Women are underrepresented in political offices almost everywhere across the globe. Figure 1 illustrates this on the basis of the share of women in national parliaments in 2019. On average, women only make up 23% of national parliamentarians. Similar levels of underrepresentation can be observed at subnational tiers of government as well.
A natural question to ask is whether this underrepresentation affects policy choices and thereby potentially reduces women’s welfare. The answer to this question has direct implications for current public debates around the introduction of interventions such as quotas to increase female representation. Such measures are more likely to be deemed necessary if political underrepresentation indeed has adverse consequences for women’s welfare.
Female councillors accelerate the expansion of public childcare
Previous literature mainly focuses on the effect of women in executive political offices on policies (Chattopadhyay and Duflo 2004, Brollo and Troiano 2017). However, the question of whether women in deliberative bodies (such as parliaments or local councils) have substantive effects on policies has been largely neglected in the economic literature.1 This is an important gap since major policy changes typically require legislative approval.
In our study (Baskaran and Hessami 2019), we address this question using the example of the municipal provision of public childcare. We analyse whether a higher share of female councillors affects the expansion of the number of local public childcare spots. Public childcare spots are a public good whose provision is planned and administered by local councils. For example, local councillors assess the need for childcare, provide municipal plots for construction and contribute to the financing of childcare facilities.
Survey evidence shows that public childcare is a higher priority for women than for men (Wippermann 2016) and that female labour supply is highly responsive to the availability of childcare (Gathmann and Sass 2018). At the same time, Figure 2 shows that demand for childcare continues to outpace supply by a wide margin in Germany.
If a higher representation of women in politics has substantive effects on policies – and in particular, ensures that women’s needs and preferences are more adequately met – one may expect that municipalities with a higher share of female councillors achieve a faster expansion of public childcare.
On the other hand, there are several reasons why a higher share of female officeholders may not affect policies after all. Since women have voting rights, male politicians in office also have incentives to take into account women’s policy preferences. In addition, even a strong increase in the share of women in local councils would not change the fact that women are a minority (on average about 20% in Bavaria and other parts of Germany). Therefore, the women in office may not be able to change policies even if they have different preferences than men.
We study the effect of female political representation on policies in the context of the 2,056 municipalities in Bavaria as this particular setting has several methodological advantages. First, Bavarian local councils are quite small, with 53% of Bavarian municipalities having between 8 and 14 seats and 89% of municipalities having 8 to 20 seats. The entry of one additional female councillor therefore has a large effect on the share of female councillors. Second, Bavarian municipalities rely on an open-list electoral system (also referred to as preference voting or preferential voting) which allows us to use an empirical strategy that captures the causal effect of female councillors on the expansion of public childcare.
To be able to conduct this study, we hand-collected micro-data on about 225,000 local council candidates that were running the Bavarian local elections in 2002, 2008 and 2014. These data include information on the gender, party affiliation, list placement, occupation, and age of individual candidates.
From a methodological point of view, the biggest challenge stems from the fact that municipalities with a large share of female councillors may systematically differ from those with a low share of female councillors. For instance, the former may be more left-leaning or more likely to be urban. We address this endogeneity problem by exploiting the fact that an open-list voting system is used. Open-list elections essentially give rise to duels between candidates who run on the same party list. Especially with regard to the last seat that accrues to a party, a duel takes place between two candidates that almost have the same number of individual votes.
Ex ante, it is impossible to predict who will win this duel, especially when it is a close outcome. This quasi-experiment allows us to run estimations based on a regression-discontinuity design. As a result, we end up comparing the expansion of public childcare across municipalities that, on average, have the same characteristics and only happen by chance to differ in their share of female councillors.
Our results show that one additional woman in the local council accelerates the expansion of public childcare by 0.4 spots per 1,000 inhabitants or by about 40%. In addition, we find that the expansion occurs mostly for childcare spots for older children (3-6 years and 6-11 years) and not for children in nurseries below the age of three or in secondary schools aged 11 to 14 years.
Council decisions are indirectly influenced by additional women
Our main results show that women do have substantive effects on the provision of childcare. But how exactly does this effect come about? To answer this question, we analyse the minutes of local council meetings held about once per month in each municipality. We hand-collected and analysed more than 7,700 minutes of council meetings. In particular, we measure how often men and women speak up and which subjects are raised.
In theory, various mechanisms might drive our main results. For instance, an increase in the share of female councillors raises the number of votes in the council that can be cast by female councillors. This might have a direct influence on policymaking if the votes on these decisions are typically split evenly. After carefully looking through the minutes we can, however, exclude this mechanism. In the majority of municipalities, these votes are not split; the councillors deliberate to find a consensus and afterwards take a vote. In most cases, these decisions are taken unanimously or with large majorities.
Our analysis of the minutes rather shows that one additional woman on the council indirectly affects the council meetings. If a woman enters the council instead of a man, all women in the council, on average, speak up more often (beyond the mechanical effect that arises due to having one more female councillor). In addition, the subject of childcare is discussed more frequently. These results show that one additional woman in the council leads all women to play a more active part in the council meetings.
One explanation could be that the extreme minority status of women in Bavarian local councils is reduced significantly when one more woman enters the council. Experimental evidence shows that women tend to be less self-confident and less willing to take on leadership when they find themselves in male-dominated environments (Born et al. 2018). By reducing the extreme minority status of women through one additional female councillor, the influence of women in a deliberative body such as a local council then rises disproportionately.
Several extensions in our study support this interpretation. For instance, we find that an additional woman only has an impact if at most three other women are otherwise on the local council. If there are more women on the local council, an additional woman has no measurable impact on policymaking. This shows that it is important for women to be able at all to play a part in the council debates. When a woman is in the council, she can steer the deliberations along a different path, for instance by bringing in new perspectives or raising other subjects. If there are only a few additional women on the council, the female influence rises disproportionately because the women can support each other and second each other’s statements in an environment in which they perceive themselves as less of a minority.
Our study illustrates that women should be adequately represented in politics to ensure that their needs and preferences are considered. A direct implication of our results is that voters should be aware that they are not only electing individual politicians in elections but that they are also thereby implicitly making substantive policy choices. Voters who care about the provision of public childcare may, for instance, want to cast their vote in favour of female candidates rather than male candidates. A further implication of our results is that interventions to increase female representation, notably quotas, may be appropriate given that underrepresentation appears to reduce women’s welfare.
Baskaran, T and Z Hessami (2019), "Competitively elected women as policy makers". CESifo Working Paper No. 8005.
Born, A, E Ranehill, and A Sandberg (2018), "A man’s world? the impact of a male dominated environment on female leadership", Working Paper in Economics No. 744, University of Gothenburg, Department of Economics.
Brollo, F and U Troiano (2016), "What happens when a woman wins a close election? Evidence from Brazil", Journal of Development Economics 122: 28-45.
Chattopadhyay, R and E Duflo (2004), "Women as policy makers: evidence from a randomized experiment in India", Econometrica 72: 1409-1443.
Gathmann, C and B Sass (2018), "Taxing childcare: effects on female labor supply and children", Journal of Labor Economics 36: 665-709.
Hessami, Z and M Lopes da Fonseca (2019), "Female Political Representation and Substantive Effects on Policies: A Literature Review", mimeo.
Wippermann, C (2016), Was junge Frauen wollen, Report Friedrich Ebert Stiftung.
1 For a more detailed discussion of previous literature on the effect of female politicians on policies, see Hessami and Lopes da Fonseca (2019).