How Large Is the Federal Government's Debt?

Policy Reports | Federal Spending | Government

No. 263
Thursday, October 30, 2003
by Liqun Liu, Andrew J. Rettenmaier, and Thomas R. Saving


  1. Though the federal government does not formally report Social Security and Medicare obligations as liabilities, it does report on the programs' solvency in a variety of publications. The dollar values of the unfunded obligations over the next 75 years and over the infinite horizon now appear in the Social Security Trustees Report. For years, the Social Security Administration has calculated the system's 75-year unfunded obligations as well as the net obligations owed to current program participants. Since 1996 it has also calculated the accrued obligations owed to current participants. The unfunded obligation for the Hospitalization Insurance portion of Medicare for the next 75 years was also reported in the most recent Trustees Report. For the last three years, the Treasury Department has examined the two programs' 75-year unfunded obligations in the Financial Report of the United States Government 2000, 2001, 2002, U.S. Department of the Treasury. The Treasury report details the components of expected tax revenues and benefit payments by current and future participants. The Office of Management and Budget provides some similar accounting in Analytical Perspectives, Budget of the United States Government, Fiscal Year 2004, Office of Management and the Budget, 2003. Further, the generational accounting literature, following Auerbach, Gokhale and Kotlikoff (1991), has grown substantially and has emphasized the importance of more complete measures of federal deficits and debts. See Alan J. Auerbach, Jagadeesh Gokhale and Laurence J. Kotlikoff, "Generational Accounts: A Meaningful Alternative to Deficit Accounting," in David Bradford, ed., Tax Policy and the Economy, Vol.5 (Cambridge, Mass.: MIT Press, 1991), pages 55-110.
  2. As pointed out by Robert Haveman, although prone to labeling games, the official deficit measure is not devoid of information on the relative burdens government is imposing on existing vs. future generations. See Robert Haveman, "Should Generational Accounts Replace Public Budgets and Deficits?" Journal of Economic Perspectives, Vol. 8, No. 1, Winter 1994, pages 95-112. The comprehensive measure of the debt would include privately held government bonds and implicit promises to current participants in Medicare and Social Security.
  3. Absent from these brief descriptions is the offsetting effect of the Social Security and Medicare trust funds. Given that redemption of the bonds in the trust funds will require future resources, we have excluded them from the discussion. However, the Trustees Reports, and unpublished Social Security tables, include the trust fund offsets in their calculation of the obligations. Additional detail on the calculation of accrued Medicare and Social Security benefits is given in subsequent discussions.
  4. There is not a historical time series for the perpetuity unfunded obligation, but there are at least several years of estimates for the other three unfunded obligation measures. The 75-year open-group and the 100-year closed-group unfunded obligation series are available beginning in 1973, while the first accrued benefits measure is available in 1996.
  5. Reported in Table IV.B6 of the 2003 Trustees Report.
  6. Trust fund offsets make sense only if the trust funds hold assets that can be used to fund benefit payments. Under current practices, every asset of the trust funds is a liability of the treasury. Summing over both parts of government leaves a net value of zero. It is in this sense that the trust funds consist of IOUs the government has written to itself.
  7. Tables I and II are based on the Financial Report of the United States Government, 2002, the 2003 Annual Report of the Board of Trustees of the Federal Old-Age and Survivors Insurance and Disability Insurance Trust Funds, unpublished estimates by the Social Security Administration's Office of the Actuary, and our estimates of Medicare's perpetuity unfunded obligation.
  8. The maximum transition cost and the 100-year closed-group estimates are from the unpublished estimates by the Social Security Administration's Office of the Actuary. A description of the maximum transition cost estimate is provided in Stephen C. Gross, "Measuring Solvency in the Social Security System," in Olivia S. Mitchell, Robert J. Meyers and Howard Young, eds., Prospects for Social Security (Philadelphia: University of Pennsylvania Press, 1999).
  9. These are estimated using the relationships between the perpetuity unfunded obligation and the 100-year closed-group obligations, and the perpetuity actuarial deficit of 3.8 percent, as reported in the 2003 Trustees Report. There, the perpetuity unfunded obligations and the 100-year unfunded obligations were both $10.5 trillion and once the trust fund offset is excluded the number rises to $11.9 trillion.
  10. The present values of benefits for all participants 15 and older are included in the calculation of accrued benefits for consistency with the age definition. However, this overstates accrued Medicare benefits because it includes those who have not yet worked 40 quarters.
  11. We assume per capita medical care spending grows at a rate of per capita GDP+1 until 2080 and then falls to per capita GDP growth by 2100, following Jagadeesh Gokhale and Kent Smetters, "Fiscal and Generational Imbalances: New Budget Measures for New Budget Priorities," American Enterprise Institute, Pamphlet, 2003. They estimate that Medicare has a $36.6 trillion perpetuity unfunded liability. Excluding trust fund offsets increases, their number rises to $36.9. This last estimate includes $6.1 trillion in projected new spending based on the Medicare modernization provisions in the 2004 budget. Without the new spending and without the trust fund offsets, their estimate of the unfunded liability is $30.7 trillion. Gokhale and Smetters also use a higher discount rate than used in this study. If we substitute a 6.25 percent long-run nominal interest rate for the 6 percent rate, our estimate of the perpetuity unfunded obligation is $30.8 trillion.
  12. Jagadeesh Gokhale and Kent Smetters, "Fiscal and Generational Imbalances: New Budget Measures for New Budget Priorities," American Enterprise Institute, Pamphlet, 2003.
  13. See Footnote 11 for further discussion of the difference between our estimates from Medicare and the estimates by Gokhale and Smetters (2003). Also note that their estimate of the perpetuity unfunded obligation for Social Security net of the trust fund is $8.3 trillion vs. the $11.3 trillion reported in Tables I and III.
  14. Alan J. Auerbach, William G. Gale and Peter R. Orszag, "Reassessing the Fiscal Gap: Why Tax-Deferred Saving Will Not Solve the Problem," Tax Notes, July 28, 2003, Brookings Institution.
  15. Note that to current retirees (i.e., those are 62 year or older), the accrued liabilities and 100-year closed-group unfunded liabilities are identical, and are, at least in the political rhetoric, "reform-proof."
  16. In practice, the Social Security Administration's estimate of the accrued liability measure, which we use here, calculates a disability benefit for all current participants and scales that benefit according to their age, so that younger workers are credited with a lower share of their disability benefit than are older workers. See Stephen C. Goss, "Measuring Solvency in the Social Security System," in Olivia S. Mitchell, Robert J. Meyers and Howard Young, eds., Prospects for Social Security Reform (Philadelphia: University of Pennsylvania Press, 1999). An alternative would be to consider past earnings up to the calculation date for all workers currently alive and calculate benefits based on the benefit formula as it is expected to exist in each worker's retirement year. Due to the graduated primary insurance amount formula which replaces a higher proportion of low-income workers' past wages, this procedure will overestimate the accrued benefits for younger workers because their few years of work produce low estimates of Average Indexed Monthly Earnings. This shortcoming can be overcome by assuming younger workers receive the average replacement rate.
  17. Herman Leonard, Checks Unbalanced (New York: Basic Books, 1986).
  18. Ibid.

Read Article as PDF