This blog is based on an article published in the Journal of Social Policy by Adam Gendźwiłł, Jan Kroszka, Julita Łukomska, and Michał Pierzgalski.
Does electing women to positions of power actually change policy outcomes? Our research provides a rather surprising answer: when it comes to childcare provision in Polish municipalities, a female mayor doesn’t necessarily make a difference.
The Descriptive-Substantive Link in Question
In our study, we examine how female executive political leaders at the local level shape policy priorities related to childcare. Political scientists have long debated whether increased female representation among elected officials translates into policies that better serve women’s interests. This relationship — drawing on Hanna Pitkin’s classical distinction between descriptive and substantive representation, where substantive representation is defined as “acting in the interests of the represented in a manner responsive to them” — underpins many efforts to increase women’s presence in elected offices. Substantive representation refers to whether representatives actually advance the interests, concerns, and priorities of the group they descriptively represent.
We focused on childcare services—nurseries and kindergartens funded by local governments—because, as has been established, these policies play a crucial role in improving women’s labor market opportunities and promoting gender equality.
Poland’s “Strong Mayor” Local Government System as a Case Study
Poland offered an ideal testing ground. Between 2005 and 2020, the country substantially expanded its childcare services, with enrollment rates for children aged 3-6 increasing from approximately 55% to 90%, and for children under 3 from less than 5% to 15%. Importantly, Polish municipalities exercise significant control over childcare provision, making them a relevant unit of analysis.
Polish municipalities operate under what has been described as a “strong mayor” system, or “local presidentialism”. Directly elected mayors control both the political agenda and oversee local administration, with considerable influence on municipal institutions, design, and implementation of local policies. Thus, mayors are able to “inject their preferences into public policies,” much as the classic citizen-candidate model suggests.
Despite increased female representation in local councils—reaching 30.3% after the 2018 elections—the gender gap in executive positions remains stark: only 12% of mayors elected for the 2018-2023 term were women.
Our Empirical Strategy
We employed a difference-in-differences (DiD) quasi-experimental design to identify causal relationships, comparing what happened when:
- A woman replaced a male incumbent mayor (313 municipalities across three electoral cycles)
- A man replaced another male incumbent (1,749 municipalities).
For each electoral term, we treated the previous term (t, t-1, t-2, t-3) as the pre-treatment period and the four consecutive years (t+1, t+2, t+3, t+4) as post-treatment periods. By excluding municipalities where incumbents were re-elected, we made treatment and control groups more similar and eliminated potentially confounding effects of incumbency.
Our dataset covered nearly 2,500 Polish municipalities across 16 years. We measured four key outcomes:
- enrollment rates in nurseries (ages 0-2),
- enrollment in kindergartens (ages 3-6),
- municipal budget expenditures for nurseries,
- municipal budget expenditures for kindergartens.
To strengthen our analysis and test the robustness of our results, we supplemented the DiD approach with a generalized synthetic control method. This more complex design allows for the presence of unobserved time-varying confounders affecting units differently. Specifically, we used matrix-completion techniques, as developed by Susan Athey and colleagues, to construct synthetic control units and impute counterfactual outcomes.
Key Finding: Gender Alone Doesn’t Drive Policy Change
The results were consistent across both methodological approaches: we found no systematic causal effects of a mayor’s gender on childcare service provision or spending.
Almost all estimates in our DiD models were statistically non-significant at the conventional 95% confidence level. The few statistically significant effects we did observe were contrary to expectations—municipalities with female mayors replacing incumbent male mayors (treatment group) saw expenditures on kindergartens increase by 165 PLN and 168 PLN per capita less than municipalities with newly elected male mayors (in the 2010-14 and 2018-22 terms respectively).
Given average annual expenditure levels on kindergartens (3,000-4,500 PLN per capita during 2010-14 and 6,000-8,000 PLN during 2018-22 term), these differences were relatively small and did not correspond to significant reductions in service coverage.
Our GSC estimates consistently yielded high p-values, with none of the coefficients exceeding the conventional 5% significance threshold, reinforcing our DiD findings.
Investigating Different Municipal Contexts
We further examined whether results varied across municipalities based on factors such as population size, financial condition, share of women councilors, and mayor’s age. The effects remained largely null.
In larger municipalities during the 2014-18 term, we did observe an increase in nursery places (approximately 2.6 percentage points per year) and expenditures (+129 PLN per child per year) when women replaced men as mayors. However, with respect to kindergartens, effects remained null, suggesting no consistent results across childcare services.
One notable finding emerged from our contextual analysis: in municipalities with below-median female representation on councils, female mayors replacing men were even less inclined to develop childcare services than newly elected male mayors.
Why Doesn’t Mayor Gender Make a Difference? Interpreting the Null Findings
Our findings challenge assumptions about the connection between descriptive and substantive representation. Several interpretations are plausible, for example:
- Shared Policy Priorities: Male and female mayors may have similar policy preferences, particularly when policies are non-controversial or incentivised by the national government.
- Conformity to Dominant Leadership Models: As a minority, female mayors may conform to prevailing models of local leadership, prioritizing traditionally “masculine” policy areas (e.g. infrastructure) over welfare services.
- Policy Inertia: Systematic differences in policy outcomes may take longer to manifest, exceeding a single mayoral term.
The Reality of Gender-Responsive Policies — The Descriptive-Substantive Link Revisited
Our study doesn’t diminish the importance of women’s descriptive representation. Rather, it reveals that the link between descriptive and substantive representation isn’t straightforward.
The evidence suggests we should look beyond the mayor’s gender alone. Our supplementary analyses showed that female mayors in municipalities with below-median representation of women among councilors were even less inclined to develop childcare services than newly elected male mayors. This finding is generally in line with the ‘critical mass’ argument, stating that sufficient descriptive representation enables substantive representation. Yet, it also suggests that with a heavy gender imbalance in local councils, female executives were even counterproductive in promoting women-friendly policy agendas.
External influences from other tiers of government are also crucial. We shouldn’t neglect the role of upper-level politicians and civil society organizations who promote policies contributing to gender equality or construct incentives for local authorities to do the same in their communities.
While we found that mayor gender alone didn’t systematically affect childcare policies, this suggests that meaningful policy change advancing gender-responsive policies likely requires female representation throughout the political system—not just among local executives.
Reference
Gendźwiłł, Adam, Jan Kroszka, Julita Łukomska, and Michał Pierzgalski. 2025. “Female Mayors Do Not Lead to Greater Childcare Provision – Evidence from Polish Municipalities.” Journal of Social Policy: 1–23. doi: 10.1017/S0047279425000121.
About the authors
Adam Gendźwiłł is Associate Professor of Sociology, University of Warsaw, Poland.
Jan Kroszka is Research Assistant at the Faculty of Geography and Regional Studies, University of Warsaw, Poland.
Julita Łukomska is Adjunct Professor in the Department of Local Development and Policy, University of Warsaw, Poland.
Michał Pierzgalski is Associate Professor in the Faculty of International and Political Studies, University of Łódź, Poland.
Discover more from Social Policy Blog
Subscribe to get the latest posts sent to your email.
