Public Roads on Private Lands: Land Costs and Optimal Road Improvements in Urban Uganda

Résumé

Despite the need for transportation infrastructure investments in developing cities, empirical evidence on their net returns is lacking due to data constraints and the common oversight of land acquisition costs. In this paper, I collect novel data to estimate the net returns of 140 km of road improvements in Kampala, Uganda, since 2017, accounting for both benefits and land acquisition costs. I conduct two surveys with real estate brokers and landowners and I exploit variation in the timing of improvements to estimate the local benefits. I then develop a quantitative spatial model to capture the city-level impacts of the policy, accounting for general equilibrium effects and heterogeneous land acquisition costs. Leveraging the coexistence of three property rights regimes in the city, I show that weak property rights are associated with lower land acquisition costs. I find that the net welfare gains from the realized road improvements were equivalent to a $119 transfer per resident but would have been negative if land had been acquired at market value, as legally mandated under eminent domain, due to the high cost of raising domestic funds. Finally, I solve for the optimal road improvements under different institutional settings and demonstrate the importance of accounting for land costs when designing, funding, and evaluating transportation infrastructure projects, particularly in low- and middle-income countries where land acquisition relies on fragile land and financial institutions.