Urban patterns, population density and optimal city dimension: The case of public infrastructure

Determination of the optimal city size underlies the economic rationality of infrastructure provision by local governments. We investigate the existence of decreasing average costs resulting from economies of scale, associated with larger urban dimensions in terms of population and housing, and economies of density, brought about by reductions in urban dispersion, and calculate optimal population densities when providing basic infrastructure. The methodology relies on novel definitions of scale and density economies and their estimation by way of flexible translog cost functions, extensively applied in the literature dealing with the provision of services—i.e., utilities, but extended here to their supporting infrastructure. Our results unveil the existence of latent economies of scale and density resulting in a cost excess in the provision of infrastructure due to the effect of urban sprawl that translates into suboptimal city sizes. Based on these findings several policy guidelines rationalizing urban development are suggested. The model is illustrated using Spanish statistical data collected from the nationwide local infrastructure and equipment survey, and prices from a new database that uses engineering cost benchmarks.