Economic Geography
ECONOMIC GEOGRAPHY
ECONOMIC GEOGRAPHY INDEX: 1) Introduction
2) Historical Path 2.1 Germanic Geometry 2.2 Social Physics 2.3 Cumulative
Causation 2.4 Local Extenal Economies 2.5 Land Rent and Land Use 3) Krugma
ECONOMIC GEOGRAPHY
INDEX:
1) Introduction
2) Historical Path
2.1 Germanic Geometry
2.2 Social Physics
2.3 Cumulative Causation
2.4 Local Extenal Economies
2.5 Land Rent and Land Use
3) Krugman’s model
3.1 What is about
3.2 The formal model
3.3 Summary
4)Conclusion
4.1 What do we learn?
4.2 Central and periphery in Europe today
4.3 Concluding thoughts
ECONOMIC GEOGRAPHY 1. INTRODUCTION
During the 1960’s and 1970’s International
trade theory was almost entirely dominated by models based on constant
returns of scale and perfect competition. Even the assumption of immobility
of factors had always represented a very distinctive characteristic of
those models. For example lets think of the Hechsher-Ohlin Theory which
we studied during our course of International Trade. It makes of all those
assumptions its fundamental base. This and other models were a sort of”counter culture” in international trade that didn’t claim that other forces
explained trade patterns.
Later new models were developed. They made
possible to write down coherent, rigorous and often elegant models of what
the economic and industrial reality was about: increasing returns and imperfect
competition. Both factors gave a new understanding of all trade patterns,
also explaining the role played by fixed costs, demand (domestic and foreign),
taste for variety and transportation costs within the inter-industry and
the intra-industry trade flows. That’s why increasing returns were no longer
something to be avoided or assumed away at all costs. Furthermore IRS and
imperfect competition seemed like a natural next step in development of
trade theory rather than a set of disparate alternatives approaches. As
a consequence all the new intellectual opportunities offered by this revolution
in theory had in turn transformed a series of other fields. Specialisation
moved to be based on increasing returns, rather than an effort to take
advantage of exogenous differences in resources or productivity (H-O and
Ricardo theories).
In the last few years the “new economic
geography” has been a movement toward applying the above concepts from
the theory of international trade, in order to analyse the location of
production and industry within countries; or more precisely within economic
units characterised by high degree of labor mobility. The latter concept
is really important, because political and geographic boundaries don’t
count anymore in economic geography. Countries are not points. Now factors
mobility and distances matter a lot. That is why in Europe it’s wrong to
consider Germany or France as an economic unit. However, is correct looking
at the area including Belgium, western Germany and northern France: that
in fact shows a region with an high degree of factors mobility, inside
trades and, especially, huge manufacturing concentration and specialisation
levels.
In order to better understand the importance
of location theory, and even more of its responses and results, let just
give a quick overview of the following figures about the European economic
geography:
the richest regions per capita are also
the most populated
the wealthy regions are clustered close
together
higher wages where the access to the markets
is better
This paper will try to give the basic
answers to those issues.
In conclusion , it’s necessary to underline
that market structure assumptions, here in terms of IRS and imperfect competition,
play a key role for location theory, while they are not so relevant for
international trade. Another divergence between the two fields is that,
for years, international trade theorists have almost worked completely
without reference to the idea of location theorists and regional scientists.
This lack in communication is also due to their different approaches, the
first ones dominated by the traditional presence of CRS, while the second
have always taken very seriously the role of economies of scale both internal
and external . That is probably the reason why many economists have supported
that studies in economic geography offer a basic rethinking of the economics.
2 HISTORICAL PATH
Even though the entire paper will be based
on Krugman’s theory and model, the most recent and well developed model
in economic geography, it will be also important to give an historical
prospective at the issue.
Five are the principal traditions in the
field of spatial economics. Four out of five represent different ways of
looking at the same thing, so they are almost rival alternatives; while
the fifth (von Thunen) largely divorces conceptually from the other approaches.
They are: Germanic Geometry, Social Physics,
Cumulative Causation, Local External Economies and Land rent and Land Use.
2.1 Germanic Geometry
It was born in Germany in the first half
of the century, concerning with a distinctive problem: the geometry of
location on a two-dimensional landscape. It had two traditional subsets:
Weber’s analysis on the location decisions of a firm serving one or more
markets and relying on one or more sources; and the central-place theory.
The latter for the first time discussed the existence, location, and role
of a manufacturing central serving a hypothetical evenly spread agricultural
population.
