towards a critique of industrial location theeory

8/2/2019 Towards a Critique of Industrial Location Theeory

http://slidepdf.com/reader/full/towards-a-critique-of-industrial-location-theeory 1/7

T O W A R D S A C R I T I Q U E OF I N D U S T R I A L L O C A T I O N T H E O R Y

I INTRODUCTION

In attempting a critique of a discipline, the fu ndamental prob-lem always arises that the very concep t and definit ion of discip-lines are themselves fun ctions of a particular ideology and

epistemology. From the point of view of the critique presentedhere, the separate existence of an entity called industrial

location theory is itself open to qu estion. In different ways,many o f the classic theories of industrial location have pro-ceeded as though the object o f stud y was an abstract firm -one without effective structural relationships to the rest ofthe economy. The specific problem of idealised abstractionwill be dealt with later. The imm ediate question is the relatedone of the p resumed separation of spatial behaviour f rom th eeconomic system as a whole. In fact, of course, the two areintimately related at all levels. In the first place it is rarelyvalid to Zetain a com plete distinction between t he specificallylocational decision o f the firm and all its other eco nomicdecisions. Secondly there is the fact th at th e nature of a firm’sbehaviour will be influenced by its position within the total

economic structure. And thirdly, at a more aggregated level,the spatial shape of the economy is the result no t only of

specifically spatial forces, b ut also of the a-spatial dynamic o fthe econo mic system having a spatial manifestation.

It is, then, impossible realistically to tr eat “the spatial” as aclosed system. Certainly industrial location theory does no thave an o bject of it s own, and in that sense there could neverbe an autonomous dynamic theory of industrial location. Notall of th e work considered here is in contradiction with th isview, but the point needs to be borne in mind throughout thefollowing discussion since many e lemen ts of the critiq uespring directly from it.

location theory”, and the spatial expression of the economic

system does have to be analysed. Given that, this critiqueseems a valid exercise to und ertake . The bulk of industriallocation theory is in fact closely related t o economics but

only in the sense that it derives very directly from neo-classicalmarginalist economic theory, sharing its ideology, and conse-quently its epistemological approach .

The threads of industrial location theopy covered in thispaper fall under fo ur major headings. F irst, there is a line ofdevelopment which derives from th e initial work of Weber(1909). This centres on the location decision o f th e individualfirm in a known locational environment, where there is nointerdependence with the locational decisions of o ther firms.In contrast, the second group o f theoretical studies focuseson small numbers of firms in locationally interdependentsituations. Much of this wor k sprang from an original article

by Hotelling (1929).Thirdly, a more “behavioural” approach ,stemming from th e work of Cyert and March (1963),hasrecently developed in response b oth to changing concrete con -ditions and to contradictions within previous approaches. Thefourth theoretical approach considered here is that of AugustLosch (1954), which was different in th at, although it startedfrom an analysis of the individual firm, its prime conc ern wasthe examination of potential whole economic landscapes. Itwas an attempt,essentially, to parallel in spatial t er n s theeconomic concept of general equilibirum.

This elationship between location theory and a-spatialneoclassical economics will form o ne of the themes of thism y . t has influenced the definition of the object for study.

Nonetheless, there is a body of knowledge called “industrial

33

Doreen Massey

the methodolog y, and th e main elements of historical develop-ment (though it is noticeable that the “capital controversy”has as yet had n o noticeable reverberations in the field oflocation theory). This element of the critique involves chal-lenging the ideology of th e approach as a whole. From “within”

that approa ch, however, a second theme, which will recurthroughout the discussion centres on internal contradictionsand problems, particularly those produced by the introductionof th e spatial dimension i nt o the a-spatial neoclassical frame-work. A further contradiction, th at between the develop-ments now taking place in location theory and those emerging(in the form of acute socio-economic problems) in the “realworld” is dealt with in the final section. Another importanttheme concerns the pattern of growth of the theory over time.Changes in theoretical sturcture are brought about in responseto the emergence of problems, but the way in which thoseproblems are “seen” determines the nature of the response tothem and the consequent theo retic development.

In order for a critique of this nature t o be effective, it isnecessary fo r the object of its study to be seen in context. If

the perspective of a whole subject is to be examined it isessential to step ou tside of th at perspective. I t is only by sucha procedure th at on e can (a) appreciate th e nature of the sub-ject for what it is and (b) envisage any po ssibility of changealong the dimension of characteristics which typifies th e sub-ject as a whole. A historical approach provides the necessaryframework for such an analysis, bot h by retaining logical cate-gories, and by refusing t o accept as d atum any “primordial” or

“natural” cond itions. However, the line of industrial locationtheory to be examined in this article takes osgiven the natureof eco nom ic organisation (essentially th at of capitalism, ofboth the nineteenth cen tury and mono poly varieties), butignores the historical context - nd, therefore, the essentialdynamic - f that form of organisation. At th e same time, thetheories themselves can be seen as part o f a historical progres-

sion, “explaining” an aspect of th at given material world fromwhich they stem. The historical perspective is thus towfold:firstly to set theories in their historical con tex t, and thereby t oillustrate both their reactive nature and the role they play inthat con text: and secondly t o see that the approach t o industriallocation - hat is the nature of the theories themselves -should also take industrial loc ation (i.e. it s object) in its his-torical perspective.

