State Capacity Under Fire: Rebel Violence and the Demand for Military Rule in Africa

This paper examines whether localized rebel violence is associated with citizens' support for military rule in Africa. We combine geocoded Afrobarometer surveys with high-resolution ACLED conflict events to build a micro-spatial dataset covering roughly 278,000 respondents in 42 countries. The empirical design compares respondents within fine geographic grids and survey waves, and is complemented by alternative exposure measures, matching, placebo reassignment tests, bandwidth checks, and an auxiliary rainfall-terrain instrumental-variable exercise. Recent rebel attacks within 50 kilometers during the 10 days before an interview are associated with higher support for military rule. In the most saturated fixed-effects specification, moving from no nearby attack to one nearby attack implies a 0.079-point increase on the five-point military-rule scale. The IV estimate is positive but imprecise, so it is treated as a directional robustness check rather than as the paper's main source of identification. Mechanism and heterogeneity results suggest that the response is short-run and security-oriented: rebel attacks are associated with lower civilian institutional trust and worse economic evaluations, but these channels do not explain away the positive association with military rule. Visible local infrastructure is associated with lower baseline support for military rule, although it does not robustly moderate the marginal response to attacks. The findings suggest that localized insecurity can weaken democratic consolidation by increasing the short-run appeal of military authority as a source of order.