Distribution Management Master Supply Chain Management
Articles Author: Andreas S. Melisse
Lecture 2 – Tuesday 04-02-2021
The evolution of spatial economics: from Thünen to the new economic geography
– Fujita, M. (2010)
1. Introduction
Spatial economics should include all branches of economics dealing with the analysis of economic processes and
developments in geographical space. The fundamental theory of spatial economics is called general locations
theory, which generally explains the geographical distribution of all agents in a given location space, together
with the associated spatial prices system and trade pattern. The general theory of location and space-economy
is conceived as embracing the total spatial array of economic activities, with attention paid to the geographic
distribution of inputs and outputs and the geographic variations in prices and costs. Or in other words, in general
location models “which are concerned with the location of all economic activities, this distribution becomes itself
a variable”. In summary, a general location model aims, either in a descriptive or normative context, to explain
the geographical distribution of all agents in a given location space, together with the associated spatial price
system and trade pattern. Here, a “given location space” may represent a metropolitan area, a nation, a system
of nations, or the entire world.
2. Thünen’s The Isolated State: The birth of spatial economics
Von Thünen imagined an isolated state where a very large town is
located at the centre of a homogeneous plain. He attempted to
determine simultaneously all variables of the economy through
competitive markets of goods, labour and land, with a special focus
on the land use pattern and land rent pattern in the agricultural
hinterland. The question he asked himself: how will the land be
allocated if there is free competition among farmers and landowners
with each individual acting according to his perceived self-interest?
This simple model indicates three lines which indicates the distance
from the Town and the Rent concerning three types of goods/crops.
Each individual line indicates the bid rent curve and the bold line is
the market rent curve.
The ‘weaker’ the product the higher the cost and should therefore be located closest to the Town so the high
transportation costs can stay relatively low. The least vulnerable product has the lowest transportation costs and
can therefore be best placed further from the Town. Two weaknesses of Thünen’s theory:
1) There is an assumption for perfect competition.
2) He had the assumption that the transportation costs were proportional to distance.
Von Thünen also investigates the forces behind industrial agglomeration (cluster). He discovered factors that
favour the location of industries in large towns:
1) Only for large-scale manufacturers it is profitable to install labour-saving machines.
2) The scale of the factory depends on the demand for its products.
3) Large-scale factories are viable only in the capital in many branches of industry.
4) Machinery is produced efficiently only in a place where factories and workshops are close enough
together to help each other work in unison.
1, 2, and 3 resembles the ‘basic story’ of Krugman (1991) and these in combination with the last factor forms
another fundamental explanation for the emergence of industrial agglomeration. If Von Thünen would combine
his monocentric spatial economy (an Isolated State) with his industrial agglomeration findings, it would become
a typical model of general location theory.
3. Non-competitive models of industrial location
In the first half of the 20th century, the Industrial Revolution was developing, and the industrial location theory
was advanced by German-writing scholars. Analyses of industrial location were presented in a partial equilibrium
1
,Distribution Management Master Supply Chain Management
Articles Author: Andreas S. Melisse
framework: under the given location of consumers/markets, locations of plants/firms where studies by
considering their invisibility of scale economies.
The first systematic study of the so-called market area analyses was presented by Launhardt (1885). He also
studied the spatial price policy of two oligopolistic firms, which turned out to be a special case of the more
general analysis by Hotelling (1229). Launhardt assumed that two firms locate at each end of a linear segment
and supply the same product.
In 1890, Marshall presented his study on industrial agglomeration, in which he examined systematically the
reasons for the concentration of specialized industries in particular localities. He described something like a lock-
in-effect: a settled industry will stay long on its chosen location and as an agglomeration is developing,
advantages, developments and creativity can be shared among other firms.
Weber (1909) developed a unique theory of industrial location in which he considered the optimal location of
the plant which minimizes the total transport cost per unit output. He did, however, not consider endogenous
determination of the prices of inputs and output (some materials are cheaper at different locations).
During the 1920s and 1930s, non-competitive economic models induced the development of new industrial
location theory based on the non-competitive behaviour of firms. Hotelling (1929) did a study on spatial
competition which became a more general framework of spatial price theory. In his study, he explains how two
firms, selling the same product on a linear market will tend to agglomerate at the centre of the market to gain
market share (but end up having more disadvantages than advantages. Remember story during lecture about ice
vendors on a beach who end up selling both in the middle of the ‘selling range’). Palander (1935) states however
that competitors will keep a distance from each other to avoid price-competition.
Lösch (1940) presents a system of equations
describing a full general equilibrium of all locations
and prices. He sets a more limited but more exiting
task of examining the concrete spatial structure of
the market area in a simple, abstract ‘economic
region’ (through interplay of purely economic forces).
The AC curve shows the average costs. Df indicates
the demand in case of the free competition (market
areas of firms do not overlap). At intersection point
N, the firm can operate with positive profit.
The price for the customer can be set by M.
Suppose that firms are uniformly
distributed over the plain with all
neighboring firms have the same distance.
