This Article Extract Figures Only Reprint (PDF) Submit a response to this article Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services E-mail this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to Citation Manager Reprints & Permissions Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (7) Citing Articles via Google Scholar Google Scholar Articles by Gilmer, T. Articles by Kronick, R. Search for Related Content PubMed PubMed Citation Articles by Gilmer, T. Articles by Kronick, R. Related Collections Insurance Coverage

Health Tracking

TRENDS

Calm Before The Storm: Expected Increase In The Number Of Uninsured Americans

Data And Methods

 
Using national health care spending data, we follow the procedure described in our earlier work to estimate per capita health spending for insured adult workers from 1979 to 1999.² We have slightly revised our methods by using data from the 1996 Medical Expenditure Panel Survey (MEPS) in addition to data from the 1987 National Medical Expenditure Survey (NMES) and the 1977 National Medical Care Expenditure Survey (NMCES) to estimate the proportion of expenditures accounted for by insured persons ages nineteen to sixty-four.³

Using data from the March supplements to the Current Population Survey (CPS) from 1980 through 1996 and the revised price series, we reestimate our 1999 model. The logistic regression model predicts the probability of coverage among adult workers as a function of per capita health spending divided by income; employment characteristics such as firm size, self-employment, part-time status, and industry; demographic characteristics such as age, sex, marital status, race, ethnicity, and education; and indicators for region and year. We use these parameter estimates, estimated per capita health expenditures, and data from the March 1997 through March 2000 CPS to predict health insurance coverage for CPS respondents from 1996 through 1999. As in the earlier paper, we restrict the analysis to adult workers not covered as a dependent or by a public program.⁴

Finally, we employ alternative assumptions about future increases in health spending and wage growth and data from the March 2000 CPS to simulate insurance coverage among adult workers from 2000 to 2009. We then use data from the March 1988 through March 2000 CPS to estimate the relationship between the percentage uninsured among adult workers not covered by public programs or as a dependent and the percentage uninsured among the entire under-sixty-five population.⁵ We use this estimated relationship to translate simulated changes in the percentage uninsured among adult workers into estimates of the number of nonelderly uninsured persons under alternative scenarios. To the extent that eligibility for Medicaid, the State Children’s Health Insurance Program (SCHIP), or other public programs changes substantially, trends in the number of uninsured persons could be different than trends in the number of adult workers without private health coverage. However, over the thirteen years from 1988 to 2000, changes in the percentage uninsured in the under-sixty-five population closely tracked changes in the percentage uninsured among adult workers not covered by public programs or as a dependent.

Results

Top Data And Methods Results Discussion NOTES

 
Per capita health spending for nonelderly insured adults divided by the median income of nonelderly adult workers declined very slightly from 8.5 percent in 1995 to 8.4 percent in 1999 (Exhibit 1). Growth in per capita health spending was modest during this time period, averaging 4.3 percent per year, and growth in median personal income was similar—4.7 percent per year. If the model presented in our earlier paper is correct, and if the distribution of other independent variables did not change substantially (for example, if there were not large changes in the number of workers employed in small firms or in the number of young workers), then there should have been a small decline in the percentage of workers who were uninsured from 1995 to 1999.

View larger version (19K): [in this window] [in a new window] EXHIBIT 1 Percentage Uninsured Among Workers And Per Capita Health Spending Divided By Median Income, 1979–1999

View larger version (14K): [in this window] [in a new window] EXHIBIT 2 Percentage Uninsured Among Workers, 1979–1999, And Predicted Percentage Uninsured, 1996–1999

Under the extremely optimistic assumption that health spending and incomes grow as they did in the 1990s, the percentage uninsured among adult workers is projected to change very little over the next decade. Based on this scenario, which assumes a slight decline (–0.5 percent) in the ratio of annual spending growth to income, the percentage of uninsured workers would decline slightly, from 23.1 percent in 1999 to 22.3 percent by 2009. Under the pessimistic, but not unlikely, assumption that health spending and income grow as they did in the 1980s, the percentage uninsured among adult workers is projected to increase from 23.1 percent in 1999 to 30.3 percent in 2009. Using SSA projections of wage growth and CMS projections of health spending increases, the middle scenario estimates that the percentage uninsured among adult workers would increase to 26.2 percent. If there are not substantial changes in eligibility for or enrollment in public coverage, under the optimistic scenario we expect that the number of uninsured persons would decline by 1.3 million over the next decade; under the pessimistic scenario the number would increase by 12.8 million; under the middle scenario, it would increase by 5.3 million.⁸

Discussion

Top Data And Methods Results Discussion NOTES

 
Many analysts have noted with dismay that even in the strong economy of the 1990s the proportion of the population that is uninsured did not decline. We have shown here that this is because the rate of growth of health spending was approximately equal to the rate of growth of personal income during the economic expansion. If economic growth slows and the rate of health spending increases, there is a widespread expectation that the number of uninsured persons will increase. Here we provide quantitative estimates of the magnitude of this relationship. If health expenditures increase at approximately twice the rate of growth of personal income over the next decade, we expect the percentage of the under-sixty-five population that is uninsured to rise from 16 percent to 21 percent.⁹ If the current geographic disparities in insurance rates persist, in California, Texas, and Florida more than one-quarter of the under-sixty-five population will be uninsured.

