Over the next 75 years, health benefits as a share of worker compensation could more than quadruple. Despite this, real cash wages per worker will be 7.5 times as much as the amount in 2008.
If current trends continue, the ratio of health-related fringe benefits to worker wages will more than quadruple. Admittedly, forecasting over a 75-year period is challenging. However, estimates of what wages will be through 2083 use the identical assumptions about growth in real (inflation-adjusted) wages per worker that are embedded in the most recent projections from Social Security trustees. Assuming that changes in these shares mirror what has happened during the past 10 years, the share of total compensation for health benefits and non-health fringe benefits can be projected,.
Using these simple assumptions, the ratio of health-related supplements per dollar of wages would grow over 75 years from 6.4 cents in 2008 to 26.3 cents (figure 20.1). Even after taking into account the growth in non-health fringe benefits as a share of compensation—projected in a parallel fashion—the amount of real cash wages will grow considerably, relative to 2008. That is, in terms of constant purchasing power, workers in 75 years will have nearly eight times as much non-health compensation as they received in 2008.
Succinctly, even though employers will have to devote a growing share of compensation to health care, these costs currently are not growing so rapidly that they will entirely displace the parallel (though slower) growth in real wages. Such a time might come, but not in the foreseeable future.
Download PowerPoint versions of figure.
Department of Commerce. Bureau of Economic Analysis.
Technology has been a far more important driver of health spending growth during the past 60 years than has population growth or aging.
Technology and other factors appear to account for approximately half or more of the growth in per capita health spending over the past 60 or more years. As figure 20.2 illustrates, the exact amount depends on the period examined and other methodological assumptions.
The exact contribution of technology is unknown (recall figure 17.3). When analysts have accounted for all the other major factors that contributed to the rise in inflation-adjusted health spending per person (for example, medical price inflation), most of the "residual" that cannot be explained is attributed to technology. Some of it is an income effect, a willingness to purchase medical services (for example, Lasik surgery) if a person has a higher income, but a person who has a lower income was not willing to purchase it. It is difficult to separate out accurately the effect of income from technology. Some would also argue that calculations such as these underestimate the role of insurance in encouraging the development of technology.
However it is defined, this "residual" is what government forecasters use to determine how fast health care expenditures will increase in the future. They can use demographic models to calculate how quickly the population is increasing and how the age-gender mix of the population is changing over time. Relative spending changes over the life cycle, depending on gender (refer to figure 12.5). Thus, if everything were static, it would be a straightforward mathematical exercise to determine how much total spending will rise because of such demographic changes.
However, the big unknown is how much per-person spending for everyone is likely to rise beyond growth in the general economy. To answer this question, fore- casters have no choice but to look to the past. They cannot avoid making assumptions about the future. The 2000 Medicare Technical Panel estimated that per capita annual growth in health care spending would be 2.2 percent a year and that long-run real per capita GDP growth would be 1.2 percent. This implies that health spending growth would exceed GDP growth by one percentage point a year. Thus, they recommended using "GDP + 1" (that is, 1.2 percent plus 1 percent) as the basis for projecting health spending over the next 75 years. The Medicare Trustees' reports from the year 2000 forward all have used this assumption in deriving long-term projections (i.e., years 25 through 75 of the projection period).
Download PowerPoint versions of figure.
Brown JD and RM Monaco. Possible Alternatives to the Medicare Trustees' Long-Term Projections of Health Spending. US Department of Treasury. Office of Economic Policy. 2004.
The excess cost ratio measures how much faster per capita health spending—adjusted for changes in age and gender—increased relative to growth in per capita GDP. Historical variations in the excess cost ratio make predicting future health spending difficult, especially over a long time.
"Excess cost growth" is the name for the residual just described. It is the difference between the U.S. per capita growth rate in age- and gender-adjusted health care costs minus the per capita growth in GDP (both rates are calculated from inflation-adjusted estimates of health spending and GDP). The level of excess cost growth has varied over time. Not surprisingly, this variation has been largest when short (five-year) time intervals are used (figure 20.3a). Over longer periods, there is less variation (but part of the reason is because the observed periods overlap).
Excess cost growth from 1940-1990 was 2.3 percent a year. The data from the longest periods illustrated in figure 20.3a (25 and 32 years) suggest that excess cost growth might have declined to less than 2 percent. However, certainty is impossible to achieve. During this period, many policy changes in Medicare (introduction of a prospective payment system for hospitals in 1983, establishment of the Medicare physician fee schedule in 1992, substantial Medicare cost-containment measures in 1997) might have had the effect of temporarily slowing growth in spending, thereby masking the underlying trend having to do with technology that might continue in the future.