With enough distance (a), the market area
(circle) does not overlap with other market
areas. In this case, each firm can choose the
optimal price. By decreasing the distance between the adjacent firms, the market areas are tangent to each other
(b) and each firm is free of competition. But, with a small change occurring, competition can occur, and the
demand curve shifts to the left. A further decrease of distance, the market areas become regular hexagons (c).
continuing further, the density of the firms becomes sufficiently high (d), the demand curve of each firm is just
tangent to the average cost curve. The demand curve that belongs to situation (d) is the most left De. the
intersection point with the AC is indicated with N’. N’ indicates the marginal cost, which is the average cost or
the marginal revenue. The profit of each firm is zero. This represents the so-called Löschian spatial equilibrium
of oligopolistic industry.
Christaller (1933) created the central place theory for an ideal system of supplying a dispersed population with
‘central goods’ or public services. The large number of central goods are ordered into groups in terms of their
‘market ranges’ (determined by transport cost and degree of scale economies in production). He produced
evidence for hierarchy: the city at the top would produce all groups of urban goods and lower-order cities fewer
groups of urban goods.
2
, Distribution Management Master Supply Chain Management
Articles Author: Andreas S. Melisse
4. Soul searching on the foundations of spatial economics and the Spatial Impossibility Theorem
The Spatial Impossibility Theorem: Assume an economy with a finite number of locations and a finite number of
consumers and firms. If space is homogeneous and preferences are locally non-satiated, then there is no
competitive equilibrium involving costly transportation. On the one hand, this theorem illuminates the
fundamental limitation of the competitive paradigm in the study of spatial economies, it implies that when space
is homogeneous and transportation is costly, the only possible competitive equilibrium is the so-called backyard
capitalism in which every location operates as an autarky. On the other hand, this theorem is also very useful in
delineating possible types of interesting general location models. It tells us that if we want to understand
interesting phenomena about the spatial distribution of economic activities, in particular the formation of major
economic agglomerations as well as regional specialization and trade, we must make at least one of the following
three assumptions: (1) space is heterogeneous; (2) externalities in production and consumption exist; and (3)
markets are imperfectly competitive.
Monopolistic competition leads to a departure from the competitive model and allows for firms to be price-
setters and to produce differentiated goods under increased returns; however, strategic interactions are weak
because one assumes a continuum of firms. Oligopolistic competition faces the integer aspects of location
explicitly. That is, we assume a finite number of large agents who interact strategically by accounting for their
market power.
5. Urban morphology
In all models introduced thus far in this section, the agglomeration forces in the city are supposed to arise from
nonmarket interactions (technological externalities) among agents. In contrast, Fujita (1988) represents the first
general location model for class C1, which demonstrates that pure market interactions alone can generate spatial
agglomeration of economic activities. Firms no longer assume that they can sell whatever they want at given
market prices. Instead, each firm is aware that its optimal choice (location and price) depends on the demand
for the variety it supplies. This demand itself rests on the spatial consumer distribution, thus showing how firms’
choices are directly affected by consumers’ choices. In turn, the optimal choice of a consumer (location and
consumption) depends on the entire firm distribution.
6. The New Economic Geography (NEG)
The NEG of Krugman (1991) has quickly emerged as one of the most exciting areas of contemporary economics.
The defining issue of the New Economic Geography is essentially the same as that of general location theory, i.e.
how to explain the formation of a large variety of economic agglomeration in geographical space. NEG is about
interaction among increasing returns, transport costs and movement of production factors. A characteristic is
the presentation of an approach that emphasizes the three-way interaction among increasing returns, transport
costs and the movement of production factors.
In early models in the NEG: agglomeration forces arise solely from monetary externalities and not from e.g.
technological externalities. The NEG is however widely open to further development instead of a narrow range
of models and issues. NEG models are divided into three groups:
1) Regional models. In which the economy is divided into a discrete set of regions.
2) International models. In which the economy consists of a discrete set of nations.
3) Urban-system models. In which the economy consists of a system of cities and their hinterlands.
The core-periphery model of Krugman provides a basic framework for the new economic geography: it illustrates
how interactions among increasing returns at the level of the firm, transport costs and factor mobility can cause
spatial economic structure to emerge and change. The core-periphery model works like the 2x2x2 models: two
regions, two production sectors (agricultural and manufacturing), and two types of labour (farmers and workers).
Manufacturing sectors produces differentiated products; agriculture sectors produced homogeneous goods
under constant return. Workers are freely mobile between regions; farmers are immobile and distributed equally
between regions. Centrifugal force is the immobility of farmers because they consume both types of goods,
whereas centripetal force is the circular causation of forward linkages (workers want to be close to the producers
of consumer goods) and backward linkages (producers want to concentrate where the market is larger). The
economy will end up with a core-periphery pattern in which all manufacturing is concentrated in one region,
and is likely to occur in three cases: (1) when transport cost of the manufacturers is low enough; (2) when
varieties are sufficiently differentiated; and (3) when the expenditure on manufactured goods is large enough.
3