It is also impor tant t o consider carefully the nature of th e“space” in which locatio n is taking place. Most industriallocatio n the ory’ deals essentially with som e form of “abstract”space. In the case of Losch, for instance, distance is th e onlyquality o f space considered as locationally significant. In factthe space of industrial location is the product of a complex

historical process. I t is also a political and ins titution al space.As a final caveat to the introduction it should be stated

th at th is essay is not in an y sense seen as a finished critiquefrom an understanding of which we can march immediatelyahead into a newly-formulated industrial location theory. Itspurpose is primarily to initiate the kind of critique which I

feel is needed, an d t o generate discussion - opefullycon ruct ve.



I I ASPECTS OF A-SPATIAL NEO-CLASSICALECO N0MICS

One of the most familiar modes of at tack o n neoclassical

economics is that which co ncentrates on the nature of abstrac-

tion. Dobb ( I 940) discusses this in so me d epth: Once theformal question of internal consistency is settled, the accept-ance or rejection of a theory depe nds on one’s view of the

appropriateness of the particular abstraction on which thetheory is based. He goes on t o discuss tw o approaches: In thefirst place, one may build one’s abstractionon the exclusionof certain feature s which are present in any actual situ ation,either because they are the more variable or becausc. they arequantitatively of lesser importance in determining the course of’

events. Such abstraction he sees as a necessary and justifiable

struct ural simplification of a reality.Secondly. however, one nlay base one’s abstraction . .

simply on the formal procedure of com bining the propertiescommon to a heterogeneous assortment of situatiorls andbuildingabstraction out of analogy. As Dobb points out, suchdistillations of common factors may form such a small part ofthe mechanics of any one situation that the real structure and

motive power is lost. Such can be th e case with neo-classical

economics. We learn, certainly, of prod ucers and consumers,but not of capitalists, workers, imperialism, and privateproper ty. The focus of theory is on an idealism. This approach

has been carried over, almost in its entirety, to much of theindustrial location theo ry discussed in this essay.

of this mode of abstraction, is paralleled by an absence ofhistorical perspective. In m ost of lo cation the ory , as inmarginalist economics, the existence .of numerous perfectly-

competitive profit-maximisers or , alternatively, of an oiligo-poly, is assumed as given - nd consequ ently (which is the

point) unalterable. The dynamic of the system as a whole isignored. Thus, for instance, although both perfectly-competitive and monopolistic or oligopolistic situations are

studied, they are analysed as separate situations, which might

obtain perhaps in different places, or in different sectors ofthe econom y. The dynamic relationships between the two.and particularly the development of one from the conditions

of the other, are ignored. Static equilibrium is the rule and theaim; internal contradictions and the dynamics of developmentare not apparent.

Putting together some of the points made abo ve, however,

we see that there is contradiction even here. On the one handthere is th e pretension to trans-system distillation, on the

other there are firm roots in the contemporary economicarrangement. This contradiction arises not because themrticular abstra ction of neo-classical economics is incorre ct.but because the whole concept of an a-historical formalism ofhuman b ehaviour is a misapprehension. Behaviour itself is aresult of historical conditions and position within the totalsystem at any point in time. Different forms of econom icsystem, and different structural positions within any one suchsystem; will lead to different forms of behaviour. Thereforethe critique is not simply either that “economic man”, fo rinstance, is an inadequate abs traction, or that there should bea simple transference fr om analysis at the level of th eindividual to a ‘structural’ ap proac h. It is necessary t o see

behaviour centrally in a historical context, and therefore tohave an overall structure containing within i t t he necessityfor the evolution of the actural variants in which it exists.

Concepts of economic man, and certain theories of the firm,fail to do this in that the sta rling poi nt fo r their analysis is an

1-historical, self-constitutive subject.

Moreover. the lack of systemic context. which is one effect

I I IT HE DEVELOPM ENT OF MAJOR APPROACHESIN INDUSTRIAL LOCATION THEORY

1 t is probably fair to say that the whole of industrial locationtheory lies within one m ajor, overall “paradigm,” fo r, thoughseparate threads of development exist, they have much in

commo n in term s of their epistemological approach and oftheir fu nction in relation to th e economic system of which

they are a prod uct. But so-called “paradigms” d o no t rise andfall in an academic or intellectual isolation. Theories do not

develop in a vacuum. Each major ap proac h bears a defin iterelationship to the material conditions of its age, to the nature

of economic organisation, and to the social relations whichmay be consequent upon those material conditions.

Marginalist economics and i ts derivative location th eory areideological in the sense that the questions which they ask per-form the function primarily of allowing the expression of

certain desired “answers”. And the particular ideology with

which they approach the stu dy of capitalism in dominantlya function of social relations. Th is ideology determ ines eventhe definition of the objects studied. Thus neo-classical

economics does not talk of surplus value or the rate of ex-ploitation. In consequence i t is possible t o assume a world of

social harmony and com monality of interest. The socially-produced ideology also determines the level of th e m ajorfocus of interest - he individual firm; the nature of that focus- he exclusive consideration of the p oint of view of theindustrialist rather than that o f th e working class (profit isthe criterion, wages are simply labour costs); and the nature ofabstraction - . which excludes the concrete power-relations of

society.

