The kindness of strangers: Community effects on the rate of employer coverage

Health Affairs

Volumn 23, Issue SUPPL., 2004, Pages

  Kronick, Richard ^a   Gilmer, Todd ^a   Rice, Thomas ^b  

^a UNIVERSITY OF CALIFORNIA   (United States)

^b UNIVERSITY OF CALIFORNIA   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

ARTICLE; EMPLOYER; HEALTH CARE POLICY; HEALTH INSURANCE; PREDICTION; WORKER;

EID: 4644362072     PISSN: 02782715     EISSN: None     Source Type: Journal    
DOI: 10.1377/hlthaff.W4.328     Document Type: Article

References (23)

1. Authors' estimates from the March Supplements to the Current Population Survey, 1996-2002.
2. Y.-C. Shen and S. Zuckerman, "Why Is There State Variation in Employer-Sponsored Insurance?" Health Affairs 22, no. 1 (2003): 241-251.
3. P.J. Cunningham and P.B. Ginsburg, "What Accounts for Differences in Uninsurance Rates across Communities?" Inquiry 38, no. 1 (2001): 6-21.
4. Supplemental Exhibit 1 can be viewed online at content.healthaffairs.org/ cgi/content/full/hlthaff.w4 .328/DC2.
5. To measure variation across states in the generosity of public programs, we used data from the March Supplements to the CPS from 1996-2001, and we estimated state effects on the likelihood that a low-income adult (under 200 percent of the federal poverty level) not covered by private insurance would be covered by Medicaid, controlling for a wide range of demographic and employment characteristics. For the United States as a whole, 35 percent of low-income adults who are not covered by private insurance have public coverage (the remaining 65 percent are uninsured). The predicted public coverage level among low-income adults varies from a high of approximately 53 percent in Tennessee, Minnesota, and Vermont to a low of approximately 25 percent in Virginia and Alabama. Details on the method of measuring public program generosity are provided in T. Gilmer, R. Kronick, and T. Rice, "Children Welcome, Adults Need Not Apply," Medical Care Research and Review (forthcoming).
6. We used 1999 data from the American Hospital Association to calculate the percent of non-Medicare hospital revenue in a state that is provided as bad debt or free care. For the United States as a whole, approximately 11 percent of non-Medicare hospital revenue is bad debt or free care, and 17 percent of the under-age-sixty-five population is uninsured; our index of the relative availability of free care for the United States as a whole averages 0.63. The index ranges from a high of 1.1 in Arkansas and Louisiana, where the percent of non-Medicare services delivered as bad debt or free care is greater than the percentage of the under-age-sixty-five population that is uninsured, to a low of 0.28 in North Dakota and Washington, where there is very little bad debt and free care relative to the number of uninsured people.
7. For hospital; physician; and other professional, pharmacy, and home health services, we estimated spending on under-age-sixty-five, privately insured people in each state by subtracting Medicare and Medicaid spending from total spending and then subtracting the portion of out-of-pocket spending estimated to be accounted for by people over age sixty-five, where the elderly's out-of-pocket spending proportion is estimated from data from the Medical Expenditure Panel Survey (MEPS) and assumed to be constant across states. We divided estimated state-level spending on people under age sixty-five by the number of people in the state in that age group to estimate per capita spending per person under age sixty-five.
8. For each indicator (dummy) variable, we computed the marginal effect by predicting the probability of coverage twice for each of the 103,000 people included in the analysis: once under the assumption that the respondent had that characteristic, and once under the assumption that the respondent did not have it. The difference in probabilities, averaged over the 103,000 people in the sample, is our estimate of the marginal effect of the variable. For the interval level variables measuring area characteristics, we computed the probability of coverage for each person, first assuming that the variable is equal to the U.S. average and then assuming that it is equal to the average plus one standard deviation.
9. Supplemental Exhibit 2 can be viewed online at content.healthaffairs.org/ cgi/content/full/hlthaff.w4 .328/DC2.
10. Supplemental Exhibit 3 can be viewed online at content.healthaffairs.org/ cgi/content/full/hlthaff.w4 .328/DC2.
11. D.M. Cutler and J. Gruber, "Does Public Insurance Crowd-Out Private Insurance?" Quarterly Journal of Economics 111, no. 2 (1996): 391-430; K.N. Rask and K.J. Rask, "Public Insurance Substituting for Private Insurance: New Evidence Regarding Public Hospitals, Uncompensated Care Funds, and Medicaid," Journal of Health Economics 19, no. 1 (2000): 1-31; R. Kronick and T. Gilmer, "Explaining the Decline in Health Insurance Coverage, 1979-1995," Health Affairs 18, no. 2 (1999): 30-47; L.J. Blumberg, D. Dubay, and S. Norton, "Did the Medicaid Expansions for Children Displace Private Insurance? An Analysis using the SIPP," Journal of Health Economics 19, no. 1 (2000): 33-60; and P.J. Cunningham, "Declining Employer-Sponsored Coverage: The Role of Public Programs and Implications for Access to Care," Medical Care Research and Review 59, no. 1 (2002): 79-98; discussion, 99-103.