Editor's Notes

 
Todd Gilmer is an assistant adjunct professor in the Division of Health Care Sciences, Department of Family Preventive Medicine, at the University of California, San Diego, School of Medicine. Rick Kronick is an associate professor and chief of that division.

NOTES

Top Data And Methods Results Discussion NOTES

 

1. R. Kronick and T. Gilmer, "Explaining the Decline in Health Insurance Coverage, 1979–1995," Health Affairs (Mar/Apr 1999): 30–47.
2. For each year, we aggregated national health expenditures for services typically included in an acute care benefit package: hospital and physician services, prescription drugs, home health services, and net administrative cost of insurance. Data from the 1977 NMCES, the 1987 NMES, and the 1996 MEPS were used to allocate a proportion of this spending to adults ages nineteen to sixty-four. Data from the March supplements to the 1980–2000 CPS were used to estimate per capita spending by insured persons ages nineteen to sixty-four.
3. Data from the 1987 NMES and 1977 NMCES show large increases over the decade in the proportion of out-of-pocket and private health spending accounted for by the elderly. Our previous analysis extrapolated this increase through 1995, removing progressively larger shares of spending by the elderly from national health expenditures to construct our estimate of total spending by nonelderly adults. Data from the 1996 MEPS, however, show that the proportion of total spending accounted for by the elderly has remained relatively stable since the 1987 survey. Our revised price series uses these new data. As a result, we estimate higher per capita spending during 1987–1995 for nonelderly adults than we did in our 1999 analysis.
4. We do this for two reasons. First, very few workers covered by public programs or as a dependent purchase coverage on their own, so including them may bias our estimates of the effect of price on purchase decisions. Second, the CPS questions on health insurance coverage changed substantially in 1988 and in 1995. The 1988 changes primarily affected coverage estimates for children age fifteen and younger and for those covered as a dependent; the 1995 changes primarily affected coverage estimates for Civilian Health and Medical Program of the Uniformed Services (CHAMPUS) and dependent coverage. Restricting our analysis to workers not covered by public programs or as a dependent minimizes the effect of the question wording changes and allows us to analyze the 1980–2000 CPS data as a time series of cross-sections.
5. We use an ordinary weighted least squares regression to estimate the relationship between the percentage uninsured among workers and the percentage uninsured for the entire population. This simple approach is highly predictive—the coefficient on the percent uninsured among workers is .73 (SE =.08), R-square =.89—and appears unbiased over the range of prediction.
6. Board of Trustees, Federal Old Age and Survivors Insurance and Disability Trust Funds, 2000 Annual Report of the Board of Trustees of the Federal Old Age and Survivors Insurance and Disability Insurance Trust Funds, House Document 106-262 (Washington: U.S. Government Printing Office, 2000); and S. Heffler et al., "Health Spending Growth Up in 1999; Faster Growth Expected in the Future," Health Affairs (Mar/Apr 2001): 193–203.
7. Using data from national health care expenditure projections, our estimate of per capita health spending is projected to increase by approximately 6.8 percent per year. Using data from the Social Security Administration, we estimate that nominal wages will increase by approximately 4.3 percent per year. Thus, we estimate that per capita health spending divided by income will increase by approximately 2.4 percent per year from 2000 to 2009.
8. These estimates of the number of insured persons hold the population constant at the 1999 level; if we were to factor in population growth, the predicted number of uninsured persons would rise accordingly.
9. Using data from a set of follow-up questions, the Census Bureau recently revised downward from 17 percent to 16 percent its estimate of the under-sixty-five population without health insurance in the March 2000 CPS. See C. Nelson and R. Mills, The March Health Insurance Verification Question and Its Effect on Estimates of the Uninsured (Washington: U.S. Bureau of the Census, August 2001).

                      What's this?

This article has been cited by other articles:

				J. Holahan and A. Cook The U.S. Economy And Changes In Health Insurance Coverage, 2000-2006 Health Aff., March 1, 2008; 27(2): w135 - w144. [Abstract] [Full Text] [PDF]

				R. Kronick Financing Health Care -- Finding the Money Is Hard and Spending It Well Is Even Harder N. Engl. J. Med., March 24, 2005; 352(12): 1252 - 1254. [Full Text] [PDF]

				T. Gilmer, R. Kronick, and T. Rice Children Welcome, Adults Need Not Apply: Changes in Public Program Enrollment across States and over Time Med Care Res Rev, February 1, 2005; 62(1): 56 - 78. [Abstract] [PDF]

				J. A. Pagan and M. V. Pauly Access To Conventional Medical Care And The Use Of Complementary And Alternative Medicine Health Aff., January 1, 2005; 24(1): 255 - 262. [Abstract] [Full Text] [PDF]

				A. Garson Jr U.S. healthcare: the intertwined caduceus of physicians, coverage, quality, and cost J. Am. Coll. Cardiol., January 7, 2004; 43(1): 1 - 5. [Abstract] [Full Text] [PDF]

				A. Astor and G. Sreenivasan Providing Free Care to the Uninsured: How Much Should Physicians Give? Ann Intern Med, November 4, 2003; 139(9): W-78 - W-78. [Full Text] [PDF]