Since the days of Weber it has not been necessary to adoptanother point of view. Minor changes have been made, whichwill be discussed, but these have been m ainly within the con-text of this overall approach. The difficulties already ex-perienced in trying to “use” location theory are mentionedin the final section of this essay. But is is also necessary to

make explic it the social role of a theoretical structure and,

in some cases, the logical con tort ion s performed as part ofthat role. At this level, the demonstration of purely technicaluntenability is also impo rtant -- viz, the “capital controversy”in economics.

Nonetheless, industrial location theory has not followedone continuous s tream of development. There have beendefinite stages and threads in its growth. The rest o f this

section will briefly elab orate those stages and their role. It is

interesting to n ote the different levels at which criticism, and

response to criticism, may occur. The critique in this paper

attempts to make fundamental points about the nature o f theapproach as a whole. When viewed as it were from within theideology, however, the inadequacies are interpr eted differently.Either they will produce minor modifications and sophistic-

ations to th e given theory, as often as not leading to a cul-de-sac of complexity. or they will stimulate a new line of thought,but still not one which breaks out of the ideological framework.

The brief histo ry which follo& attempts to illustrate thisdevelopmen t. and its predominan t characteristics. Thu s, forinstance, thre e levels of critiq ue are described o f the initialwork of Weber. Firstly, there was the line of developmentwhich restricted itself to modification and sophistication ofdetail in the original approa ch. Secon dly, there was a reactionto more fundamental inadequacies, producing what may beseen as new threads of theoretical development - tudies ofinterdependence, and a more behavioural approach. Thirdly,

there is an att em pt t o show how even those developments

34



are incorrect insofar as they fail to cIiiIIenge certain assu mp-tions of the analysis.

Th e Weberian approachThe generally accepted foun ding father of industrial location

theory was Alfred Weber (1909). His analysis was of single

firms. and was con fined t o minimising costs, in the generalcase specifically transp ort costs. W eber himself did not form -

ulate the problem in the framework o f economic theor y, butinstead saw it in terms o f a physical analogy as a problem ofthe resolution of various ‘location pulls’ or ‘forces’. Laterdevelopments, however, through lsard (1956) and Moses(1958) have show n how Weberian conc epts can be incorporat;dwithin the existing marginalist substi tution analysis.

At the most detailed level of critique. this analysis wasjudged far too simple, both in terms of the numb er and com -plexity of factors which might operate as forces on the loca-

tion of a plant, and in the fact th at to o many relationships or

states (production coefficien ts, for instance, and size ofoutput) were taken to be fixed. It was these criticisms whichled to further development within this line of theory. And thepoints are of course perfectly valid within a given cont ext. Thefact that the consequ ent line of development has no t greatly

increased industrial location theory’s ability to understand. ordeal with , the real-world situ ation it faces is not because these

criticisms are incorre ct, but because they are th e wrong ones:the critique was made within too narrow a perspective. Thusa thread of development was produced from Weber’s initialanalysis. Authors such as Hoover (1937) and S mith (1966)introduced new factors, made some implications explicit.and transformed locational ‘forces’ in to cost and revenue sur-faces. On the oth er hand , Predohl started the process ofintroducing the m ethods of su bstitution analysis into the

Weberian approach. This was further elaborated by Isard

(1956) wllo, however, kept ‘transport inputs’ as separate iden-tifiable entities, and thus limited the dimensions of possiblesubstitution. With Moses’ (1958) article, this branch of locat-

ion theor y was fully integrated in to marginalist economics.Such a line of development thus involved a process of in-

creasing sophistication and detail. Moses in so me sensesrepresents the end o f the line, for although he did introduce

new conceptual approaches from Isard, and his attack wastherefore very much a reformulation of th e problem, none-theless his results, showing the total interdependence o f the

system an d the requirement o f locational ad justment for everychange in any other variable are certainly not useful either inoperational application or as a framework for understanding

the forces at work in shaping the space-economy. Attem ptsb e een made, such as with Smith’s (1966) spatial margins,to extract a simplified structur e from this approach. But even

here th e emphasis o n the d ecision of an idealised single firmstultifies th e approach as an analytic tool.

Two points should be made at this juncture. First of all,this metho d of analysis is not withou t use. Some of the ideasderived from the Weberian approach, for ins tance, are genuin-ely important conceptualisations. The second point, however,is indicative of basic problems in the development of loc ation

theory as a whole. A considerable amou nt o f the comp lexitythat w as finally injected into the Weberian approach was dueto the simple introduction of th e spatial dimension. In thiscase the problems arose just because another variable wasadded but i t will be shown later that the attempt to introducethe specifically spatial dimension has rendered some of the

easy conclusions of a-spatial econ omic s muc h more difficultto draw.

