From Stylized to Quantitative Spatial Models of Cities

This paper describes the progression from a standard monocentric model of a city to its analog in the quantitative spatial framework recently reviewed by Redding and Rossi-Hansberg (forthcoming). In this progression, we preserve the basics of preferences, technology, and endowments across models. The monocentric model allows for many of a city's characteristics to be endogenous, including size, population, wages, and commercial and residential land rents, but it is also highly stylized. In contrast, quantitative spatial models impose far fewer restrictions in the way that these variables capture a city's structure. In particular, they allow firms and residents to potentially locate in any part of the city and residents to commute between any two locations. Quantitative spatial models, therefore, can more accurately capture the distribution of economic activity across space. We describe how to match widely available urban microdata to the spatial model we lay out, an exercise that ensures that any counterfactual policy experiment is grounded in a framework consistent with the city's current allocations and prices.