摘要：The traditional trade theory—and indeed the theory of
the firm—begins with three critical assumptions that
make geography irrelevant to the choice of location for
most industries. These are constant returns to scale, full
and costless information and the absence of externalities.
In a world of constant returns, production is
highly divisible, and there are no penalties associated
with setting up a small plant to serve a local market.
With full and costless information, it is not important to
be physically close to purchasers or suppliers. In the absence
of externalities, there is little to be gained by being
in close proximity with other producers, and possibly
something to be lost to increased competition by
locating near firms in the same industry. While these
assumptions used to be standard, they are now
recognized as seriously misleading. In the real world of
industry and international trade, firms locate in
industrial clusters and in cities. The relevance of that
phenomenon for trade and industrial development has
only recently been rediscovered as part of the “new
economic geography” (Krugman, 1991; Fujita, Krugman
and Venables, 1999).
