Geography Reference
In-Depth Information
otherwise very useful attempt to distinguish between dif erent forms and degrees of path dependence. He
states: 'of course a process need not attain an equilibrium distribution or outcome, but for the purposes
of this essay I restrict attention to that case [of equilibrium], with one exception' (p. 92).
8. More explicitly, David dei nes path dependence in terms of non-ergodic Markov chain models that possess
two or more absorbing states. An absorbing state is one from which there is a zero probability of exit or
movement to another state. In such models, the limiting distribution depends on the initial distribution,
that is on where the system started: dif erent starting distributions will result in dif erent limiting (equilib-
rium) distributions: hence the idea of multiple equilibria. We should note here, however, anticipating the
discussion that follows, that absorbing Markov chains are only one form of non-ergodic process, and that
in general the latter do not have to possess any long-term limiting distribution or stable equilibrium state.
9. The literature on NEG models is now extensive, but for representative expositions see Fujita et al. (1999),
Baldwin et al. (2003), Henderson (2005) or Brakman et al. (2009). For a recent assessment of these models
from a 'proper' economic geography perspective, see Martin (2010a).
10. Although Setteri eld makes this argument in the context of deterministic systems, it applies equally to
stochastic systems like the absorbing Markov chain processes that David uses to conceptualise path
dependence.
11.
Of course, it could be argued that equilibrium economics has long moved on beyond traditional equi-
librist methodology, and while retaining its focus on the existence and achievement of equilibrium, has
come to embrace the idea that history matters through the construction of models of sequence econom-
ics, temporary equilibrium, and non-tatonnement processes. But as Setteri eld retorts, these extensions
actually constitute 'nothing more than glorii ed conventional disequilibrium analysis and are roughly in
keeping with equilibrist methodology . . . the language of the analysis may be that of sequences and tem-
porary equilibria, but the substance does not involve path dependency, as the economy is held ultimately
to converge to a predei ned, “fully adjusted” equilibrium position' (Setteri eld, 1997, p. 62).
12.
Bassinini and Dosi's
discussion falls victim to this last problem when they seem to invoke radical techno-
logical shifts - between steam power and electricity, and between thermionic valves and semi-conductors
- as examples of 'temporary' resting states.
13.
As Schumpeter (1944) insisted, transformation arises from
within
the socio-economic system, and adap-
tive development is the primary process by which this occurs.
14.
Ramlogan and Metcalfe make the point that if knowledge ever did reach an equilibrium state then all
economic change and development would cease.
15.
Complexity theorists usually use the term hysteresis rather than path dependence. Though not identical,
the two notions are related.
16.
The rise of agent-based models within some quarters of neoclassical economics can be argued to be a
manifestation of recognising the value of assigning more importance to economic process than to end
states (see, for example, Axtell, 2007).
17.
We prefer the phrase 'non-equilibrium' here, rather than 'disequilibrium' or 'out-of-equilibrium', since
both of the latter continue to imply that there is some latent equilibrium outcome to which path depend-
ence would lead if only the system in question were free to do so. The term 'non-equilibrium' is intended
to imply that no such latent equilibrium may exist at all.
18.
It is analogous in a sense to Paul Davidson's (1991) post-Keynesian economics in which decision-making
is so non-ergodic that the past is no real guide to the future.
19.
In evolutionary economics, adaptation refers to a process of response to the selection environment so
that adapted entities i t the environment and are apt or i t for purpose (Metcalfe, 2005). Adaptability,
on the other hand, refers to 'the potential to adjust to changing circumstances in an appropriate way:
it is about the capacity to respond to changes in the selection environment: to maintain good design'
(Metcalfe, 2005, p. 414). However, economic adaptation remains a layered and contested concept: does
it refer, for instance, to the process in which entities ensure that they are i t for market survival or to the
process of using resources appropriately and satisfying needs, as clearly the two may overlap but are not
synonymous? It is important therefore to clarify the precise i rm strategy taken in response to a specii ed
environment and to examine its implications for adaptability.
20.
The exploration of alternative approaches to path dependence being developed in political science and
historical sociology is taken up in more detail in Martin (2010b).
References
Abbott, A. (2001),
Time Matters: On Theory and Method
, Chicago: University of Chicago Press.
Agarwal, R. (1998), 'Evolutionary trends of industry variables',
International Journal of Industrial Organization
,
16
(4), 511-25.
Agarwal, R. and M. Gort (1996), 'The evolution of markets and entry, exit and survival of i rms',
Review of
Economics and Statistics
,
78
(3), 489-98.
