taxes in the following way. First, the payroll taxes are estimated using the statutory payroll tax rates. Second, the net general revenue funding requirements for Medicare (all parts) and Social Security are assumed to be financed by all nonpayroll tax federal revenues. An estimated 58 percent of the revenues are from labor taxes. Third, the nonpayroll tax revenues are assumed to be equal to the 50-year average historical share of GDP less the current level of GDP devoted to Supplementary Medical Insurance (SMI) funding. This treatment of income taxes overstates the income tax burden on an average worker given the relatively higher tax burden on higher-income workers.

Read Article as PDF