Developing countries face massive urbanization and slum upgrading is a popular policy to improve shelter for many. Yet, preserving slums at the expense of formal developments can raise concerns of misallocation of land. We provide causal long-term impacts of the 1969-1984 KIP program, which provided basic upgrades to 5 million residents covering 25% of land in Jakarta, Indonesia. We assemble high-resolution data on program boundaries and 2015 outcomes to address program selection bias. On average, KIP areas today have lower land values, shorter buildings, and are more informal, per a novel photographs-based slum index. The negative effects are concentrated within 5km of the CBD. We develop a spatial equilibrium model to characterize where the welfare implications of KIP are the largest. Counterfactuals suggest 77% of the welfare effects from removing KIP stem from land in the center and highlight how to mitigate losses to displaced residents. JEL Classifications: R14, R31, R4.