Eligibility Under State Children's Health Insurance Programs

Abstract

Objectives. This study analyzed associations between income eligibility criteria under the State Children's Health Insurance Program (SCHIP) and state characteristics.

Methods. We used multivariate methods to explore relations between eligibility expansions under SCHIP and percentages of uninsured children from low-income families, per capita income, and political characteristics.

Results. Proportions of uninsured children, per capita income, and states' preexisting eligibility thresholds were statistically associated with changes in eligibility thresholds, whereas only per capita income was associated with overall SCHIP eligibility thresholds. Political dynamics were not statistically related to SCHIP expansions.

Conclusions. State demographic characteristics were associated with changes in eligibility from preexisting levels but rarely were associated with SCHIP eligibility thresholds.

In August 1997, Congress enacted the State Children's Health Insurance Program (SCHIP) to expand health insurance coverage for children of low-income families. As of September 1999, all states received federal approval to participate in the program. In this report, we examine factors associated with state decisions to expand health insurance coverage for children under SCHIP. Although extensive literature exists on state eligibility decisions in cash assistance programs, this is among the first reports to assess factors associated with state eligibility decisions under SCHIP. Previous literature showed that higher federal matching rates,1 greater fiscal capacity,2 and higher degrees of democratic control3 are correlated with more generous eligibility in state social programs.4 We examine here the extent to which income eligibility expansions under SCHIP are associated with states' percentages of low-income uninsured children, wealth, preexisting income eligibility levels, changes in financial matching requirements (relative to Medicaid), and party affiliations of state governors and state legislatures. In addition to examining the change in income eligibility thresholds—pre-SCHIP and post-SCHIP—we examine associations between state characteristics and states' eligibility thresholds as of June 2000.

METHODS

This study assessed the relation between state characteristics and state income eligibility expansions under SCHIP in a multiple regression framework. The data contain observations from each state and the District of Columbia. The primary variable of interest—changes in income eligibility thresholds under SCHIP relative to preexisting eligibility thresholds—was compiled from an analysis of SCHIP plans and source data from the National Conference of State Legislatures and the National Governors' Association.5 We derived changes in state income eligibility thresholds by subtracting the pre-SCHIP threshold from the post-SCHIP threshold; income eligibility thresholds are presented as an average for all age groups.

We generated income eligibility thresholds by state, pre- and post-SCHIP, by assigning an income eligibility threshold per age of child (aged 0 to 19 years), summing the income eligibility thresholds, and then dividing by 19. For this analysis, the pre-SCHIP thresholds relied on levels in effect as of June 1997, whereas post-SCHIP eligibility thresholds included those that had been enacted as of June 2000. The dependent variable for the primary analysis was a state's change in income eligibility threshold.

The explanatory variables were taken primarily from government sources. State percentages of low-income uninsured children were based on 3-year averages of the March 1996, 1997, and 1998 Current Population Surveys, conducted by the US Bureau of the Census. Income (per capita personal income) for 1998 is presented as compiled by the US Department of Commerce.6 The change in matching rate variable is the percentage point increase in a state's federal matching rate under SCHIP, relative to the state's federal matching contribution under the state's Medicaid program. Matching rates for Medicaid and SCHIP are published by the US Department of Health and Human Services.7

Finally, political variables were included. Governors' affiliations in 1998 were coded as dummy variables: a “1” represents a Democratic (or Independent) governor, and a “0” represents a Republican governor. We calculated the percentage of Democratic state legislatures by adding the percentage of a state house that was Democratic and the percentage of the state Senate that was Democratic in 1998 and dividing by 2.

RESULTS

States responded to SCHIP by substantially raising income eligibility thresholds for children; between June 1997 and June 2000, the average state raised its eligibility threshold by 85 percentage points, from 121% of the poverty level under Medicaid to 206% under SCHIP.