Forecasts of the year in which the Medicare Trust Fund is exhausted have varied substantially over 20 years (figure 20.3b). This is not to suggest that government forecasters are incompetent. First, by their nature, official projections of Medicare spending by Medicare Trustees or of Medicare, Medicaid, and total health spending by the CBO reflect only current policy. Policy can and did change in response to these forecasts. Remember this when using hypothetical forecasts in the following pages to explore whether current trends in health spending are sustainable in the future. Uncertainty is an unavoidable feature of long-term forecasts.
Download PowerPoint versions of both figures.
Boards of Trustees, Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds, The. 2010 Annual Report. US Government Printing Office. August 5, 2010.
Department of Health and Human Services. Centers for Medicare and Medicaid Services.
Over the next 50 years, the country can "afford" growth in health spending that exceeds growth in the general economy only if the difference is not too much. Continuing historical rates of excess growth in health spending could result in a decline in real GDP per capita within 30 years.
After deducting amounts for health care, the inflation-adjusted amount of personal consumption spending per American more than tripled in the 80 years since 1929 (figure 20.4a). Despite enormous growth since 1929 in health spending and the size of government—both of which reduced the amount that otherwise would have been available for personal consumption—what Americans had left for everything else still was able to grow. Admittedly, what was left would have been twice as large in a world without any spending on health care or government. No one aspires to live in such a hypothetical world. The purpose of figure 20.4a is to illustrate that the growth in health care (and government) has not been so rapid that it reduced the American standard of living from generation to generation.
However, the United States cannot afford a continuation of historical rates of excess cost growth unless Americans are comfortable with a decline in non-health GDP per capita within the next 75 years. How quickly that occurs will depend on which historical excess cost growth number is used.
Using actual experience from relatively recently (1990-2008), annual excess cost growth was 1.4 percent. Continuing this into the future would produce a downturn in inflation-adjusted non-health spending per capita in the year 2070 (figure 20.4b). However, extrapolating the experience of 1975-2008 (1.9 percent excess cost growth) would result in a decline in real GDP per capita by 2053. This downturn would be so sharp that health care would absorb all of GDP by 2090.
Because people have to have food, clothing, and shelter, 100 percent of GDP logically can never be fully devoted to health care. These projections are not predictions because they lack assumptions about changes in policy or behavior. Their use lies in demonstrating how much of a change in experience is required to "bend the cost curve" enough to avoid an undesirable outcome.
Download PowerPoint versions of both figures.
Author's calculations.
Congressional Budget Office.
Department of Commerce. Bureau of Economic Analysis.
Department of Commerce. Bureau of the Census.
Department of Health and Human Services. Centers for Medicare and Medicaid Services.
Social Security Administration.
Worthington NL. National Health Expenditures, Calendar Years 1929-73. Research and Statistics Note No 1. Office of Research and Statistics 1975.
Much uncertainty exists in 75-year forecasts of health spending; yet if even 1 percent excess cost growth persists, almost 90 percent of annual GDP growth will be devoted to health care by 2085.
Uncertainty is pervasive in long-term forecasts of health spending. An approach called "stochastic modeling" is used by Medicare Trustees and other analysts to try to quantify the extent of that uncertainty.
Because perfect prediction is impossible, the goal of analysts is to identify some boundaries within which the true value of future spending is likely to lie. A 95 percent probability interval (also termed a confidence interval) is a typical approach to articulating these boundaries. Given uncertainty in future population, health status, and growth in health-status-adjusted per-beneficiary spending, the actual level of future spending is expected to fall within such an interval 95 percent of the time. There is more than a four-to-one ratio between the estimated Medicare share of GDP at the top of this interval compared with the bottom (figure 20.5a).
Even official forecasts of Medicare spending vary dramatically. In its 2010 annual report, the Medicare Trustees predicted that total Medicare spending in the year 2080 would be 6.4 percent (figure 20.5b). The CBO, using different assumptions about growth in GDP and growth in Medicare spending, concluded that Medicare by 2080 would account for 11 to 12 percent of GDP. Which prediction is "true" is less important than understanding that even highly skilled experts using reasonable variations in assumptions can produce widely disparate estimates of 75-year costs. Even a seemingly small difference in assumptions (for example, GDP will grow 0.1 percent more slowly) accumulates into a vast difference when compounded over decades.
All the official forecasts assume a decline in the excess cost growth rate over the 75-year projection period. After 75 years, the Trustees assume that excess cost growth will be zero, that is, that Medicare will grow at the same pace as the general economy. The CBO, in contrast, assumes that excess cost growth for all other components of health spending (including Medicaid, CHIP, and exchange subsidies) will be zero, but that the excess cost growth for Medicare in 2085 still would be one percentage point. Were excess cost growth of one percentage point to continue for the entire health system, it would imply that by 2085, 90 cents of every extra dollar of GDP would flow into the health system (figure 20.5c).
