Evaluation of Three States’ Reforms to Expand the Availability of Children’s Health Insurance Coverage

Evaluation of Three States’ Reforms to Expand the Availability of Children’s Health Insurance Coverage

• 1.
  Evaluation of ThreeStates’ Reforms to Expand the Availability of Children’s Health Insurance Coverage Carole Roan Gresenz (RAND) Sarah Edgington (UCLA) Jos é Escarce (UCLA and RAND) November 2009 Preliminary Results—Not for Citation Funded by SHARE (State Health Access Reform Evaluation), an initiative of the Robert Wood Johnson Foundation
• 2.
  IL, PA andWA Are Among States that Have Led the Way in Expanding Access to Public Health Insurance for Children In 2009, 40 states and DC cover children with family incomes >=200% FPL, compared to only 3 states in 1997. State (Year) Eligible with no premium contribution Eligible with sliding scale premium contribution Eligible with full cost premium contribution Key Restrictions Illinois (2006) <150% FPL 150-400% FPL >400% FPL Pennsylvania (2007) <200% FPL 200-300% FPL >300% FPL Citizens, legal residents & refugees Washington (2007) <200% FPL 200-250% FPL (to be implemented)
• 3.
  Project Uses Mixed-Methodsto Study 3 States’ Expansions; Presentation Focuses on Quantitative Analyses Qualitative and quantitative approaches to address study goals: What are states’ approaches to ensuring affordability, maximizing take-up, and minimizing crowd-out? What are threats to the sustainability of the reforms? What are reforms’ effects on coverage rates, take-up of public coverage and crowd-out? What are potential effects on out of pocket costs for uninsured children in newly eligible families?
• 4.
  Expansions of PublicCoverage to Higher Income Children Have Raised Concerns about “Crowd Out” Crowd out is extent to which expansions of eligibility for public insurance might result in the transfer of privately insured children to public insurance Central to 2007 debate over SCHIP reauthorization In response to concerns about crowd out, states have established variety of policies Waiting periods Employer premium support programs Premium contributions for public insurance
• 5.
  Mixed Findings AboutCrowd Out in Previous Research; Many Analytic Issues Magnitude of crowd-out has been difficult to pin down, as estimates vary depending on the data and methods used Likely to differ across states and target populations, over time, and for specific expansions. Literature suggests crowd-out rates are lower for low-income than for higher-income children But studies based on limited income range because extensions of eligibility to higher income children are recent.
• 6.
  Approach Data andsample: 2002-2009 CPS; children 0-17 in all states in families with income <400% FPL Dependent variables: Child has public insurance Child has private insurance Child has no insurance Key independent variable of interest Eligibility for public insurance Controls: Individual characteristics State characteristics Program characteristics State fixed effects Year fixed effects
• 7.
  Estimation Separate regressionfor each dependent variable Instrument for eligibility Eligibility measured with error and some argue potentially endogenous Instrument using percentage of children (national draw) in particular age groups eligible for public insurance in a given year and state (given eligibility rules in that state/year) State characteristics control for state specific factors that affect insurance outcomes as well as potential “policy endogeneity” Lagged and current unemployment; per capita income; percentage of children in poverty; percent of workers unionized; price of private insurance
• 8.
  Estimation Comprehensive setof individual-level control variables Child age, gender, race Family structure (number of children, number of parents in household) Highest education of either parent, family income Nativity of child and parent MSA/nonMSA Parental employment status (FT, PT, public sector, self employed, firm size, occupation)
• 9.
  Estimation Issues Differenteffects for non-group privately insured vs privately insured through group? CPS insurance questions: measure last year or as of interview? Classifying individuals who report public and private Use data as reported; no recoding of public+private to either private or public Measuring price of private insurance Various approaches: individual level, state level, derived as probability of offer*price as offer
• 10.
  Future Analyses Effectof public premium contribution Interactions of eligibility with family income Exploring effects among immigrants by legal status Analysis of Washington state survey data (WSPS)
• 11.
  Preliminary Findings—Effect ofEligibility for Public Insurance on Insurance Status Gold font indicates statistically significant, p<.01 PRELIMINARY RESULTS √ -0.098 -0.082 0.212 Spec 4 √ -0.097 -0.080 0.210 Spec 3 √ √ -0.100 -0.081 0.214 Spec 2 √ √ -0.098 -0.079 0.210 Spec 1 State Unempl. Rate (t-1, t-2) State Unempl. Rate (t, t-1) Private premium (state level) No Ins. Private Public
• 12.
  Preliminary Findings—Effect ofEligibility for Public Insurance on Insurance Status Gold font p<.01; Blue font p<.05, Gray font not significant PRELIMINARY RESULTS -.051 -.053 -.052 -.053 Non Group Private √ -0.098 -.031 0.212 Spec 4 √ -0.097 -.027 0.210 Spec 3 √ √ -0.100 -.030 0.214 Spec 2 √ √ -0.098 -.027 0.210 Spec 1 State Unempl. Rate (t-1, t-2) State Unempl. Rate (t, t-1) Private premium (state level) No Ins. Group Private Public
• 13.
  Conclusions These statesare likely to be bellwethers for the nation, as state-level reforms to insure all children are growing in popularity due to their political appeal and relatively low cost. Evidence suggests important effects of expanded eligibility on take-up (increases) and uninsurance (decreases) Evidence further suggests that substitution away from private coverage to public coverage occurs primarily among those enrolled in non-group private coverage (i.e., individual market)
• 14.