تقاضا برای بیمه سلامت وابسته: هزینه پوشش خانواده چقدر مهم است؟

From the mid-1980s to the mid-1990s, the proportion of non-elderly Americans with employment-based health insurance declined. Roughly 80% of this decline was due to the loss of coverage by dependent family members. During this period, workers became increasingly responsible for the costs of family coverage, while expanded Medicaid coverage provided low-income working families with an alternative to employment-based insurance. We examine the role of out-of-pocket premiums and expanded Medicaid eligibility in households’ demand for employment-based family coverage. Cross-sectional results reveal that demand is affected by both factors. We find that between 1987 and 1996, the increase in out-of-pocket premium costs accounted for nearly half of the decline in dependent coverage while expanded Medicaid eligibility represented 14% of the decline.

From the latter part of the 1980s to the mid-1990s, the proportion of non-elderly Americans covered by employer-sponsored health insurance (ESI) declined markedly, from 70.1% in 1987 to 64.8% in 1996 (Fronstin, 2003). What was particularly striking about this reduction in coverage, and what distinguishes it from subsequent periods, was that it was dominated by a reduction in coverage for dependent family members. While the percentage of non-elderly Americans with ESI in their own name declined slightly over this period (from 34.3% in 1987 to 33.3% in 1996), the reduction in the proportion of individuals covered as dependents on such health plans fell from 35.8% to 31.5% (some 3.1 million persons), and represented 81% of the total decline.1 While much research has focused on the overall decline in ESI since the mid-1980s, there has been less effort directed to identifying the precise mechanisms associated with the sizeable decline in dependent coverage over the period cited. Since the late 1980s, two factors have been frequently cited as affecting the decisions of some households to enroll dependents in ESI. First, employees have become increasingly responsible for bearing the cost of family coverage and this cost remains well in excess of that required for single-person coverage.2 At the same time, the ability of many low and middle-income families to afford the increased out-of-pocket premium costs for family coverage has deteriorated as real earnings for such households remained relatively stagnant. Next, the expansion of Medicaid eligibility during the late 1980s and early 1990s also provided low-income working families with an alternative to ESI at no out-of-pocket cost. Such enhanced access to public insurance, together with the rise in out-of-pocket premium costs for ESI, may have induced some households to substitute public for private coverage, especially for dependent children. While estimates of the magnitude of private insurance crowd out vary widely, there has been less effort directed at identifying the precise mechanisms through which such substitution may have occurred. Finally, while the decline in employment-based dependent coverage together with the increased Medicaid enrollment did not increase the overall proportion of uninsured children (Weinick and Monheit, 1999), the shift from private to public sources of coverage raises questions about the equity and efficiency of policies seeking to expand health insurance to disadvantaged groups. In this paper, we address a specific gap in research related to the decline in ESI by examining factors that affect household decisions to obtain ESI for dependent family members. Our primary focus is to consider whether the insurance decision of a household with dependent family members is sensitive to a worker's out-of-pocket premium costs for single and family coverage. By examining the variation in these out-of-pocket costs across households, we can assess the price sensitivity of the demand for coverage and also gain insight into how increases in worker responsibility for health insurance premiums have affected the insurance status of dependents. We extend this inquiry by assessing the contribution of changes in out-of-pocket premiums and expanded Medicaid eligibility to the decline in rates of ESI and family coverage based on estimates from our 1987 and 1996 cross-sectional data. Our analysis reveals that both factors contributed to the decline in ESI for households with dependent family members over this period.

An important feature of the decline in the proportion of non-elderly Americans with ESI from the late 1980s to the mid-1990s has been the diminished rate of coverage for dependent family members. Rising health insurance premiums, increasing employee contributions, and stagnant earnings for low-skilled workers over this period have focused attention on whether certain groups of workers could afford ESI. Research assessing the impact of the Medicaid expansions has suggested that such initiatives may have reduced the rate of ESI through a crowd out of private coverage. Despite this focus, little effort has been directed to identifying the precise mechanisms that have led households to reduce the rate at which they insure dependents through ESI. Toward this end, we have applied cross-sectional data from the 1987 NMES and 1996 MEPS in an attempt to document the decline in dependent coverage, to quantify the role played by the out-of-pocket cost of single and family coverage and the availability of alternative coverage in decisions to obtain any ESI as well as family coverage, and to draw implications for the changes observed over time. In descriptive analyses, we find striking evidence that households with a worker-policyholder of ESI were less likely to insure all family members over our study period. This change occurred for all family types but was most pronounced for two-parent HIEUs with multiple children and for single-parent HIEUs. While the likelihood of having an uninsured child increased for all HIEUs between 1987 and 1996, the increases were greatest for lower income two-parent households. The dependent children who experienced relatively large declines in ESI and increases in uninsured rates were those less than age 4, those aged 18–23 who were full time students, those eligible for Medicaid, and those in households with incomes less than twice the poverty line. To answer the question posed at the outset of our analysis, we do find that the out-of-pocket cost of coverage affects household demand for family coverage. Our results for both 1987 and 1996 indicate that increases in such costs reduce the likelihood that a household selects ESI or a family health plan, although the effects are small. Our empirical work further suggests that the availability of alternative health insurance can contribute to household decisions to obtain any ESI and family coverage in particular. We find that having a family member eligible for Medicaid reduces the likelihood of obtaining ESI and family coverage as does the presence of alternatives to ESI such as CHAMPUS/CHAMPVA. In assessing the contribution of each of these factors to the changes in the rate of family coverage among households with ESI between 1987 and 1996, we find that changes in the cost of coverage, especially the increase in employee contributions for family coverage, account for about half of the decline in this rate and that expanded Medicaid eligibility represents 14% of this decline. Our findings suggest that efforts to reduce the growth in health care costs, which have fueled the rise in health insurance premiums and out-of-pocket premium costs, may make family coverage relatively more attractive to families, some of whom may have access to public programs. Such efforts may contribute to the multiple aims of reducing the number of uninsured dependents, possibly lessening the effect of any Medicaid crowd out, and enhancing the availability of public funds to be used on behalf of the more disadvantaged.