Document file:///D|/Export1/www.netlibrary.com/nlreader/nlreader.dll ...

Document
Page 105
What we get from this static exercise is an indication of how the parameters of the model ought to affect
the balance between the centripetal forces that favor agglomeration and the centrifugal forces that
oppose it. Agglomeration is favored by low transport costs (low τ), a large share of manufacturing in
the economy (high µ) and strong economies of scale at the level of the firm (low σ).
This is, however, only an analysis of a static, two-location equilibrium. We now turn to dynamics in a
multi-location example.
Dynamics in a Multi-Location Model:
The Economy as a Self-Organizing System
We now turn to a dynamic, multi-location model. In doing so we find that paper-and-pencil analysis
will no longer suffice, and must be supplemented with numerical methods. Since the model has only a
few parameters, however, it is not hard to use numerical methods to explore its properties fairly
thoroughly; and the numerical results are easy to understand given the intuition developed in the twolocation
case.
We assume, then that there are J › 2 locations, and we return to the assumption that agricultural workers
are equally distributed among the locations, with a share 1/J in each.
In a many-location model it is necessary to specify the matrix of distances between locations. I choose
the simplest setup that preserves symmetry: the locations are equally spaced around a circle, with
transportation possible only along the circle's circumference. We let the distance between any two
neighboring locations equal 1. In the numerical examples described shortly, we consider in particular
the case of 12 locations, laid out like a clock face.
file:///D|/Export3/www.netlibrary.com/nlreader/nlreader.dll@bookid=409&filename=page_105.html [4/18/2007 10:31:53 AM]