Economic Geography

An evaluation of the Krugman model
One of the obvious failures of earlier neoclassical theories in economic geogra-
phy and regional science is that their commitment to perfect competition and
62
Allen J. Scott

---

constant returns induced them to overemphasize the divisibility of economic
activities, leading in turn to a radical underemphasis of agglomeration as a force
in shaping the economic landscape. Fujita and Thisse (2002) describe the space-
economy as seen through neoclassical spectacles as a tending to a system of ‘back-
yard capitalism’.
The originality and value of the Krugman model as an approach to spatial
analysis is its formulation of the problem in terms of monopolistic competition
and increasing returns within the firm. The notion that agglomeration has its
roots at least partially in monopolistic competition is particularly interesting, and
corresponds well with the character of much industry today. Modern sectors
such as high-technology manufacturing, business and financial services, cultural
products, and so on, are especially prone to form distinctive clusters, and it is
exactly in such sectors that we find the high levels of product variety, intra-sectoral
trade, and the drive to market extension that characterize monopolistic competi-
tion. At the same time, the core model breathes new life into the notion of pecu-
niary externalities as originally formulated by Scitovsky (1954). The emphasis on
agglomeration as an outcome of the complex pecuniary effects of Chamberlinian
competition and internal economies of scale is unquestionably the model’s prin-
cipal claim to theoretical significance, and it is all the more interesting because it
sets these within a framework of multi-region interdependencies.
Once all of this has been said, many reservations remain. At the outset let me
state that I do not share the inclination of some geographers to discard the new
geographical economics simply on the basis of its commitment to a priori forms
of deductive theorizing. In practice, of course, such theorizing sometimes turns
in upon itself in highly dysfunctional ways, and economists are notorious for their
cultivation of an ingrown professional culture focused on displays of bravura but
vacuous analytics (cf. McClosky 2002). The Krugman model and its derivative
expressions certainly suffer from this syndrome, especially in view of the implau-
sibility and arbitrariness of many of its assumptions, where enormous compro-
mises with reality are made so as to ensure that numerical solutions can be
generated,
3
and it is tempting to reject the model out of hand on the grounds of
its unrealistic assumptions alone. However, I think it better to issue the main chal-
lenge from the basis of a related but slightly different perspective. In other words,
what is this a model of? It may well be a description of life on some planet some-
where in the universe, but what exactly is its relevance to an understanding of
economic realities on planet earth at any time in the past or the foreseeable future?
This question is underlined by the fact that the core model puts the emphasis
on market-driven pecuniary relationships in an equilibrium Chamberlinian
framework. In so far as it goes, this point of departure has the merit of making
inter-regional competitive forces an explicit element of the analysis (as befits its
intellectual origins in international trade theory). Its principal deficiency is that
it fails adequately to grasp at the notion of the region as a nexus of production
relationships and associated social infrastructures from which streams of external
economies of scale and scope continually flow, even in single-region economies, and
even in cases where competition in final product markets is non-monopolistic.