Information about Western Hemisphere Hemisferio Occidental
Chapter

IV Long-Run U.S. Fiscal Imbalance: An Intergenerational Analysis

Author(s):
Martin Mühleisen, and Christopher Towe
Published Date:
January 2004
Share
  • ShareShare
Information about Western Hemisphere Hemisferio Occidental
Show Summary Details
Author(s)
Cardarelli Roberto and Towe Christopher 

The recent dramatic deterioration of the U.S. fiscal position has heightened long-standing concerns about the extent to which the retirement and health care systems are prepared to cope with the pressures of an aging population. These concerns have been echoed in the administration’s FY2004 budget, which described these programs as being on “an unsustainable path” and stated that the “resources of these programs are insufficient to cover their long-range shortfalls.”1

The insolvency of the Social Security and Medicare programs is not a new phenomenon. It has been highlighted on a regular basis in the annual reports of the Old Age, Survivors, and Disability Insurance (OASDI) and Medicare Trustees, with the 2003 reports suggesting that the combined actuarial deficit of these programs is in the range of 160 percent of current GDP, if calculated over a 75-year horizon.2

More recent analysis, however, has flagged that actuarial estimates, which tend to focus only on the Social Security and Medicare systems, may provide a misleading picture of the effect of demographics on the broader fiscal imbalance. For example, Auerbach, Gale, and Orszag (2003) illustrate that, at least to some degree, tax revenues may be boosted by withdrawals from tax-deferred instruments by retirees. Similarly, OECD’s (2001) analysis has shown that estimates of the fiscal impact of population aging also needs to take into account the full gamut of age-related spending, including on education, which may partly offset the rise in spending on the elderly.

At the same time, these broader analyses also suggest that the standard actuarial estimates of the Social Security and Medicare imbalances tend to offer too optimistic an assessment of the fiscal imbalance, while obscuring the intergenerational inequities involved. For example, the 75-year horizon used in assessing the position of the OASDI system would suggest that the payroll tax increase required to “balance the system” would be only 1.87 percentage points. However, while such a tax hike would meet the system’s obligations over the next 75 years, in net present value terms, it would still leave a significant cash flow deficit at the end of the period. Moreover, closing the fiscal gap can be accomplished through a variety of instruments (e.g., tax hikes, spending cuts, and so on) and in varying degrees of urgency (e.g., immediate tax hikes, or tax hikes phased in over years or decades). These alternatives have significantly different implications for the economic well-being of different generations, which are not captured by typical actuarial measures.

These considerations have led to a renewed emphasis on fiscal gap estimates that take into account longer horizons and the intergenerational transfers that are involved. Most recently, Gokhale and Smetters (2003) have prepared estimates of the U.S. fiscal imbalance using an intergenerational accounting framework that encompasses the entire federal fiscal system over an infinite horizon. Their results are alarming, suggesting that the U.S. fiscal imbalance is over 400 percent of current GDP (or 6.5 percent of the present value of future GDP), and that filling this fiscal gap would require measures that yield a 70 percent immediate and permanent increase in personal and corporate income tax revenues—that is, the ratio of GDP to federal income tax would have to be raised from just under 10 percent at present to 16 percent.

In this section, these worrisome conclusions are confirmed by constructing similar estimates of the U.S. intergenerational fiscal accounts. The results presented below suggest that the fiscal imbalance is as high as $47 trillion, nearly 500 percent of current GDP, and that closing this fiscal gap would require an immediate and permanent 60 percent hike in the federal income tax yield, or a 50 percent cut in Social Security and Medicare benefits. The analysis also illustrates that this gap is associated with a severe intergenerational imbalance—the burden that future generations will be required to bear to close this fiscal gap will be significant and will only increase further if measures are delayed.

Constructing Intergenerational Fiscal Accounts

The estimates are based on an intergenerational accounting framework, similar to that used by Gokhale and Smetters (2003). This approach first requires an estimate of the distribution of taxes and transfers across age cohorts in the current year.3 Long-term fiscal projections are then constructed by assuming that taxes and transfers paid, and the level of other government outlays received, by the average person in each age cohort increase in line with productivity, unless programmatic or policy commitments dictate a different growth path.4

The demographic projections used in the baseline scenario are derived from the intermediate series of the 2003 OASDI Trustees’ report, which represent the Social Security Administration’s (SSA’s) best estimate of future demographic trends. The projections cover the period 2002–80 and are extended beyond this horizon by using the SSA’s terminal year fertility, mortality, and immigration assumptions.5 During the next 10 years, the age distribution of the U.S. population is expected to remain roughly unchanged. After 2010, however, the old-age dependency ratio (the ratio of retirees to those in the labor force) begins to increase rapidly from around 20 percent in 2010 to 37 percent in 2036, stabilizing at around 40 percent after 2050. Labor productivity (defined as real GDP per hour worked) is assumed to grow at a rate of 1.7 percent per year. This rate equals the average annual rate achieved by the U.S. economy over the past 40 years and is in line with the assumption used by Gokhale and Smetters (2003) and the intermediate assumption in the 2003 OASDI Trustees’ report.