The results said that the market areas
should be hexagonal (Loch) with a hierarchical role played by the central
(Christaller). These kinds of results were often consequences of the strong
and, even sometimes unrealistic, assumptions of distribution of demand,
and the relationships between costs of transportation and distances.
If, on one hand the Germanic traditions
introduced for the first time new issues, economic parameters and trade-off
(economies of scale vs. transportation costs), on the other hand they revealed
some problems. First of all, Losh was blurry about who was making the decisions
and how decisions of individuals might effect one other. That will be the
key point for the further theorems (i.e. cumulative process).Second, market
structure description wasn’t fully provided. Not even forgetting the difficulties
of accessibility for non-German speaking.
Nevertheless, both served a sort of schematic,
a way of organise thoughts and data about urban systems.
2.2 Social Physics
This new approach faced the location matters
indirectly, as the solution to some maximum or minimum problem. It introduced
dynamics in the location process and decision as well, doing geography
in analogy with physics, in applying the old neo-classical economic analysis
(i.e. the Zipf’s law that gives the distribution of city sizes via incorpora
CorelEquation s , where N is the population of city j, R its rank and
b an exponent close to one; or a very useful index like the market potential
one incorpora CorelEquation s , that depends on k the market, Y the
income or purchasing power of the market, D the distance between two markets,
and g(.) some declining function).
All of this offered a new scientific point
of view for special economics, guaranteeing some striking empirical regularities
and useful basis for empirical work.
2.3 Cumulative Causation
In this fields the two pioneers were Harris
and Lowry with their two papers, one on the US manufacturing belt , and
the other on the numerical land use.
At the basis of their works there was a
circularity relationship between two economic parameters not really new
for spatial economics (market potential analysis): market size and firm
location. Basically they affirmed that firms want to locate where market
potential is high, and markets tend to be larger where lots of firms locate.
The results explain the existence of multiple equilbria.
This approach is very similar to the Big
Push type-stories, one of the bottom lines of the more recent and “advanced”
Krugman’s economic geography theory and model.
In 1980 Murphy, Shleifer, and Vishny formalised
the Big Push, analysing how, in an imperfectly competitive economy, the
industrialisation could move from a bad to a good equilibrium, with a particular
emphasis in case of developing countries.
The main assumption is that the industrialisation
process involves primarily all the sectors with large economies of scale
potential advantages. In order to fully exploit the above advantages it’s
necessary to have a large sized domestic market and also free-cost trade
(that means high exports). Nonetheless, in a imperfect competitive economy
with large fixed costs, there even are other sources of multiplicity of
equilibria, due to pecuniary externalities (IRS external to the firm) and
demand spillovers created by simultaneously increasing returns technologies
in profitable sectors.
Another important assumption is that the
industrialised sector doesn’t keep for itself all 100% of the profits,
because profits are the only channel of spillover.
In this type of economy the investments
profitability in a specific sector is not the only regard. A correct evaluation
of an investment depends as well on being enough for the other related
sectors via demand spillover. That’s why the coordination degree of investments
is an important measure. In these terms infrastructure (intermediate common
goods) shows a quite high degree.
In conclusion, regarding government’s exploitation
of Big Push effects, two are the conditions in order to reach higher levels
of industrialisation and multiple equilibria: free market policies, that
means lower transportation costs and more exports, and coordinated investments
in infrastructures.
The concept of circular and cumulative
causation, could also be associated with Pred’s story, a variant on Big
Push. Briefly, he supposed an import substitution with local production
due to economies of scale. That would have expanded regional employment,
local market and thus induced a cascade of growth with circular relationships
between market size and range of industries within the region.
Definitely all the concepts above offer
a very useful set of boxes and arrows for a better understanding of the
location patterns, decisions and growths.
2.4 Local External Economies
The third tradition is probably the most
close to mainstream economics. The basic idea is that clustering of producers
in a particular location yields advantages, and in turn these advantages
explain such clustering.
These advantages are mainly due to external
economies. They differentiate into technological external economies, which
are pure spillover, and peculiar externalities mediated through the market.
The latter, however, have only the reason to exist and to create advantages
in a specific market structure: a world with increasing returns to scale
at the level of the firm, and with imperfect competition (while the first
ones don’t).
It was Alfred Marshall who presented the
classic economic analysis of the phenomenon. He identified three distinct
reasons for the localisation. First, by concentrating a number of firms
in an industry in the same place, an industrial center allows a pooled
market for workers with specialised skills. It benefits workers and firms,
ensuring a lower probability of unemployment and a lower probability of
labor shortage. Second, localised industries can support the production
in greater variety and at lower costs of nontraded inputs specific to the
industries. Third, because information flows locally more easily than over
greater distances, an industrial central generates technological/informational
spillovers.
