You are on page 1of 14

This article was downloaded by: [University of Winnipeg]

On: 04 September 2014, At: 17:07


Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer
House, 37-41 Mortimer Street, London W1T 3JH, UK

Regional Studies
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/cres20

The Regional Economy, Spatial Structure and


Regional Urban Systems
John B. Parr

School of Social and Political Sciences (Urban Studies) , University of Glasgow ,


Glasgow , G12 8QQ , UK E-mail:
Published online: 18 Jun 2013.

Click for updates


To cite this article: John B. Parr (2013): The Regional Economy, Spatial Structure and Regional Urban Systems, Regional
Studies, DOI: 10.1080/00343404.2013.799759
To link to this article: http://dx.doi.org/10.1080/00343404.2013.799759

PLEASE SCROLL DOWN FOR ARTICLE


Taylor & Francis makes every effort to ensure the accuracy of all the information (the Content) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of
the Content. Any opinions and views expressed in this publication are the opinions and views of the authors,
and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied
upon and should be independently verified with primary sources of information. Taylor and Francis shall
not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other
liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or
arising out of the use of the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions

Regional Studies, 2013


http://dx.doi.org/10.1080/00343404.2013.799759

The Regional Economy, Spatial Structure and


Regional Urban Systems
JOHN B. PARR
School of Social and Political Sciences (Urban Studies), University of Glasgow, Glasgow G12 8QQ, UK.
Email: John.Parr@glasgow.ac.uk

Downloaded by [University of Winnipeg] at 17:07 04 September 2014

(Received July 2011: in revised form March 2013)


PARR J. B. The regional economy, spatial structure and regional urban systems, Regional Studies. In approaching the concept of the
regional urban system, attention is initially drawn to the better-known types of economic region. The distinctive nature of the
regional economy is next examined, and it is argued that its spatial structure represents an important dimension. Spatial structure
can be characterized in a variety of ways, the most comprehensive of which employs the perspective of an urban system. This is
examined rstly in terms of particular models from location theory, which provide important points of reference, and then within
the setting of the present-day city-region.
Regions

Regional economy

Spatial structure

Regional urban systems

PARR J. B.

PARR J. B. Lconomie rgionale, la structure spatiale, et les systmes urbano-rgionaux, Regional Studies. Pour aborder la notion de
systme urbano-rgional, on porte une attention particulire dans un premier temps aux typologies de rgion plus connues. Il
sensuit un examen des particularits de lconomie rgionale, et on afrme que sa structure spatiale constitue un facteur
important. On peut caractriser la structure spatiale de diffrentes manires, dont la plus exhaustive est du point de vue dun
systme urbain. On lexamine primo partir des modles particuliers puiss dans la thorie de la localisation, qui fournissent
dimportants points de repre, et secundo dans le cadre de la cit-rgion daujourdhui.
Rgions

conomie rgionale

Ossature spatiale

Systmes urbano-rgionaux

PARR J. B. Regionalwirtschaft, Raumstruktur und regionale Stadtsysteme, Regional Studies. In dieser Untersuchung des Konzepts
des regionalen Stadtsystems richtet sich der Augenmerk zunchst auf die bekannteren Typen der Wirtschaftsregion. Anschlieend
werden die charakteristischen Merkmale der Regionalwirtschaft untersucht, und es wird die These aufgestellt, dass es sich bei ihrer
Raumstruktur um eine wichtige Dimension handelt. Die Raumstruktur lsst sich auf vielfltige Weise charakterisieren; bei der
umfassendsten dieser Methoden kommt die Perspektive eines Stadtsystems zum Einsatz. Diese wird zunchst mit Hilfe
bestimmter Modelle der Standorttheorie untersucht, die wichtige Anhaltspunkte bieten, und anschlieend innerhalb der
Umgebung der heutigen Stadtregion.
Regionen

Regionalwirtschaft

Raumstruktur

Regionale Stadtsysteme

PARR J. B. Economa regional, estructura espacial y sistemas urbanos regionales, Regional Studies. Al analizar el concepto del sistema
urbano regional, prestamos atencin a los tipos ms conocidos de la regin econmica. A continuacin se examina la naturaleza
distintiva de la economa regional y se postula que su estructura espacial representa una dimensin importante. La estructura espacial
puede caracterizarse de varias formas, y la ms completa de ellas emplea la perspectiva de un sistema urbano. Esto se analiza en
trminos de modelos especcos de la teora de la ubicacin que ofrecen puntos de referencia importantes y luego en el
entorno de la regin metropolitana de hoy da.
Regiones

Economa regional

JEL classications: R, R1, R12

2013 Regional Studies Association


http://www.regionalstudies.org

Estructura espacial

Sistemas urbanos regionales

John B. Parr

Downloaded by [University of Winnipeg] at 17:07 04 September 2014

INTRODUCTION
The intention of the following discussion is to consider
some of the more important aspects of the urban system
within the context of a regional economy. There can be
no question that the regional urban system occupies a
prominent position within market economies, in
much the same manner that the metropolitan area
came to play a comparable role during the course of
the nineteenth century. The comments to follow are
based on four interrelated propositions, which collectively form the background motivation for this examination of regional urban systems.
First, despite the upsurge of interest in the region
within the social sciences over recent decades
(STORPER , 1995), the concept of the region remains
shrouded in uncertainty, with opinion sharply divided.
Some view the concept of the region as an inherently
vague one that dees precise denition (WREN ,
2009). Others (probably representing the majority
view) embrace the concept as a exible and therefore
convenient spatial scale that lends itself to a wide
variety of interpretations and applications, not recognizing perhaps that this very exibility may be a source of
imprecision. Second, the regional economy represents
a particular scale of economic organization. It is not to
be seen simply as an element of a disaggregated national
economy: the region contains an internal spatial structure of varying complexity, which is inuenced by the
course of economic and social change, but which may
also exert an inuence over such change. The nature
of the regional economy can only be fully appreciated
if there is adequate scrutiny of its spatial dimension.
Third, in understanding the signicance of regional
spatial structure, the urban system is an especially appropriate perspective, since it is able to incorporate a
number of different approaches. In this connection
much is to be gained from visiting or revisiting those
branches of location theory concerned with urban
systems. To a large extent such frameworks tend to be
overlooked, probably because these are thought to
have little relevance in a world that has undergone
rapid economic and technological change during the
last sixty years. Fourth, the urban-system view of
spatial structure takes on a particular signicance in the
case of the city-region, an increasingly important
feature of the national space economy. Although cityregions display a diversity of forms, both within the individual nation and among nations, striking commonalities are revealed when city-regions are seen through
the prism of the urban system.
In the following discussion the perspective adopted
will be largely of an economic character. As a consequence, important aspects of a political, sociological
and institutional nature are not emphasized. Each is sufciently substantial as to merit separate treatment, and
under ideal circumstances these various strands would
be drawn together. Such a synthesis has yet to emerge,

however; the disciplinary tradition dies hard! In order


to simplify the argument, the major concern will be
with conditions in nations of the economically developed world, and the weaknesses of such a restriction
are acknowledged. The discussion proceeds as follows.
After briey surveying several forms of economic
region, attention is drawn to the regional economy
and the importance of spatial structure in its operation
and development. The nature of the urban system is
next examined, with the initial focus on particular
location-theory frameworks. These provide a platform
for examining conditions within the city-region, a
major feature of the contemporary economic landscape.