During 2002–08, federal revenues are projected to grow in line with the estimates contained in the July 2003 Mid-Session Review of the FY2004 budget. During 2008–75, revenues from payroll tax contributions and premiums to the Social Security and Medicare trust funds are derived from the intermediate projections contained in the 2003 OASDI and Medicare Trustees’ reports. All other revenues, including Social Security and Medicare revenues after 2075, are projected by assuming that per capita taxes and contributions increase in line with labor productivity, taking into account the impact of projected shifts in the age and gender distribution of the population.6

The simulations implicitly assume, therefore, that the tax cuts of recent years are made permanent, in line with the stated policy of the administration. In the long run, the ratio of tax revenues to GDP evolves in response to changes in the composition of the population. For example, revenues from individual income taxes attributed to labor are projected to remain constant as a share of output, as both these revenues and GDP are affected by the decline of the working-age population. Conversely, revenues from taxes on nonlabor income increase as a proportion of GDP because of the increase in the elderly population relative to those in the working-age cohorts (Figure 4.1).7 Relative to GDP, Social Security contributions decline slightly, as taxable payrolls decline as a share of GDP.

Figure 4.1.Tax Revenues

(In percent of GDP)

Source: IMF staff estimates.

During 2002–08, federal outlays are derived from the estimates contained in the Mid-Session Review of the FY2004 budget. After 2008, both discretionary and mandatory outlays are projected assuming that per capita spending, on an age- and gender-adjusted basis, grows in line with labor productivity.8 The exception is spending on Social Security and Medicare, which is assumed to grow during 2008–80 in line with the intermediate projections contained in the 2003 OASDI and Medicare Trustees’ reports. These reports have Medicare expenditure growing rapidly—at an average annual rate of 6½ percent in nominal terms—until 2012, reflecting recent trends and the impact of specific statutory provisions. During 2012–25, the Trustees’ reports have Medicare expenditure growth slowing somewhat and after 2025, age- and gender-adjusted, per beneficiary expenditure increases 1 percent faster than labor productivity, reflecting continued shift in the relative price of health care.9 To calculate the estimates described below, health care costs after 2080 are assumed to rise in line with productivity so that, after this year, changes in Medicare spending as a share of GDP reflect only a change in the demographic composition of the population.10

Based on these assumptions, spending on Social Security is projected to accelerate rapidly from about 2010 to 2030 as the baby boom generation reaches retirement age, rising from 4½ percent of GDP in 2002 to around 7 percent of GDP in 2080 (Figure 4.2). Spending on Medicare (net of premiums) is also projected to increase rapidly, from around 2¼ percent of GDP in 2002 to 5¼ percent of GDP in 2035 and to around 9 percent of GDP in 2080 (Figure 4.3).

Figure 4.2.Social Transfers

(In percent of GDP)

Source: IMF staff estimates.

Figure 4.3.Health Care Spending

(In percent of GDP)

Source: IMF staff estimates.

To illustrate the sensitivity of the projections to assumptions regarding the Medicare system for the longer-term solvency of the federal budget, an alternative scenario is constructed that assumes that the 1 percentage point gap between health care costs and productivity growth is eliminated from 2008 onward. In this case, Medicare spending as a share of GDP would increase only as a result of shifts in the age distribution of the population, rising to only around 5¾ percent by 2080.

Using the projected path for revenues and expenditures, the fiscal primary balance, debt service costs, and the change in the net debt of the government sector (gross federal debt held by the public less federal government financial assets) are then calculated.11

Debt Dynamics and Fiscal Sustainability

In this section, we present the implications of these assumptions for various indices of the long-term fiscal position, including the standard debt- and deficit-to-GDP ratios, as well as measures of the fiscal imbalance and the fiscal gap. The fiscal imbalance represents the net present Value of future government revenues less noninterest expenditures and current government net debt, and it is roughly equivalent to the government’s actuarial balance.12 The fiscal gap is a calculation of the immediate increase in income taxes that would be required to close the intertemporal budget imbalance.13

The long-term fiscal projections that are derived from these assumptions confirm the perilous situation facing the U.S. economy. As earlier studies have illustrated, the effect of demographic pressures on both retirement and health care systems, as well as the rapid increase in the price of health care services, would cause government debt and deficits to explode over the next century. In this case, the primary deficit rises sharply and reaches nearly 11½ percent of GDP by 2080, with government debt exceeding 600 percent of GDP by 2080 (Table 4.1).