If, on one hand the Local External Economies
reinforced the circular causation creating concepts like backward and forward
linkages; on the other hand it focused only on answering why a particular
industry is concentrated, instead than explaining the reasons of the dualism
between a manufacturing core and an agricultural periphery, as the more
recent economic geography will do.
Another critic regards of the monopolistic
competition models they used, models in which economies of scale were purely
external. Those appear to be limited for three reasons. First, the disembodied,
abstract nature of pure external economies may have seemed unsatisfactory.
Second, external economy models of trade typically seemed to yield a bewildering
variety of equilibria, leaving the model with a taxonomy, rather than a
clear set of insights. Finally, much of the traditional literature on external
economies models regarded of external economies as being one source of
modifying or distorting the pattern of specialisation away from that implied
by resource abundance. That’s why external economies are by no means always,
or even usually, consistent with the integrated-economy approach, and the
economic tradition.
2.5 Land Rent and Land Use
It derives directly from von Thunen and
his pioneering Isolated State. His idea is based on the existence of an
agricultural plain supplying a variety of products to an isolated central
city. The configuration of this model is made by a series of concentric
rings in which different crops would be cultivated and/or different farming
methods adopted. That would be a consequence of high transportation costs.
Stays in fact that there would be high rent land near the central which
would be reserved for crops with high costs of transportation and/or crops
yielding high value per acre (certainly not land intensive cultivation).
That means rents would decrease with the increase in distance from the
central.
Its contribution to spatial economy is
important, especially in terms of analysing the internal structure of urban
areas and its equilibrium. However the von Thunen model doesn’t say much
about the central issue, in terms of avoiding the backyard capitalism and
explaining the forces that spread economies activities away from the central
(centrifugal forces), keeping others together (centripetal forces).
As a consequence, economic geography needed
to revitalise, bringing back the other missing half of the story, trying
to legitimise and make sense of the insights of the outcast approach.
All five traditional references above certainly
offered an interesting set of core ideas, and make considerable sense in
light of more recent economic geography analysis. They introduced new important
economic parameters and concepts, also giving a dynamic process for the
interpretation of location and spatial patterns.
However these ideas resulted to be unacceptable
to mainstream economics. The main reason was that they could not at that
time be modelled. The problem was amplified by the fact that even the peculiar
market structure (increasing returns to scale and imperfect competition
), acute problem in this type of matter, wasn’t fully formalised by that
time. Thus, due to the impossibility of modelling macrobehavior explanations,
these economists settled for what they could do, interacting micromotives
via schematic descriptions of data and organising principle that made intuitive
sense.
The methodological problems that afflicted
this field, brought the narrow minded mainstream economists to think that,
what the first ones could not formalise, had to be almost ignored. This
was the reason why the economic literature talked about of exile of economic
geography.
Although economists can often be remarkably
obtuse, failing to see things that are in front of them, sometimes a bit
of obtuseness in not entirely a bad thing. That was the case, because modelling
presents costs but also lots of benefits, in primis the ability of reduce
complexity of methodological problems, simplifying especially through negligible
assumptions and available techniques (nevertheless costs in formalization
too).
In conclusion, however, the sad intellectual
exile of economic geography had not villains, although the unwillingness
to grant even one page in a thousand of fairly sensible efforts and to
make sense of an important subject seemed extreme. Sometimes the temporary
evolution of ignorance may be the price of progress, an inevitable part
of what happens when someone tries to make sense of the world’s complexity.
3 KRUGMAN’S MODEL
3.1 What is about
In response to the inability of economists
to produce models of economic geography that satisfied the profession’s
ever growing demand for rigor ( inability that was in turn, as it was said
before, essentially tied to the problem of modelling specific market structures),
Krugman, thanks to the efforts of industrial and trade theorists over the
past twenty years, was in 1991 able to face geographical matters as rigorously
as it was requested even by the economic journals’ requirements. In such
a way, he claimed that the model demonstrated the feasibility of telling
the kind of stories that are needed to do meaningful economic geography
in a way that mainstream economists can live with.
Krugman’s model is, indeed, simply an attempt
to capture most of the classical writing traditions seen above, within
a formal framework which he draws on the insights and the technical tricks
of the new trade theory.
