In this paper, the stylised assumption that one single ‘optimal’ city size exists for all cities – achieved when marginal location costs equal marginal location benefits – is abandoned, as well as the opposite view that each city operates on its own cost and production curves, defining a specific optimal size. Instead, this work maintains the comparability among cities and demonstrates that urban specificities in functions performed, quality of life, industrial diversity and social conflicts shift up and down the benefits and costs linked to pure physical size, leading to different ‘equilibrium’ sizes for cities. A model of equilibrium urban size is set up, and empirically estimated on a sample of 103 Italian cities with data at NUTS3 level. Empirical results verify the empirical model on the analysed sample; results hold both with standard OLS estimates as well as with the use of instrumental variables in order to correct for the possible endogeneity of some of the variables in the model. Differences between predicted and real city size are interpreted with good or bad governance, thereby suggesting future strategies for more efficient urban planning.
Keywords: equilibrium city size, spatial equilibrium, urban functions, city networks