35

It was tw o ot her , interrelated, criticisms of Weber whichled to th e development o f new strand s within the overall con-text of location theory. They were indeed much more funda-mental criticisms, but the way in which they were treated

meant that they still did not essentially break out of the

existing ideological framework. In the first place, Weberassumed perfect competition. This. however. is at variance withthe inclusion of the spatial dimension. There cannot be per-

fect Comp etition over space. The causes and effects of this fact

will be examined in greater detail later, but it should be notedthat its implications are important, and that they affect thedegree to which location theory can play the same sustaining

role as economi cs. In spite of these implications of the need fora major conceptual re-think, the thread of development t owhich this realisation gave rise did not pro duce the jolt to thedevelopment of location theory that it might have clone.Neither did it have an y significant feedback effect o n a-spatial

economics. Because of the nionopolistic element introducedinto markets by space the policies of firms are interdependent.

An oligopolistic situation prevails. Locational interdependencebecame the next area for research. Another line of develop-ment, closely related, arose from the realisation of the need torelax the assumption of “economic man”. This is again a

recognition of a genuine inadequacy. However, the alternative

“behavioural” approaches which have most commonly heensought reveal a real misapprehension of the reasons why an

assumption of “economic man” is incorrect.

Locational interdependenceThe interdependence of prod ucers is in part a result of the

monopolistic elemegt in their competition. There are two dis-tinct ways such an element of monopo ly may en ter a spatial

economy. In the first place there is that which results fromthe function o f distance itself, and secondly there is the

normal a-spatial concept of mon opoly as the con centration ofcapital into a few hands. The development of the branch ofindastrial location th eory which concentrates on locationalinterdependence was probably t o some degree a reaction to

bot h, but certainly much of its stimulus came from the theoriesof imperfect competition developing in a-spatial economics.

And this developm ent in economics was in tern a response t o

changing real conditions. As a descriptive device, perfect com-petition was becoming an increasingly unconvincing stru cture.But although much o f the development of theories of location-

a1 interdependence was thus undoubtedly an indirect result of

the actual m onop olisatio n of capital, with a-spatial economics

providing many of the tools of the trade, the resultant theoryhas focussed almo st exclusively on certain aspects of that ele-ment of monopoly which is solely the result of distance. In

fact, very little explicit atten tion h as ever been paid in locat-

ion theory to the actual monopolisation of capital. In largemeasure this is once again a result of t he essentially s tatic andpartial nature of most of that theory. System ic development

is normally excluded from analysis.

of mono poly conferred by distance has, as mentioned ahovebeen of a fairly restricted nature . The main emphasis has beenonlndividual locational-interdependence situations, examinedeither thro ugh market-area analysis (e.g. Fet ter, 1 924 andHyson and Hyson, 1950), or through an approach more alliedto game theory (see, for instance, Hotelling, 1929 and Stevens,

1961).These analyses have produced som e interesting results,but the restriction to this level of detail has prevented the realinferences being drawn. As will be shown below, these affectthe whole tenability of Pareto-optimal equilibrium solutions.

Moreover, even the e xamination of the effect of th e element



Furthermore. even within the analysed situations, the actualresults in terms of l ocatio n decisions are usually inde termin ate,

both in some specific individual situation s and as a generalcharacteristic of this class of problems. This point is well made

by Chisholm (1971), and is again essentially a result of th emonopolistic character of the situation. For while market

mechanisms are the controlling force in perfect competition.in situations of imperfect competition it becomes possible for

the actors to influence the environment. “Strategies” become

impo rtant. Given those, the result of any specified game sit-uation may be predictable, but the problem is now that the

choice of strategies at this micro level is indetermiante. In-creasing com plexity of this approach as an analytic tool on ce

again led to a cut-d-sac.

Beha vioural alterna rivesAF should be evident from the preceding discussion, the“behavioural” approach to industrial locatio n developed asthe culmination of a number of tendencies. On the one hand,i t was a reaction t o changing material con dition s - n otherwords to the development of monopoly capital. On the otherhand it was also the result of consideration of the eff ects of

space. Assumptions of perfect knowledge, as required by theoriginal Weberian models, are obviously unrealistic as soonas concepts of distance are introduced, while, as pointed o ut,in a spatial conte xt firms are necessarily locatio nally inter-

dependent, and consideration of their behavioural strategiestherefore becomes important.

I n fact, the critique ‘from outside the ideology’ should bemade on very different lines, as will become apparent duringthe following discussion. This fundamental misconception of

the nature o f the issues in question mean t tha t on ce again,although a new “approach” in industrial location studies hasemerged, it still contains within it the same fundamentalproblems.

The reaction to these problems, as they were in fact seen,

has taken a variety of form s, which can b e characterised (or,more fairly, caricatured) along an axis of which the two poles

are purely descriptive behaviouralism an d th e behaviouraltheory of the firm. These two approaches are not in p ractice,

of course, often separable, but they d o seem to represent pos-sibly conflicting tendencies. This itself signals an emergingcentral tension in industrial location studies, between the

evident need to deal with considerable variety, and the desireto have one central theoretic structure.

