Labor Supply Differences between Married Heterosexual Women and Partnered Lesbians: A Semi-Parametric Decomposition Approach

Article excerpt

I. INTRODUCTION

Partnered lesbian women have substantially higher labor supply than married women (Leppel 2008; Tebaldi and Elmslie 2006). What might account for the sexual orientation gap in partnered female labor supply? Everyday conversation and casual empiricism suggests that partnered lesbian women have a stronger attachment to the labor market relative to married women because of their choice to have fewer children than their married counterparts; yet this has not been formally analyzed. The primary goal of this study is to determine the role observable characteristics, particularly children, play in explaining the differences in labor supply between partnered lesbian women and married women.

While empirical research on the determinants of partnered lesbian women's labor supply is limited, (1) the determinants of married women's labor supply have been studied in great detail (see Blundell and MaCurdy 1999 for a survey of the literature). In general, married women's labor supply is found to be positively related to own wages, negatively related to spouse wages and nonlabor income, and negatively related to the presence of children, particularly young children, in the household. (2) Importantly, the same control variables are not found to have the same effect on married men's labor supply. In particular, men are not responsive to their wife's earnings and tend to have higher labor supply because of the presence of children.

These findings are consistent with a traditional division of labor into market and household work which results in married men being viewed as the primary earners and married women being viewed as the secondary earners. (3) In particular, Becker (1981) argues that increasing returns from investments in specific human capital encourages a division of labor in market work and household work among household members. The sexual division of labor, however, arises from intrinsically different comparative advantages of men and women (e.g., in the production of children) which would determine the direction of the division of tasks by gender. This in turn leads to differences in specific human capital accumulation which reinforces the intrinsically different comparative advantages of men and women.

Even in the face of rising female labor force participation, increased divorce rates, and lower fertility, male and female earnings would not be equalized (Becker 1985). Specifically, Becker argues that married women with household responsibilities (e.g., child care and food preparation) would expend less energy on market work, make lower investments in market-oriented human capital, face lower hourly earnings, and choose less demanding jobs/ occupations than married men (even when they work the same number of hours) because household responsibilities are time and effort intensive relative to leisure and other nonmarket uses of time by men. Moreover, married women would have lower labor force participation than their husbands because of the lower earnings they would face (because of less energy expended on market work and lower investments in market-oriented human capital) and a full equilibrium could involve complete specialization by married women in household production.

Although married men have increased their time spent on child-care production, there is strong evidence that married women continue to spend more time on child-care production than their spouses (Bianchi, Robinson, and Milkie 2006; Drago and Lee 2008a, 2008b; Kalenkoski, Ribar, and Stratton 2005, 2007; Lundberg, Pabilonia, and Ward-Batts 2007). In addition, the sociology literature argues that married women are more likely to identify themselves in the context of family and market work while married men are more likely to identify themselves in the context of market work alone (see, e.g., Bielby and Bielby 1989). Finally, there is a large literature that tests the specialization hypothesis (Bardasi and Taylor 2008; Daniel 1992; Gray 1997; Hersch and Stratton 2000; Kenny 1983; Loh 1996; Stratton 2002). …