The entire model is based on the central-periphery
dilemma and relationship,and characterised by the interaction of increasing
returns and transportation costs, so explaining the uneven regional development
at a grand level. It provides at least a rough synthesis of two theories:
the central-place theory and the cumulative causation theory. For both
the formation of cities is driven by economies of scale at the level of
individual firm. The process of city formation is one of cumulative causation,
but the eventual locations of cities tend to have a roughly central-place
pattern. Even for Krugman’s model, those circular relationships among economic
parameters are central, and in many cases they represent the source of
multiple equilibria.
Beside IRS and transportation costs, a
very important economic variable of the model is the demand, especially
in terms of domestic demand. Is the interaction between these three that
drives the cumulative process of regional divergence through circular causation.
In these terms manufacturing will move toward more desirable sites and
away from less desirable ones, but in doing that will change the market
potential map, typically reinforcing the advantage of al ready favoured
location. Thus, the market potential becomes part of the story of circular
and cumulative causation. An example, given sufficiently strong economies
of scale, each manufacturer wants to serve the national market from a single
location. To minimise transportation costs, it chooses a location with
large local demand. But local demand will be large precisely where the
majority of manufacturers choose to locate (loop between manufacturers
location and demand).
If on one side, the model presents a relevant
dynamic process, that is mostly a consequence of a kind of pecuniary external
economy, description certainly not really inconsistent with Marshall’s
one. So that demand externalities and clustering together are indeed two
of the main sources of the entire cumulative path. On the other side, however,
history and accidents (initial advantages and points in time) have a decisive
role and matter a lot. Moreover they perhaps sheer self-fulfilling prophecy,
transforming location from a result of transitory advantages to a result
of sustained advantages (in primis IRS exploitation). The key role of history
depends more precisely on three parameters: in addition to the two mentioned
before, IRS and transportation costs, the third is the share of “footloose”
production not tied down by natural resources. With a sufficiently large
share, production and trade concentration/specialisation patterns would
not depend on comparative advantages, in terms of exogenous differences
in productivity and/or in resources endowments (Ricardo and H-O theorems).
To conclude this brief presentation of
Krugman’s model , it’s interesting to notice two similarities. One, how
closest in spirit to this model is the literature on “market potential”
index, begun by Harris. Second, how through the model is possible to carry
out a number of simulation experiments with a highly stylised economy,
in which locations are lined up symmetrically around a circle. That appears
to be really close to the concepts of spacing the central imagined by Chrisataller
and Losch, and earlier mentioned. In fact it shows a quite defined agglomeration
shadow, where the central could be single or multiple-rivals depending
on the parameter values, that also influence the distances between the
centers (nevertheless always more and more regular).
3.2 The formal model
The purpose of a fully specified and general
equilibrium core-periphery model is to lay out a version of the same story
presented in the previous paragraph, but that doesn’t have any loose ends.
The difference from the traditional urban
approach is that any special direct assumptions are made about localised
external economies or nontradeability, indeed cities are not primitive
concepts in the model. However, as we have learned in class, assumptions
play a key role within empirical testing and modelling, and even here they
do so.
Externalities emerge as a consequence of
market interactions involving economies of scale at the level of individual
firm. Thus it’s necessary to model an imperfectly competitive market structure.
The workhouse model of this kind, also used by Krugman, is the Dixit-Stiglitz
model of monopolistic competition. If, on one hand, it is grossly unrealistic
and involves a kind of layering of implausible assumptions, on the other
hand, it is tractable, convenient, and it has turned out to be a remarkably
flexible tool of analysis.
Let consider a country that contains two
locations, East and West, and produces two kind of goods (two sectors),
agricultural and manufactured. Agricultural production is homogeneous,
produces under constant returns of scale and in perfect competition. Manufactures
consists of a number of differentiated products, each produced subject
to economies of scale, with a monopolistically competitive market structure,
and everyone in the economy is assumed to share tastes, sharing the same
Cobb-Douglas function of consumption of aggregate goods:
incorpora CorelEquation s
where incorpora CorelEquation s
is the share of manufactured goods in expenditures. Manufactures, however,
is a composite of a large number of symmetric products varieties, with
a constant elasticity of substitution between any two varieties ( incorpora
CorelEquation s ), so:
incorpora Equation.2
There are as well two factors of production,
each of which is specific to a particular sector. Labor and possibly capital
are the mobile factors, while land is the immobile factor, mobile and immobile
both used in both sectors. The peasant population, who produces agricultural
goods, is assumed completely immobile between regions, with a given peasant
supply of (1-u)/2 in each region; workers are however mobile, moving to
whichever location offers them a higher real income; and the total supply
is constant:
incorpora CorelEquation s
Farming takes place under CRS, thus farm
labor used in producing any given quantity of agricultural goods can be
set equal to production:
incorpora CorelEquation s
Manufacturing, however, is characterised
by IRS and, thus, involves fixed costs and constant marginal costs:
incorpora CorelEquation s
Because the economy-wide supply is fixed,
if incorpora CorelEquation s is the force of farm labor in the location
j ( share taken exogeneously ), and at any point in time a share incorpora
CorelEquation s of manufacturing labor is also in location j, then
these shares will evolve to:
incorpora CorelEquation s incorpora
CorelEquation s
Finally, let introduce transportation
costs. For simplicity, some unrealistic assumptions about these costs.