THE ECONOMIC REGION


The notion of the economic region is a relatively new
one. It came to be used during the 1920s in the
United States within the elds of agricultural economics
and economic geography, where agricultural regions,
crop cultivation regions and manufacturing regions represented early examples. More rened denitions subsequently appeared.
A preliminary classication

A useful starting point in any discussion of modern-day


economic regions is the tripartite classication proposed
by MEYER (1963). He argued that for the purpose of
economic analysis the following typology was useful:
homogeneous (or uniform) regions; nodal regions; and
programming (that is, policy-related) regions. The
homogeneous region probably comes closest to the
popular view of an economic region as a space throughout which there is a similarity of economic characteristics, and which by implication is distinguishable from
other regions. The regions proposed by OHLIN (1933)
and by NORTH (1955) fall into this category (PARR ,
2008).
The nodal region, by contrast, is based on the internal
structure of its economy. Reduced to its essentials, the
nodal region is composed of two components: the
core or node and the hinterland. The core may represent a service centre or point of supply for the hinterland, which would comprise its market area.
Alternatively and/or additionally, the core can be seen
as a point of demand on the part of households or
rms for goods and services produced in the hinterland
which would function as its supply area. In both cases,
however, the core and the hinterland may contain
economic activities that are primarily related to areas
beyond the region. A general characteristic of the
nodal region is the pronounced interdependence
(direct and indirect) between the core and the hinterland (HOOVER and FISHER , 1949), not simply with
respect to trade, but also in terms of income levels,
capital movements, migration, commuting, etc.1

The Regional Economy, Spatial Structure and Regional Urban Systems


MEYER s (1963) third type of region the policy
region represents a space over which an ofcial body
exercises a jurisdiction in the implementation of a particular policy. While this could represent an existing
political or administrative unit, it might also be a
special-purpose region such as the area of operation of
the Appalachia Regional Commission in the United
States. This tripartite classication of regions is a
common one, and most academic discussions of the
region are concerned with one of the three types.
There is a degree of overlap between the three types
of region, as Meyer clearly recognized. For example,
the policy region could be based on a homogeneous
region or a nodal region.

Downloaded by [University of Winnipeg] at 17:07 04 September 2014

Other regional types

The Meyer classication is incomplete, however,


and mention is made of two other regional types. One
is the region based on a watershed or river basin
(KRUTILLA and ECKSTEIN , 1958). There is a case for
regarding this as a homogeneous region, to the extent
that there may be a dominant pattern of agricultural
land uses. On the other hand, the river has been
regarded as a barrier to interaction, often forming the
basis for a political or administrative boundary.
However, because of such trends as rail and road
communications replacing waterways, and resource
development declining in relative importance, regions
based on river basins have tended to lose much of
their signicance in the organization of space.
Another regional type not tting comfortably within
the Meyer classication is the polycentric urban region
(PUR) (DAVOUDI , 2003; MEIJERS et al., 2003). This has
also been referred to as the dispersed city (BURTON ,
1963) and network city (BATTEN , 1994). Examples
include the Ruhr District of Germany, Randstad in the
Netherlands, and the Research Triangle (Raleigh,
Durham, Chapel Hill) of North Carolina in the United
States. As originally proposed, the PUR represented an
area containing a cluster of urban centres, none of
which had a pronounced dominance over the others,
and each of which had one or more specializations,
usually giving rise to a well-developed pattern of interaction among the centres.2 Relevant here was the
concept of borrowed size, by which smaller urban
centres would be able to benet from their proximity
to a larger centre (ALONSO , 1973). Note that the presence of clustered urban centres is a necessary, though
not a sufcient, condition for the existence of a PUR.
Again, there is a case for regarding the PUR as a homogeneous region, largely on the grounds of its overwhelming urban character, but this characterization
would fail to capture certain key facets of such a
region. An important aspect of the PUR is the extent
to which it is integrated, both morphologically and functionally, a question recently explored in the case of
Randstad by BURGER and MEIJERS (2012). These two

forms of integration are not specic to the PUR, and


may be applied to other types of economic region.
Variation in scale and hierarchical structure

It will be apparent that each of the regional types discussed above may display a considerable variation in territorial extent or relative scale, so that a given regional
type may be seen as occupying a particular position
along a continuum. To emphasize this point, consider
the case of the nodal region. Toward the lower end of
the continuum there exists the functional urban
region (FUR) (KARLSSON and OLSSEN , 2006),
which is often used in labour-market analysis. As with
all nodal regions, the FUR has a corehinterland structure. The core is regarded as a centre of employment,
and is taken to be either the central municipality of an
urban area or the entire urban area, while the hinterland
consists of a series of directly or indirectly adjacent
municipalities. Various methods have been used to
dene this type of FUR. One method is to include
each municipality in the hinterland if it has a
minimum number of jobs per hectare, and if there are
more commuters travelling to the relevant core than
to any other (CHESHIRE and HAY , 1989, p. 15). A less
sophisticated method is to include a hinterland municipality, if at least a certain percentage of its workforce is
employed in the core, this percentage being as low as
10% in a study of European FURs (BRUNEL , 2002).
Toward the other end of the continuum of nodal
regions is the major city-region, a sub-national unit
comparable with those based on urban concentrations
(and not simply their political cities) such as Barcelona,
Birmingham, Bordeaux, Milan and Munich in
Western Europe, and Boston, Denver, Kansas City,
Seattle and Vancouver in North America. Here the
core is represented by a larger city or metropolitan
area that acts as an important centre of ownership,
decision-making and communications, as well a provider of goods and services to the hinterland, sometimes
of considerable extent. Simultaneously, the hinterland,
which contains an urban as well as a rural population,
supplies the node with agricultural goods, raw materials,
and semi-manufactured or nished goods as inputs to
production, and is increasingly associated with the provision of services involving recreation, amenity and
retirement. Under no circumstances, however, should
the city-region be viewed as a closed system or selfsufcient entity, particularly since it invariably has
extensive interregional as well as international linkages.
The term city-region appears to have been rst used
by DICKINSON (1947), and the concept was also examined in considerable depth by BOGUE (1950). Bogue
referred to it as the metropolitan community, and
somewhat later DUNCAN et al. (1960) used the
expression metropolitan region. Whereas Dickinson
and Duncan et al. tended to regard the hinterland in
general terms as an area dominated by the core,

John B. Parr

Downloaded by [University of Winnipeg] at 17:07 04 September 2014

Bogue examined the structure of the hinterland in more


detail, with attention given to the size-frequency structure of urban centres. From time to time it is argued that
what is deemed to be a PUR might be better viewed as a
series of city-regions, each based on a major centre of
the supposed PUR. The work of MUSTERD and VAN
ZELM (2001) and VAN OORT et al. (2010) may be
cited as examples. It can be expected that the greater
the territorial extent of the PUR, the more likely will
be the validity of this alternative interpretation. The
penultimate section of this paper will return to the
concept of the city-region.
This differentiation of a given type of region may also
be viewed in hierarchical terms. Continuing with the
example of a nodal region, the FUR may represent a secondary city-region, situated within some primary cityregion (of a scale similar to that indicated above). In
view of the obvious advantage in having a political or
governmental region coincide with an economic
region, this hierarchical arrangement of economic
regions could form the basis of a two-tier system of
administration. Under actual conditions, however, it is
not unusual for a secondary city-region to straddle the
boundary between two (sometimes three) primary cityregions. While this would not constitute a problem in
the supply of goods and services by the private sector
(nor, more generally, an obstacle to the efcient functioning of the relevant regional economies), it would
pose a problem for the spatial structure of administration.
In particular, the secondary city-region could not form
the basis of a secondary administrative region, since this
would be run counter to a fundamental tenet of political
organization, which requires that a secondary unit must
be wholly contained within a primary unit (PARR , 2007).
Having considered the principal forms of economic
region, the obvious question arises as to which is the
most appropriate type. Unsurprisingly, this will depend
on a number of factors, including the scale of the area
being studied, the kinds of research issues being investigated, the nature of the problems being addressed, etc.
Nevertheless, some variant of the nodal region appears
to have become the most useful and the most commonly
employed. To a greater extent than is possible with other
types of region, the nodal region (in its various forms) is
able to provide insights into the operation of its economy
over space and into the spatial interrelations of the various
elements of this economy. These particular strengths do
not, of course, amount to a rejection of other types of
region in all circumstances. In the balance of the paper,
however, the emphasis will be on the nodal region,
with the scale specied as appropriate.