At on e extre me, descriptive behaviouralism is a simplerecognition of variety and will not ofitself produce a new

theoretical formulation. At th e other extrem e, however, th eattempt to develop the theory of the firm is just as m uch a mis-

apprehension of the problem - nsofar, tha t is, as it is an

attempt simply to replace one abstract formalism (“economicman”) with another (“behavoural theory of the firm”).

In other words, we arrive again at the points made in the

first two sections of the paper. On the o ne hand, ‘theory’ do esnot have to be represented by a-historical abstraction. On the

other, spatial behaviour cann ot be isolated from the overalldevelopment fo the econ omic system.

General spatial equilibriumIn the last part of this section it is useful to consider the con-tribution of August Losch (1954). Losch was one of the firstpeople to attempt a spatial counterpart to econo mic generalequilibrium. From an initial even distribution of population hederived from each sector a net of hexagonal m arkets which ful-

filled the zero excess profit co nditio ns. By combining sets ofmarket areas for different goods he produced the well-knownsectoral and semi-hierarchical economic landscapes w hich ful-filled certain welfare criteria such as maximising the number ofpurchases made locally a nd minimising the sum of the minimumdistances between production sites. The economic system con-sidered explicitly w as one of competing profit-niaximising

firms under conditions of free entry, and with f.0.b. pricing.The implication is, therefore, that such a market system can

produce the result described.

a general equilibrium solution , and the co nformity of the latterwith the marginal conditions o f Pareto optima lity has always

been a major weapon in a-spatial economics’ armoury ofapologetics. The purpose is to prove that a completely decent-

ralised (perfectly -com petitiv e) price mechanism -will producean allocation which is optimal, or, as Bramhall(l969) says in

his “An introduction to spatial general equilibrium”, it is an

affirmation of the existence of internal control mechanisms inthe social order.

the derivation of general equilibrium have reached heroic

proportions ( eg . see Graaf, 1957), and the implications o f the

theory of second best and of A rrow’s theorem are far-reaching- ee Hunt (1972). Nonetheless the introd uction of the spatialdimension is critical. Indeed t he assu mption of m uch theoret-

ical economics tha t life takes place on the head of a pin shouldbe regarded as one of its less convincing abstra ctions .

In the first place some of th e assum ptions already dubious

in a-spatial economics become even less tenable wh en space is

introduced. Thus the whole set of assumptions ab out economicrationality and perfect knowledge become impossible to main-tain over distance, even ignoring, as economic theory tradit-

ionally does, an y institutional constraints. Even within onecity, jobs in the suburbs go unfilled while potential employees

remain o ut o f wo rk in th e city centre. Similarly, locationalinertia prevents the adjustm ents, which “ideally” are neces-

sary to changes in other variables (and th e num ber of van,-

ables to which locational adjustments might be req uired, evenwithin the contex t o f a single plant, is indicated in Moses’(1 958 ) analysis).

They are not theoretical arguments, in t hat they d o no tprove an y logical impossibility in th e attainm ent thro ugh themark et o f a spatial equilibrium satisfying certain welfarecriteria - ave of course th at t o overcome them it is neces-sary to assume tha t (continuous) space has no effect. It wouldstill be possible t o argue tha t the m arket would produce th edesired result if only we allowed it to ope rate ‘perfectly’.

There are, however, serious theoretical problems caused bythe consideration of space.

Consider first the specific system designed by Losch,where un der a given set o f assumptions, profit-maximising

firms arrived at an equ ilibrium satisfying some sort of welfarecriteria. (These criteria, as will be men tioned again later, areanyway no t those of Pareto optimality.) Even within this

simple framew ork, however, Mills and Lav (1964) have point;ed o ut a critical inconsistency. Thu s, among Losch‘s assump-

tions are the following: (i) that the firms a re profit-maximisers,and (ii) th at th e locations are so numerous th at the entirespace is filled. Mills and Lav show tha t these two assumptions

are no t necessarily consistent; that in fact the second is a

theorem w hich should be examined in relation t o the first.

They proceed to examine it, and find t hat it will not necess-arily always hold. In very similar vein, though m uch more

The ability of th e decentralised price mechanism to sustain

Even in a-spatial economics the ass ump tions necessary for

Nonetheless, such objections in fact ask only for “realism”.

36



rarely pointed out, is the problem of the arrangement of thefmal multi-com modity econom ic landscape. The idea is thatthe singlecomm odity sets of hexagonal market areas, havingestablished one point at which they all have a plant, thenrotate about this central metropolis until the num ber of

locational coincidences is maximised. Obviously su ch was n otthe econom ic process intended; Losch was explicitly design-ing an optimal system with norm ative implications. But asRtlander has shown . ..agglomeration economies cannot

cause towns to formunless the model postulates somemechanism whereby firms can co-ordinate their choices.(Webber, 1972). In oth er w ords, although Losch's system wasin some sense normative , it did nonethe less assume tha t, giventhe right conditions, the market economy could produce thedesired result, The above arguments show tha t even w ithin thisvery restricted form ulation this is not necessarily the case.Both of the p oints above imply a necessity for econ omicmechanisms of co ntrol or co-ordination not included withinthe model. It is often precisely to DISprove the need for suchmechanisms that equilibrium is cited in a-spatial economics.

