Rural cooperative arbitration panel seating quotas and women's contract-enforcement claims in Rwandan sectors: a regression-discontinuity analysis of mediation-threshold assignment

Rural cooperative arbitration panels in Rwanda operate under a seating quota that reserves a share of panel positions for women, but the quota is triggered only when a dispute crosses a mediation-threshold caseload. Because no micro-data on cooperative arbitration are available, the paper is methodological and conceptual: it specifies the running variable, the cutoff, the bandwidth, the local polynomial estimator and the assumptions required for a credible discontinuity, and it derives the conditions under which quota activation can be separated from caseload composition. The design is illustrated with a real macro-panel of 560 country-year observations from 24 African countries (2000–2023) drawn from World Bank World Development Indicators, estimated by OLS with heteroskedasticity-robust (HC1) standard errors, yielding R-squared = 0.591, adjusted R-squared = 0.5895 and F = 184.764. The mediation threshold is a plausible source of quasi-random variation in panel composition, but that its credibility depends on the absence of manipulation around the cutoff and on the exclusion restriction that threshold crossing affects claims only through quota activation. The contribution is a transferable design template for evaluating gender quotas in customary and cooperative dispute institutions where random assignment is infeasible.