THE REGIONAL ECONOMY AND ITS


SPATIAL STRUCTURE
The economic region (whatever its type) has a number
of similarities with the economy of the nation. As with

the national economy, social accounts provide important information on the regional economy. In output
terms the gross domestic product of a region (corresponding to the gross domestic product of the nation)
is the value of all goods and services produced within
a region. By contrast, the gross regional product (corresponding to the gross national product) is equal to the
sum of gross domestic product and net external payments, that is, the dividends, rents, wages received by
residents and rms of the region from outside the
region and transfers from the central government,
minus comparable payments made by residents and
rms to outside the region, and payments made to the
central government.
Against these similarities between regional and
national economies need to be set the equally important
differences. For example, the divergence between gross
domestic product of the region and gross regional
product is usually more pronounced than the divergence at the national level. In addition, the economic
performance of the region in terms of the growth rate
of income or employment is likely to be above or
below that of the nation, depending inter alia on the sectoral composition of its economy and the performance
of its individual sectors. Similarly, the temporal uctuations of the regional economy (with respect to the
length and amplitude of the trade cycle) can be expected
to differ from those of the national economy.
As an economic unit the region typically represents a
small, open economy. The relative openness of the
regional economy is reected by the fact that exports
(visible and invisible) from the region usually account
for a greater proportion of its gross domestic product
than is the case for the nation in which the region is
located. Associated with this relative openness of the
regional economy is the high level of interregional
factor movement, particularly in the case of capital and
skilled labour. Such a characteristic, together with the
existence of economies of scale in production, causes
trade among regions to take place on the basis of absolute advantage, a tendency that partially explains the
presence of problem regions. These are unable to capitalize on their comparative advantage, and are vulnerable to competition from successful regions.
Perhaps the most fundamental difference between
the regional economy and the national economy lies
within the area of economic sovereignty. Here the position of the region is greatly constrained. Even if the
region has a government or administration (and this is
the exception rather than the rule), it is unable to
pursue a meaningful macro-economic policy. Thus,
there can be no monetary policy, simply because the
region lacks the ability to issue currency, and as a
result the region cannot follow an external economic
policy, involving exchange-rate adjustment. While
under certain conditions it is possible for a region to
pursue a scal policy, the prerogative of taxation and
expenditure has been increasingly surrendered to (or

The Regional Economy, Spatial Structure and Regional Urban Systems


usurped by) a higher authority, usually the central or
federal government (in the UK the recent publicnance proposals for Scotland and Northern Ireland
are something of an exception to this trend). The limiting effects of this lack of economic sovereignty are mitigated to some extent by the existence of interregional
nancial transfers from the relatively prosperous
regions to the relatively poor regions. These transfers
are orchestrated by the central government (sometimes
according to strict formulae), and have the effect of partially offsetting the lack of competitiveness of the
problem regions and the absence of adequate interregional mobility of parts of the workforce.

Downloaded by [University of Winnipeg] at 17:07 04 September 2014

The question of spatial organization

By the same token that the regional economy displays a


variation over time (in terms of economic development
and trade-cycle uctuation, for example), it also exhibits
a variation across space. Lsch (LSCH /WOGLOM and
STOLPER , 1944/1954) regarded this latter type of variation as anything but random, and referred to it as rumliche Ordnung (spatial order or spatial ordering), that is,
the spatial structure of the regional economy. In
addition to the locational pattern of economic activity
and infrastructure endowment (including communications and other elements of economic and social overhead capital), spatial structure is concerned with the
organization and functioning of markets for goods and
factors of production, an aspect considered in the following section. Spatial structure highlights the fact that
the regional economy does not operate at a single
point and is not evenly distributed over space, the
actual spatial structure of a region representing a conguration intermediate between these two extremes.
The spatial structure of a region may be seen as
reecting the locational characteristics of economic
activities located there. For example, a region with an
economy based on extensive commercial agriculture
will have a different spatial structure from that of a
region with an economy dependent on extractive
industries and the processing of raw materials, which
in turn will differ from the spatial structure of a regional
economy reliant on engineering and other fabricative
manufacturing. The contrasting spatial structures are
likely to be inuenced by a number of factors. These
include the orientation of economic activity (to
markets, raw materials, energy sources, etc.) and the
strength of agglomeration economies, both internal
and external to the rm, based on the advantages of
scale, scope and complexity (OHLIN , 1933; PARR ,
2002a). A useful distinction is sometimes made
between static and dynamic agglomeration economies:
static agglomeration economies are generally concerned
with the efcient locations of rms under a given set of
economic conditions, while dynamic agglomeration
economies draw attention to their inuence on innovation and productivity, leading to the growth of

particular cities within a region or nation (GLAESER ,


2008). There is, in addition, a more general factor in
operation. This concerns the fact that the spatial structure at a given point in time will be determined to a substantial degree by the spatial structure that existed during
the preceding era of economic development. It is
seldom the case that an existing spatial structure of a
region has emerged tabula rasa.
The relationship between economic activity and
spatial structure may operate in the opposite direction,
so that spatial structure exerts an inuence on the level
of economic activity. For example, if for technological
and/or transport-cost reasons the locational requirements of the economic activities forming part of the
regional export base were to undergo a substantial
change, continued operation of these activities might
require an identiably different spatial structure. The
problem here is that the spatial structure may not be
able to respond adequately to these changed conditions,
thus inhibiting the continued existence of the dominant
economic-activity set, and possibly preventing the
emergence of replacement economic activity that
might be viably located within the region. Although
adequate supplies of labour and elements of infrastructure are available within the region, these may exist at
inappropriate locations and/or in insufcient quantities
to be attractive to domestic or external sources of
capital that would otherwise be invested within the
region.
The inuence of a pre-existing spatial structure on
regional economic performance need not be adverse,
however. It is sometimes neutral, and on occasions
may be positive, as will be argued in a later section.
The spatial structure of a region may thus be conducive
to improved productivity, innovation and creativity, as
well as its ability to adjust to external shocks. Considerable debate has taken place on the net advantages of
different spatial structures, although no consensus has
been reached. Successful economic performance
appears to be associated with a range of signicantly
different spatial structures. It becomes evident that
while the spatial structure may be seen as a reection
of regional economic activity, it can also act as an important inuence on the nature and the level of such
activity, an inuence that may be positive or negative.
Characterizing spatial structure

After considering the signicance of spatial structure


within the regional economy, the various ways in
which spatial structure can be specied are now briey
reviewed. One fairly obvious approach is to view
spatial structure as a series of areal distributions. These
would involve such major aggregates of the regional
economy as population, workforce, employment,
unemployment, etc. In certain cases the areal distribution of xed investment, raw materials, energy
sources, and the land available for industrial or

