Recent Trends in Children's Health Insurance

Originally published by the Center for Studying Health System Change (HSC). HSC was a nonpartisan policy research organization funded principally by the Robert Wood Johnson Foundation.

No Gains for Low-Income Children

Peter J. Cunningham, Michael H. Park

Despite targeted expansions of public coverage through Medicaid and the State Children's Health Insurance Program (SCHIP), low-income children saw no improvement in their overall insurance rates between the mid- and late 1990s. According to data collected in 1996-1997 and 1998-1999 by the Center for Studying Health System Change (HSC) through the Community Tracking Study (CTS), public coverage among low-income children grew during this period, but private insurance enrollment dropped by a nearly equal amount, leaving uninsurance rates essentially unchanged.

This Issue Brief examines the shifts in both public and private coverage, explores factors behind declining employer-sponsored insurance among low-income families, and considers policy questions about whether public expansions are substituting for private coverage or serving as a necessary safety net for families losing access to employer plans.

Policy Changes Affecting Children's Coverage

Several policy developments during this era were expected to reshape how low-income children obtained health coverage. The passage of SCHIP in 1997 stood as the most prominent initiative, designed to reach children in families earning above the poverty line who had previously been left out of Medicaid. At the same time, welfare reform at both the state and federal level was believed to have reduced the share of poor children enrolled in Medicaid. And these policy shifts were playing out against a backdrop of rising private insurance premiums, which put additional pressure on coverage rates for low-income families.

The CTS Household Surveys from 1996-1997 and 1998-1999 captured the most current estimates of children's insurance status during this period and made it possible to examine coverage trends during a time of major change. Between the two survey rounds, 33 states had begun implementing SCHIP, while welfare reform continued to alter the landscape for public insurance enrollment.

Documenting the Coverage Shifts

The household survey data revealed a clear pattern among children in families earning below 200 percent of the federal poverty level: Medicaid and other state-sponsored coverage rose from 29 percent to 33 percent, while private insurance fell from 47 percent to 42 percent. The net result was no statistically meaningful change in the overall uninsurance rate for this population. Most of the private coverage loss came from employer-sponsored plans.

Children in families earning 200 percent of poverty or above experienced virtually no change in their coverage patterns during the same period. Their private insurance rates remained at 89 percent, and employer-sponsored coverage stayed at roughly 80 to 81 percent.

The gains in public coverage and losses in private coverage were concentrated among families earning between 100 and 199 percent of the poverty line -- precisely the group that SCHIP and recent Medicaid expansions were designed to help. For children below the poverty line, public coverage held steady, and declines in private insurance were smaller.

Private insurance coverage also eroded for the parents of low-income children during this period. Because adults had fewer options for public coverage than their children did, parental uninsurance rates climbed from 31 percent to 35 percent between the two survey periods. Notably, this erosion of private coverage was limited to parents with children; low-income adults without children saw no comparable change.

Declining Offer and Take-Up Rates

The drop in employer-sponsored coverage among low-income children resulted from two compounding factors: fewer children had a parent who was offered workplace insurance, and among those who did have access, fewer parents were enrolling their children in those plans.

The share of low-income children with access to employer coverage through a parent edged down from 48 percent to 46 percent across the two surveys. This decline was not caused by fewer parents working; rather, fewer employed parents were being offered health benefits by their employers.

Among low-income children whose parents were offered and eligible for employer-sponsored insurance, enrollment fell from 72 percent to 66 percent. Meanwhile, the proportion of these children with access to employer coverage who were instead enrolled in Medicaid or other state programs grew from 10 percent to 14 percent. Approximately 11 percent of children with access to employer-sponsored insurance remained uninsured, a figure that did not change between survey periods.

Further analysis showed that about two-thirds of the overall decline in employer-sponsored coverage among low-income children was attributable to lower take-up rates when coverage was available, while the remaining third reflected reduced access to employer plans through a parent.

What Drove the Changes in Offer and Take-Up Rates

Several factors may have contributed to the decline in both the availability of and enrollment in employer-based coverage among low-income families. While the share of low-income children with employed parents held steady, there was a notable shift in the types of firms employing those parents. The percentage of low-income children whose parents worked in the public sector or at companies with 100 or more employees dropped from 62 percent to 58 percent, while employment at firms with 10 to 50 workers rose from 14 percent to 18 percent. Because smaller employers are much less likely to offer health benefits, this shift partly explains the reduced offer rates -- though it did not account for the full decline.

Welfare reform and a strong economy may have pushed some families' income above thresholds that changed their eligibility for public programs and their relationship to private coverage. However, changes in demographic characteristics such as race and age between the two surveys were small and could not explain the coverage shifts. The sizable changes in both private and public coverage would have occurred even if these sociodemographic characteristics had remained constant.

