The Rising Burden of Health Spending on Seniors

Studies | Health

No. 297
Thursday, February 01, 2007
by Liqun Liu, Andrew J. Rettenmaier and Zijun Wang


  1. This study is motivated by a series of papers by Victor Fuchs.  In those papers, Fuchs projected health care spending by the elderly, examined the degree of income inequality, noted the importance of transfers from the young in providing consumption for the elderly and pointed to the need for additional savings and work prior to retirement.  See Victor R. Fuchs, "Provide, Provide: The Economics of Aging," in Andrew J. Rettenmaier and Thomas R. Saving, eds., Medicare Reform: Issues and Answers (University of Chicago Press: 1999), pages 15-36.  See also Victor R. Fuchs, "The Financial Problems of the Elderly:  A Holistic Approach," National Bureau of Economic Research, Working Paper No. 8236, April 2001. 
  2. The health-related summary measures are from the Centers for Medicare and Medicaid Services' 2003 Medicare Current Beneficiary Survey, as well as the 2003 Health and Nutrition Examination Survey. Income data are from the Federal Reserve's 2004 Survey of Consumer Finances (SCF) and the U.S. Bureau of Labor Statistics' 2004 Consumer Expenditure Survey. Wealth data is derived from the 2004 SCF data file. Total medical care spending is from the Centers for Medicare and Medicaid Services' 2004 National Health Expenditure Accounts.
  3. Medicare's total percentage of gross domestic product (GDP) is reduced by the total estimated spending on disabled beneficiaries to arrive at the series identified as Medicare spending on the population age 65 and above. Total health care spending for this age group is computed assuming medicare will pay for approximately 51 percent of the retired population's health care spending in all future years.
  4. Total potential consumption is calculated by adding together seniors' total per capita income and total per capita health care spending, less the amount seniors pay from their own income (out-of-pocket health care spending, and premium payments for Medicare and privately purchased medigap insurance).  For the purpose of this examination, taxes are not taken into account.
  5. Victor Fuchs' definition of "full income" is similar to the terms "total potential consumption" or "spending" in this paper.  Fuchs also projects that health care spending as a share of full income will range from 40 percent to 52 percent by 2020 when the growth rate in per capita health care spending is between 1 and 3 percentage points higher than the growth rate in per capita GDP excluding health care. This paper estimates that, by 2020, health care will comprise 48.8 percent of total potential consumption for older Americans.  For more information, see Victor R. Fuchs, "The Financial Problems of the Elderly: A Holistic Approach," National Bureau of Economic Research, Working Paper No. 8236, April 2001.  Total income is estimated based on projected Social Security income per retiree.  Social Security is assumed to represent 38.6 percent of average income in all future years.  Setting the Social Security share to 38.6 percent for all retirees is based on the Social Security Administration 2004 Income of the Aged Chartbook.  Calculating the Social Security share from the Survey of Consumer Finances results in about a 33 percent share of average per capita income while the 2004 Current Population Survey produces a higher 45 percent share.  Seniors' total health care consumption is derived from the Medicare projections from the 2006 Trustees reports assuming that Medicare spending accounts for 51 percent of seniors' total health care spending in all future years.
  6. We estimate that 27 percent of total health care spending is paid by retirees either in the form of out-of-pocket spending or as premium payments.
  7. Spending patterns for the elderly from the Consumer Expenditure Survey are presented in a later section of this paper.  The 2004 Consumer Expenditure Survey estimate of health care spending as a percent of total income for all households headed by individuals age 65 and older is 11 percent. Including the insurance purchases category raises the total to 15.7 percent of income.  Because the Consumer Expenditure Survey is limited to the non-institutionalized population, it therefore omits the 7 percent of the population who are in a long-term care facility sometime during the year.  This paper uses the authors' estimates of health care spending as a percent of income by age group and adjusts the "all other" category from the Consumer Expenditure Survey.  In 2006, we estimate that 17.2 percent of seniors' incomes is allocated to health care spending in the form of out-of-pocket spending and Medicare and individually purchased medigap premium payments. Figure II combines shares from the Consumer Expenditure Survey data for individuals age 65 and older with the estimates of total income and health care spending which underlie the other estimates in this paper.