John B. Parr

Downloaded by [University of Winnipeg] at 17:07 04 September 2014

commercial development may also be included. This


approach would not simply be concerned with inventories of the various elements; certain of the interrelations among these various elements could be
analyzed with the aid of geographical information
systems (GIS) and the application of spatial statistics.
It is also possible to view the spatial structure of the
region as a set of ows among specied locations
within the region, making use of the extensive literature
on network analysis, an early example being the work of
GARRISON (1960). An important category of intraregional ows involves those that can be expressed in
economic terms such as trade ows (transactions
among rms as well as those between rms and households), capital movements, wages and salaries paid by
rms to commuters, and (where relevant) transfers of
funds to and from the regional government. Unfortunately, the lack of reliable data on these ows often
makes it necessary to rely on surrogate ows such as
those relating to shopping trips, freight trafc, commuter ows, telephone connections, etc.
Additional methods of characterizing regional spatial
structure are worthy of mention. One of these involves
city-size distributions in terms of population and
employment (LASUEN et al., 1967; MALECKI , 1980).
Another approach treats spatial structure as a surface
(HAGGETT , 1968; KRUMBEIN , 1959), an important
perspective if the concern is with dynamic spatial structure. A particular type of surface is the density function.
Though usually employed at the scale of a metropolitan
area, the density function may also be applied at the scale
of the region, centred on the regional core (BOGUE ,
1950; VINING , 1955). Undoubtedly, the most comprehensive characterization of spatial structure makes use of
an urban-system perspective. Such a view of spatial
structure is concerned with the hierarchical differentiation of centres in terms of their size, spacing, frequency and functional composition, and also includes
other approaches already mentioned, notably the economic ows among centres and their size distributions.
This urban-system perspective on the spatial structure
will be explored in the next two sections.

REGIONAL URBAN SYSTEMS:


THEORETICAL BACKGROUND
In considering the internal structure of a region in terms
of an urban system, attention is rst given to several
important models developed in location theory. As
with models in general, a model of the urban system is
an abstraction, which is not intended to replicate
reality. It seeks to explore the processes by which particular variables are related or interrelated, and attempts
to formulate a structure which reects this. Models of
the urban system can thus be employed not only as
benchmarks in the examination of actual conditions,
but also as the means by which working hypotheses

can be generated. The construction of a model of the


urban system requires many variables to be held constant, often involving the imposition of strong (even
heroic) assumptions. Inevitably, therefore, urbansystem models contain a simplicity or starkness, which
not all researchers in the area nd acceptable. With
these considerations in mind, attention is now focused
on several models of the urban system, starting with
the Christaller framework (CHRISTALLER /BASKIN ,
1933/1966). This is considered in some detail since it
will be the standard against which other models are
compared.
The Christaller urban system

The Christaller contribution model was based on an


empirical study of the central places of Southern
Germany in the 1920s. The primary focus was on
three regions (the L-systems centred on Munich, Nuremberg and Stuttgart, each being a Landstadt or regional
city).3 Although not without a number of precursors,
the Christaller framework was an early attempt at
viewing the urban system as a set of hierarchically differentiated service centres, each of these central places
having one or more exclusive market areas to which
the various goods and services were supplied.
The Christaller urban system is concerned exclusively
with central goods which are supplied to a dispersed
market, and which are provided by the public sector
as well as the private sector.4 In terms of household
demand the central goods involve retail goods, consumer services, the output of certain types of manufacturing such as job printing, newspaper production, and
various kinds of food processing. In the case of intermediate demand (that is, the demand exerted by
rms), central goods include wholesaling, transport
and communications, storage, business and nancial services, as well as numerous repair, maintenance and
engineering services. The various goods and services
can be grouped to form N sets, each of which may be
regarded as a composite good of level m where m = 1,
2, , N. This grouping is based on the frequency of
points required for the viable supply, whether private
or public. Although not undertaken by Christaller, it
is possible to introduce a level 0 set of goods, which
includes agricultural commodities, foodstuffs, minerals
and other raw materials, etc., and which is supplied
only from level 0 of the hierarchy, that is, rural areas.
Structure of the Christaller urban system

Four major inuences underlie the form of the Christaller urban system. First, the demand for all central goods
is dispersed, and in varying degrees is sensitive to distance from the point of supply. Second, the supply of
these goods is closely related to this demand, and takes
place at points centrally located within their respective
market areas. The frequency of these points (and thus

Downloaded by [University of Winnipeg] at 17:07 04 September 2014

The Regional Economy, Spatial Structure and Regional Urban Systems


market areas) depends on transport costs borne by the
consumer and on economies of scale in the production
of the good in question. A third inuence reects the
propensity of consumers to minimize travel costs, by
engaging in multipurpose trips, both with respect to
goods within a given set, but also in the case of goods
belonging to different sets. The fourth inuence is the
tendency of suppliers of a given good to be drawn to
the location of suppliers of other goods, whether in
the same set or different sets. To some extent this is
determined by the previous inuence, but it also
permits suppliers to benet from agglomeration economies of the urbanization type such as access to infrastructure and public utilities. This advantage tends to
increase with hierarchical level, although it is likely to
be partially offset by increasing location rents.
These four inuences give rise to an urban system in
which there is a correspondence between spatial
demand and spatial supply, although this correspondence is not of the same strength for all sets of
goods.5 A prominent feature of the Christaller urban
system is the successively inclusive hierarchy. It is
assumed that there are N levels of centre, N levels of
market area and N sets of goods (as dened above).
The successively inclusive hierarchy is such that a
centre of level m supplies a characteristic set m, together
with sets 1 to (that is, up to and including) m 1. More
specically, the centre of level m supplies sets 1 to m to
itself and to its market areas of levels 1 to m, respectively, where its level m market area represents the
characteristic market area. For a level 1 centre the
characteristic market area is wholly rural (part of level
0), while the characteristic market area of a centre of
level m > 1 consists of a rural component (part of
level 0) plus particular centres (of decreasing frequency)
belonging to levels 1 to level m 1. The existence of
this successively inclusive hierarchy places a limitation
on the number and the sizes of the various market
areas.6 It also implies a distinctive pattern of trade
among centres. Thus, a centre of level 1 < m < N
exports to centres of levels 1 to m 1, and imports
from centres of levels m + 1 to N. A balance-of-payments problem does not arise, if it is assumed, reasonably enough, that centres of particular levels are being
supplied with the level 0 goods from level 0 (rural
areas). Under actual conditions, the balance-of-payments problem is also avoided by the existence of
non-trade ows such as capital movements, government expenditures, and the wages and salaries associated with commuting.
The overall structure of the Christaller system may be
summarized with respect to centres belonging to any
given level: the higher the level of a centre, the
greater the number of sets of goods supplied from it;
the greater its population; the greater the size of its
characteristic market area; and (because of this) the
lower the frequency of such centres and the wider
their spacing. In addition, centres of level m are evenly

spaced on a triangular or some alternative lattice (strictly


speaking, the even spacing of level m centres refers to all
centres of levels m to N, a consequence of the successively inclusive hierarchy). This even spacing stems
from the underlying assumption of a uniform density
of rural population. Such an assumption is easily dispensed with, as demonstrated by RUSHTON (1972),
and the resulting pattern of variable centre spacing is
obviously more in keeping with actual conditions.
Limitations of the system