Rising health insurance costs appear to be a more significant factor. Employer-sponsored premiums rose 3.3 percent in 1998 and 4.8 percent in 1999. Increases were steeper for small firms (5.2 percent and 6.9 percent, respectively). The employee share of premiums for family coverage climbed from an average of $122 per month in 1996 to $145 in 1999, while the employee share for single coverage actually fell slightly. This difference may help explain why private coverage declined for low-income families with children but not for low-income adults without children, who are more likely to carry single coverage.

The Substitution Question

A central concern was whether the expansion of public coverage through SCHIP and Medicaid was directly displacing private insurance -- a phenomenon known as crowd-out or substitution. If public programs were drawing families away from employer coverage, the cost per newly insured child would be considerably higher than anticipated.

Substitution could manifest in several ways: employers might drop coverage or increase cost-sharing in response to public alternatives; families might voluntarily leave employer plans for free or subsidized public coverage; or workers might choose higher-paying jobs without health benefits, banking on public coverage for their children.

Analysis of CTS data, however, found that among low-income children enrolled in Medicaid or other state coverage, only about 2 percent had transitioned directly from private insurance at any point during the preceding year. This rate did not change between the two survey rounds. In both periods, approximately 85 percent of children on public coverage had been enrolled continuously for the prior year, and another 13 percent had been uninsured immediately before enrolling.

That said, the transition from private to public coverage may unfold over a longer timeframe -- families might experience a spell of uninsurance between dropping employer coverage and enrolling in a public program. SCHIP legislation also required states to institute waiting periods to discourage substitution. Given that most states had been implementing SCHIP for less than a year at the time of the second survey, any longer-term substitution effects might not yet be visible in the data.

Policy Implications

The period between the two surveys yielded no net improvement in coverage for low-income children, and their parents actually lost ground -- with uninsurance climbing from 31 to 35 percent. These findings, however, captured only the earliest stages of SCHIP rollout. As enrollment grew, the number of uninsured children might still decline.

The critical policy question was whether private coverage losses reflected substitution driven by public expansions, or whether they occurred independently due to market forces like rising premiums. If substitution was the primary driver, it would mean public dollars were largely shifting children from one form of coverage to another rather than reducing the ranks of the uninsured. But if private coverage was eroding on its own -- driven by premium increases, changes in employer behavior, and the financial burden on low-income families -- then public expansions were serving their intended role as a safety net, preventing what would otherwise have been a sharp rise in uninsurance.

Other HSC research demonstrated that low-income workers were five to ten times more likely than higher-income workers to decline employer coverage, and that lower-wage employees typically faced steeper premium costs relative to their income. As health insurance costs continued to climb -- a trend widely predicted at the time -- the financial strain on low-income families would intensify, making it increasingly difficult to distinguish between crowd-out effects and the natural erosion of private coverage under market pressure.

The findings underscored that simply expanding eligibility for public programs would not automatically reduce the number of uninsured children. The interplay between public and private coverage, employer decisions about benefits, and the affordability of family insurance plans all shaped outcomes in ways that made incremental coverage expansions both necessary and complicated.

Data Sources

This Issue Brief drew on findings from two rounds of the CTS Household Survey, a nationally representative telephone survey of the civilian, noninstitutionalized population. In-person interviews supplemented the phone survey to capture households without telephone service. The 1996-1997 round included nearly 33,000 families and 60,000 individuals; the 1998-1999 round covered about 32,000 families and 59,000 individuals, with an overall family response rate of 65 percent. Each round included data on approximately 10,000 children under age 18.

Health insurance status was measured as of the interview date and classified into mutually exclusive categories: employer-sponsored coverage, other private insurance, Medicaid and other state programs (including SCHIP), other public coverage (Medicare, military coverage, Indian Health Service), and uninsured.

Sources and Further Reading

Kaiser Family Foundation and Health Research and Educational Trust. Employer Health Benefits: 1999 Annual Survey, Menlo Park, CA, and Chicago, IL, 1999.

Ginsburg, Paul, "Tracking Health Care Costs: Long-Predicted Upturn Appears," Issue Brief No. 23, November 1999. Center for Studying Health System Change, Washington, DC.

Cunningham, Peter, et al., "Who Declines Employer-Sponsored Health Insurance and Is Uninsured?" Issue Brief No. 22, October 1999. Center for Studying Health System Change, Washington, DC.

This Issue Brief is adapted from "Tracking Recent Changes in Health Coverage for Low-Income Children with the Community Tracking Study, 1996-1997 and 1998-1999" by Peter J. Cunningham and Michael H. Park, Research Report No. 4, April 2000.