How does rebel governance affect long-term development? Rebel forces have controlled territory and imposed their own institutions in many countries over the past decades, affecting millions of people. We investigate the economic, social, and political consequences of temporary territorial control by guerrillas during the Salvadoran Civil War. During that time, guerrillas displaced state authorities and created informal institutions that encouraged autonomy and self-sufficiency. Using a spatial regression discontinuity design, we show that areas once under guerrilla control have experienced worse economic outcomes over the last 20 years than adjacent areas controlled by the state. In these areas, higher social capital coexists with negative economic consequences. The fact that rebel institutions developed as an alternative to the state generated mistrust of outsiders and isolated these areas from the rest of the country, resulting in over-dependence on subsistence farming and disengagement from post-war governments. Results are larger in areas where rebel governance initiatives were stronger and do not revert despite increased postwar public investment in formerly guerrilla areas. This study shows that when non-state actors develop alternative governance institutions, these can prompt adverse development effects through lasting norms of distrust of out-groups.