While the Christaller structure is able to reect several


important facets of actual urban systems, it nevertheless
contains a number of serious weaknesses. Perhaps the
most serious of these is the fact that attention is conned
to goods and services of a market-orientated nature,
where centrality with respect to the consuming
market (whether comprising households or rms) is of
paramount importance. Consideration of goods with
other orientations is excluded. Another weakness of
the model is that it wholly neglects the input structure
of production: it is as if all inputs to production are ubiquitous, that is, available everywhere at approximately
equal cost.
A further deciency is concerned with the directional pattern of trade. Thus, a centre of a given level
does not export to (that is, supply) a centre of a
higher level, and does not import from (that is, is not
supplied by) a centre of a lower level. The reasons
for this pattern may be explained as follows. A centre
of level m < N, for example, is exporting goods of
levels 1 to m, but these same goods are also exported
from centres of level m + 1 to N. Since these higherlevel centres can assumed to be self-sufcient in such
goods, there is no basis for exports from the centre of
level m to centres of higher levels. At the same time
this centre of level m < N is self sufcient in goods of
levels 1 to m 1, so that there is no basis for imports
to this centre from centres of levels 1 to m 1. By
similar reasoning there is no basis for trade among
different centres of level m. As will be argued, these
restrictions on the pattern of trade (though logical in
terms of the assumptions of the model) represent an
inadequacy of the Christaller structure in the context
of the modern conditions.
Finally, the Christaller model places an excessive
premium on proximity. Thus, when a centre (or a
location in a rural area) is supplied with a good of a
given level m, this is always from the nearest centre of
level m or higher.7 Such a requirement results from
the assumption that demand for goods and services is
extremely sensitive to distance. No allowance is made
for a person travelling to a centre other than the
nearest relevant one, in order to buy a particular style
of clothing or brand of consumer good, nor for a rm
failing to select the closest relevant centre, in order to
obtain an input to production.

Downloaded by [University of Winnipeg] at 17:07 04 September 2014

John B. Parr

The economic region of Lsch

The Tinbergen contribution

It fell to Lsch (LSCH /WOGLOM and STOLPER , 1944/


1954) to provide the micro-economic underpinnings of
the Christaller model. The objective of the Lsch analysis was to construct from rst principles a regional landscape of production points, giving rise to an urban
system. In several important respects the Lsch approach
was similar to the Christaller model: attention was conned to central goods (though there is more concern
with manufactured goods); no consideration was given
to the location of inputs; the supply of a good to a particular location was always from the nearest relevant
centre.
The approach involved the analysis of demand over
space with freight-on-board (FOB) pricing, increasing
returns to scale in production, and spatial competition
among suppliers under conditions of free entry. In
the case of a single good the spatial equilibrium is specied for each identical rm in terms of price, output
and market-area radius. This outcome for a single
good is broadly equivalent to that in the Christaller
structure. The analysis was then extended to examine
the outcome for all goods. However, instead of imposing the successively inclusive hierarchy in the manner
of Christaller, Lsch allowed a wider variation in
market-area size. In this multi-good setting the
primary concern was with minimizing the overall
number of production points. What emerged was a
continuous economic landscape in which an economic
region could be identied. Interestingly, Lsch
(LSCH /WOGLOM and STOLPER , 1944/1954,
pp. 130131) was aware of the Christaller urban
system, and regarded this as a special case of his own
more complex system, a case in which the number of
market-area sizes is limited by the number of levels
of political or administrative organization.
The economic region derived by Lsch contains
three major features not present in the Christaller structure. First, the centres have a wider diversity of functional complexity and therefore a greater range of
population sizes (STOLPER , 1955). Second, a relatively
small centre (one supplying few goods) can export to
a large centre (one supplying relatively many). Third,
the regional economic landscape is dominated by a
single metropolitan centre, from which all goods are
supplied, and from which extend six centre-rich
sectors and six centre-poor sectors. The centre-rich
sectors of adjacent regions form the bases for transport
axes between pairs of neighbouring metropolitan
centres. This explanation for the common phenomenon
of inter-metropolitan corridors is imaginative, although
the analysis contains too many inconsistencies and contradictions for such a result to be accepted. Moreover,
within an actual regional urban system it is debateable
whether centre-rich sectors gave rise to transport corridors or whether the causation was in the reverse
direction.

The work of TINBERGEN (1961, 1968) stands in contrast to the ChristallerLsch tradition. TINBERGEN
(1968, p. 65), who was apparently unimpressed with
previous attempts at modelling the urban system,
offered an alternative structure. This considers N
goods (or N sets of goods), N levels of centre, as well
as N levels of market area (where m = 1, 2, , N).
Also included is a level 0 good, which is supplied from
the rural area (level 0) to the rest of the system. Tinbergen introduced a successively inclusive hierarchy that
resembled the Christaller structure, but only to the
extent that a centre of level m supplies goods of levels
1 to m to itself. As far as exports are concerned, the
centre of level m only supplies the level m good to its
single market area (its level m market area), comprising
part of level 0 (the rural area) and centres of levels 1 to
m 1. This assumption results in an arrangement of
centres and market areas different from that in the Christaller system, and also a distinctive pattern of trade
(MULLIGAN , 1982). To appreciate this, we may consider a household or rm within a rural area in the
immediate vicinity of a centre of level m. By the assumptions of the model this rural location would be supplied
with a level m 1 good from the nearest centre of level
m 1, and not from the closer centre of level m, despite
the fact that this centre would be supplying the level m
1 good to itself.
Such a pattern indicates a relaxation, but not the
abandonment, of the requirement of centrality,
although it is not at all obvious how such a pattern
could be sustained in a market economy. What the Tinbergen model does illustrate, however, is that centres of
level m of the hierarchy have a particular export specialization, namely, the level m good. This is consistent with
the observed tendency for certain industries to be
located at particular size classes of centre. It is also consistent with the presence of agglomeration economies of
the localization type (ISARD , 1956). These factors
suggest that the model is best suited to explaining the
location of intermediate goods that have a particular
input supply, that is, where the input is supplied from
level 0 or from a higher-level centre. As with the Christaller model any centre of a level m < N only exports to
lower levels and only imports from higher levels. No
balance-of-payments problem arises, however, since it
is specically assumed that there exists a rural area
(level 0), which exports the level 0 good to one or
more of the urban levels of the hierarchy.
Before moving to the next section, a general
comment is included, regarding a particular aspect of
the models outlined above. It will be observed that
each model treats the urban system as existing within
an unchanging equilibrium. The principal reason for
this is that in the development of these models it was
convenient for time to be held constant, so that attention could be focused on spatial structure. Obviously,

The Regional Economy, Spatial Structure and Regional Urban Systems


urban systems can be expected to become modied over
time, and none of the models discussed can be considered inherently stationary. Moreover, the apparent
static bias of the models is largely overcome, if these
are reworked in dynamic terms or as multi-static
models. The various attempts along these lines are discussed elsewhere (PARR , 2002b).