Table 4.1Long-Term Projections as a Share of GDP
2002200820202030204020602080
Social Security4.44.25.46.46.76.97.1
Old-age and survivors insurance3.73.54.55.45.65.75.9
Disability insurance0.60.70.91.11.11.21.2
Medicare2.22.63.44.65.57.09.0
Part A1.41.82.23.03.74.96.6
Part B1.00.81.21.61.82.12.5
Medicaid1.41.82.22.52.83.54.4
Other expenditure9.69.09.610.110.310.610.9
Total expenditure17.617.620.623.625.328.031.4
Total revenues17.717.918.719.319.519.720.0
Primary balance0.10.2–1.9–4.3–5.8–8.3–11.5
Net interests1.61.92.03.76.414.527.3
Overall balance–1.5–1.6–4.0–8.0–12.2–22.8–38.8
Net debt128.435.247.486.0147.7331.3620.3
Sources: OMB, Mid-Session Review, Budget of U.S. Government, FY2004 (July 2003); OSADI 2003 Trustees’ report; and IMF staff estimates.

Net financial debt at the end of FY200I (debt held by the public less federal government financial assets) (U.S. Government, 2003, Table 3.1, p. 37).

Sources: OMB, Mid-Session Review, Budget of U.S. Government, FY2004 (July 2003); OSADI 2003 Trustees’ report; and IMF staff estimates.

Net financial debt at the end of FY200I (debt held by the public less federal government financial assets) (U.S. Government, 2003, Table 3.1, p. 37).

These estimates suggest that U.S. federal government faces an intertemporal deficit of $47 trillion, nearly five times the present level of U.S. GDP (Table 4.2). Social Security transfers (net of payroll contributions) account for a third of the imbalance, while Medicare transfers (net of payroll taxes and premium payments by the elderly) account for the remaining two-thirds, with the rest of the federal government showing a small surplus. Closing this fiscal gap would require a 60 percent immediate and permanent increase in personal and corporate income tax revenues or, alternatively, an immediate and permanent 50 percent cut in Social Security and Medicare outlays.14 By comparison, the tax cuts enacted since early 2001—if assumed permanent—have lowered the income tax ratio by about 15 percent, implying that the tax cuts have expanded the fiscal gap by roughly one-third.

Table 4.2.Indicators of Fiscal Imbalance in Alternative Economic and Policy Scenarios
Fiscal Imbalance (in trillions of U.S. dollars)Increase in Income Taxes for All Generations That Closes the Imbalance (in percent)Decrease in Social Security and Medicare Benefits for All Generations That Closes the Imbalance (in percent)
Baseline47.261.1–53.4
Demographic projections: low-cost scenario47.553.0–51.0
Higher productivity (2 percent)61.166.0–55.3
Higher productivity (2 percent) and no commensurate increase in expenditure133.636.3–38.0
Tighter Medicare spending228.837.4–37.8
Noncontributory transfers and government expenditure indexed to prices37.810.1–9.5
Adjustment delayed to 201047.267.9–57.1
Adjustment delayed to 202047.282.2–63.7
Sources: OMB, Mid-Session Review, Budget of U.S. Government, FY2004 (July 2003); OASDI 2003 Trustees’ report; and IMF staff estimates.

Per capita revenues from income tax, excise tax, and custom tax rise at 2 percent per annum compared with 1.7 percent in the baseline scenario, while all other revenues and expenditure remain unchanged.

Medicare and Medicaid transfers per capita grow at the same rate as labor productivity from 2008 onward.

All federal consumption and noncontributory transfers per capita are indexed to prices only.

Sources: OMB, Mid-Session Review, Budget of U.S. Government, FY2004 (July 2003); OASDI 2003 Trustees’ report; and IMF staff estimates.

Per capita revenues from income tax, excise tax, and custom tax rise at 2 percent per annum compared with 1.7 percent in the baseline scenario, while all other revenues and expenditure remain unchanged.

Medicare and Medicaid transfers per capita grow at the same rate as labor productivity from 2008 onward.

All federal consumption and noncontributory transfers per capita are indexed to prices only.