  8. Medicare has three parts.  Medicare Part A (Hospital Insurance) benefits are funded by a payroll tax on wages.  Medicare Part B (Supplemental Medical Insurance) is funded three-fourths by taxpayers and one-fourth by premiums paid by beneficiaries.  The newly created Medicare Part D (prescription drug benefits) is funded 77 percent by taxpayers and 23 percent by beneficiary premiums, although some of these premiums are paid by the states for dual-eligible beneficiaries.
  9. A recent study by Keehan, Lazenby, Zezza, and Catlin (2004) allows one to infer the sources of health care payments for older Americans.  See Sean P. Keehan, Helen C. Lazenby, Mark A. Zezza and Aaron C. Catlin, "Age Estimates in the National Health Accounts," Health Care Financing Review , Web Exclusive, Vol. 1, No. 1, December 2, 2004; available at  Because of the addition of prescription drug benefits to the Medicare package with the passage of the Medicare Modernization Act in 2003, the sources of payment data are adjusted to reflect this change.  The age tables include out-of-pocket spending, private insurance, other private sources, Medicare, Medicaid and other public payment sources.  Spending is also broken down by types of services including: hospital, physician services, prescription drugs, nursing homes, dental, other professional, other personal, home health, nondurables and durables.  The shares by payment source associated with physician services are assigned to the prescription drugs category and the shares are recomputed within each age classification.  In doing so, Medicare's share rises from 46 percent to 51 percent of total spending for the population 65 and above, based on the most recent estimates (1999).  For comparison, Medicare's share of total health care spending by the retirement age population was 46 percent in 1987 and 49 percent in 1996, based on CMS estimates.
  10. The shares of private insurance spending allocated to individually purchased medigap insurance is based on the age distribution of private insurance coverage for community residents from Table 1.6 in Health and Health Care of the Medicare Population: Data from the 2002 Medicare Current Beneficiary Survey (Rockville, Md.: Westat, 2006).
  11. These amounts are denominated by average Social Security benefits across all retirement-age beneficiaries and by average total retiree income where total income is estimated as before.
  12. These estimates have been based on applying multiples to projected per capita Social Security and Medicare benefits.  They illustrate the implications of the current projections that health care expenditure growth is faster than the expenditure growth on other goods.  It is possible that both health care expenditures and spending on all other goods rise with retirees' incomes, even though the share spent on health care can rise while the share on other goods falls.  But if health care cost sharing rises too quickly, such that retirees are faced with the possibility of falling real consumption on other goods, they will opt for less health care spending growth so that both their health care spending and other spending rise with rising incomes.  
  13. These results do not take into account the behavioral responses on the part of retirees that will result from increased cost sharing, nor do they include the possible supply responses on the part of health care providers.
  14. "The Characteristics and Perceptions of the Medicare Population: Data from the 2003 Medicare Current Beneficiary Survey," Centers for Medicare and Medicaid Services.  Available at 
  15. Kevin M. Murphy and Robert H. Topel, "The Value of Health and Longevity," Journal of Political Economy , Vol. 114, 2006.
  16. David M. Cutler and Elizabeth Richardson, "Measuring the Health of the United States Population," Brookings Papers on Economic Activity: Microeconomics, 1997.
  17. Data for this section is derived from the Consumer Expenditure Survey (CEX) sponsored by the Bureau of Labor Statistics and the Survey of Consumer Finances sponsored by the U.S. Federal Reserve.  The CEX is primarily used to define the basket of goods and their weights in the Consumer Price Index (CPI). The CEX data is summarized by consumer units, or groups of individuals who share financial resources. These groups are typically families or households. The age used in the figures based on the CEX data is the age of the reference person, or the person in the consumer unit who owns or rents the residence. 
  18. For a discussion of the effects of this change on the labor market participation among older workers, see Liqun Liu and Andrew J. Rettenmaier, "Work and Retirement," National Center for Policy Analysis, Policy Backgrounder No. 162, November 3, 2006; available at
  19. The Survey of Consumer Finances sponsored by the Board of Governors of the Federal Reserve System is a triennial survey of the balance sheet, pension, income and other demographic characteristics of U.S. families.
  20. Consumer Expenditure Survey, U.S. Bureau of Labor Statistics, 2004.  In the Consumer Expenditure Survey, health care primarily refers to out-of-pocket spending on medical care and does not include pretax employer premium payments or Medicare reimbursements on behalf of retirees.  See footnote 6 for additional discussion of the Consumer Expenditure Survey.  

Read Article as PDF