1. The Impact of Same-Sex Marriage on
Health Insurance Coverage:
Evidence from Four States
Gilbert Gonzales, MHA
University of Minnesota
Association for Public Policy Analysis & Management
Policy Research on Same-Sex Marriage
Saturday, November 9, 2013

2. Background
Click to edit Master title style
Same-sex couples styles
Click to edit Master textare less likely to have health
insurance, especially through employers
Second level
Third level
NHIS level (Heck et al. 2006)
Fourth
Fifth level
CPS
(Ash & Badget, 2006)
BRFSS (Buchmueller & Carpenter, 2010)
ACS
(Gonzales & Blewett, 2014)
2

3. Most to edit Master covered through a
Click Americans are title style
family member’s employer health plan
Click to edit Master text styles
 Second level
“Legal” spouse
 Dependent children
Third level
Fourth level
Fifth level
3

4. Employersedit Master title style benefits to
Click to often required to extend
same-sex spouses after same-sex marriage
Click to edit Master text styles
Second level
Third level
Fourth level
Fifth level
Source: National Conference of State Legislatures
4

5. Click to edit Master title style
Research Question
Click to edit Master text styles
Second level
What is the effect of same-sex marriage on
Third level
health insurance coverage among same-sex
Fourth level
couples? Fifth level
5

6. American edit Master title style
Click to Community Survey, 2008-2011
Healthedit Master text styles in 2008
Click to insurance added
• Employer-Sponsored Insurance (ESI)
Second level
• Directly Purchased / Individual
Third level
Fourth level
• Medicare
Fifth
• Medicaidlevel
• Uninsured
Large sample size!
•
•
•
3 million people each year
Supports state level research
Leading data resource for same-sex couples
6

7. Difference-in-Differences style Analysis
Click to edit Master title (DD)
Pre-Post research design
Click to edit Master text styles
• Target group: Same-sex couples in 4 states (CT, IA, NH, VT)
Second level
• Comparison groups unaffected by policy change
Third level opposite-sex couples in same state
1) married
Fourth level
2) same-sex couples in neighboring states
Fifth level
Linear probability models controlling for:
• Race/ethnicity, age, income, educational attainment,
employment status, industry, related child in household,
citizenship, state fixed effects
Separate models for:
• State with no provisions adopting marriage (IA)
• States replacing civil unions/domestic partnerships with
marriage (CT, NH, VT)
7

8. Difference-in-Differences style Analysis
Click to edit Master title (DD)
Click to edit Master text styles
Yi = β0 + β1Treat + β2Post + β3Treat*Post + Xi + ε
Second level
Third level
Fourth level
Fifth level
8

9. Difference-in-Differences (DD)
From No Provisions totitle style Marriage
Click to edit Master Same-Sex
Click to edit Master text styles
12.0%
Second level
10.0%
8.0%
6.0%
+ 9.3%
+ 8.2%
Third level
Fourth level
Fifth level
4.0%
2.0%
0.0%
Men
Women
Comparison Group 1:
Married Opposite-Sex Couples in Iowa
Adjusts for race/ethnicity, age, employment, industry, income, state, citizenship, minor child.
Source: American Community Survey, 2008-2011. * indicates p<0.05
9

10. Difference-in-Differences (DD)
From No Provisions totitle style Marriage
Click to edit Master Same-Sex
Click to edit Master text styles
12.0%
Second level
10.0%
8.0%
6.0%
+ 9.6%
+ 9.3%
+ 8.2%
Third level
+ 7.7%
Fourth level
Fifth level
4.0%
2.0%
0.0%
Men
Women
Comparison Group 1:
Married Opposite-Sex Couples in Iowa
Men
Women
Comparison Group 2:
Same-Sex Couples Neighboring Iowa
Adjusts for race/ethnicity, age, employment, industry, income, state, citizenship, minor child.
Source: American Community Survey, 2008-2011. * indicates p<0.05
10

11. From Civil Unions to Same-Sex Marriage
Click to edit Master title style
Difference-in-Differences (DD)
Click to edit Master text styles
Second level
8.0%
6.0%
4.0%
+ 7.6%
*
Third level
Fourth level
Fifth level
2.0%
0.1%
0.0%
Gay Men
Lesbian Women
-2.0%
Comparison Group 1:
Married Opposite-Sex Couples in
CT, NH, VT
Adjusts for race/ethnicity, age, employment, industry, income, state, citizenship, minor child.
Source: American Community Survey, 2008-2011. * indicates p<0.05
11

12. From Civil Unions to Same-Sex Marriage
Click to edit Master title style
Difference-in-Differences (DD)
Click to edit Master text styles
Second level
8.0%
6.0%
4.0%
+ 7.6%
*
*
+ 7.5%
Third level
Fourth level
Fifth level
2.0%
Men
-1.4%
0.1%
0.0%
Men
Women
Women
-2.0%
Comparison Group 1:
Married Opposite-Sex Couples in
CT, NH, VT
Comparison Group 2:
Same-Sex Couples Neighboring
CT, NH, VT
Adjusts for race/ethnicity, age, employment, industry, income, state, citizenship, minor child.
Source: American Community Survey, 2008-2011. * indicates p<0.05
12

13. Key Findings
Click to edit Master title style
Click to edit Master text styles
1) Same-sex marriage leads to ~7.5% increase in health
Second level
insurance coverage for women in same-sex
Third level
relationships
Fourth level
Fifth level
2) Limited evidence that same-sex marriage improved
health insurance coverage in Iowa or for men
3) Detectable gains in coverage by replacing civil union
laws with legal same-sex marriage
13

14. Click to Gilbert Gonzales, MHA
edit Master title style
Click to edit Master text styles
PhD Student
Second level Graduate Research Assistant
Third level
gonza440@umn.edu
Fourth level
Fifth level
University
of Minnesota
School of Public Health
Division of Health Policy & Management
www.shadac.org
@shadac

15. Discussion
Click to edit Master title style
Click to edit Master text styles
Second level
Third level
Fourth level
Fifth level
15

16. Click to edit Master title style
GLB Population in the ACS
• Same-sex spouses / unmarried
Click to edit Master text styles partners
Second level
Third level
Fourth level
Fifth level
16