Experimental Design
We conduct a three-arm randomized controlled trial with 180 joint-liability microfinance groups from four branches in La Paz, El Alto, and Patacamaya, Bolivia. Groups are randomly assigned in equal proportions (1/3 to each arm) to: (1) Leadership training only (60 groups), (2) Leadership training plus performance incentives (60 groups), or (3) Control (60 groups).
Treatment Arms:
• T1 (Leadership Training): Peer-led training on leadership skills, in-person informal meeting, and 4-month WhatsApp mentoring
• T2 (Training + Incentives): Same as T1 plus performance-based non-monetary rewards tied to group repayment rates
• Control: Standard microfinance operations
Randomization: Stratified by loan officer and months to loan maturity using re-randomization to achieve covariate balance on baseline characteristics including leader demographics, group loan amount, loan cycle number, group size, and member characteristics.
Timeline: Training delivered between November 17, 2025 and November 30, 2025, with 6-month follow-up period capturing complete loan cycles. Data collection includes baseline administrative data, post-training surveys, and 6-month follow-up surveys, and high frequency monthly administrative data on loan repayment from the microfinance institution.
Eligibility: Groups with at least six loan payments remaining until cycle completion.