This paper develops a quantitative model of internal city structure to analyse the consequences and costs of weak urban land rights in a developing country city. The land tenure system in different areas of Kampala, Uganda, strongly correlates with the density of different types of households and firms. Using a spatial Computable General Equilibrium Model, we develop a simulated version of a city with different firms and households optimising their location decisions across the urban space, when faced with transport costs for shipping goods, commuting costs, and a fixed supply of urban land. The model is used as a benchmark against which evidence from Kampala can be compared. Calibrating the model on data on household and firm location reveals, through the land rents, the absolute advantage or disadvantage of each area of the city. Productivity and amenity parameters that vary across the city are also measured through the calibration, measuring the comparative advantage firms or households of different types have in a particular area. These parameters are then explained using city-wide variation in land tenure systems. The evidence suggests that the presence of customary land and Mailo land reduces economic activity, particularly in tradable services, while increasing residential densities, particularly low-income informal neighbourhoods. A land market reform which converted Mailo land to leasehold land could have a very large impact on the welfare of the city through improved land allocation. Joint with Tony Venables and Louise Bernard.