Welfare Analysis of Cities : Urban Sprawl, Transportation Pricing, and the Optimal Rawlsian Town

This paper studies the Rawlsian first-best allocation in a monocentric city model using a unifying framework of land ownership. We show that a Rawlsian planner would not choose the market outcome, except for the extreme case of public land ownership in which all the differential rent is transferred in lump-sum fashion to residents. In any other case, there is an outcome with equal utility for all city residents that brings higher welfare than the market outcome. In particular, it holds in the traditional textbook formulation with an absentee landlord that owns the land. We also show that the first-best scenario can be decentralized with a revenue-neutral combination of location-specific taxes and subsidies. This instrument may produce a Rawlsian first-best city that is more extended than the market city. Thus, depending on the structure of land ownership, in the absence of externalities, the market equilibrium city may be inefficiently compact. Then, when externalities are present, policies aimed to restrict urban sprawl should take this effect into consideration. To study the relevance of our results, we assess welfare-maximizing transport pricing policies in the absence of location-specific taxes. For public transport, we show that the fare that decentralizes the first-best scenario is below marginal cost, and thus the system should be subsidized, even in the absence of externalities. In the case of car congestion pricing, we show that the welfare-maximizing toll may be non-monotonic, yielding a city that is more extended and with more aggregated mileage than the unpriced city