Geography Reference
In-Depth Information
produced by repeated iteration of the system over time. In this sense, there is no single ex
ante unique distribution of the economic system, as in general equilibrium economics:
the actual limiting distribution depends on history, and particularly where the economic
system (for example, the technological or industrial structure) started. Further, according
to David, once the processes of path dependence have locked the system (technology or
industry) into one of these alternative stable equilibria (or 'attractors'), it requires an exter-
nal shock to break the particular development path or trajectory in question. The evolu-
tion of the technological, industrial and institutional structure of an economy in this path
dependence model would thus seem to be akin to 'punctuated equilibrium', a (stochastic)
pattern of historically contingent evolution in which particular paths of technological,
industrial and institutional development become 'selected', locked-in to stable equilibria,
and, at some point, eventually disrupted and broken up by external shocks. As David puts
it:
Sudden shifts in structure, corresponding to the new evolutionary biologists' notion of 'punctu-
ated equilibria' . . . may open up a way for the formulation of dynamic models that are compat-
ible with 'stage theories' of development. (2005a, p. 187)
What is interesting about this 'path-dependent equilibrium analysis' is that it resonates
with how notions of 'path dependence' and 'history' have been used in the so-called 'new
economic geography' (NEG) that has arisen over the past two decades. Adherents of this
approach to analysing the space economy are explicitly equilibrist in orientation, assert-
ing that it is 'the general equilibrium modelling of an entire space economy which sets
the approach apart from traditional location theory and economic geography' (Fujita
and Mori, 2005 p. 380). Moreover, while committed to this equilibrist representation of
the economic landscape, NEG theorists simultaneously claim to recognise that 'history
matters' and that their models incorporate 'path dependence'. The way this is achieved is
precisely by constructing (deterministic) models that yield multiple equilibria outcomes
(in this case equilibrium spatial patterns of industry and employment) depending on the
'initial conditions' (such as the level of transport costs, the relative mobility of labour and
capital, the initial distribution of industry between regions, and the number of regions)
specii ed in the models.
9
For example, the basic two-region 'core-periphery model', the
foundation of NEG theory, possesses i ve such possible equilibria (three stable and two
unstable). Because the 'initial conditions' determine which of the alternative equilib-
rium landscapes emerges, NEG theorists feel able to claim that 'history matters' in their
models. And because these models generate their equilibrium outcomes via processes
of self-reinforcing agglomeration processes, it is also claimed that they embody path
dependence (or 'locational hysteresis').
Now the notion that 'initial conditions' matter is certainly an improvement over the
timelessness of conventional equilibrium economics, since potentially it at least draws
attention to one aspect of the past history of a system in the determination of future
outcomes. However, to our mind, characterising path dependence as a problem of mul-
tiple equilibria (whether stochastic, or deterministic as in NEG models) is to restrict the
concept and its relevance for analysing the evolution of regional and local economies.
In David's work, as in NEG models, multiple equilibria are conceived as representing a
system that possesses a variety of 'locally stable attractors' (to use his term), from which
the i nal equilibrium position is 'selected' on the basis of the system's initial starting
