What is the effect on local economies when the state intervenes to capture its own territoriesback from non-state actors? In 2008, the government of Rio de Janeiro, Brazil implemented apolicy to take control offavelasthat were previously dominated by criminal organizations. We use day and night luminosity to assess the effects of this program on economic growth. The difference-in-differences design shows that state intervention has a significant and negative average treatment effect on the favelas that received the intervention. We further a mechanisms to explain this economic downturn: institutional replacement. Based on crime data, we demon-strate that this effect is caused by the destruction of local markets, especially illicit activities.The data highlight the perils of order transition, even when OCGs are removed by state actors.Furthermore, this paper reinforces the need for policies that are mindful of the externalities ofinstitutional shifts