Do women and men pursue different policy once they have a seat in the cabinet? A theory and empirical test of when descriptive representation is associated with substantive representation

In some presidential cabinets women hold a third to half of the posts, and women are appointed to diverse types of posts including leading departments that handle stereotypically masculine policy areas, making it possible to study whether and how increased diversity in descriptive representation is associated with increased diversity in substantive representation. Toward that end, this paper has two purposes. First, we develop a theory based on interests, incentives, and capacity to derive predictions about when cabinet ministers could be expected to expand representation. This theory merges work from feminist institutionalism and from formal models of how heads of government avoid cabinet drift to generate new implications about when, why, and for whom members of the cabinet represent the interests of women and of other historically marginalized groups. “Interests” theorizes about how the backgrounds and beliefs of cabinet ministers (e.g., feminist v. traditional, occupation, socioeconomics, race/ethnicity) may provide incentives for representation of historically marginalized groups. “Incentives” refers to how institutions, parties, and future career opportunities (e.g., specialist v. generalist and political v. technical norms for ministers, party ideology, coalition government) enhance or limit rational actor incentives to represent diversity in government. “Capacity” examines how backing from key actors (e.g., the president, party factions, the finance minister, the minister’s group alliances or electoral base of supporters) expand or limit opportunities for a minister to deliver on substantive representation of women and other historically marginalized groups. Second, we propose an empirical framework for systematically studying how cabinet ministers do their jobs, enabling comparison of decisions by men and women. A perennial challenge to studying cabinet ministers is that policy deliberations often take place “behind closed doors” in cabinet sessions or other private meetings among top government leadership. To circumvent this problem, we focus on observable actions by cabinet ministers that occur outside of those deliberations. In this paper we provide a preliminary empirical test of this new theory of interests, incentives and capacity, with data about executive decrees (regulations) in Costa Rica issued by government departments. We compile the relevant text using supervised machine learning algorithms, and this text-as-data allows us to compare policy initiatives drafted under female and male ministers during two presidential administration (2010-18). The executive branch administers laws passed by the legislature, which often requires the relevant department to issue regulations providing detail about how policy will be implemented. This regulation is essential to policy implementation, but it is less publicly visible than new legislative initiatives proposed to the congress. Due to its more bureaucratic nature, we expect executive decrees (regulations) will offer cabinet ministers more latitude to pursue their own policy preferences, and thus that they provide an optimal arena for examining whether men and women in the cabinet take action to represent different groups in society. Costa Rica offers an interesting pilot case due to the country’s relatively long history of extensive incorporation of women into government.