Yet even more evidence on the spatial size of cities: Urban spatial expansion in the US, 1980–2000

a r t i c l e i n f o JEL classification: R14 R52 Keywords: R14 Land use patterns R52 Land use and other regulations This paper expands empirical testing of the predictions of the standard monocentric urban model to examine the size of urban spatial expansion for all US metropolitan regions for the years 1980, 1990 and 2000. Until recently, the lack of temporally and spatially-consistently interpreted data on urban extent for U.S. metropolitan areas has limited the ability to estimate panel models of land use change. This paper first describes the creation of a consistent data set on urbanized land area for all U.S. metropolitan areas. It then estimates cross-sectional models of the determinants of urban spatial extent. For purposes of comparison, results are compared to temporally-consistent satellite-derived measures of developed land area from the National Land Cover Database. The paper then presents panel-model estimates of changes in urban land area, finding that both population and income elasticities are substantially less than unity. Tests for the structural stability of urbanized land area determinants across metropolitan size categories and polycentric regions are presented. Measuring and understanding the drivers of the spatial expansion of urban land development is of great interest not only to urban and regional economics, but also to fields as diverse as landscape ecology, water quality, transportation, and climate change (Kalnay and Cal, 2003). Climate models, infrastructure spending, transportation models, and water quality models all involve some projection of future land uses, densities, and the spatial patterns of urban develop-ment—all without any empirical consensus as to relative elasticities of urban land demand. For example, a recent National Research Council (2009) report presented a number of future urban density and land scenarios in its climate model, with no agreement among experts as to what future urban land use would be, while suggesting the need for additional research to estimate finer-grained changes in urbanized land areas in metropolitan areas. The traditional monocentric " Alonso–Mills–Muth " (AMM) model of urban spatial structure continues to serve as a core theoretical and empirical concept of urban land development within and across metropolitan areas because of its elegance, simplicity, and empirical Because of the difficulty in acquiring historically consistent measures of urbanized land area extent, however, the model has only been empirically tested a handful of times. For the US, Brueckner and Fansler (1983) first estimated the monocentric model with cross-section data …