For Love and Money? Earnings and Marriage among Same‐​Sex Couples

While the impact of marriage has been widely studied among heterosexual individuals, little research has investigated the impact of being married on the economic outcomes of gay men and lesbian women.

March 10, 2021 • Research Briefs in Economic Policy No. 254
By Michael E. Martell and Peyton Nash

Same‐​sex couples have long lived in marriage‐​like relationships. These relationships, however, have only recently been accompanied by the legal rights and benefits that have long been afforded to different‐​sex couples. The extension of the legal rights and benefits of marriage to same‐​sex couples began in Massachusetts in 2003 and concluded across the United States in 2015. Access to legal marriage was a goal of rights advocates in part because marriage has the potential to increase the social and economic well‐​being of those who enter into the institution. While the impact of marriage has been widely studied among heterosexual individuals, little research has investigated the impact of being married on the economic outcomes of gay men and lesbian women.

Social scientists have long observed a large earnings premium associated with marriage among heterosexual men, with estimates ranging from 10 to 50 percent. Historically, heterosexual women have experienced a marriage penalty, but recent work suggests that heterosexual women also enjoy a marriage premium. These premiums highlight the possibility that marriage remains an important contributor to economic and social security. We examine the relationship between earnings and marriage among a population that, until recently, was not able to enter into marriage. Our analysis, therefore, combines two related fields of literature: economic effects of marriage and the economic experience of sexual minorities.

Our knowledge of the economic experiences of sexual minorities has grown rapidly in recent years. This literature has documented robust evidence that gay men earn less than similarly situated heterosexual men and that lesbian women earn more than similarly situated heterosexual women. While the bulk of research investigating earnings differentials for sexual minorities utilizes data from the United States, similar patterns have been observed in Sweden, Canada, the Netherlands, the United Kingdom, and Greece. These earnings differentials are large, typically ranging from 10–20 percent for gay men and as large as 30 percent for lesbian women. However, recent work suggests that these differentials are getting smaller.

The source of the earnings effects of sexual orientation continues to be debated. The difficulty of reconciling the asymmetric earnings effects of sexual orientation as well as the relative invisibility of homosexuality with standard theories of discrimination motivated researchers to investigate several nondiscriminatory explanations. For example, the earnings effects could be explained by important but unobserved characteristics that vary by sex and sexual orientation that are rewarded in the labor market, such as masculinity. If gay men are less masculine and lesbian women are more masculine than their heterosexual counterparts, the earnings differentials they experience could be due to their characteristics, not their sexual identity. However, the inclusion of a wide variety of individual and family characteristics (including markers of masculinity) does not affect the size of earnings‐​differential estimates for gay men or lesbian women. Preferences may also play a role. The earnings differentials are consistent with preferences for labor motivating gay men to work less and lesbian women to work more than their heterosexual counterparts. However, the asymmetric earnings effects of sexual orientation persist when controlling for labor market attachment via estimating hourly wages instead of annual earnings. Furthermore, varying preferences for labor may also arise as a response to discrimination or workplace stress related to sexual orientation.

Evidence in favor of discrimination, particularly against gay men, as an explanation of the earnings effects of sexual orientation is growing. The wage penalty gay men experience is smaller in geographic areas with higher levels of tolerance of homosexuality. Furthermore, the sorting of gay men into occupations that allow a greater degree of independence translates into smaller earnings penalties because these occupations allow gay men to avoid involuntary disclosure of their identity and discrimination. Similarly, disclosing a gay or lesbian identity on job applications leads to lower likelihood of receiving an invitation to interview for a job. Evidence of discrimination in hiring complements evidence of compensation discrimination for gay men but not lesbian women. Discrimination may have a larger observed effect on hiring than it does on compensation for several reasons. First, anti‐​homosexual animus is more acutely articulated against gay men than it is against lesbian women and should therefore translate into larger negative differentials for gay men. Second, lesbian women are more likely to work in male‐​dominated occupations and have more education than heterosexual women. Educational and occupational attainment explains much, but not all, of the lesbian earnings premium. The lesbian premium may coincide with discrimination against lesbians based on sexual orientation if the educational attainment and occupational sorting of lesbians arise as a response to anticipated discrimination in the workplace or discrimination in access to rights, such as marriage and anti‐​discrimination laws, afforded to many heterosexual individuals.

Using data from the American Communities Survey from 2013 through 2017, we find that, in terms of earnings, marriage matters. Marriage increases the earnings of gay men by approximately 3 percent and the earnings of lesbian women by approximately 6 percent relative to their unmarried cohabiting counterparts. These are meaningfully large, but the premium for gay men is smaller than the premium for heterosexual men. The premium is similar in size for heterosexual and lesbian women. The effect of marriage relative to single gays and lesbians is nearly three times larger. This financial return to marriage is in addition to the social, symbolic, and personal benefits that may accompany the institution. Thus, for same‐​sex couples, marriage may bring love and money.

The marriage premiums we estimate seem to arise as a result of marriage increasing investments in relationships and, therefore, household specialization. The fact that marriage increases the wages of heterosexual men via an increase in household specialization has long been argued to explain the marriage premium among heterosexual men. Our results are consistent with the increased return to joint investments motivating greater labor market attachment in same‐​sex households. However, the increase in labor force attachment appears larger for primary earners, as same‐​sex households increase their household specialization. Thus, the effect of marriage on specialization appears universal given our results that document it occurring in gay and lesbian households, which comprise a previously unexplored population.

The higher earnings that marriage affords to gay men and lesbian women suggest that marriage can reduce economic vulnerability for this population. This reduction in vulnerability is important, particularly among lesbian households, which face higher risks of poverty than gay or heterosexual households. In addition to contributing to an understanding of the likely impact of marriage on economic vulnerability, these results also contribute to an understanding of the impacts of the quickly changing legal environment on economic outcomes of sexual minorities and how sexual minorities organize their work and professional lives.

This research brief is based Michael E. Martell and Peyton Nash, “For Love and Money? Earnings and Marriage among Same‐ Sex Couples,” Journal of Labor Research 41, no. 3 (2020): 260–94,–09305-4.


Download the Research Brief in Economic Policy

About the Authors
Michael E. Martell

Bard College

Peyton Nash

Bard College