They know each other, but do they trust each other? Social capital and selected beneficiaries of community-based development projects: A lab-in-the-field in rural Zimbabwe

Community-based development (CBD) projects have long emphasized a bottom-up approach. For CBD initiatives to succeed, communities must harness their social capital, organize themselves, and actively engage in development processes. While CBD proponents highlight the promotion of social capital through community-based projects, critics argue that their effectiveness relies on pre-existing levels of trust, trustworthiness, and community interactions. To contribute to this debate, we investigate the selection bias regarding social capital induced by the recruitment strategy of an NGO in Zimbabwe. We look at differences between selected beneficiaries and non-beneficiaries in terms of pro-social behaviors, measured by incentivized games, and in terms of social networks. We also use this information to test whether being part of the same networks translates into increased trust, altruistic behaviors, and willingness to participate in collective action. Our study, conducted in 2022 in the rural district of Murehwa, Zimbabwe, comprised a survey and lab-in-the-field experiments (trust game, public good game, dictator game) involving 341 subjects. Findings showed that selected beneficiaries exhibit higher network density than non-beneficiaries. However, except for a partial experimental measure of trustworthiness, we observed no significant differences in prosocial behavior between the two groups before project implementation. The results suggest that although selected beneficiaries are more socially connected, they do not exhibit higher prosocial behaviors. These findings shed light on the common selection process used by development agencies and the inherent bias they introduce. To address this bias, development agencies should reconsider recruitment strategies that prioritize existing social ties, as they may unintentionally exclude less-connected community members. Instead, they should explore alternative selection approaches, such as the use of field data to ensure inclusiveness. Additionally, integrating trust-building activities at the beginning of projects could enhance cooperation among participants.