Panel Paper: Marriage Equality Legislation Had a Differential Effect on Health Insurance in ACA Expansion and Nonexpansion States

Friday, November 9, 2018
8216 - Lobby Level (Marriott Wardman Park)

*Names in bold indicate Presenter

Jae Downing, University of South Carolina and Paulette Cha, University of California, Los Angeles

Same-sex marriage has become legal in all 50 states since the Supreme Court decision in June of 2015. Nearly two-thirds of same-sex, cohabiting couples were married in June of 2017, compared to a third of same-sex couples before the Supreme Court decision. Prior to marriage equality, same-sex couples were less likely to have any health insurance and access to ESI than different-sex couples. Gonzales et al found that same-sex couples were less likely to have ESI compared to married different-sex couples, yet the gap was smaller in states that had legalized same-sex marriage. In California, Ponce at al find that among partnered and married people, lesbians and gay men were less likely than heterosexuals to have dependent employer-sponsored health insurance.

Private and public health insurance rates for lesbian, gay, and bisexual adults increased from summer 2013 to winter 2014-15. During that period, nearly half of states passed marriage equality and concurrently, many states expanded Medicaid. Our study tests whether the effect of marriage equality on health insurance coverage had differential effects across Medicaid expansion and nonexpansion states.

We use the American Community Survey, a large, nationally representative dataset with demographic and insurance variables that allows for state-level analysis, making it ideal for this study. Our main study population includes adults aged 26 to 64 since this group is age-ineligible for Medicare and the parental private health insurance expansion under the ACA. Our study will include approximately 14 million observations over 2008-2016. We retain individuals in the analysis sample regardless of relationship status or sexual orientation, and our analyses estimate intent-to-treat effects.

We use a difference-in-differences approach to estimate the effect of state marriage equality legislation on health insurance status. Our outcomes are uninsurance, public insurance, and private insurance. We also use a triple-differences model to estimate separate effects for ACA Medicaid expansion states and non-expansion states. We operationalize our analyses as linear probability models, and conduct inference using heteroskedasticity-robust standard errors clustered by state. All analyses control for individual demographics and state-year economic conditions, and include full sets of state and year fixed effects.

Our preliminary results show a 1.76 percentage point decrease in uninsurance nationwide as result of marriage equality legislation, mostly accounted for by a 1.66 percentage point increase in public health insurance coverage. We also find a decrease in uninsurance in states that did not participate in the ACA Medicaid expansion, but there the decrease was attributable to an increase in private health insurance coverage. Results for men are similar to the full sample, yet women also experienced an increase in public health insurance in expansion states.

These findings suggest that the marriage equality legislation led to decreases in nationwide uninsurance rates. Legal recognition of their unions through marriage allows members of same-sex married couples to access dependent health insurance coverage through employers, ACA marketplaces, or other private insurance. Recognizing same-sex marital status can also alter eligibility for public health insurance, notably Medicaid, through changes in family size and income. Next steps include investigation of the mechanisms at play.