Some Aspects on the Interregional Spatial Distribution of Local
Sector Activities (preliminary version)
Arnstein Gjestland, IngeThorsen, and Jan Ub0e
Stord/Haugesund College, Bj0rnsonsgate 45, N-5528, Haugesund, Norway}
Abstract
This paper focuses on the spatial distribution of economic activities that serve intraregional demand.
The level of local sector activities is measured by employment per inhabitant. The basic hypothesis
relates to how this proportion varies systematically over space, from high values in a central business
district, through low values in suburban areas, and asymptotically approaching the average regional
level as the distance from the center increases. This hypothesis is examined both analytically and
through simulation experiments where the location decision of firms is assumed to reflect the net
effect of agglomeration economies, economies of scale and transportation costs. We further discuss
to what degree the relevant hypothesis is consistent with specific assumptions on the distribution of
wages (transport cost) between consumers, and the spatial distribution of the prices on goods and
services. The model formulation we propose is useful as a part of economic base modeling for
predicting regional development.
1. Introduction
In the literature one can find extensive results on the modeling of locational decisions of firms and
households. Complementary to this, much research is directed towards the modeling and prediction
of traffic flows. Surprisingly little attention, however, is paid to the important issue of treating
location decisions and traffic flows simultaneously within the same modeling framework.
One area of research where this issue has been addressed, is the formulation of large-scale models for
urban structure and development. In almost a decade after Lee's “Requiem for large-scale models”
(Lee 1973), the activity within this area of research was very low. After this period there was a
renaissance of large-scale models, see for example Boyce (1988). This renaissance was a result of
progress within mathematical methods, solution algorithms and computers, estimation procedures
and data, and the theoretical framework. Concerning the theoretical framework it is particularly
important that standard methods of spatial interaction analysis has been demonstrated to be
consistent with utility theory, see for example Ben Akiva and Lerman (1985). During the last decade
a multitude of large-scale models has been constructed and applied for specific metropolitan areas,
for reviews see Wegener (1994, 1998), or Batty (1994).
Many of the operational large-scale models are based on ideas from the Lowry model (Lowry 1964).
Anas (1987) offers a more recent presentation and evaluation of this modeling tradition. The Lowry
model is pivoted on the central idea of economic base theory, where production is split between local
and basic production sectors. The activity level in local production sectors is determined by demand
that originates within the study area, while production in basic sectors is exogenously given,
independent of the relevant intraregional components of demand. Increased activity in basic sectors
attracts workers to the region. This further increases the demand for locally produced goods, and a