Modeling Agglomeration and Dispersion in City and Country: Gunnar Myrdal, Francois Perroux, and the New Economic Geography - Critical Essay

American Journal of Economics and Sociology, The,  Jan, 2001  by Stephen J. Meardon

• E-mail
• Print
• Link

Benson and Hartigan's extension thus allowed them to maintain the intra-industry trade result of other new trade theorists while locating their model in a slightly more real (one-dimensional rather than no-dimensional) spatial setting. Incorporating space also allowed them to address topics other than intra-industry trade, namely the spatial distribution of the effects of a tariff on the real incomes of consumers. Relating their work to the new economic geography, it would appear they met criteria (2) and (4). They missed (1) only because in their free-entry version of the model there is no economy-wide resource constraint. They did not meet (3) or (5). This is not to suggest however that their work was somehow less evolved than the new economic geography. But, they did not at all aspire to write in that genre, not only because it did not exist at the time, but also because, like Losch, they wished to address different questions.

The same is true of products of the new trade theory, many of which resemble the new economic geography even more closely than Benson and Hartigan's models. Indeed there are some works--notably by Krugman (1980), Avinash Dixit and Victor Norman (1980, ch. 9), and Elhanan Helpman (1981)--that meet all the criteria but (5). These too are addressed to different questions, however: again, chiefly intra-industry trade and other consequences to international trade of increasing returns to scale. Like the models of Losch and Benson and Hartigan, then, it would be inappropriate to lump these in with the new economic geography. At this point, though, it is important to note that we are now considering literature from which the new economic geography differs by little more than a tweak of the model and a change of emphasis. In fact that is precisely how the new economic geography evolved.

In sum, the five criteria listed previously for the definition of the new economic geography serve not only in themselves to distinguish the literature from others; in addition they reflect the questions towards which the literature is directed, which also distinguish it from others. Many other "spatial" works, though they employ models that are similar in many respects, do not share the new economic geography's concerns: their models are intended to help explain market areas, or spatial competition, or intra-industry trade. The new economic geography asks instead: aside from the usual "geographical" determinants like rivers and mountains, what determines the spatial agglomeration or dispersion of economic activity?

The first step towards answering the question is to define "agglomeration" in the context of the model. Agglomeration is taken to mean that more monopolistically competitive firms are located in one region than in another. Increasing returns to scale at the firm level implies that each monopolistic competitor will be the sole producer of its own variety, which it can choose to produce in one region or another. If firms agglomerate in one or a few regions, they do so impelled by pecuniary externalities that arise from the interaction of increasing returns with transportation costs between regions.