Do agricultural producers forgo otherwise profitable investments due to civil conflict? Answering this question is crucial to our understanding of the costs of violence, but requires the ability to measure farmers’ willingness to invest and access to exogenous variation in conflict intensity. We exploit a unique administrative dataset from Colombia’s largest agricultural bank and the 2016 demobilization agreement between the Colombian government and insurgent group FARC to overcome these challenges. A difference-in-difference analysis yields three main findings: First, credit to small producers increases after the agreement in municipalities with high FARC exposure (17% over sample mean). Higher loan applications drive this increase, with no change in supply-side variables. Second, a simple theoretical framework combined with rich information on characteristics of loan applicants and projects (including credit scores and loan outcomes) suggests that changes in project returns, but not in risk, underlie the increase in credit demand. Third, conflict is not the binding constraint on investment in areas with low access to markets. Higher investment, unchanged default rates and additional evidence of increased nighttime luminosity after the end of conflict imply an overall positive economic impact.