In fact, of course, the end-state produced by the Loschianequilibrium itself does not fulfil even those w elfate criteriaa F e table to a-spatial neoclassical econo mics. F.0.b. prices

qual average cost, and delivered price, therefore , exceedsmarginal cost plus transp ortation cost. This is again the resultof the mono poly element inheren t in any discussion of spacein these terms. (F.0.b. pricing is here considered thro ughout,though a Loschian system can be derived w ith uniform pricing-see e.g. McCrone (1969) -w ith, or course, different dis-tributional implications.) Thus Boventer (1962) writes:

f i e main difficulw is location theoly is that in order for ageneral equilibrium model to have an optimal solution whichthe market or theoretical solution process necessarilyapproaches and which firlfils the usual welfare conditions ofproduction, it is necessaty to assume linear homogenouspro-duction pnctions. Indivisibilities and agglomeration economies,which are basic for locational analyses, in particularfo r

urban analysis,can not be incorporated in such a model. I fthey am included, the substitution principle,if it is applied atthe margin only, losses muchof its force and becomes uselessin finding the optimal spatial structure.For this reason, themarginal principles have to be supplemented by the total con-ditions of equilibrium.

And Richardson (1969), w riting on a general theo ry of location,says:

nus,the dilemma is that it is difficult to formulate a deter-minate general theow of location without adopting the peda-gogicdevice of the equilibrium concept,yet if this concept isadopted complications arise from the probabiliw that generalequilibrium s inconsistent with the implicationsof the spaceeconomy.

equaPaverage cost, and delivered price, therefore, exceeds

IV TH E SPATIAL DIMENSION, AND CONTRADIC-TIONS W ITHIN ECONOMIC THEORY

Some elements of the relationsh ip in ideology, and hence inepistemology, of industrial location theory to neoclassicaleconomics emerged in the last section. Industrial locationtheory is thus vulnerable to many of th e same criticisms ascan be made of tha t (Very vulnerable) subject. For locationtheory, however, this s only the beginning of the problems.

A furthe r line of argument began to emerge in section three,about the crucial nature of the introduction of the spatialdimension into neoclassical economics. In the discussion ofLosch, for example, it was shown tha t his system could notguarantee to produce the equilibrium he described withouthigher level control m echanisms: that the aggregation of in-dividual decisions cannot be guaranteed t o produce the re-quired solution . Moreover, men ifsuch a solution were pro-duced, it would not, without very restrictive assumptions,

fulfd Pareto optimality criteria. In both these cases, the intro-duction of the spatial dimension produces internal incon-sistencies within th e usual body of neoclassical econom ictheory. The purpose of t h is brief section is to exp lore this inmore structured form.

In the first place it was seen in th e discussion of Moses(1958) how the introduction of space, simply as an extravariable, enorm ously complica ted the analysis and, mo stparticularly, its applicability. With the introdu ction of aconcrete sp atial dimension, in fac t, the need really does be-come apparent for a rigorous but structured form of analysiswhich does not depend o n the over-precision of m arginalism.

Another way in which the spatial dimension complicatesthe arguments ofeconomics isthrough ts introduction of an

element of monopoly. This has already been me ntioned at anumb er of po ints and in some detail. Most significantly inthe context of this section it was seen how the apparently corn-forting w elfare conclusio ns of a -spatial general equilibrium be-come much less easy to draw. This element of monopoly may,however, be illustrated in a variety of ways. In the p recedingparagraphs the aspect on w hich attention was focussed was theexistence of interdependence between producers. But thereare other ways of demonstrating the point, and which link nwith other theoretical criticisms already made.

In the first place, each point in space has an absolutequality. Land is not in locational terms an undifferentiatedgood, and private possession of a parcel of it endows theowner w ith exclusive rights over a unique entity. The alloc-ation of such a good is therefore critical. This ine of argument

is of cour se closely tied to that of rent theory and, to the ex-tent that industrial location theory has managed frequently toremain distinct from tha t theory, isoutside the scope of thispaper. The point t o be made here is simply that in any integ-rated (i.e. structural and inter-sectoral) analysis of the spatialarrangement of society at inter-regional evel, industriallocation theory cannot be divorced from (reformulated) renttheory.

'Ihe degree of monopoly conferred by space is particularlyimportant t o industrial location theory (as opposed, for in-stance, to residential location) in terms of the control itallows of the area surrounding the po int of location. In thiscase it is the intrasectoral context which is most important;th is is the geography of mark et areas. The level of analysismay be intra-urban, or inter-regional, or international; the

point is the same. One of the conditions of perfect competitionis tha t, in equilibrium, consumers will be indifferent betweenproducers- hey all charge the same price and their productsare indistinguishable. Such con ditions can n o longer holdwhen the economy is distributed over space. For that space,and its concomitant distance and transport costs, amount forthe consumer to a form of product differentiation. Theproducers may, therefore, raise prices without losing all oftheir custom, extract monopoly profits, and indulge themore easily in various forms of price discrimination.

