Gender and public service provision: Evidence from Sierra Leone
A field experiment in Freetown, Sierra Leone, shows how a concentration of wealth in favour of men could lead to the provision of public goods preferred by men. This implies that wealth disparity between men and women can have gendered spillover effects for political participation and the demand for public services.
In 2020, women made up 50% of the population in Sierra Leone, which is also the case for Western Area, the region where the capital Freetown is located. Despite the balanced distribution of gender, we find that women own only approximately 33% of the properties in the city Freetown. This suggests that women control disproportionally less wealth – proxied by property ownership – than men. The unequal distribution of wealth is not only alarming in terms of gender parity in property ownership, but also has the potential to influence related areas like the demand for public services and political participation in a gendered way.
Property tax income directly links property ownership to the provision of public goods. Historically, government demand for taxes has been shown to induce increases in the demand for public goods (Schumpeter, 1918; Tilly, 1985). According to an entitlement mechanism, taxpayers expect reciprocity from the government when paying taxes (Weigel, 2020). Where wealth (proxied by property ownership) is skewed towards men, this is likely to produce a male-bias in the distribution of the demand for public services: men will feel more entitled to influence how tax money is spent. This means that more men will engage with the political entities in charge to express their preferences for the public services they want.
In our baseline survey, we observe that female citizens are less likely to engage politically on various levels. Figure 1 shows that a significantly lower share of women participate in community meetings or meetings with politicians, discussions on social media, the radio or TV, or engages with their ward councillor or member of parliament. The only type of engagement for which no significant differences by gender are observed is participation in political marches. Overall, evidence indicates that men use more opportunities to express their political opinion, including public service preferences.
Figure 1: Women are less likely to participate in political discussions across types of engagement
While gender bias in political participation alone is concerning, studies from several settings also suggest that women and men have different public service preferences (Chattopadhyay & Duflo, 2004; Olken, 2010; Khan, 2021). In places where political participation is biased towards male citizens, different preferences are likely to result in a gender bias in actual service provision. In our survey, we capture several layers of public service preferences. We first look at individual preferences, asking respondents to choose the service they would most want to be provided. Strikingly, at this level, we do not find any differences in the ranking of services between women and men (see Figure 2). Although the percentage of women stating access to water as their first preference is slightly higher amongst women, women and men agree on the primary importance of access to water.
Figure 2: Individual service preferences for women and men
However, when asking respondents about the service that would most improve the lives of women and men, respectively, the ranking of services differs by gender. While women and men agree on the importance of healthcare for women, education seems to be the most important service for men. A potential reason for the differences between the ranking of preferences on the individual level and on a more aggregate level by gender is the overarching importance of water for all citizens. When asked for services that improve the lives of a specific group of individuals, respondents most likely emphasise services which matter a lot for women, but less so for men, and vice versa. This means that a male-bias in the demand for public services is likely to represent the citizens’ need for water, but potentially fails to properly reflect the secondary needs of women as compared to those of men.
These findings are an interesting starting point for further research on the gendered aspects of the relation between public service preferences, representation, and political participation. In the second phase of our project, we plan to provide a random subset of individuals with information on services that are currently provided by the Freetown City Council. If the provision of public services follows the pattern of service demand, and as such is biased towards male preferences – a hypothesis that is yet to be confirmed – we expect female citizens to react to our treatment by increasing their political participation disproportionally more than men. This hypothesis is based on the idea that more engagement is necessary to ensure service provision in line with own preferences if these preferences are not yet satisfied. While it is beyond the scope of our project to directly influence the patterns of public service provision, we hope to mitigate imbalances in the service distribution by nudging female citizens to increase their engagement with the political process and demand service provision according to their preferences.
This article is part of our gender equality series.