Download PowerPoint versions of all figures.
Author's calculations.
Boards of Trustees, Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds, The. 2010 Annual Report. US Government Printing Office. August 5, 2010.
Congressional Budget Office.
Lee R and T Miller. An Approach to Forecasting Health Expenditures, with Application to the US Medicare System. Health Services Research 2002; 37(5):1365-86.
In today's dollars, the long-term unfunded liabilities associated with health entitlements exceed $66 trillion—approximately four times as much as the unfunded liabilities related to social security.
The unfunded liabilities for entitlements exceed $80 billion over the long term (figure 20.6a). Only approximately 20 percent has to do with Social Security. The rest ($66 trillion) is due to health entitlements. If it works as planned, the new health plan will reduce unfunded liabilities for Medicare by tens of billions of dollars.
However, both the Medicare actuary and the CBO have raised questions about whether some of the law's underlying premises will be maintained. For example, all the CBO projections used in this section assume that physician fees under Medicare will be cut by approximately 30 percent. With looming physician shortages, few believe such a drastic cut is desirable even though current law technically requires it. As it has for the past eight years, Congress is expected to keep deferring this scheduled cut indefinitely or to change the law to eliminate it. The CBO has developed an alternative fiscal scenario in which this and several other policies designed to limit spending would not continue. Under these alternative assumptions, Medicare spending as a percent of GDP in 2080 (as estimated by the CBO) would be two percentage points higher than illustrated in figure 20.5b.
Calculations for unfunded liabilities assume that current payroll tax levels for Social Security and Medicare Part A remain in effect. Another assumption is that the current level of Medicaid and Medicare Parts B and D spending (as a percentage of GDP) reflect societal willingness-to-pay for these programs. The "unfunded" amount equals the increase in the burden (relative to current levels) required to sustain these programs.
To give some sense of whether these entitlements are "affordable," figure 20.6b expresses the funding shortfall as a percent of payroll. Essentially, the Social Security (FICA) payroll tax of 15.3 percent would have to more than double by the year 2080 simply to bankroll health-related entitlements. In present-value terms, the long-term unfunded liability is five times the U.S. national income (GDP) (figure 20.6c). This is approximately equivalent to assuming a mortgage equal to five times a family income. The unfunded liability is 1.5 times as much as the country's net worth. This is approximately equivalent to borrowing 1.5 times a family's net worth. Countries are not families, but these comparators at least provide a rough sense of just how large the problem of funding future entitlements has become.
Download PowerPoint versions of all figures.
Author's calculations.
Boards of Trustees, Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds, The. 2010 Annual Report. US Government Printing Office. August 5, 2010.
Congressional Budget Office.
Department of Commerce. Bureau of Economic Analysis.
Social Security Administration.
The projected 75-year increase in mandatory federal health spending exceeds current revenues from the three largest sources of federal tax revenue.
The size of the projected increase in mandatory federal health spending dwarfs the amount currently collected through the three largest sources of federal revenue (figure 20.7a). Thus, filling the fiscal gap through higher taxes would imply more than doubling individual income taxes, quintupling the amount collected in corporate income taxes, or tripling current payroll taxes (all approximations). Alternatively, because the long-run increase in mandatory federal health spending amounts to 67 percent of federal tax revenue, this would imply that a minimum 67 percent increase in federal taxes across the board would be needed to tax ourselves out of the health entitlements burden. Of course, given the inevitable behavioral response that would result from increasing taxes by this magnitude, the increase in tax rates would have to be even higher than these multipliers suggest.
Even if taxpayers could find a way to bankroll the enormous increases in health spending reviewed in this section, it is less clear that beneficiaries of public insurance programs such as Medicare can absorb their share of rising health costs. By 2085, the out-of-pocket burden of just Medicare Parts B and D will approximately double for the average Social Security recipient (figure 20.7b). By 2085, premiums and out-of-pocket expenses will absorb more than half of the average Social Security check, assuming that such checks are not trimmed as part of the efforts to save Social Security. This does not even count any cost sharing associated with Medicare Part A. The CMS Office of the Actuary has developed an illustrative alternative scenario that recognizes that some of the Medicare spending reductions contemplated by the health reform law might not occur. For example, the cuts in physician fees required by the sustainable growth rate formula (SGR) now have been overridden by Congress for nine consecutive years. Under this alternative scenario, premiums and out-of-pocket expenses for Medicare Parts B and D would equal four-fifths of the average Social Security check by 2085.
Burdens of this magnitude will amplify pressures for the federal government to subsidize such individuals even more to make their health care affordable. This would then further escalate the amount of federal health spending beyond the levels already described. Finding ways to reduce spending, rather than raising revenues, at least offers the prospect of averting this bind on the elderly who have fixed incomes.