12. D.M. Cutler and J. Gruber, "Does Public Insurance Crowd-Out Private Insurance?" Quarterly Journal of Economics 111, no. 2 (1996): 391-430; K.N. Rask and K.J. Rask, "Public Insurance Substituting for Private Insurance: New Evidence Regarding Public Hospitals, Uncompensated Care Funds, and Medicaid," Journal of Health Economics 19, no. 1 (2000): 1-31; R. Kronick and T. Gilmer, "Explaining the Decline in Health Insurance Coverage, 1979-1995," Health Affairs 18, no. 2 (1999): 30-47; L.J. Blumberg, D. Dubay, and S. Norton, "Did the Medicaid Expansions for Children Displace Private Insurance? An Analysis using the SIPP," Journal of Health Economics 19, no. 1 (2000): 33-60; and P.J. Cunningham, "Declining Employer-Sponsored Coverage: The Role of Public Programs and Implications for Access to Care," Medical Care Research and Review 59, no. 1 (2002): 79-98; discussion, 99-103.
13. D.M. Cutler and J. Gruber, "Does Public Insurance Crowd-Out Private Insurance?" Quarterly Journal of Economics 111, no. 2 (1996): 391-430; K.N. Rask and K.J. Rask, "Public Insurance Substituting for Private Insurance: New Evidence Regarding Public Hospitals, Uncompensated Care Funds, and Medicaid," Journal of Health Economics 19, no. 1 (2000): 1-31; R. Kronick and T. Gilmer, "Explaining the Decline in Health Insurance Coverage, 1979-1995," Health Affairs 18, no. 2 (1999): 30-47; L.J. Blumberg, D. Dubay, and S. Norton, "Did the Medicaid Expansions for Children Displace Private Insurance? An Analysis using the SIPP," Journal of Health Economics 19, no. 1 (2000): 33-60; and P.J. Cunningham, "Declining Employer-Sponsored Coverage: The Role of Public Programs and Implications for Access to Care," Medical Care Research and Review 59, no. 1 (2002): 79-98; discussion, 99-103.
14. D.M. Cutler and J. Gruber, "Does Public Insurance Crowd-Out Private Insurance?" Quarterly Journal of Economics 111, no. 2 (1996): 391-430; K.N. Rask and K.J. Rask, "Public Insurance Substituting for Private Insurance: New Evidence Regarding Public Hospitals, Uncompensated Care Funds, and Medicaid," Journal of Health Economics 19, no. 1 (2000): 1-31; R. Kronick and T. Gilmer, "Explaining the Decline in Health Insurance Coverage, 1979-1995," Health Affairs 18, no. 2 (1999): 30-47; L.J. Blumberg, D. Dubay, and S. Norton, "Did the Medicaid Expansions for Children Displace Private Insurance? An Analysis using the SIPP," Journal of Health Economics 19, no. 1 (2000): 33-60; and P.J. Cunningham, "Declining Employer-Sponsored Coverage: The Role of Public Programs and Implications for Access to Care," Medical Care Research and Review 59, no. 1 (2002): 79-98; discussion, 99-103.
15. D.M. Cutler and J. Gruber, "Does Public Insurance Crowd-Out Private Insurance?" Quarterly Journal of Economics 111, no. 2 (1996): 391-430; K.N. Rask and K.J. Rask, "Public Insurance Substituting for Private Insurance: New Evidence Regarding Public Hospitals, Uncompensated Care Funds, and Medicaid," Journal of Health Economics 19, no. 1 (2000): 1-31; R. Kronick and T. Gilmer, "Explaining the Decline in Health Insurance Coverage, 1979-1995," Health Affairs 18, no. 2 (1999): 30-47; L.J. Blumberg, D. Dubay, and S. Norton, "Did the Medicaid Expansions for Children Displace Private Insurance? An Analysis using the SIPP," Journal of Health Economics 19, no. 1 (2000): 33-60; and P.J. Cunningham, "Declining Employer-Sponsored Coverage: The Role of Public Programs and Implications for Access to Care," Medical Care Research and Review 59, no. 1 (2002): 79-98; discussion, 99-103.
16. In the model including only individual-level variables, the weighted average regional effect is 1.4 percent age points. In the model including both individual and community-level characteristics, the weighted average regional effect is only 0.5 percentage points, indicating that these variables largely account for the observed regional differences in coverage rates.
17. See Note 10.
18. Excluding the outliers of Alaska and Hawaii, the weighted standard deviation of the state-level residual (that is, the difference between the actual and predicted levels of ESI in each state) is 0.012; the weighted standard deviation of the actual level of ESI is 0.057, which indicates that the factors included in the model account for most of the geographic variation in the level of ESI.
19. A.W. Dick, "Will Employer Mandates Really Work? Another Look at Hawaii," Health Affairs 13, no. 1 (1994): 343-349.
20. Complete results, by state, are available from the authors. Contact rkronick@ucsd.edu.
21. The decomposition results are computed separately for each state. In Exhibit 3 we present results averaged over groups of states because the sources of variation are similar for the states in each group.
22. The effect of income in our analysis is much greater than in the Shen-Zuckerman results. This is primarily because Shen and Zuckerman entered income as a single variable - implicitly assuming a linear relationship between income and probability of coverage - while our model uses a series of 0-1 indicator variables and more sensitively estimates the effects of income.
23. Some have argued that mandates to offer insurance may reduce the level of employment-increasing insurance coverage among workers who have jobs but reduce the number of workers overall. Our analysis does not directly address this controversy.