First, assume that they apply only to manufactured goods, and second assume
that they take the iceberg form introduced by Paul Samuelson. That means
only a fraction, y;1, of any manufactured goods shipped melts away on
en route (arrives). It has two advantages as a model trick. First, it eliminates
the need to introduce transportation as an additional sector. Second, it
implies that the elasticity of demand with respect to a firm’s f.o.b price
is the same as that with respect to its c.i.f price, eliminating many potential
complications. Moreover, assuming that agricultural goods’ transportation
is costless, is very conventional because it ensures that the wage rate
of farmers and the price of those goods is the same in the two locations.
Specifically, let x be the amount of some goods shipped from j to k , and
let z be the amount that arrives, so:
incorpora CorelEquation s
where y is the inverse index of transportation
cost and incorpora CorelEquation s is the distance between the two
locations. y will be also the final parameter determining whether regions
converge or diverge.
Now, turning to the behaviour of firms,
let describe some of the features of short-run equilibrium. There are a
large number of manufacturing firms, each producing a single product. The
profit-maximising strategy is to set a price as a fixed mark-up over marginal
costs:
incorpora CorelEquation s where
w is the wage rate within the location.
If there is free entry of firms into manufacturing,
profits will be driven to zero. The zero-profit condition may be written
as:
incorpora CorelEquation s
price condition, where the price is equal
to the average cost. This means that the ratio of average costs to marginal
costs (one measure of economies of scale) is incorpora CorelEquation s
, thus equilibrium economies of scale are function only of incorpora
CorelEquation s , a parameter of taste rather than of technology, nonetheless
acts as a sort of inverse index of importance of IRS.
The zero-profit and pricing condition imply:
incorpora Equation.2
Furthermore output per firm is the same
in each region, irrespective of wage rates, relative demand, and so forth.
Since all varieties are produced at the
same scale, the number of varieties produced at any given location is simply
proportional to that location’s manufacturing labor forces:
incorpora CorelEquation s
The latter equation plays a crucial role
in the whole analysis of this approach. Within it stays visible the increasing
returns parameter that is crucially on the logic of the model. That makes
profitable to produce each variety in only one location, that means different
locations produce differentiated bundles of products. When a location gains
labor it doesn’t produce more of the same mix of products, but adds new
products. This quantization of production is the only way in which increasing
returns enter the solution, but is perhaps enough.
The short-run equilibrium is also possible
to be represented as the solution of four sets of equations: the income,
the price index, the equilibrium wage rate, and the real wage.
The income of each location is equal to
the sum of farmers’ (share of 1-u/2, same in each location) income and
workers’ (share of u just in one location, the manufacturing concentrated
for hypothesis) income:
incorpora CorelEquation s
In the two region model (all equations
easily solved by the computer), the income of the manufacturing concentrated
location will be Y=(1+u)/2, while in the other location, that only has
its immobile farmers, the income will be Y=(1-u)/2.
The price index of the manufactures aggregate
to consumers in each location will be function of the fixed mark-up on
its marginal cost, which in turn is proportional to wages, of the transportation
costs and thus of distances. Given the constant elasticity of substitution
function the price index will therefore be:
incorpora Equation.2
Staying always into the two region model,
the price of manufactured goods in the “only farms” location will be 1/y
times as high as that in the other location. It means that the overall
price index, which is a geometric average, will thus be incorpora CorelEquation
s times as high.
