What explains low female political representation? Evidence from survey experiments in Japan

There is no evidence that demand-side factors explain women’s low representation in politics. Instead, results point to supply-side factors such as heavy family duties and expectations that prevent women from running in the first place.

Politicians’ gender is not an important factor in Japanese voters’ voting decisions. 

• Using a conjoint analysis survey, male and female politicians are viewed similarly in most attributes. Factors such as partisanship, policy positions, age, prior experience, and marital status do not systematically disadvantage female relative to male candidates.  

• There is no bias in terms of how Japanese people view female versus male capabilities. Survey participants were asked to evaluate a piece of writing attributed to either a female or male author. Respondents did not punish a female author more critically, suggesting that no gender bias exists among the respondents in terms of their view towards authors’ capabilities.   

The heavy burden of family duties and societal expectations discourage women from seeking political office, leading to women self-selecting out of running for public office. 

• Women of reproductive age are less likely to run for office. This is more pronounced in single-member districts (SMD) as this electoral system requires politicians to spend more time campaigning, which could discourage women with young children from seeking office. 

• Offering help for household work had the greatest effect on women’s likelihood of running for office. Such results were evident for respondents between the ages of 25 and 34 – precisely the range of age when many men entered politics, but women do not.  

This study suggests that women will not make significant headway in Japanese politics until society reduces women’s, and raises men’s, obligations in the family. While childcare support such as subsidies may help increase women’s political representation, it is unlikely that additional childcare support can make up for the “expectations of around-the-clock working hours” (pg. 305). The authors suggest that quotas may increase women’s representation across political parties, given that center-right parties are found to have more severe gatekeeping effects on women’s representation than center-left parties. However, the feasibility of a quota system in Japan calls for additional research. 

Demand-side study: 

The demand-side experiment is designed to evaluate whether voters in Japan harbor gender bias against women political candidates. It consists of a two-wave survey, starting from a conjoint survey exploring voters’ preference for a political candidate and a survey exploring people’s perceptions of female and male capabilities. 

A conjoint experiment aims to examine if participants prefer certain attributes over others, and whether this preference varies by gender. Survey participants (N = 1,611) were asked basic demographic questions, including general attitudes towards women. Participants were reinterviewed and asked to choose between pairs of hypothetical politicians with different attributes, such as gender, age, education level, and other relevant information. Respondents completed this conjoint exercise six times for a total of 19,332 observations.  

As a robustness check, authors asked if Japanese believe women are less competent than men (N = 202). Participants evaluated a memo written by a hypothetical college graduate seeking a first job. The gender of the hypothetical author is randomly assigned. Comparing evaluation scores for each hypothetical author allows researchers to test if women are judged more harshly than men.  

Supply-side study: 

To test women’s willingness to run for public office, authors survey all political candidates and run an additional survey of the general public.  

The first study (N = 1,132) measures candidate demographics to test if family obligations hinder women’s willingness to run. The authors leverage Japan’s mixed electoral system to test the impact of supply-side factors. The authors theorize that women of reproductive age should be less likely to run for office in single-member districts (SMDs) than PR districts, given that SMDs require greater time commitments. Therefore, fewer women of reproductive age running in SMDs supports the claim that women’s political representation is impacted by family obligations. This theory is reiterated through a poll of the House of Representatives where 51% of female respondents indicates caretaking responsibilities prevent women from pursuing politics. 

To better understand the mechanism behind women’s reluctance to run for public office, authors surveyed a nationally representative sample of Japanese adults in 2016 (N= 3,206). The survey consists of basic demographic questions, key covariates, and an experimental component. In the experimental component, respondents are randomly assigned to receive a version of stimulus information to encourage running for a hypothetical general election. The experiment tests whether the existence or absence of support for household chores impacts respondents’ willingness to run for office.