Downloaded by [University of Winnipeg] at 17:07 04 September 2014

REGIONAL URBAN SYSTEMS: THE


CONTEMPORARY CITY-REGION
In the light of these theoretical approaches to the urban
system, not all of which were originally presented by
their authors in an explicit manner, it is now possible
to explore the extent to which the models are able to
cast light on the modern-day city-region. Although
the major concern has been with the internal factors
inuencing the organization of a regional urban
system, it is important to consider some of the changing
external forces that have continued to shape it. Over virtually any period of economic history during the last
two to three hundred years, urban systems have been
undergoing transformations, affected in part by external
conditions. This is as true in the present era as it was in
the nineteenth century, for example, which saw the
continuation of the Industrial Revolution, the rise of
steam power and the expansion of overseas trade.
The decades following the Second World War
experienced considerable economic expansion, innovation and modernization. Associated with these developments were fundamental changes in the technologies
of production and communications, overall improvements in living standards, shifts in consumer preferences,
vastly different patterns of international trade (not only
in terms of volume, but also with respect to the location
of markets and sources of supply), substantial increases in
the international movement of capital, changed patterns
of corporate ownership and control both nationally and
internationally, etc. These various changes exerted an
inuence (direct as well as indirect) on regional economies. But despite such a dramatic modication of the
general economic environment the regional urban
system endured, though inevitably in a modied form,
and this cannot be emphasized strongly enough. The
remainder of this section will be concerned with the
urban system of the major city-region, as discussed
above.
The location of economic activity within the city-region

The urban system of the major city-region typically


possesses certain features of the L-system of Christaller,
particularly in terms of the hierarchical structure of
urban centres.8 One such feature is the size, spacing
and frequency of centres, which distinguishes the cityregion from other types of region, for example, the
PUR. Although Christaller identied seven hierarchical

levels of centre, only three of these had centre populations that could be considered urban in any practical
sense, that is, in excess of 10 000 people (ULLMAN ,
1941). With some exceptions, the present-day hierarchy
of urban centres typically consists of three or four levels.
The successively inclusive hierarchy continues to be
present, but it is less clearly dened, being partially
obscured by the supply of non-central goods.
It will be recalled that in the urban system of Christaller the location of all non-rural economic activity is
governed exclusively by the principle of centrality
with respect to the location of demand. As a consequence a given level of urban centre only supplies
central goods, with the number of sets of goods supplied
corresponding to the number of market areas being
served. For the modern city-region this pattern tends
to hold for centres of lower levels. Centres of
medium-to-high levels, however, commonly supply
particular non-central or specialized goods (to national
and international markets), in addition to central
goods (largely to regional markets). It will be argued
shortly that the structure of the hierarchy is not immutable, and may thus be subject to change.
Non-central or specialized goods comprise a highly
diverse and evolving group of activities, which includes
mineral exploitation, a wide array of manufacturing
industry, a range of business and commercial services,
as well as numerous specialized consumer services. In
the case of non-central goods the locational orientation
of production or supply is not primarily to the market
but variously to sources of raw materials and manufactured inputs, to major ports and international airports,
to supplies of skilled or semi-skilled labour, etc. (BECKMANN , 1968). However, the predominant location of
these kinds of economic activities is in urban centres.
This is largely a reection of the co-location of rms,
in order to gain agglomeration economies, not simply
of the urbanization type but also of the activitycomplex type (PARR , 2002a). In addition, there is a tendency (consistent with the Tinbergen analysis) for
certain activities to be associated with particular levels
of the urban system, suggesting that the relevant rms
benet from agglomeration economies of the localization type (BERRY and HORTON , 1970; EVANS ,
1972; ISARD , 1956). This tendency is not present in
all city-regions, however.
Economic ows within the city-region

Since the nature of economic activity and its overall


locational pattern in the modern city-region are more
complex than in the Christaller structure, it is not surprising that more intricate patterns of interaction are
present. Consumer accessibility continues to be important for many goods, but its inuence is no longer as
decisive as in the past. Improved mobility and the extension of product differentiation have meant that supply
from the nearest relevant centre is no longer a

Downloaded by [University of Winnipeg] at 17:07 04 September 2014

10

John B. Parr

requirement. And in the case of trade among rms,


reductions in the cost of transport have reduced the
importance of centrality in gaining access to markets
and inputs to production. Under these conditions it is
certainly possible for a low-level centre to export to a
high-level centre (and therefore for a high-level centre
to import from a low-level centre), a feature of the
Lsch system. It is also possible for centres of approximately the same level to trade with each other, regardless of their respective locations with the city-region.
In recent decades, information ows have rightly
been accorded an importance among many researchers.
Unfortunately, ows of this kind are difcult to acquire,
hard to measure and not readily interpreted (HALL ,
1991; HALL and PAIN , 2006). Of particular interest
here is the extent to which information ows are sensitive to distance, and the evidence on this is not clear. It
should not be forgotten that the city-region exists
within the wider territory, be this the nation, a trading
bloc or the world economy, and in discussions of the
city-region external linkages sometimes go unrecognized or underemphasized. Of relevance at the national
level is the location of a city-region in relation to other
city-regions. One aspect of this, which was considered
by Lsch, is the development of inter-metropolitan
axes or corridors between the dominant metropolitan
cores of adjacent city-regions, and the accompanying
interaction based on trade ows, the movement of
people and information, as well as the diffusion of
innovations.
Another category of ows in the urban system of a
city-region (though one not present in the models discussed above) concerns the daily movement of labour
from centres of one level to centres of a higher, a
lower or the same level. For each of these ows of
labour there is a reverse ow of wages and salaries,
which may be an important part of the economic base
for many urban centres in the city-region. Daily ows
of labour to (and also from) the dominant metropolitan
core continue to be important, but there is an increase in
ows among centres within the hinterland, which are
unrelated to the core. There is, of course, some upper
limit on the daily distance travelled between centres
(whether between levels or within a level), but this
has exhibited a steady increase over recent decades due
to the expansion of car ownership, reduced travel
costs and shorter travel times.
Commuting represents an important lubricant for
both labour-market exibility and residential adjustment. It not only allows the separation between residence and workplace, but also permits changes in the
location of the workplace, with the location of residence
remaining unaffected, as well as changes in the location
of residence with the location of the workplace unaffected. In the typical city-region of Northwest Europe
and the Eastern Seaboard of North America, where
population density is relatively high and the accompanying spacing of urban centres is correspondingly low,

adjustments of this kind are possible, particularly since


commuting may be a temporary or even a permanent
substitute for migration (TERMOTE , 1978). In addition,
the joint-location problem of dual-income households
(VAN OMMEREN et al., 1999) becomes a less pressing
one. In this way the spatial structure of the city-region
permits or even facilitates economic adjustment. It
must be stressed, however, that this spatial-structure
advantage in the process of adjustment would be less
prominent and perhaps even non-existent in cityregions that are more extensive, with low overall population densities.
Spatial-structure change in the city-region

It will be apparent that the urban system of the modern


city-region is to be seen as something other than a static
structure, although there may be intervals when the
pace of change is relatively slow. As already mentioned,
the urban system of a city-region is subject to the effects
of external as well as internal change, these taking the
form of shocks. Obviously, there can be no question
of the urban system being rebuilt de novo. Instead,
there is a process of self-adjustment on the part of the
system, which takes place over the short-to-medium
run. For example, as the result of a given shock, an individual good may now need to be supplied from fewer
centres or from more centres. Under these conditions
there is a hierarchical redenition of the relevant sets
of goods, but the hierarchy remains unchanged in
terms of its number of levels and the number of
centres of each level. However, the shock may be sufciently great that the hierarchy, itself, is transformed.
One case involves the formation of a new level. In
another case the level continues to exist but becomes
less extensive (with fewer centres) or more intensive
(with a greater number of centres), while in a further
case there is complete disappearance of a level, this
former level being combined with the next higher or
next lower level. It follows that these different adjustments all involve changes in the structure of trade
ows. The effect of such shocks is to modify the size distribution of centres within the regional urban system.
Within most city-regions of the developed world the
spatial structure continued to undergo a process of concentration well into the rst quarter of the twentieth
century, up to which time the metropolitan core grew
at a faster rate than its hinterland in terms of population,
employment and income. In more recent times,
however, the trend has been toward deconcentration,
with the growth rate of the hinterland (most of this
growth being in urban centres) exceeding that of the
core. In part, this has resulted from households and
rms relocating from the metropolitan core to locations
beyond it, often as part of a mutually reinforcing shift of
population and employment. Such shifts include relocations among levels of the urban hierarchy, which increasingly take place from the highest level (the metropolitan

The Regional Economy, Spatial Structure and Regional Urban Systems


core) to lower levels, and also from the highest level to
urban centres located within a particular distance range.
There seems to be every indication that the nature of
economic and technological change will continue to
favour the broad trend toward deconcentration.

