This book explores the implications of the new economic geography for the theory of economic
policy. As the authors note, much of the interest in this topic comes from its potential importance
for policy, yet previous writers, including the originators of the approach, have avoided explicit
discussion of policy issues. (See, for example, Fujita et al. (1999).) The authors argue persuasively
that there are two reasons for this. First, the original new economic geography models cannot be
solved analytically, so recourse to numerical methods is needed even to determine their properties
when countries are ex ante identical, far less to solve the inherently more complex problems that
arise in applying them to policy issues. Second, the central externalities in new economic
geography models imply that “Agglomeration is unambiguously good for you”. (The authors
generously credit this argument to Neary (2001), but I am sure it has been implicit in the literature
from the beginning.) Firms want to locate near their customers, workers want to locate where the
cost of living is low. Taken together with the standard assumptions of new economic geography
models (Dixit-Stiglitz preferences, increasing returns, and iceberg transport costs for
manufactures), these imply that prices and/or costs will be lower in more agglomerated locations.
This may be plausible in general, but its implications for policy are “too stark to be true”: any policy
(such as protection) which encourages in-migration of mobile factors will lower prices and raise
welfare. Hence the policy implications of standard new economic geography models seem
alarmingly like a restatement of import-substitution-led industrialization, which we know from the
sad experience of many countries is rarely successful in practice.