Absence of conflict of interest.
Citation
Highlights
-
The study's objective was to examine the impact of state employment nondiscrimination acts on employment and earnings of cohabitating gay men and lesbian women.
-
The study used a difference-in-differences analysis to assess changes in employment and earnings for gay men and lesbian women associated with state employment nondiscrimination acts (ENDAs). Using data from the American Community Survey, the U.S. Census 5 percent sample, state laws, the Williams Institute, and Government Accountability Office, the author used statistical models to observe changes in earnings and employment related to ENDAs.
-
The study found ENDAs were associated with an increase in employment, annual income and hourly wages for gay men and a decrease in employment, hours worked, and annual income for lesbian women. ENDAs addressing punitive damages were associated with a decrease in annual income and hourly wages for gay men and with a decrease in hours worked for lesbian women. ENDAs addressing employer size were associated with an increase in employment for gay men.
-
This study receives a low evidence rating. This means we are not confident that the estimated effects are attributable to ENDAs; other factors are likely to have contributed.
Intervention Examined
State Employment Nondiscrimination Acts (ENDAs)
Features of the Intervention
Policies that prevent discrimination against groups of people have been associated with employment outcomes for these groups. However, there is also research suggesting that these nondiscrimination policies lead to increased costs for the employer (e.g., hiring and terminating these protected groups) and this increase in cost may lead to reduced hiring of people in these protected groups. ENDAs are passed at the state level, and they aim to protect gay men and lesbian women in the workforce. ENDAs differ in content by state (e.g., groups protected, process for resolving claims, and availability of damages for plaintiffs).
Features of the Study
The study used a difference-in-differences analysis to assess changes in employment and earnings for gay men and lesbian women associated with ENDAs. The study data sources included: American Community Survey, the U.S. Census 5 percent sample, state laws, the Williams Institute, and Government Accountability Office. The data used in this study were drawn from a period during which ENDAs increased in use (from three states in 1990 to 21 states in 2014). The ENDAs applied to everyone living in the United States between 1990-2014.
The study groups included in the analysis were determined by sexual orientation. People were identified as gay men or lesbian women if the householder and the partner of the householder had the same sex. People were identified as heterosexual if they were married to someone of the opposite sex. The sample included married heterosexual men (6,128,130) and women (5,016,785) and cohabiting gay men (44,143) and lesbian women (44,275) between the ages of 22-65. The author used statistical models to analyze changes in employment and wages in heterosexual people and gay men and lesbian women after ENDAs were passed, as compared to states in which ENDAs were not passed. The author also looked specifically at ENDAs that allowed for compensatory and punitive damages, recouping attorney fees, and ENDAs that addressed the minimum size of employers and the statute of limitations for filing complaints.
Findings
Earnings and wages
-
The study found that ENDAs were significantly associated with a 2.7 percent increase in hourly wages for gay men, but no significant association between ENDAs and hourly wages for lesbian women was found.
-
The study found that after controlling for the effects of different provisions of ENDAs, ENDAs were significantly associated with a 11.4 percent decrease in annual income for lesbian women. However, there were no significant associations for gay men.
-
The study found that ENDAs addressing compensatory damages were significantly associated with an 8.3 percent increase in annual income and a 12.7 percent increase in hourly wages for gay men, but the study found no significant associations for lesbian women.
-
The study found that ENDAs addressing punitive damages were significantly associated with a 4.4 percent decrease in annual income and an 8.1 percent decrease in hourly wages for gay men, but the study found no significant associations for lesbian women.
-
The study found no other significant associations between annual income or hourly wages of gay men and lesbian women and ENDAs or different provisions.
Employment
-
The study found that ENDAs were significantly associated with a decrease in employment and hours worked for lesbian women, but the study found no significant associations between ENDAs and employment or hours worked for gay men.
-
The study found that after controlling for the effects of different provisions of ENDAs, ENDAs were associated with a significant decrease in hours worked by lesbian women, but the study found no significant association for gay men.
-
The study found that ENDAs addressing employer size were associated with a significant increase in employment for gay men, but the study did not find a significant association for lesbian women.
-
The study found that ENDAs addressing punitive damages were associated with a significant decrease in hours worked by lesbian women, but the study found no significant association for gay men.
-
The study found no other significant associations between employment or hours worked of gay men and lesbian women and ENDAs or different provisions.
Considerations for Interpreting the Findings
Although the author included demographic control variables in the analysis, he does not provide information suggesting that the groups being compared had similar pre-intervention trends of earnings or employment. Preexisting differences in earnings or employment—and not ENDAs—could explain the observed differences in outcomes. Therefore, the study is not eligible for a moderate causal evidence rating, the highest rating available for nonexperimental designs.
Causal Evidence Rating
The quality of causal evidence presented in this report is low, because the author did not account for trends in outcomes prior to the study. This means we are not confident that the estimated effects are attributable to ENDAs; other factors are likely to have contributed.