Given the price index, the equilibrium
wage rate can be solved as:
incorpora Equation.2
The last equation, however, only determinates
the wage rate in terms of agricultural goods, but because workers are more
interested in terms of real wage (consumption basket). Thus, it would be
more correct write:
incorpora CorelEquation s
Moving again to the two region model,
where for assumptions incorpora CorelEquation s and incorpora CorelEquation
s , normalising the distance between the two locations to 1, the real
wages will turn to:
incorpora CorelEquation s incorpora
Equation.2 (*)
At this point, it’s possible to bear out
the intuition that agglomeration is due to the circular relationship between
the location of the market and the location of manufacturing. If manufacturing
were a very small part of the economy, u close to 0, the real wage in location
two would be:
incorpora Equation.2 which is less
than 1 because of Jensen’s inequality. That means, it would be advantageously
to move away from any concentration of manufacturing. This shows the centrifugal
force in the model, that pushes toward dispersal of workers. Assumed the
immobility of farmers, if there were no economies of scale , workers would
do of course have an incentive to move toward locations where they were
a relatively scarce force. However, manufacturing sector (huge economies
of scale) is a significant part of the economy. Therefore there are two
other forces, centripetal, which work to hold the agglomeration. First,
the region that has the a larger population of workers will also offer
a larger market for manufactured goods (a kind of Hirschman forward linkage).
Second, workers will, other things equal, receive a higher real wage if
they are located closer to the suppliers manufactured goods (backward linkage).
Mathematically, it means that the first term in equation (*) becomes less
than one, and the expression inside the brackets involves an higher weight
on the component that is less than one and a lower weight on the component
that is greater than one.
Regarding the second parameter y, the transportation
costs, concentration is more likely when it is low. When y=0 , incorpora
CorelEquation s , no transportation costs location doesn’t matter.
With y close to 0
incorpora CorelEquation s is less
than 0, while it becomes greater with the increasing of y. Moreover when
incorpora CorelEquation s (the first term is the ratio of hence average
cost to marginal cost and price-a measure of economies of sale) than as
y grows the real wage in location two must exceed one. It means that if
scale economies and the manufacturing share are too sufficiently large,
workers will prefer to cluster together even with prohibitive transportation
costs. Illustrating the relationship between real wage and transportation
cost, the shape would be the one showed in allagate 1. Notice that there
is a critical value of y (y*),under which concentration is an equilibrium.
Considering the effect of other two parameters
on the critical value y*, it possible to show, holding first incorpora
CorelEquation s and then incorpora CorelEquation s , that:
incorpora CorelEquation s incorpora
CorelEquation s
They mean that concentration would be
more likely the higher is the share in manufacturing, and that an high
elasticity of substitution(equilibrium degree of economies of scale)works
against agglomeration. In conclusion, all it has been presented in the
previous static analysis shows nothing more that the effects of centripetal
and centrifugal forces on agglomeration decisions.
3.3 Summary
The framework developed in the model is
extremely special and unrealistic. Nonetheless, it has the useful feature
of being relatively tractable in an area that has long seemed very resistant
to formal modelling. And rather than displacing the traditional insights
in that field, this effort at modelling seems to confirm their usefulness.
Thus, one of the pleasant surprises of applying the methods of new trade
theory to location issues is that they appear both to validate and to unify
a number of seemingly disparate traditions in economic geography, especially
the cumulative process, with circular causation with the market potential
functions ,with central-place theory.
Costs of transactions across space and
economies of scale in production are the two key facts in the entire story.
Because of economies of scale producers have an incentive to concentrate
production of each good or service in a limited number of locations. Because
of the costs of transacting across distance, the preferred locations for
each individual producer are where demand is large or supply of inputs
is particularly convenient. Thus concentration of industry, once established,
tend to be self-sustaining. In terms of economic parameters, agglomeration
is favoured by low transportation costs, low incorpora CorelEquation s
, by a large share of manufacturing in the economy- expression of a
self-sustaining manufacturing concentration due to forward and backward
linkages -, high incorpora CorelEquation s , and strong economies
of scale at the level of the firm, low incorpora CorelEquation s .
However, there are other very important parameters that indicate key determinants
of the nature equilibrium in location, and which influence and are influenced
by the whole cumulative and circular process. First, the share of non agricultural
goods expenditures, that is positively correlated to agglomeration; second,
the initial distribution of population, positively correlated, and third,
the footloose production/demand not tied down by natural resources, even
positively correlated.
Further thoughts ought to be spent on transportation
costs’ issues. First of all, something about the misleading impression
occurred between the Weber-type story of transport cost minimisation and
the central-periphery model by Krugman. Weber suggested that localised
industrial complexes will emerge only if it is more costly to transport
intermediate inputs than final goods, conveying the impression that localisation
due to the clustering of suppliers occurs only in this special case. In
the Krugman’s model, however, the localisation will tend to occur unless
the costs of transporting intermediates are particularly low compared with
those of transporting final goods. Moreover, a general reduction in both
transport costs will ordinarily tend to encourage localisation rather than
discourage it. To see why, it is useful to consider a tricky sort of model
where intermediates and final goods are the same thing then all will appear
clear following the usual model.
