12th Annual IGSS Conference • October 28-29, 2021

Integrating Genetics and the Social Sciences 2021

The Power of the Pill: Evidence from Gene-environment Interactions in the UK Biobank.

Dilnoza Muslimova, Department of Applied Economics, Erasmus University Rotterdam

Background and Objectives. In this paper, I explore the role of education genetics in the effect of the access to the birth control pill and abortion on women's education and labor market outcomes. Specifically, I ask if there is a gene-environment interaction between the birth control policies (environment) and genetic predisposition for education (genes) in driving the above outcomes. Early research on the "power of the pill" shows that the introduction of birth control pill in early 1970s empowered women in the US to make independent decisions about childbearing and career planning (Goldin & Katz, 2002; Bailey, 2006). This has been followed by a plethora of studies in different countries with various outcomes. One of the most recent examples investigates the power of the pill in the Netherlands (Zwiers & Marie, WP ). I add to this literature by investigating if and to what extent such liberalization in policy in the United Kingdom (UK) helped women to realize their genetically driven potential for education in terms of educational attainment, career trajectories, and income. Theoretically, in line with the social compensation model (see for an overview: Mills, Barban, Tropf, 2020), I expect a positive gene-environment interaction since introduction of a birth control pill is a liberating policy, which enabled women to plan for children, education and qualifications, and would potentially allow them to realize their genetic potential for education. Methods & Results. I apply the gene-environment interaction framework in the UK Biobank (UKB). The environmental component is measured with an exogenous policy change in the access to the birth control pill and abortion law introduced in the UK between 1960-1970, and it is analyzed using regression discontinuity approach. For the identification strategy, I follow previous literature and use birth place coordinates of women in the UKB to explore regional variation in the access to the pill using data on illegitimacy rates (Baker, WP), religiosity (Zwiers & Marie, WP ) and the proportion of women taking up the pill in the area (De Cao, Barban, Francesconi, WP ). The genetic component is measured with polygenic score for years of education (EA PGS). To increase the power of the polygenic score, I use the UKB's subsample of men as a discovery cohort for the GWAS and construct the polygenic score for UKB's women in the holdout sample. Additional EA PGS is constructed based on the meta-analysis of the above summary statistics with the educational attainment GWAS by Okbay et al. (2016). I conduct within sister pair analysis for causal interpretation of the polygenic score. Moreover, I correct for the measurement error in the polygenic score with the ORIV method (van Kippersluis, et al., 2021). Discussion. This paper contributes to at least two strands of literature. First, it adds to the flourishing literature on gene-environment interactions. There are only a handful of studies with both exogenous genes and environment. This study is the first one to bring causal evidence on the interaction between fertility policies and the genetic endowments for education within the context of the UK. The second contribution is to the literature on the effects of the access to birth control methods. I contribute to this literature by studying heterogeneity in the effect of these policies with respect to genetic endowment for education. The potential interaction has important implications in terms of modifying the resulting inequalities, which could carry over to the next generations.

Presenter's website