In-situ Upgrading or Population Relocation? Direct Impacts and Spatial Spillovers of Slum Renewal Policies

Slums are widely viewed as disadvantaged areas characterized by substandard housing and negative neighborhood spillovers that constrain local economic development. Governments typically respond with one of two place-based interventions: (a) in-situ upgrading that improves infrastructure and housing conditions within existing settlements, or (b) relocation of slum households to formal housing elsewhere in the city, freeing valuable urban land for alternative uses. We compare the direct effects and local spatial spillovers of these two approaches in Chile. We construct a 20-year national panel of slum polygons that integrates satellite imagery, census microdata, construction permits, administrative records, crime reports, and property-tax data. Using this dataset, we estimate how each policy affects the physical characteristics of slums and adjacent areas, and whether either intervention shifts the socioeconomic composition of residents. Our empirical strategy relies on Synthetic Difference-in-Differences for causal identification. We find that both policies reduce the share of land devoted to housing within the original slum perimeter — in-situ upgrading by reallocating land toward neighborhood infrastructure, and population relocation by reducing the number of slum households and housing structures. However, only in-situ upgrading generates durable improvements in housing quality and the socioeconomic status of residents. In addition, upgrading produces substantial positive spillovers in nearby neighborhoods, increasing formal housing investment and reducing crime. Moreover, administrative cost data show that in-situ upgrading is roughly one-third less expensive per household than population relocation, underscoring its greater cost-effectiveness.