Duke 2008 | ASHE | Equity and Efficiency in Health and Healthcare

A growing literature has shown that SCHIP significantly improved public coverage for low-income children (LoSasso and Buchmueller 2004;Cunningham et al. 2002; Hudson, Selden et al. 2005). Most studies computed an overall average effect. This study examines heterogeneous effects of SCHIP on health insurance including spillover on Medicaid enrollment, over time, by program type (Medicaid expansion, separate, or combination program), and by children's age.

We use data from the 1995-2002 National Health Interview Survey (NHIS) linked to state identifiers. The NHIS collects a variety of health-related information from a nationally-representative sample of the U.S. civilian non-institutionalized population annually. We restrict the sample to children whose family income is below their state SCHIP eligibility limits. These children are most likely to be affected by SCHIP implementation. The analysis sample consists of 108,253 children age 0-18 years old. We estimate health insurance coverage including any public, Medicaid, private, and uninsurance as a function of SCHIP availability in linear probability models, controlling for state and time fixed-effects and child and family characteristics. SCHIP availability is a binary variable indicating whether SCHIP had been implemented in a child's state of residence at the time of interview. To estimate SCHIP's evolving effect over time, we create 21 quarter dummies, each representing the number of quarters since SCHIP implementation. To estimate the spillover effect of SCHIP on Medicaid, we further restrict the sample to those ages 0-5 and in families with income below the Federal Poverty Level. This sub-population is only eligible for Medicaid in any state. We estimate SCHIP's differential effects by children's age and program type by adding interactions of program availability with indicators of age group and program type to the baseline model.

Results show that the effects of SCHIP increased steadily over time. SCHIP increased public coverage from 2.6 percentage points in the second quarter to 10.2 percentage points by the sixteenth quarter. After SCHIP had been available for 8 quarters the probability of being uninsured was 7 percentage points lower than pre-SCHIP period. Most of the changes in private coverage over time cannot be precisely estimated. The average spillover effect of SCHIP on Medicaid enrollment among the youngest and poorest children is about 4 percentage points.

States with Medicaid expansion programs improved children's insurance status the most. SCHIP implementation significantly decreased the uninsured rate (?0.039) for low-income children whose states chose to have Medicaid expansion SCHIP programs, but the effect is virtually zero for those whose states implemented separate SCHIP programs only. For children in states with combination programs, the reduction in the uninsured rate was about half less (?0.021). Older children had greater gains in health insurance than younger children. Under SCHIP older low-income children (age 6+) were more likely to gain public insurance by 2.7 percentage points and less likely to be uninsured by 1.8 percentage points compared to the younger ones.

Our study provides a unique and more informed set of national estimates of effects of the SCHIP program on low-income children's health insurance.