Correcting this fiscal imbalance will involve a substantial intergenerational redistribution of wealth that will only be magnified if the adjustment is delayed.15 For example, if the adjustment is postponed until 2010, income tax revenues would have to increase by nearly 70 percent relative to baseline, while deferring adjustment until 2020 would require an increase of over 80 percent. The size of the expenditure adjustment would similarly increase if the adjustment were delayed. The larger adjustment that would be required in later years would mean that a greater proportion of tax increases (or expenditure cuts) would have to be borne by younger workers or those not yet in the workforce, while the burden faced by the current elderly, and others in the workforce, would be correspondingly lessened.

The results mentioned above, while alarming, appear relatively robust in the face of alternative economic and policy assumptions. Alternative demographic projections do not significantly affect the broad picture that emerges—for example, using the SSA’s low-cost alternative, which embodies a higher fertility rate, slower improvements in mortality, and higher immigration flows, still yields a large imbalance. Similarly, higher productivity growth does not improve the fiscal imbalance, principally because retirement and health care spending are assumed to increase commensurately.

At the same time, measures to contain the growth of spending can significantly improve the fiscal imbalance. Simply containing Medicare spending to the rate of productivity growth lowers the gap to $28 trillion and roughly halves the immediate and permanent increase in income tax revenues required. Substantial reductions in the imbalance are also obtained if noncontributory transfers and federal consumption expenditure are indexed only to prices, rather than to real wages and/or productivity growth (see Table 4.2).

Conclusions

As noted in the introduction to this section, the results above underscore the well-known point that demographic and cost pressures on health care programs will—in the absence of policy measures—place an enormous burden on the U.S. federal budget. The specific estimates above suggest a fiscal gap of $47 trillion, or nearly 500 percent of current GDP. Closing this gap would require massive adjustments in either tax or spending programs. The longer-term fiscal deficit is also associated with a severe intergenerational imbalance; the estimates clearly illustrate that the longer the action is delayed, the greater will be the adjustment required and larger the burden that will be placed on future generations.

Although these results are undeniably worrisome, the discussion above suggests that the U.S. fiscal problem is manageable if action is taken soon. Most importantly, early steps to contain the growth of spending on entitlement programs, by enacting reforms to slow the growth of Medicare spending and contain the projected growth of Social Security benefits, would significantly narrow the fiscal gap. In the absence of action on these fronts, significant tax measures may be required, including a willingness to allow the cuts of recent years to expire as scheduled.

References

    Auerbach, A.J.,W.G.Gale, and R.R.Orszag,2003, “Reassessing the Fiscal Gap: Why Tax-Deferred Saving Will Not Solve the Problem,” Tax Notes (Washington: Urban Institute, Brookings Institution),July28.

    • Search Google Scholar
    • Export Citation

    Congressional Budget Office (CBO), 2000, The Long-Term Budget Outlook (Washington: U.S. Government Printing Office).

    Gokhale, J., and K.Smetters,2003, Fiscal and Generational Imbalances: New Budget Measures for New Budget Priorities (Washington: AEI Press).

    • Search Google Scholar
    • Export Citation

    Cardarelli, R.,J.Sefton, and L.J.Kotlikoff,2000, “Generational Accounting in the U.K.,” Economic Journal, Vol. 110, No. 467, pp. 547–74.

    • Search Google Scholar
    • Export Citation

    Medicare Trustees (Board of Trustees, Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds), 2003, Annual Report (Washington: U.S. Government Printing Office).

    • Search Google Scholar
    • Export Citation

    OASDI Trustees (Board of Trustees, Old-Age, Survivors, and Disability Insurance Trust Funds), 2003, Annual Report (Washington: U.S. Government Printing Office).

    • Search Google Scholar
    • Export Citation

    Office of Management and Budget, 2003, Mid-Session Review, Fiscal Year 2004 (Washington: U.S. Government Printing Office),July.

    Organization for Economic Cooperation and Development (OECD), 2001, OECD Economic Surveys: United States (Paris: OECD).

    U.S. Government, 2003, Budget of the United States Government, Fiscal Year 2004: Analytical Perspectives (Washington: U.S. Government Printing Office).

    • Search Google Scholar
    • Export Citation
1

See U.S. Government (2003), Chapter 3.

3

For another example of how this approach was applied to the United States, see the Congressional Budget Office’s study on the long-term fiscal outlook (CBO, 2000).

4

These projections do not incorporate any feedback between the fiscal position and macroeconomic variables, such as productivity growth or interest rates. Therefore, they simply provide an indication of the magnitude of long-term pressures for fiscal balances if current tax and spending policies remained unchanged, given long-term economic and demographic assumptions.

