The economics of urban size.

Abstract

An aggregative economic approach to the theory of city size is presented along with some empirical findings which suggest that even the largest cities have not yet reached excessive sizes from the point of view of growth and productivity. Urban magnitude is no simple 1 dimensional phenomenon. Modern urban centers are surrounded by very large, diffuse zonal boundaries, within which there are marked variations in the proportion of firms and people associated with that center, and in the intensity of the association. In sum, population does not constitute a conventional, countable set. In general, population will be considered as the basic magnitude and as a conventionally definable number. Most approaches to city size have emphasized the presumed diseconomy of urban scale and have sought to establish that population at which costs per capita are least, regarding this as optimal. It is argued here that both the logic and the factual basis of this approach are faulty. The argument of minimum costs is insufficient in its own terms. Such an objective is reasonable only if output per capita is constant, but it appears that output per capita is an increasing function of urban size. In that case, a more sensible objective of public policy would deal with the relation of outputs and inputs, rather than only with inputs. In every country for which evidence was found, local product per capita (or some index for it, such as income or wages) rises with urban size, and where comparable figures on cost are available, these rise far more slowly if at all. Although all of the data desirable are not available for any single country, the overall pattern is clear. Possibly the most surprising element in the data is the marked decline with increasing density in Social Overhead Captial Stocks (SOCS) per capita. This runs counter to common belief that the bigger the city the more infrastructure per capita is needed and may be the result of such effects as the greater linear quantities of roads per capita necessary in low density areas. Since it appears that the biggest cities are not too big from the perspective of economic efficiency, it may be that higher average incomes of bigger cities do not mask sharper inequalities among their citizens, so that efficiency is gained at the cost of equity. This does not appear to be the case at least for the US. On the contrary, some studies indicate that there is less poverty and a more equal distribution of incomes in big cities than in smaller cities. There are some weak indications of a downturn in the product curve at the largest urban sizes, but even should the downturn be real, this would not be inconsistent with the efficiency of those larger sizes in a hierarchical system of cities.