Saving Teens: Using a Policy Discontinuity to Estimate the Effects of Medicaid Eligibility Laura Wherry 1 Bruce Meyer 2 1 UCLA David Geffen School of Medicine 2 University of Chicago Harris School of Public Policy AcademyHealth ARM June 26, 2017
Introduction Policy Discontinuity Data and Methods Results Conclusions Introduction U.S. policy has largely focused on expanding public health insurance to address disparities in child health Strong evidence that public insurance increases use of medical care, but evidence of impact on health is more limited See comprehensive review in Howell and Kenney (2012) Only recently have we been able to evaluate the long-term effects of public health insurance coverage for children 2 / 16
Introduction Policy Discontinuity Data and Methods Results Conclusions Public Health Insurance May Affect Long-Term Health Payoffs from certain types of medical care might not be evident until later e.g. preventive services protect healthy children from future risks In addition, linkages to other social services or freeing up of resources for other investments may have long-term impacts Long-term effects are an important potential program benefit 3 / 16
Introduction Policy Discontinuity Data and Methods Results Conclusions Overview of Study We use a quasi-experimental design that approximates random assignment to identify the effect of public health insurance on health We examine immediate and longer-term effects of childhood Medicaid on mortality for black and white children Findings indicate increase in public coverage decreased longer-term mortality for black children Just one study among many.... 4 / 16
Introduction Policy Discontinuity Data and Methods Results Conclusions Medicaid Expansions for Children Medicaid eligibility for children historically linked to cash welfare - primarily single-mother families with very low incomes In 1984, Congress began to expand Medicaid eligibility for all children with family incomes below poverty To phase in the expansions, Congress specified that only applied to children born after September 30, 1983 5 / 16
Average Years of Childhood Public Eligibility by Birth Cohort and Family Income Average Years of Childhood Eligibility 0 2 4 6 8 10 12 14 16 18 Oct−79 Jan−80 Apr−80 Jul−80 Oct−80 125−150% FPL 100−124% FPL 75−99% FPL 50−74% FPL 25−49% FPL 0−24% FPL Jan−81 Apr−81 Jul−81 Oct−81 Jan−82 Apr−82 Jul−82 Oct−82 0.43 years 4.57 years 3.44 years 2.01 years 0.22 years Jan−83 Apr−83 Birth Cohort Jul−83 Oct−83 Size of discontinuity = 0.19 years of eligibility Jan−84 Apr−84 Jul−84 Oct−84 Jan−85 Apr−85 Jul−85 Oct−85 Jan−86 Apr−86 Jul−86 Oct−86 Jan−87 Apr−87 Jul−87
Average Public Eligibility for Each Age of Childhood by Birth Month Cohort 1.0 Share of Birth Cohort Eligible for Public Health Insurance Sept. 1983 Oct. 1983 0.8 0.6 0.4 0.2 Ages 4 − 7 Ages 8 − 14 Ages 15 − 18 0.0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Age in Years
Introduction Policy Discontinuity Data and Methods Results Conclusions Differences in Exposure by Child Race We examine effects of expansions separately by child race Black children more likely to be poor than white children (45.6% vs. 16.1% in 1991) Percent Average Gain (in Years) Average Gain (in Years) Gaining for Children Gaining for Total Child Eligibility Eligibility Population Black children 17.13% 4.53 0.82 White children 8.18% 4.48 0.37 8 / 16
Introduction Policy Discontinuity Data and Methods Results Conclusions Outcomes and Data Construct rates of death for birth cohorts: Oct. 1979 - Sept. 1987 Number of deaths divided by total population at risk (per 10,000) Data sources: Multiple Cause Mortality Files, 1979-2011 (restricted) Natality Files, 1979-1987 Distinguish between internal and external causes of death 9 / 16
Introduction Policy Discontinuity Data and Methods Results Conclusions Empirical Analysis Estimate effect of Medicaid expansions on mortality at ages 8-14 (immediate effect), and ages 15-18 and 19-23 (longer-term effects) Use a regression discontinuity approach to compare outcomes for cohorts born just before and after September 30, 1983 Graphs and regression analyses model trends in mortality unrelated to Medicaid expansions Quadratic function used in estimates presented here, alternative specifications reported in paper Estimated discontinuity at cutoff provides estimate of effect of Medicaid 10 / 16
Introduction Policy Discontinuity Data and Methods Results Conclusions Internal Mortality for Black Children (a) Ages 8-14 (b) Ages 15-18 (c) Ages 19-23 Black, Ages 8 − 14 Black, Ages 15 − 18 Black, Ages 19 − 23 2.0 ● ● ● ● ● ● 4.5 ● ● ● ● ● ● ● 1.8 3.0 ● ● ● ● ● ● ● ● ● Internal Mortality ● ● ● ● Internal Mortality ● ● ● ● ● ● ● ● ● ● 1.6 ● Internal Mortality 4.0 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 2.5 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 1.4 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 3.5 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 2.0 ● ● ● 1.2 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 3.0 ● ● ● ● ● 1.0 ● ● ● ● ● ● ● 1.5 ● ● ● ● ● ● 0.8 ● 2.5 Jun − 80 Feb − 82 Oct − 83 Jun − 85 Feb − 87 Jun − 80 Feb − 82 Oct − 83 Jun − 85 Feb − 87 Jun − 80 Feb − 82 Oct − 83 Jun − 85 Feb − 87 ˆ ˆ ˆ β = − 0.094 β = − 0.448 ∗ ∗∗ β = 0.052 ( 0.087 ) ( 0.125 ) ( 0.125 ) ∗∗∗ p < 0.01, ∗∗ p < 0.05, ∗ p < 0.1 11 / 16
Introduction Policy Discontinuity Data and Methods Results Conclusions Internal Mortality for White Children (a) Ages 8-14 (b) Ages 15-18 (c) Ages 19-23 White, Ages 8 − 14 White, Ages 15 − 18 White, Ages 19 − 23 1.3 1.6 1.2 ● ● ● ● ● ● ● ● ● 2.2 ● ● ● ● ● ● ● ● ● ● ● ● ● 1.1 ● ● ● ● ● ● ● Internal Mortality ● ● ● ● ● ● ● ● ● ● ● Internal Mortality ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● Internal Mortality ● ● ● ● ● ● ● ● ● 1.4 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 2.0 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 1.0 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 0.9 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 1.8 ● ● ● ● ● ● ● 1.2 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 0.8 ● ● ● ● ● ● ● ● ● ● ● ● ● 1.6 ● 0.7 1.0 ● 1.4 Jun − 80 Feb − 82 Oct − 83 Jun − 85 Feb − 87 Jun − 80 Feb − 82 Oct − 83 Jun − 85 Feb − 87 Jun − 80 Feb − 82 Oct − 83 Jun − 85 Feb − 87 ˆ ˆ ˆ β = 0.014 β = 0.022 β = − 0.007 ( 0.028 ) ( 0.046 ) ( 0.054 ) ∗∗∗ p < 0.01, ∗∗ p < 0.05, ∗ p < 0.1 12 / 16
Recommend
More recommend