Table 1 ▶ presents an analysis of how state characteristics are related to these changes in income eligibility levels. The analysis found that 3 variables—percentage of low-income uninsured children, per capita personal income, and income eligibility thresholds before SCHIP—were statistically significant, at the 90% level, in their relation to changes in state income eligibility thresholds. States with higher proportions of low-income uninsured children had greater increases in income eligibility thresholds. Consistent with the preexisting literature, states with higher per capita income had greater increases in income eligibility thresholds—the analysis suggested that a $1000 increase in a state's per capita personal income was associated with a 3.75 percentage point increase in the SCHIP eligibility threshold. Conversely, states with higher income eligibility thresholds before SCHIP had smaller increases in income eligibility thresholds post-SCHIP, with other factors being equal. For example, states with income eligibility thresholds in the lowest quintile before SCHIP were projected to increase their eligibility thresholds by 94 percentage points, whereas states in the highest quintile were projected to raise their income eligibility thresholds by 63 percentage points.

TABLE 1—

Predictors of Changes in Eligibility Thresholds

Regression statistics
R2	0.35
Adjusted R2	0.26
Observations	51

	Mean	Coefficients	P	95% CI
Note. CI = confidence interval; SCHIP = State Children's Health Insurance Program.
Intercept	. . .	–41.62	.35	–129.51, 46.26
Low-income uninsured children, %	21.91	1.60	.05	–0.01, 3.22
Per capita income, $ thousands	25.36	3.75	.05	–0.02, 7.53
Income eligibility threshold pre-SCHIP (percentage of federal poverty level)	121.27	–0.35	.02	–0.62, –0.07
Change in federal matching rate, %	11.8	1.83	.57	–4.65, 8.32
Governors' affiliations in 1998	0.37	–1.09	.92	–22.25, 20.06
Democrats in state legislature, %	0.52	31.8	.34	–34.87, 98.47

The 3 remaining variables in the analysis—change in federal matching rate and the political affiliations of states' governors and legislatures—were not found to be significantly related to changes in states' income eligibility thresholds.

The results of an analysis of how state characteristics are related to the income eligibility levels under SCHIP are shown in Table 2 ▶. In this analysis, just 1 variable—per capita personal income—was significantly related to average income eligibility thresholds under SCHIP. State percentages of low-income uninsured before SCHIP and political dynamics were not statistically significant in this analysis. (For an analysis that focuses on the relation between ideologic or political factors and levels of eligibility in SCHIP, see Beamer.8)

TABLE 2—

Predictors of SCHIP Income Eligibility Thresholds

Regression statistics
R2	0.27
Adjusted R2	0.21
Observations	51

	Mean	Coefficients	P	95% CI
Note. SCHIP = State Children's Health Insurance Program; CI = confidence interval.
Intercept	. . .	42.01	.40	–58.28, 142.30
Low-income uninsured children, %	21.91	0.16	.87	–1.71, 2.02
Per capita income, $ thousands	25.36	5.67	0	2.58, 8.75
Governors' affiliations in 1998	0.37	1.59	.90	–24.25, 27.42
Democrats in state legislature, %	0.52	30.9	.43	–47.81, 109.62

One limitation of these analyses was that some of the variables analyzed were highly correlated. For example, the change in federal matching rate is, by definition, driven by state per capita income. By including both of these variables in our analysis, the statistical significance of each is weakened. Furthermore, we were limited by the small number of observations—a total of 51—representing each state and the District of Columbia.

DISCUSSION

States have expanded eligibility for children substantially in response to SCHIP, from 121% to 206% of the poverty level. Under SCHIP, most states' eligibility thresholds hover around the median, 200% of poverty. As a result, eligibility levels across states under SCHIP are more similar than they were in states' preexisting Medicaid programs. However, the range of eligibility levels under SCHIP is greater because some states, such as New Jersey at 350% of poverty, have significantly expanded coverage beyond the thresholds of other states. The results of this study reinforce descriptive findings that higher proportions of low-income uninsured children, lower pre-SCHIP eligibility levels, and higher state per capita personal income are associated with larger increases in income eligibility thresholds under SCHIP.9 The finding that states with lower pre-SCHIP eligibility thresholds expanded coverage the most suggests that states may be coming closer together in their coverage policies for children—that states with previously low levels of coverage may be “catching up” to those with historically more generous programs. The effects of these expansions on the actual numbers of children who will become eligible for SCHIP and subsequently enroll in the programs are critical areas for further research.