Monday, February 19, 2018

Some Patterns for Same-Sex Households

A couple of decades back, it was  hard to find systematic and reliable information on the economic and family characteristics of same-sex couples. However, the US Census has been collecting baseline data for some years now, and just released an annual report, Characteristics of Same-Sex Couple Households: 2005 to Present. The report is just a few statistical tables. I've taken one of the tables, cut some of the rows, left out the columns showing statistical significance, and cut the footnotes. But there's enough left to note some patterns.

  • For the US as a whole, The Census folks counted 56.4 million married opposite-sex couples, 6.8 millions unmarried opposite sex couples, and almost 900,000 single-sex couples, more or less evenly split between male-male and female-female couples. 
  • By average age, the same-sex couples are much closer to the opposite-sex couples than to the unmarried opposite-sex couples. 
  • Same-sex couples are much more likely to be interracial than married opposite-sex couples, especially male-male couples. 
  • Same-sex couples tend to have higher education levels than opposite-sex couples, who in turn have  higher education levels than unmarried opposite-sex couples. 
  • In terms of employment, unmarried opposite-sex couples rank highest, followed by the same-sex couples, while married single-sex couples are lowest. 
  • Married and unmarried opposite-sex couples are about equally likely to have children in the household, while same-sex couples--especially male-male couples--are much less likely to have children in the household. 
  • In terms of median household income, the rank from highest to lowest is: male-male couples, married opposite-sex couples, female-female couples, and unmarried opposite-sex couples. 

A table like this one is useful for showing some factual patterns, but it would be extremely unwise to draw any conclusions about causality from it. For example, it is certainly not true that if the unmarried couples were to marry, they would immediately become older, better educated, and with higher paychecks.

[Note added 3/1/18: Those interested in this topic might also want to check this Research Report from the Tax Policy Center: "Same-Sex Married Tax Filers After Windsor and Obergefell," by Robin Fisher, Geoff Gee, and Adam Looney (February 28, 2018) It looks at evidence on same-sex married couples from tax returns, and how the patterns shifted after key Supreme Court rulings affecting same-sex marriage in 2013 and 2015.]

A decade ago, Dan A. Black, Seth G. Sanders, and Lowell J. Taylor. wrote about "The Economics of Lesbian and Gay Families." in the Spring 20017 issue of the Journal of Economic Perspectives (21:2,  53-70). They noted that many of the observed patterns of same-sex couples are consistent with economic reasoning. For example, if a smaller share of same-sex couples are likely to have children, this will influence patterns of work, willingness to live in urban areas, and other factors. They wrote:
"The economic approach to thinking about families—with its focus on deliberative choices made by people with reasonably stable preferences—stresses the importance of constraints. That is, when economists seeks to understand systematic differences in the behavior of gays, lesbians, and heterosexuals, we do not start with a hypothesis of innate differences in preferences, but instead seek to understand how differences in constraints systematically alter incentives faced by gay, lesbian, and heterosexual people. This paper uses the economic approach to organize our thinking about gay and lesbian families, and thus implicitly assumes that, aside from sexual preference, other preferences do not systematically differ by sexual orientation. Based on that approach we provide evidence pertaining to some initial questions. 
"For example, do differing biological constraints faced by gay, lesbian, and heterosexual couples affect choices over children? Clearly they do; there are far fewer children in lesbian and gay families than heterosexual-partnered families. Gays and lesbians also face higher costs for adoption, which further reduces opportunities to raise children. 
"Do differences in fertility (or anticipated fertility), again owing to differences in constraints, influence where people live? We argue that because gay couples face a relatively high price for children, they will consume more non-child goods, and this increased consumption often takes the form of locating in expensive highamenity locations like San Francisco or New York. 
"Do same-sex couples have patterns of household specialization that differ in predictable fashion from heterosexual couples? Available evidence indicates that lesbian women have higher wages and greater labor force attachment than heterosexual women, while the opposite situation pertains for gay men relative to heterosexual men. A variant of Becker’s (1991) model of household specialization would predict this pattern. 
"Of course, many other interesting and fundamental questions await analysis, from economics and other social science disciplines, as better data sources become available."