State withdrawal and the (relative) democratization of many developing countries have encouraged the proliferation of voluntary, self-help organizations, such as micro-finance and farmer associations. Unfortunately, only very little is known about the factors influencing those groups' ability to provide services to their members. The central question motivating this study is whether observed variation in voluntary associations' provision of services is a function of the rules governing their leaders' selection.
More specifically, studying the performance of farmer associations that were created as part of Uganda's largest development program, the research examines the causal effect of different rules for selecting leaders on members' behavior (e.g. participation-level) and leaders' behavior in office (e.g. effort-level and rent-seeking). In self-help associations, rules for selecting senior executives are usually clustered around two opposing types. One selection rule is Direct Vote, i.e. all association members vote directly for their preferred candidate. Alternatively, members elect representatives who are imbued with the powers to select the association's senior executives. Proponents of direct vote argue that this rule increases grassroots participation and the legitimacy of leaders. It has also been argued that direct vote reduces the likelihood that the management be captured by narrow self-interests. By contrast, proponents of representatives' vote argue that representatives are better positioned to elect able and honest candidates. This, they argue, is because ordinary members tend to be myopic, under-appreciate the qualifications needed to perform complex managerial tasks, and lack information on candidates.
Obviously, resolving this debate has important theoretical and practical implications. Alas, testing the relative effectiveness of leaders' selection rules is empirically challenging, since in almost all naturally occurring settings, electoral rules are chosen endogenously through a complex political process.
To overcome the selection-bias problem, the proposed study relies on a unique situation in which leaders' selection rules were externally assignment to newly-formed farmer associations by project facilitators. Conditional upon identifying well-matched control groups, this external assignment, allows the estimation of the causal effects of selection rules on the behavior of both members and leaders, and ultimately on the provision of services to group members.
The study makes several contributions. First, by using micro-level and by examining dynamics within small groups in which members' interests are closely aligned, this study is well positioned to test some core political science hypotheses regarding the relation between voting rules, voters' information and participation and leaders' accountability and responsiveness. In addition, due to its powerful identification strategy, this study contributes to theory-building in topics such as social choice, participatory democracy, elite capture, rent-seeking and political selection. Third, this study makes several methodological contributions, for example the integration of measures of managerial abilities from organizational psychology and economics into political economy models, and the use of IV at the cluster level to measure outcome variations at the individual level.