Another instance in which th e m isconceptions of ec onomi&become critical when applied to spatial location theory relates

37



to the static and partial nature of t he conceptual appraoch. Ina.spatial economics th e normal restriction t o a static viewprevents analysis of th e dynamic forces within capitalism.

for the spatial and a-spatial concentration of capital may beseen as mutually reinforcing mechanisms. In th e first place,the a-spatial conce ntratio n of cap ital itself has spatial effects.Thu s Parsons (1972), for instanc e, illustrates th e reinforce-ment of tendencies to spatial agglomeration which are prod uc-

ed b y t he increasing dominance of large corporations. On theother ha nd, space itself may be the active element in th edynamic. In the con text of industrial location theory, o ne ofthe most im por tant of these “spatial forces” is agglomerationeconomies. These are, of course, considered by locationtheory, an d considered t o be important. But they are usually

studied within a conceptual structure which allows their fullsignificance t o be missed. In most cases the structure istha t of partial analysis, with only one, o r at most an oligo-polistically small numbe r, of firms being studied . Such is thecase wifh Weberian analysis, for instance. We see the likelihoodof individual firms to agglomerate, but do n ot stud y theeffects of this in terms of the inp ut-outpu t system of econom icactivity as a whole. b s c h , on the other ha nd, does considerlocation at a more aggregate level, and does include the m oremacro-level effects of (intersectoral only ) agglomerationeconomies. Apart from the oth er problems already mentioned,however, (including tha t of th e means by which numerou sprivate decision-makers can actually take maximum advantageof agglomeration economies) this approach fails, because ofits static equilibrium nature, t o appreciate th e cumulativeeffects of such economies.

eration economies, are not part o f the main thread of highlyprecise mathematical neoclassical economic formulations.This must be a criticism of any the ory o f the economy . Butfor location theory/spatial economics this is more tha n merelysome ‘aspect’ or ‘imperfection’ ignored. Externalities are, orshould be, central to any theory of locational arrangement.Once

again t he transfer from a-spatial to spatial involves areinforcement o f the inadequacies o f neo-classical economics.These misconceptions and contradictions are not totally

ignored in industrial location literature. The co ntribution ofMills and Lav has already been cited, for instance. There are,moreover, a numb er of rigorous demonstrations of the de-ficiencies of th e decentralised price mechanism in achievingand maintaining a non-trivial equilibrium solution t o certainproblems in spatial allocation (e.g. see Koopm ans and Beck-mann, 19 57, for the case where profitabilities are dependentnot only on individual location b ut are interdependent).Again, Beckmann (1972) demonstrates in each typical caseof the single firm monop oly/multiple firm mo nopoly/freeentry con tinuum how t he m arket falls short of welfaremaximisation. Much of this work is, however, given only

peripheral attention. Ko opmans and Beckmann started off adebate (see, for instance,Mills 1970), but few of the centralimplications of this literature for the standard body of locationtheo ry are ever fully evaluated. If quoted at all, they are

usually presented as problems implying a need for m odificationof the basic theory. Infucr in many cases they constituteelements of a much more fundam ental technical critique.

But the restriction is even more crucial in a spatial con text,

Externalities, as a general category which inclu des agglom-

V CONCLUSIONS: T H E PRESENT POSITION

It is appropriate t o stress again the tentative nature o f thiscritique. Many other approaches, consistent with this one,remain to be covered. Moreover there exist certain thread s oflocation theory (the Dutch analyses of spatial dispersion, forexample, and the work on growth pole theory and thresholdanalysis), omitted here because of no t falling within the strictlineage of ‘classical’ locatio n theo ry, which are im mun e to

some of t he criticisms made he re. Nonetheless it is possibleto draw a few conclusions from th e points m ade in this essay.

This critique has attemp ted to dem onstrate three mainthemes. Firstly it shows the confusion implicit in attemptingto conceive of a comp letely autono mou s location theory.Similarly, critiques of the standard body of econom ic theoryhave concentrated on the econom ic but non-spatial aspectsof th e development of capitalism. It is impo rtant to understandalso the spatial dimension o f its development and o f strategy.Secondly, the critique points out tha t tha t body of knowledgewhich is known as industrial location theory suffers from thesame idealistic misconceptions as neoclassical economics. It isthis overall viewpoint which both determines its object and itsform, and enables it to perform a political and ideologicalfunction. This function may be articulated through theelaboration of specific constructions which justify theecon omic system (as can be th e case with general equilibrium),through a concentration on individual industrialist rationalityand a co nsequent ignoring of the effects of this ‘rationality’,or through an ex treme acceptance of “whatever th e locatorwants we must provide” (even when it falls in the most non-econ omic realms of ‘psychic income’), an at titud e exemplifiedby much latter-day market-research-type descriptive be-haviouralism. Thirdly, this critique has aimed t o demo nstratetha t, even within its own terms of reference, location th eorycontains fundamental theoretical inconsistencies in relationto ‘a-spatial’ economics.

