Many communities on the urban fringe are implementing a range of policies to preserve farmland and open space, cluster residential development, and guide development to areas with existing infrastructure. These efforts are an attempt to control overall growth and the concomitant loss in open space and also to counter a trend toward the so-called large lot development that often takes place in these areas. Planners have argued that policies to manage density are the most important local policy focus for urban areas in the coming years. It is possible that large lot development and sprawl are themselves the result of government policy. Most local governments use zoning to establish minimum acreage requirements for each residential dwelling unit; in ex-urban localities, these limits are often quite high. Developers might build a subdivision with average lot sizes greater than the minimum but they cannot by law go below it. Some researchers have argued, however, that the spatial patterns of development are simply the natural result of household preferences and market forces. In this paper, we address the question of whether zoning limits are the primary cause of lowdensity, sprawling development or whether market forces tend to dictate this outcome. If zoning limits account for low-density development in at least some cases, how would development patterns be different if there had been no such rules? We begin by constructing a simple model of the developer decision about the density of new development. The subdivision is the unit of observation, and developers must weigh both demand and cost considerations in choosing density, in addition to complying with zoning restrictions that vary across parcels. We apply the model using parcel-level data from a region where zoning rules vary but are exogenous to the period under study. Calvert County, Maryland, near Washington, DC, is an historically rural county that has experienced rapid growth in recent years. The county has a transferable development rights (TDRs) program that has led to a great deal of variability in the intensity of development across properties. We are able to not only examine the extent to which zoning has contributed to large lot development but also to determine the economic forces that underlie density decisions. Finally, we are able to forecast how density would have been different in the absence of zoning rules by estimating a Tobit equation that is censored for the observations constrained by zoning.