Association of the New York State Marriage Equality Act With Changes in Health Insurance Coverage

This study used data from the 2008-2012 American Community Survey, a nationally representative mail survey conducted annually by the US Census Bureau, with an annual sample size of approximately 3 million housing units and a 97% response rate.⁴ Follow-ups for nonresponse are conducted by telephone and in-person interviews. This survey does not ascertain sexual orientation; same-sex couples were identified when the primary respondent identified another person of the same sex as a husband, wife, or unmarried partner.⁵

The primary respondent reports current health insurance status for each household member. If multiple sources of coverage were reported, the primary source of insurance was assigned in the following order: (1) Medicare; (2) ESI through a current or former employer, TRICARE or other military health care; (3) Medicaid; (4) insurance purchased directly from an insurance company; and (5) uninsured.

Two-tailed t tests and a difference-in-differences approach were used to compare changes in health insurance coverage between adults in same-sex relationships and opposite-sex relationships. Linear probability models adjusted for factors associated with coverage, including age, race/ethnicity, citizenship, disability status, the presence of a child in the household, and year fixed effects. The coefficient of interest was the 2012 × same-sex relationship interaction term, reflecting the net change in coverage for adults in same-sex relationships in 2012 vs 2008-2010, minus the underlying trend among adults in opposite-sex relationships.

All data from 2011 were excluded as a washout period; 2013 was not included because the US Supreme Court ruled portions of the federal Defense of Marriage Act unconstitutional, making it easier for LGBT workers to add a same-sex partner to ESI. All samples were restricted to New York adults aged 26 to 64 years with a partner aged 26 to 64 years. Regression models were estimated separately for men and women using Stata version 12 (StataCorp) and sampling weights. Two-sided P < .05 was considered statistically significant. This study was deemed exempt by the University of Minnesota institutional review board because data were obtained from publicly available sources.

The 2848 adults in same-sex relationships were less likely to have children and reported higher educational attainment compared with the 228 470 adults in opposite-sex relationships (20% vs 55% had children, respectively; 63% vs 39% had college degrees). Both groups had parallel trends in ESI coverage until the implementation of same-sex marriage; ESI coverage increased significantly among adults in same-sex relationships in 2012 (Figure).

Figure. Percentage Covered by Employer-Sponsored Health Insurance Before and After New York State’s Marriage Equality Act Data are from an analysis of the 2008-2012 American Community Survey. aThe reported percentages reflect survey weighting and have not been adjusted for covariates. bEmployer-sponsored insurance coverage was significantly different between adults in same-sex relationships and adults in opposite-sex relationships in 2012 at the .05 level, 2-tailed t test.

The implementation of New York’s Marriage Equality Act was associated with substantial increases in ESI and smaller reductions in state-funded Medicaid assistance for men and women in same-sex relationships. Limitations include the observational design, short follow-up, missing health information, unknown generalizability to other states, and possible inaccuracy or response bias in self-reporting a same-sex partner (particularly after same-sex marriage was legalized); however, the American Community Survey remains the predominant source for measuring same-sex households⁵ and health insurance in the United States.

The US Supreme Court will soon determine whether states are required to recognize same-sex marriages. Based on New York’s experience, a favorable decision may extend access to ESI to LGBT couples.