Downloaded by [University of Winnipeg] at 17:07 04 September 2014

CONCLUSIONS
Attention may now be drawn to some of the main issues
that have emerged from this discussion, beginning with
the notion of the region. This represents an important
scale of analysis (intermediate between the locality or
urban centre and the nation), although its signicance
and utility have been called into question. Unfortunately, the region as a unit of analysis is not always
well dened and can thus become a source of confusion.
At the very least, it would seem desirable that whenever
the term region is being used, some working denition
should be provided. More generally, it would be helpful
if urban and regional analysis had at its disposal a
language (or simply a grammar) by which this confusion
could be avoided. Various elds of the medical and biological sciences, for example, have availed themselves of
such a facility for centuries, and to good effect.
It has long been maintained that for economic (and
other kinds of) analysis the nation is too unwieldy a
unit, which should be disaggregated into regions, in
order to understand more adequately the spatial and sectoral variability of economic performance. By extension, it may be claimed that the region is too crude a
focus, and that its economy should be considered at a
sub-regional scale or preferably in spatial-structure
terms, so as to gain insight into its internal operation
and evolution over the long run. As was argued,
spatial structure represents a partial determinant of
regional economic change as well as a reection of it,
the latter emphasis being the more conventional view.
Consideration of spatial structure thus goes some way
to understanding the nature of the regional economy
and its ability to adjust to external inuences.
In the analysis of spatial structure the contributions of
location theory are substantial, although there has been a
tendency for these to be underestimated. While particular aspects of location theory have assisted our understanding of urban systems, the formulation of a general
theory has nevertheless proven elusive. Further progress
can be expected, and the continued development of the
New Economic Geography (NEG), stimulated by the
work of KRUGMAN (1995, 2009) and FUJITA et al.
(1999), is of potential importance in this connection.
It must be added, however, that while NEG models
are able to provide powerful insights into the operation
of an economy over space, the treatment of the region
requires further elaboration and specication, not least
in terms of its spatial structure (GARRETSEN and
MARTIN , 2010; OVERMAN and IOANNIDES , 2001).

11

For the developed world, and to an increasing extent


the developing world, the city-region has emerged as a
dominant spatial form, and it is common for a national
space economy to comprise a set of city-regions,
usually hierarchically differentiated. Over recent
decades the city-region has become the arena in
which the processes of economic development and
economic change are being played out, the implications
of which are likely to be considerable. As a consequence
the spatial structure of the city-region (particularly when
viewed as a regional urban system) will assume an
obvious importance in public as well as private
decision-taking. With regard to the future structure of
the city-region, the outcome is anything but certain.
HALL (2009, pp. 808811), while warning of the
dangers of prediction, has offered a thoughtful account
of the more important tendencies (what he terms
changes in the basic parameters) that are likely to inuence the form of the city-region in coming decades.
These range from environmental modication and
demographic evolution to the impacts of economic
and technological change, along with greatly altered
patterns of consumption. While there is an obvious
need for information about future trends (in order to
assist the policy-maker, the planner and the analyst), it
would probably be imprudent (and also very difcult)
to pursue matters concerning the future beyond this
point.
Acknowledgements The author wishes to thank
D. Batten, C. Wren and two anonymous referees for their
comments on earlier versions of this paper. Thanks are also
due to the Guest Editors of this special issue of Regional
Studies, and to the Deputy Editor-in-Chief of the journal.

NOTES
1. As an economic unit, the origins of the nodal region can be
traced back to the work of von Thnen (VON THNEN /
WARTENBURG , 1842/1966). In Part I of this study his
Isolated State consisted of a single urban core and a hinterland, the core representing the single market for agricultural commodities produced in the hinterland, to which
it supplied goods and services.
2. Currently, the term polycentric tends to be employed
indiscriminately. Naturally, all regions can be regarded as
polycentric, in the narrow sense that each contains a
number of (sometimes many) urban centres, and to this
extent the term is redundant. On occasions the term polycentric is used as a shorthand means of describing a region
that is not monocentric, that is, not excessively dominated by a particular centre.
3. Two further L-systems were included: those based on
Frankfurt and Strasbourg. Since these extended beyond
the boundary of the area of the case study (Southern
Germany), only part of the L-system was included in
each case.
4. Christaller was certainly aware that goods and services
were produced in Southern Germany which were not

Downloaded by [University of Winnipeg] at 17:07 04 September 2014

12

John B. Parr

central goods, and that many goods were manufactured


there (particularly in the regional capitals) for national
and international markets. Given the paucity of prior
work on urban systems, Christaller appears to have
excluded non-central goods in order to isolate the inuence of centrality in the process of urbanization and in
the formation of the urban system.
5. This correspondence is particularly strong when demand
for a set of goods is inuenced by high transport costs,
and supply is subject to minimal economies of scale, so
that the frequency of centres is high, and each of their
market areas is relatively small. By contrast, the correspondence is far less strong (though still present) when
the demand for a set of goods is only mildly affected
by transport costs, and economies of scale are substantial,
in which case the frequency of centres is low, and each
of their market areas is large. Not all goods tend to
one or other extreme, and this is a basis for a hierarchy
of more than two levels, which will be considered
shortly.

6. In Christallers theoretical model of the urban system, the


size of the characteristic market area increases from one
level to the next by a constant factor, usually denoted as
K. This is consistent with, though not required by, the existence of a successively inclusive hierarchy (PARR , 2002b).
7. This argument does not apply in the case of certain kinds of
multipurpose trips, for example, when a consumer travels to a
level m centre to purchase a level m good, but uses this opportunity to buy a level m 1 good, which could have been purchased at a centre of m 1 located closer to the consumer.
8. For this reason the city-region is sometimes likened to a
central place model of the Christaller type (CAMAGNI
and CAPELLO , 2004), while BATTEN (2011) has referred
to it as a central place region. Such characterizations of
the city-region may create the wrong impression,
however. As will be demonstrated, the urban system of
the modern city-region departs from the Christaller
central place structure in several fundamental respects,
while also displaying most of the attributes of the
CamagniCapello urban network model.

