Association for Public Policy Analysis & Management

Research Objective: To examine the impact of prior state expansions that increased parental eligibility for public programs with and without premiums, or provided premium subsidies for private insurance, on parent and child insurance status. 

Study Design: We use variation in states’ expansions over time (1999-2012) to estimate their effects on the probability that: 1) parents are insured; 2) both parent and child within a family unit are insured, 3) type of coverage (public/private), and 4) child insured. We create parent/child ‘dyads’ by selecting a random child for each parent in the Current Population Survey (CPS) 2000-2013. We analyze the effects of parental expansions of ≥25 percentage points (pp) of Federal Poverty Level (FPL) using within-state difference-in-difference models with higher income dyads as controls. All analytic models control for child and parent demographic factors, parental employment characteristics, and local economic conditions. We estimate cross-state models to examine the impact of expansions by program type including: 1) Public Programs (with- and without premium); 2) Premium Assistance programs; and 3) Premium Assistance-like programs that require enrollment in a state-offered health insurance or managed care plans. We use logistic regression models to estimate the probability of parent’s insurance and multinomial models to examine parent/child insurance with parents in non-expansion states as controls. All analyses incorporate state and year fixed effects and cluster standard errors by state/year. 

Population: Parents (n=54,193 within-state models; 285,498 across–state models) of children under age 19 and parent/child dyads (n=45,995 within-state models; 267,221 across–state models) with incomes below 300% FPL. 

Findings: Based on within-state models, parent eligibility for any expansion increased their coverage by 3.1 pp, child coverage by 1.8 pp, and reduced the likelihood of both being uninsured by 1.8 pp. Based on cross-state models, eligibility for public programs without premiums had a significant impact on the probability of parent coverage (4.0 pp), and child coverage (1.8 pp), and reduced the probability of both being uninsured by 1.8 pp. Premium assistance-like expansions also had a significant impact on parent coverage, increasing coverage by 4.2 pp. Among insured dyads, eligibility for public programs without a premium increased public coverage for both (3.1 pp) and decreased mixed coverage (2.2 pp), whereas premium assistance-like expansions only increased public coverage for both (1.3 pp).  

Policy Implications: Expansions of public coverage without premiums for parents are successful at not only insuring parents, but also increasing the likelihood their children are insured. States expanding Medicaid for newly eligible adults under the ACA may see greater increases in insurance coverage for children than non-expanding states, even with similar CHIP programs. For states with 1115 waivers to purchase premium assistance-like coverage for low-income residents through exchanges, results suggest these programs will increase parental coverage as effectively as public programs without a premium. Given that subsidies under ACA are individual rather than family based, state designed buy-in options could moderate the bias against family coverage inherent in current IRS regulations regarding ACA subsidies – the so called ‘family glitch’.