Download PowerPoint versions of both figures.
Author's calculations.
Boards of Trustees, Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds, The. 2010 Annual Report. US Government Printing Office. August 5, 2010.
Congressional Budget Office.
Organisation for Economic Co-operation and Development.
Potetz L and J Cubanski. A Primer on Medicare Financing. July 2009. http://www.kff.org/medicare/upload/7731-02.pdf (accessed August 10, 2010).
U.S. competitors have already had to confront the challenge of an increasing number of dependents—both children and elderly—per working adult. For the United States, much of this challenge still lies in the future.
The U.S. elderly-dependency ratio is much lower than in any other G7 nation, despite having risen during the past 50 years (figure 20.8a). The elderly-dependency ratio measures the number of people age 65 and older relative to a "working-age" population (those ages 15-64). The U.S. ratio is lower, and in almost all comparisons, it increased less rapidly over the past 25 years (absolutely and relatively) than among the nation's major competitors. Japan's ratio, for example, was less than the U.S. level in 1984, yet by 2007 was almost twice as high.
The OECD projects that this U.S. margin of advantage will persist through 2050, because all these countries will experience sharp rises in their dependency ratios. However, due to the "Baby Boomers," the U.S. ratio is expected to reach almost 45 in the year 2040 before declining to approximately 40 in 2050. Thus, the nation is doing better in relative terms, but it still must face the fiscal challenges posed by the dependency ratio more than doubling over the next few decades. In fact, the number of covered workers per Social Security beneficiary is expected to decline even under the most optimistic assumptions (figure 20.8b).
The U.S. demographic margin of advantage is offset considerably when differences in the relative burden of elderly health spending are taken into account. In the United States, per capita elderly health spending amounts to approximately 45 percent of the average annual compensation for manufacturing workers (figure 20.8c). In the rest of the G7, this fraction is approximately only half as much. Manufacturing compensation is considered a "good" wage in all countries, and standardized cross-national estimates of hourly compensation for such workers are readily available. Because U.S. manufacturing productivity is higher than in these other countries (contributing to its higher wages), this comparison should favor the United States. To observe such a large differential despite a comparison tilted in the nation's favor is quite striking. It implies that relative to a competitor, the United States could face a comparable burden of financing elderly health care even if its elderly-dependency ratio were only half as much.
Download PowerPoint versions of all figures.
Boards of Trustees, Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds, The. 2010 Annual Report. US Government Printing Office. August 5, 2010.
Organisation for Economic Co-operation and Development.
Projected increases in government-related health spending could, within 50 years, eradicate the margin of advantage that the United States currently has over its major European competitors in terms of the burden of government.
The United States holds a considerable margin of advantage over its major European rivals in terms of the government share of GDP (figure 20.9a). Compared with countries such as Italy, France, and the UK, that margin of advantage is approximately 10 percent of GDP; in the case of Germany, it exceeds 5 percent of GDP. The size of government (all levels) in Canada and Japan is much more comparable to that of the United States.
Although much of U.S. federal spending has been deficit-financed in recent years, that situation cannot continue indefinitely. Eventually, increased government spending as a percent of GDP translates into relatively higher taxes and the hidden efficiency costs they impose. At the margin, these efficiency losses in the United States amount to more than 40 cents on the dollar. All other things being equal, a country that can minimize such efficiency losses will outperform a nation that has a larger government.
However, absent changes in policy, this margin of advantage could be more than eradicated over the next 50 years simply through increases in mandatory federal health spending for the aging population in the United States (figure 20.9b). Other nations also face a rising tax burden for aging populations. The size of their burden is less, for two reasons. First, many countries are ahead of the United States in terms of the share of their populations that is elderly. Thus, they have comparatively less future aging to address. Second, in nations with universal coverage, government already has been financing their health care in advance of retirement. The incremental increase in tax burdens associated with an already-covered 64-year-old becoming 65 is far less than for a U.S. retiree whose health care for the first time is a federal responsibility.
The United States has been able for decades to enjoy the highest standard of living in the world only by maintaining one of the world's most highly competitive economies. Unless it can address its entitlements crisis without impairing its own competitiveness, the nation might find itself unable to afford either the health care or other goods that its residents desire. Conversely, by addressing entitlements in a responsible fashion, the United States can continue to afford investments in health care that provide good value for the money, with the attendant increases in longevity and years of healthy life that are reasonable to expect from a high-performing health care system.
Download PowerPoint versions of both figures.
Hagist C and L Kotlikoff. Who's Going Broke? Comparing Healthcare Costs in Ten OECD Countries. NBER Working Paper No. 11833. December 2005.
Organisation for Economic Co-operation and Development.