Malawi is a hybrid regime. While elected local governments hold considerable de jure authority, in practice they are circumscribed by incomplete fiscal decentralization and by more powerful actors—including members of parliament, unelected district executives, and traditional authorities—who view councilors as potential political threats. These power struggles, and the presence of competitive clientelism have severely hampered the ability of local government to deliver services in response to community priorities. This has led to significant citizen disillusionment with local councilors. In 2016-2017, two-thirds of respondents disapproved of their ward councilor’s performance and only 15% of incumbents were reelected in 2019. The unpopularity of ward councilors is in part due to a lack of citizen input in local decision-making: only 5% of Malawians felt that their councilor receives input from ordinary citizens.
While citizen engagement and satisfaction with local governance is generally low, women in Malawi participate even less in local politics than men: On the most recent Afrobarometer Malawi survey, 60% of women report “never” discussing politics (versus 43% of men), only 11% of women report having contacted their local councilor in the last year (versus 21 % of men), and only 17% of women report knowing about recent land reforms (versus 37% of men). In addition, women hold just 13% of local council seats, and face significant cultural barriers when running for elected office. Against this backdrop, the aim of this project is to develop and test interventions to increase the quantity and quality of women’s non-electoral forms of political participation. The intervention is designed to increase the demand women have for political knowledge and skills, as well as the supply of the relevant skillsets to participate effectively.
Intervention Date: February – April 2023
Our intervention – the common treatment arm across Metaketa V experiments – is a series of collective efficacy training sessions delivered to women participants of women’s action committees. In the Malawian case, these women’s action committees are pre-existing village savings and loan associations (VSLAs). VSLAs in Malawi provide one of the few sources of financing for rural residents. The facilitated collective efficacy sessions are a group-level treatment, delivered to women’s-only VSLAs in treatment communities. We will measure the effectiveness of our intervention in reference to members of VSLAs that are randomized into the control group, who will not receive these facilitated sessions. Members of all groups will attend an initial “burn-in” meeting, at which we will provide all group members with information on local council mandates and funding. After the burn-in meeting, we will block randomize groups into either treatment or control based on baseline-collected information. Treatment groups will then receive additional facilitated sessions meant to foster a sense of collective efficacy. Control groups will receive no further training sessions, but will continue to meet for VLSA matters, as before.
In addition, we will cross-randomize incentives to join the VSLA across the treatment and control groups at the individual level. We will recruit new members to both treatment and control VSLAs through the recommendations of existing VSLA members. This design allows us to test how simply joining a non-political women’s group (i.e., a VSLA) affects women’s political participation, and whether and how collective efficacy interventions increase the magnitude of these effects.
We anticipate that the intervention will increase women’s gender-based identification, feelings of relative deprivation, and perceptions of collective efficacy at the individual level, relative to the control group. We also expect that members of the treated VSLAs will be more likely than those in the control group to have a shared understanding of their ward’s key needs and to attempt to communicate those needs and demands to their ward councilor. Furthermore, we expect the substantive focus of the treated VSLA members’ demands to their ward councilor to be different from those in the control VSLAs. At the ward level, we thus expect that the substantive content of priorities communicated to councilors in treated VSLAs will be more focused on women’s issue areas, compared to the control.
We expect the trainings also will improve participants’ political interest and internal and external efficacy compared to members of the control group. At the ward level, we anticipate that participation in these trainings will improve the quality of requests to councilors, making them both more realistic and more relevant to local government decision making. Finally, we anticipate that councilors will be more responsive to women’s priorities in treated versus control VSLAs.
While we hypothesize participating in collective efficacy sessions will increase women’s political engagement, we also leave open the possibility that just participating in a women’s group such as VSLA could have a similar effect. Our individual level randomization allows us to test for this possibility: if just joining a VSLA has such an effect, we should see treatment effects at the individual level (i.e., those randomly incentivized to join the VSLA), but not at the group level.