Second thought, given a central location
in which the wage rate and hence production costs are high but which has
good access to markets, and a peripheral location in which labor costs
are low but access to markets is less good, a reduction in transportation
costs will have two effects. On one hand, it will facilitate locating production
where it is cheapest, on the other it will also facilitate concentration
in one location in order to realise economies of scale, and that will happen
in the location with higher costs but better access. All of this in order
to understand that, if the goal is minimise the sum of production and transport
costs, production costs (in terms of wage) are not the central matter.
What counts is the trade off between transportation costs and economies
of scale. Strictly regarding transportation costs, there are three cases:
high , medium, and low(zero). When they are high, production will take
place in both location, the opposite when they are low (zero), it will
be total concentrated in low wage location. Peculiar is the case when they
are medium, decreasing from the high status. Here the production will shift
away from the low wage location, because access to the market will become
the key determinant of location. This fully explain the u shape of the
graph in allegate 1 rather the monotonic: transportation costs matter in
locating and in concentration decisions whenever are small, but exist.
4 CONCLUSION
4.1 What do we learn?
Came to the end of this paper, covering
a quite new “branch” of international economy and trade (stays in fact
the available literature on these issues is not abundant at all. Indeed
that creates some difficulties due to the impossibility to confront different
sources and points of view), the things we learned a basically four.
First, we saw how important modelling is
in economics. In chapter 2 we faced different traditional approaches to
economic geography, which even offered few goods and correct ideas/concepts.
However, the inability of modelling by spatial economists, caused that
those had been unacceptable to mainstream economics. Definitely the benefits
of modelling are remarkable, cutting through the complexies of a situation.
Nevertheless, we also observed more the role played by the assumptions
in every model. Whichever they are: negligible, domain, or heuristic, they
always are the key factor in order to make significant a theory. It doesn’t
matter if they are false or unrealistic, because their purpose is to create
a flexible set of tools of analysis of the phenomenon or to be just a set
of steps toward more precise predictions.
Second, both ,the early tradition in spatial
economics and the Krugman’s central-periphery model, are mainly based on
the concept and principle of circular and cumulative causation. All the
economic parameters we mentioned interact one to each other, through forward
and backward linkages. The “loop” between domestic demand and manufacturing
location is probably the core of the entire model. Although even levels
of transportation costs effect the dynamics in manufacturing concentration
and specialisation, nevertheless within the circularity history matters,
and a lot. Over every circle of equilibrium there is a initial point, an
initial advantage, an accident, that is the primary origin of the whole
economic and trade pattern. That’s why dramatic changes can happen at any
time, we must just be prepared to, be flexible.
Third, whenever we talk about economic
geography and inter-regional trade (even international trade if we consider
the new concept of nation as it has been presented at the beginning of
the paper), we mostly refer to the pattern of intraindustry trade, a two-way
trade within the x industry.
If the varieties of x were produced under
constant returns, it would be possible to reproduce the integrated economy
by dividing the production of each variety among countries, and there would
be not need to engage in intraindustry trade to satisfy consumers’ taste
for variety. But is IRS that prevent each country from producing the full
range and that explain concentration and specialisation within the x industry.
So, the essential reason for the intraindusty trade is the existence of
economies of scale..
Krugman’s model takes widely in consideration
this economic parameter as well, in terms of degree of differentiated products
( incorpora CorelEquation s ), reinforcing that IRS (always linked
to demand/market size) are today the leading parameter in relocation for
manufacturing production, in both directions, toward the core and the periphery.
That’s why Volvo and Volkswagen exist and why Sweden, although periphery,
is still leader in pulp and paper production. However, attention, transportation
costs still distort the trade pattern because they still exist, even though
the world seems to be more and more integrated.
