Very large urban centers are a conspicuous feature of many developing
economies, yet the subject of the size distribution of cities (as oppo
sed to such issues as rural-urban migration) has been neglected by dev
elopment economists. This article argues that some important insights
into urban concentration, especially the tendency of some developing c
ountries to have very large primate cities, can be derived from recent
approaches to economic geography. Three approaches are compared: the
well-established neoclassical urban systems theory, which emphasizes t
he tradeoff between agglomeration economies and diseconomies of city s
ize; the new economic geography, which attempts to derive agglomeratio
n effects from the interactions among market size, transportation cost
s, and increasing returns at the firm level; and a nihilistic view tha
t cities emerge out of a random process in which there are roughly con
stant returns to city size. The article suggests that Washington conse
nsus policies of reduced government intervention and trade opening may
tend to reduce the size of primate cities or at least slow their rela
tive growth.