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

The model described above differs only slightly from the one in Krugman's (1991a) influential paper. [7] Indeed, most work in the new economic geography consists of theoretical wrinkles on the preceding story. There is some, however, which also endeavors to be a bit more applied: one example is Krugman and Livas (1996), who view Mexico through the lenses of the new economic geography, attempting "to explain why so much population and industry concentrated in Mexico City in the first place" (p. 138). Their basic set-up is similar to the model described above insofar as the Dixit and Stiglitz monopolistically competitive market structure underlies everything, fulfilling criteria (1) through (3). The geometry that fulfills criterion (4) is different, though: there are three regions situated at the apexes of an isosceles triangle pointed upwards, the uppermost apex representing the United States and the lower two representing Mexican regions. Moreover, while labor is mobile between the two Mexican regions, allow ing the model to meet criterion (5) in the same fashion as the model described above, the dispersing force is different: in this case, bearing in mind the reality of Mexico City, it is the cost of congestion in the more-agglomerated region that constitutes the dispersing force, not a partially dispersed output market.

The model's insight is that the high tariff barrier maintained by Mexico prior to 1986 may have played a role in promoting excessive agglomeration in Mexico City. Cut off from the United States, immense agglomeration in one of the two Mexican regions was practically inevitable. Under free trade, firms would have wanted to locate wherever they could serve the U.S. market more cheaply; but as it happened, given high trade barriers, they wanted to serve the Mexican market--which was to be found wherever Mexican firms had already located.

The insights the model shares with others in the new economic geography are the cumulative logic of agglomeration once it gets started; the possibility of multiple equilibria, meaning that if there is to be agglomeration it could just as well be in one region as in another; and given the latter possibility, the dependence of the actual equilibrium on initial conditions or "history". Mexico City rather than Puebla is Mexico's primal city, in other words, not because of contemporary advantages to Mexico City's geographical location (to the contrary!), but because of a small difference in the cities' relative sizes that appeared in the late nineteenth century and then expanded cumulatively and spectacularly. [8]

The question of whether these insights into agglomeration are profound, relevant, and novel--or instead "obvious, wrong, and anyway they said it years ago"--remains to be addressed. If not a definitive answer, an advantageous perspective can be gained by comparing the approaches of Gunnar Myrdal and Francois Perroux to the same subject.

II

Perroux

ZAFIROVSKY (1999, p. 596) LISTS FRANCOIS PERROUX'S Economie et Societe (1960) [9] among the works of a "distinct economic school (in France) guided by the premises of economic sociology," namely the premise that economic behavior is "embedded in institutions, norms, and values." Of course few economists would disagree that such factors are important, but nevertheless most have proved willing, for the sake of mathematical modeling, to set them aside or make severe simplifying assumptions about their influence. Francis Perroux (1903-1987) was an exception. His analysis, not only in Economie and Societe but throughout his oeuvre, was based on the belief that neoclassical economics has consigned itself to irrelevancy because it abstracts away from the most important aspects of economic behavior. One of the most fundamental and omnipresent characteristics of market interactions, claimed Perroux, is asymmetry of power between the actors. Asymmetrical power relations are present in interactions between similar decis ion making units--whether individuals, firms, regions, or nations--just as commonly as between dissimilar units. In this way Perroux's ideas about regional agglomeration stemmed from his economic sociology: to him, regional agglomeration resulted from, and was itself a manifestation of, power exercised in geographical space.