Spatial Inequality and Infection Risk: Urban Determinants of COVID-19 Exposure

This paper investigates how spatial inequalities in urban infrastructure shape the risk of COVID-19 infection. We develop a stylized urban sorting model with regulated housing markets to formally link residential density, healthcare accessibility, and residence in low-income areas under special planning regulations to heterogeneous contagion risks. The model yields testable propositions, which we evaluate using georeferenced individual-level data from Recife, Brazil. To address endogeneity, we estimate an instrumental variable Probit model, using novel instruments based on 19th-century railway lines, early-2000s building density, and the share of designated low-income areas within each census tract. Results show that greater distance to health services, higher residential density, and residence in low-income neighborhoods each increase infection risk. The effects of distance and density are strongest among individuals aged 20–30, while the low-income area effect is more pronounced among those under 20; combined vulnerabilities amplify risks for older adults. Additional analyses reveal that risk varies systematically across land-use contexts. Our findings demonstrate that urban form and the spatial distribution of public services play a causal role in epidemic exposure, offering lessons for public health strategies in rapidly urbanizing and unequal cities worldwide.