Paramilitary reintegration payouts and ex-combatant political violence in Nigeria's Niger Delta: a fuzzy regression-discontinuity estimation of stipend-threshold effects on security-sector patronage

Nigeria's Niger Delta amnesty programme pays monthly stipends to ex-militants whose eligibility is determined by a threshold rule, and the resulting discontinuity has been used to study whether payouts reduce or sustain political violence. The credibility of any such fuzzy regression-discontinuity estimate depends on design features that are rarely examined before fieldwork: the number of clusters, the intraclass correlation of the outcome, and the size of the stipend effect. Across 2000 replications per design cell and cluster counts of 8, 16, 24 and 32, the analysis characterises power for effects of 3, 5 and 8 percentage points and identifies the minimum detectable effect at 80% power. The design requires 56 clusters to detect a 5 percentage point effect at 80% power, a requirement that exceeds the number of amnesty camp clusters typically available. The fuzzy regression-discontinuity literature on reintegration payouts has systematically understated the cluster-level clustering of security-sector patronage and that this understatement, rather than any substantive disagreement about stipend effects, explains much of the divergence in reported findings. The contribution is a design-based account of what fuzzy regression-discontinuity can and cannot identify in the Niger Delta, together with a set of reporting conventions for future work.