REFERENCES
ALONSO W. (1973) Urban zero population growth, Daedalus 109, 191206.
BATTEN D. F. (1994) Network cities: creative urban agglomeration of the 21st century, Urban Studies 32, 313327.
BATTEN D. F. (2011) On creative network cities, in ANDERSSON . E., ANDERSSON D. E. and MELLANDER C. (Eds) Handbook of
Creative Cities, pp. 284306. Edward Elgar, Cheltenham.
BECKMANN M. J. (1968) Location Theory. Random House, New York, NY.
BERRY B. L. J. and HORTON F. (1970) Geographic Perspectives on Urban Systems. Prentice-Hall, Englewood Cliffs, NJ.
BOGUE D. (1950) The Structure of the Metropolitan Community. Horace H. Rackham School of Graduate Studies, University of
Michigan, Ann Arbor, MI.
BRUNEL D. (Ed.) (2002) Cahiers de lInstitut dAmnagement et dUrbanisme de la Rgion le-de-France (IAURIF) Number 135,
Economic Performance of the European Regions. IAURIF, Paris.
BURGER M. and MEIJERS E. (2012) Form follows function? Linking morphological and functional polycentricity, Urban Studies 49,
11271149.
BURTON I. (1963) A restatement of the dispersed city hypothesis, Annals of the Association of American Geographers 63, 285289.
CAMAGNI R. and CAPELLO R. (2004) The city network paradigm: theory and empirical evidence, in CAPELLO R. and NIJKAMP P.
(Eds) Urban Dynamics and Growth, pp. 495529. Elsevier, Amsterdam.
CHESHIRE P. C. and HAY D. G. (1989) Urban Problems in Western Europe: An Economic Analysis. Unwin Hyman, London.
CHRISTALLER W. (1933/1966) Die zentralen Orte in Sddeutschland [1933]. Gustav Fischer, Jena; trans. BASKIN C. (1966) as Central
Places in Southern Germany. Prentice Hall, Englewood Cliffs, NJ.
DAVOUDI S. (2003) Polycentricity in European spatial planning: from an analytical tool to a normative agenda, European Planning
Studies 11, 980999.
DICKINSON R. E. (1947) City, Region and Regionalism. Routledge & Kegan Paul, London.
DUNCAN O. D., SCOTT W. R., LIEBERSON S., DUNCAN B. and WINSBOROUGH H. H. (1960) Metropolis and Region. Johns Hopkins
Press, Baltimore, MD.
EVANS A. (1972) The pure theory of city size in an industrial economy, Urban Studies 9, 4977.
FUJITA M., KRUGMAN P. R. and VENABLES A. J. (1999) The Spatial Economy: Cities, Regions and International Trade. MIT Press,
Cambridge, MA.
GARRETSEN H. and MARTIN R. (2010) Rethinking (new) economic geography: taking geography and history seriously, Spatial
Economic Analysis 5, 2760.
GARRISON W. L. (1960) Connectivity of the interstate highway system, Papers and Proceedings of the Regional Science Association 6,
121137.
GLAESER E. L. (2008) Cities, Agglomeration, and Spatial Equilibrium. Oxford University Press, Oxford.
HAGGETT P. (1968) Trend surface analysis in the interregional comparison of intra-regional structures, Papers of the Regional Science
Association 20, 1928.
HALL P. (1991) Moving information: a tale of four technologies, in BROTCHIE J., BATTY M., HALL P. and NEWTON P. (Eds) Cities of
the 21st Century: New Technologies and Spatial Systems, pp. 121. Longman, Melbourne, VIC.
HALL P. (2009) Looking backward, looking forward: the city-region of the mid-21st century, Regional Studies 43, 803817.
HALL P. and PAIN K. (Eds) (2006) The Polycentric Metropolis: Learning from Mega-City Regions in Europe. Earthscan, London.
HOOVER E. M. and FISHER J. L. (1949) Research in regional economic growth, in Problems in the Study of Economic Growth, pp. 173
250. National Bureau of Economic Research (NBER), New York, NY.
ISARD W. (1956) Location and Space-Economy. Technology Press of MIT and Wiley, New York, NY.

Downloaded by [University of Winnipeg] at 17:07 04 September 2014

The Regional Economy, Spatial Structure and Regional Urban Systems

13

KARLSSON C. and OLSSON M. (2006) The identication of functional regions: theory, methods and applications, Annals of Regional
Science 40, 118.
KRUGMAN P. R. (1995) Development, Geography and Economic Theory. MIT Press, Cambridge, MA.
KRUGMAN P. R. (2009) The increasing returns revolution in trade and geography, American Economic Review 99, 561571.
KRUMBEIN W. C. (1959) Trend surface analysis of contour-type maps with irregular control-point spacing, Journal of Geophysical
Research 64, 823834.
KRUTILLA J. V. and ECKSTEIN O. (1958) Multiple Purpose River Development. Johns Hopkins Press, Baltimore, MD.
LASUEN J. R., LORCA A. and ORIA J. (1967) City size distribution and economic growth, Ekistics 24, 221226.
LSCH A. (1944/1954) Die raumliche Ordnung der Wirtschaft, 2nd Edn [1944]. Gustav Fischer, Jena; trans. WOGLOM W. H. and
STOLPER W. F. (1954) The Economics of Location. Yale University Press, New Haven, CT.
MALECKI E. J. (1980) Growth and change in the analysis of rank-size distributions: empirical ndings, Environment and Planning A
12, 4152.
MEIJERS E. J., ROMEIN A. and HOPPENBROUWER E. C. (Eds) (2003) Planning Polycentric Regions in North West Europe: Value. Feasibility and Design. Delft University Press, Delft.
MEYER J. R. (1963) Regional economics: a survey, American Economic Review 53, 154.
MULLIGAN G. F. (1982) Tinbergen-type central place systems, International Regional Science Review 7, 8391.
MUSTERD S. and VAN ZELM Z. (2001) Polycentricity, households and the identity of places, Urban Studies 38, 769796.
NORTH D. (1955) Location theory and regional economic growth, Journal of Political Economy 63, 243258.
OHLIN B. (1933) Interregional and International Trade. Harvard University Press, Cambridge, MA.
OVERMAN H. G. and IOANNIDES Y. M. (2001) Cross sectional evolution of the US city size distribution, Journal of Urban Economics
49, 543566.
PARR J. B. (2002a) Missing elements in the analysis of agglomeration economies, International Regional Science Review 25, 151168.
PARR J. B. (2002b) The location of economic activity: central place theory and wider urban system, in MCCANN P. (Ed.) Industrial
Location Economics, pp. 3282. Edward Elgar, Cheltenham.
PARR J. B. (2007) On the spatial structure of administration, Environment and Planning A 39, 12551268.
PARR J. B. (2008) Cities and regions: problems and potentials, Environment and Planning A 40, 30093026.
RUSHTON G. (1972) Map transformation of point patterns: central place patterns in areas of variable population density, Papers of the
Regional Science Association 28, 111129.
STOPLER W. F. (1955) Spatial order and the economic growth of cities: a comment on Eric Lampards paper, Economic Development
and Cultural Change 3, 137146.
STORPER M. (1995) The resurgence of regional economies, ten years later: the region as a nexus of untraded interdependencies,
European Urban and Regional Studies 2, 191221.
TERMOTE M. (1978) Migration and commuting in Lschs central place system, in FUNCK R. and PARR J. B. (Eds) The Analysis of
Spatial Structure: Essays in Honour of August Lsch. Karlsruhe Papers in Regional Science, Vol. 2, pp. 8390. Pion, London.
TINBERGEN J. (1961) The spatial dispersion of production: a hypothesis, Schweizerische Zeitschrift fr Volkswirtschaft und Statistik 97,
412419.
TINBERGEN J. (1968) The hierarchy model of the size distribution of centres, Papers of the Regional Science Association 20, 6568.
ULLMAN E. L. (1941) A theory of location for cities, American Journal of Sociology 46, 835864.
VAN OMMEREN J., RIETVELD P. and NIJKAMP P. (1999) Job moving, residential moving, and commuting: a search perspective,
Journal of Urban Economics 46, 230253.
VAN OORT F., BURGER M. and RASPE O. (2010) On the economic foundation of the urban network paradigm: spatial integration,
functional integration and economic complementarities within the Dutch Randstad, Urban Studies 47, 725778.
VINING R. (1955) A description of certain spatial aspects of an urban system, Economic Development and Cultural Change 3, 147195.
VON THNEN J. H. (1842/1966) Der isolirte Staat in Beziehung auf Landwirthschaft und Nationalkonomie, 2nd Edn [original spelling,
1842]. Leopold, Rostock; trans. WARTENBURG C. M. (1966) Von Thnens Isolated State, Intro. by P. HALL . Pergamon,
Oxford.
WREN C. (2009) Onward the spatial: an essay on the nature and relevance of regional economics, Spatial Economic Analysis 4,
447465.

You might also like