American Political Science Review
State-Building in the City: An Experiment in Civilian Alternatives to Policing
2026-04-06
We helped design and evaluate a statebuilding intervention in Medellín, Colombia. The municipal government dramatically intensified nonpolice state presence in 40 neighborhoods over 20 months. On average, perceptions of security and legitimacy changed negligibly, suggesting that returns to statebuilding investments are generally low, at least within electoral cycles. Prespecified heterogeneity analysis, however, reveals significant increases in security and legitimacy where state governance began relatively higher, while impacts were null or possibly negative where it began lower. This suggests increasing rather than diminishing returns to statebuilding. The divergence apparently resulted from city officials under-delivering in initially lower-governance sectors. One reason might be “start-up costs” in statebuilding. Alternatively, both initial state penetration and incentives to implement new programs might depend on neighborhoods’ ability to hold agencies accountable. Whatever their source, increasing returns could drive persistent “neglect traps”—channeling political attention and investment to areas where state penetration is already robust, reinforcing existing disparities.