5

Under the intermediate projections, the total fertility rate is assumed to gradually decline from the estimated level of 2.05 in 2001 to the long-term rate of 1.95 by 2027; life expectancy at birth is assumed to rise from 74.2 in 2002 to 81.6 in 2080 for men, and from 79.5 to 85.5 for women; and total level of net immigration (both legal and illegal) is assumed to decline gradually from the estimated level of 1.2 million persons in 2002 to 900,000 persons in 2023 and for each year afterward.

6

The gender and age distribution of tax payments is calculated on the basis of data from several sources, including the Current Population Survey and the Survey of Consumer Finances. Social Security and Medicare future outlays and revenues are expressed in 2002 dollars by using the projections of the consumer price index for urban wage earners and clerical workers (CPI-W) contained in the 2003 report of the OASDI Trustees.

7

Taxes on labor are the labor income part of the individual income tax revenues. This is estimated based on the average share of labor income in net national income over the last 10 years. Taxes on capital are the capital plus corporate income part of individual income tax revenues. These revenues are allocated across cohorts based on the age profile of wealth, which is skewed toward older cohorts.

8

The gender and age distribution of outlays, as above, is estimated on the basis of various sources. Note that the expenditure projections deviate from a strict “current policy” assumption, since many noncontributory income security transfers—including income support, food stamps, aid to families with dependent children—are at present indexed only to CPI inflation. However, assuming CPI indexation over the long run implies that these payments would decline steadily as a share of GDP per capita, which is unrealistic and inconsistent with past experience in which ad hoc increases in the benefit rates have been introduced. A similar strategy has been followed by the CBO (2000), which terms the policy assumption in its baseline projections as “prevailing,” rather than “current.”

9

This assumption is aimed at capturing the fact that historically per capita health expenditure has increased significantly more than wages, as a result of the increase in the utilization and complexity of health services. Deflating nominal outlays using the CPI yields that real spending on Medicare has grown at an average annual 4¾ percent between 1990 and 2002, almost 2 percentage points faster than real GDP. Such a pace is higher than the one assumed in the simulations, as the projections in the Medicare Trustees’ report imply an average growth rate of real spending on Medicare of about 3½ percent for the period 2002–75.

10

Spending on Medicaid is assumed to grow during 2008–80, based on the Medicare long-term assumption that age- and gender-adjusted, per beneficiary expenditure will increase at a rate of 1 percent faster than labor productivity. Again, this assumption appears more conservative than historical experience suggests. While real spending on Medicaid has grown annually at an average of 8½ percent between 1990 and 2002, based on the above assumption it is projected to grow at an average annual rate of 3¾ percent during 2008–75.

11

For the present calculations, the effective nominal interest rate on net government debt is assumed to be constant at 4.7 percent, which is equal to the average annual effective interest rate projected by the FY2004 budget for 2002–08.

12

In calculating the present values of future receipts and outlays, the baseline scenario assumes a discount rate of 3.6 percent, which is the average real yield on 30-year Treasury bonds in recent years, as in Gokhale and Smetters (2003).

13

It is thus different than a common interpretation given to the “fiscal gap” measure, namely, the immediate increase in taxes and/or decrease in spending necessary to assure that the debt-to-GDP ratio would return to the initial level by a certain time period. Clearly, restricting the analysis to a finite horizon period would not provide a complete picture of the long-run soundness of the federal budget if the overall fiscal balance continued to worsen at the end of the period.

14

These estimates are much larger than the 75-year period actuarial imbalances reported by the SSA, for the reasons explained in footnote 13, but are broadly consistent with the $44 trillion estimate of the fiscal imbalance prepared by Gokhale and Smetters (2003). Differences between the Gokhale and Smetters estimates and those described above may reflect the weakening of the longer-term fiscal position that has occurred during 2003 and different assumptions regarding the extension of the normal retirement age.

15

One way to illustrate this redistribution is to compare the present value of the net lifetime taxes paid by a person born in the current year versus the taxes that would have to be paid by future generations, assuming that the burden of adjustment is borne by these future generations. The estimates suggest that this gap—in present value terms—would be around $270,000. In other words, future generations would have to pay almost twice as much income tax during their lifetimes than those alive at present, compared with a 60 percent increase in tax burdens that would be required if the increase applied to all generations immediately. A similar redistribution would occur if the adjustment takes place on the expenditure side. See Cardarelli, Sefton, and Kotlikoff (2000); and Gokhale and Smetters (2003) for a discussion of alternative indicators of intergenerational fiscal equity.

    Other Resources Citing This Publication