4.2 Central and periphery in Europe today
It’s very interesting to evaluate the
European Community integration in terms of economic geography in order
to better understand the patterns of manufacturing localization and competitiveness
between the two “poles”: central (Belgium, France, and Germany) and periphery
(Southern and Northern Europe). First of all, the main reason for keeping
manufacturing relocation away from the low-wage entering countries (reasonable
following the comparative advantages approaches) is due to the importance
of transportation costs and, even more, of the easier access to the market
which the central has. Both factors, as we have seen before, play a crucial
role in cumulative-circular causation, and thus in the model. That especially
happens, because simultaneously with integration doesn’t occur any complete
elimination of those costs, but only a shift from high to medium cost case,
indeed reinforcing the central’s role. Definitely that is the case of Europe,
where moderate barriers, transport costs, difficulties of communication,
and cultural differences as well as governmentally imposed costs, interacting
with economies of scale have encouraged concentration in high-cost locations
with good market access.
The degree of spacialization of European
nations is however less than US regional one. The US industry is far more
localized. Obviously the reasons are due to that in Europe still exist
barriers of trade (tariffs in primis), and that we are at a different point
in the cumulative process. The latter is even supported by completely different
role played by history and accidents in the last decades, especially in
terms of factor mobility and trade.
The population distribution of Europe is
as well nothing like the unevenness of the American distribution, even
though Europe is characterised by a strong center-periphery pattern more
considering purchasing power(income), than population. That means there
is a more casual relationship running from peripheraly to income.
Another characteristic, that differentiates
Europe from the US, is the multiple European equilibria, instead than a
single core equilibrium typical of the US. An explanation could be that
European economic integration took more the form of intraindustry, rather
than interindustry, as happened in the past in the US.
In conclusion, we should say that, on one
hand in Europe center-periphery pattern is primarily not the result of
forces that the model stress. On the other, even applying the model, integration
and improved access might actually hurt, not help, peripheral industry
due to high and substantial natural barriers.
4.3 Concluding thoughts
Two last brief remarks. First, the entire
Krugman’s center-periphery model is defined in such a way that location
matters and a lot for manufacturing success. Nevertheless there are other
way of thinking. An example is the one in which infrastructure and human
capital investments would count more in determining whether a country will
expand or shirk its manufacturing sector. In other words, being located,
one could shift from the periphery to the center by accumulating human
capital and physical capital. That could be seen as a new way of interpreting
the future development of peripheral countries in Europe, certainly opposite
to the negative path drawn by Krugman’s economic geographical approach.
Second, transportation costs, that are
an important key determinant of “our” model, are also central to model
multinational enterprise theory. Models in which firms place operations
in different countries for comparative advantage reasons, are today unsatisfactory
as a complete explanation for the actual pattern of foreign direct investment.
The alternative view indeed is that firms go multinational in order to
improve their access to markets, avoiding transportation costs or other
barriers to trade in their products. For example the Brainard and Horstmann,
and Markunsen have recently offered models of MNE in which the decision
of go multinational reflects the trade-off between the loss of economies
of scale associated with multiple plants, and the reduction of transport
costs they can achieve by producing locally for each market. These models
are however fairly general. But factor proportion as well as transport
costs may motivate overseas production and firms can make a trade-off between
fixed and variable costs. MNE theories represent a quite new field of international
economics, even very interesting in these days, when globalization is very,
sometimes too much, fashioned. Unfortunately now there is any time and
space in order to face those issues further.
References
Dicken, P. and Lloyd P. (1993), Nuove
Prospettive su Spazio e Localizzazioni-Le piu recenti interpretazioni
geografiche dell’ economia- a cura di G. Rizzo e C. Robiglio, Franco Angeli
(titolo originale: Location in Space. Theoretical Perspective in Economic
Geography, traduzione unica di R. Gasperoni)
Krugman, P. and Venables, A. (1990), “Integration
and the competitiveness of peripheral industry”, in C. Bliss and J. Braga
de Macedo, eds., Unity with Diversity in the European Community, Cambridge:
Cambridge University Press
Krugman, P. (1991), Geography and Trade
(MIT Prees, Cambridge, MA)
Krugman, P. (1991), “Increasing returns
and economic geography”, Journal of Political Economy 99:483-499
Krugman, P. (1993), “On the number and
location of cities”, European Economic Review 37:293-298
Krugman, P. (1995), Development, geography,
and economic theory (MIT Press, Cambridge, MA)
Krugman, P. (1995), “Increasing returns,
imperfect competition and the positive theory of international trade” (MIT
Press, Cambridge, MA) Handbook of International Economics, vol. III, Edited
by G. Grossman and K. Rogoff, Elsevier Science B.V.
Murphy, K., Shleifer, A. and Vishny, W.
(1989) “Industrialization and the Big Push”, Journal of Political Economy
97:1003-1026
Musgrave, A. (1991), Unreal Assumption
in Economic Theory, Kyklos