Lastly, a major and deepening contradiction emerges fromthe preceding discussion: th ere is a widening disparity between

the current d irections of development o f location theory andthose o f spatial prob lems (or socio-economic problems with aspecific spatial manifestation) in the ‘real world.’

We have seen that some of the more recent lines of develop-

ment in location theo ry appear t o be leading it into a desert ofdetail - f either th e highly mathem atical or th e behavioural

kind. Some of these developments have been discussed.Meanwhile, however, c oncre te problems are growing worse.Within cities the differentials between central core and suburbsincrease; mult ination al firms make nonsense of policiesdesigned to prom ote regional equality. Most of the analyticaltools available in present industrial locatio n theory are able toexplain individual-level decisions as in som e way rational, butare helpless in face o f th e resultant systemic-level irrationality.The fact th at planning c ont rol is relatively soweak in th is sec-

tor is onl y a reflection o f the same reverence fo r the individualcapitalist’s decision as we fin d in th eoretic al industrial locationanalysis. Our o peratio nal regional models recognise the funda-mental importance of the industrial base and its intersectoraleffects. What are emerging as “locational problems”, w hetherintra-urban, interregional o r international, are the spatialmanifestations of the contradictions of capitalism.

At this point, t he argumen t has returned t o the fact thatan auton omo us industrial location theory canno t be constitut-ed. Spatial development can only be seen as part of the over-all development o f capitalism. However it is also true tha tmany o f th e emerging contradictions of th e economic system

38



both take o n a specifically spatial form , and are exacerbate dby the existence of the spatial dimension. To this extent, con-sideration of “th e spatial elemen t” is essential t o all effectiveeconomic analysis. But most of existing industrial locationtheory is placed within an ideology which defines its obje ctand mode o f analysis in a way which makes effective analysisimpossible.

REFERENCES

b c k m a n n M J 1Y72 “Equilibrium versus optimum: spacing of firms,and patterns of market areas” in Reccnt developments in RegionalScience R Funck (ed.) Pion, Lon don.

Boventer E von 1962 “Towards a United Theory of Spatial EconomicStructure” PPRSA 10, 163-80. Reprinted in Dean, Leahy and McKeeSpatial Econom ic Theo ry Macmillan New York 1970.

in Karaska G J and Bramhall D F Locational analysis for manu-facturin g a selection of reedings MIT Prcss.

Chisholm M 1Y71 “In search of a basis for location theory” in Progressin Geography 2 , 11 1-133.

Cyert R M and March J C 1963 A behavioural theory of the firmPrentice H all.

Dobb M 1940 “The trend of modern econ omics” in PoliticalEconomy and C apitilism (Routlcdge and Kegan Paul)

Fetter F A 1924 “The Economic La w of Market Areas” Quarterly

Journal of Economics, vol 39, 520-29.J der Graaff 1957 Theoretical W elfare Economics Cambridge

University Press

Hoover E M 1937 Location Theory and the Shoe and Leatherindustries.

bramhall D F 1969 “An introduction to spatial general equilibrium”

Hotellin H 1929 “Stability in competition” El vol 39 March 1929pp 4f-57 Reprinted in Dean etal; and in Stigler and Boulding AEAReadings in Price Theory.

Hunt E K 1972 “Economic scholasticism and capitalist ideology” inHunt and Schwarz kds .) A critique of economic theory. PenguinModern Economic Readings.

Hyson C D and Hyson W P “The Economic La w of Market Areas”Quarterly Journal of Economics vol64 pp 3 19-24. Reprinted inDean et al Spatial Economic Theory pp 165-169

lsard W 1956 Location and Space Economy, MIT Press

Koopmans T C and Brckmann M J 1957 “Assignment Problems

and the Location of Economic Activities” Econ ometrica 25, 53-76.Losch A 1954 The Economics of Location Y.U.P.

McCrone G 1969 Regional policy in Britain Allen and Unwin.

Mills E S 1970 “The Efficiency of S patial Competition” PPRSA, 25, 7

Mills E S nd Lav M R 1964 “A model of market areas with free71-82.

entry” Journal of Political Economy vol 72 , pp 278-288. (reprintedin Dean, Leahy and McKee).

Moses L N 1958 “Location and the theory o f production”. QuarterlyJournal of Economics, 73 No. 2 (also in Dean, Leahy and McKee).

Parsons G F 1972 “The giant m anufacturing corporations andbalanced regional growth in Britain”. Area vol4, 2 .99-103

Richardson H W 1972 Regional Economics Weidenfeld and Nicolson

Smith D M A 1966 “A theoretical framew ork for geographicalstudies of industrial location” Economic Geography 42.95-1 13

Stevens B H 1961 “An application of Game Theory to a problem inLocation Strategy”. PPRSA 7, 143-158

Webber M J 1972 Impact of Uncertainty on L ocation MIT Press

Weber A 1909 Uber den S tandort de r Industrien; translated asTheory of th e Location of industries (Chicago 1929).

WHY SHOULD WORKERS PRODUCE FOR IDLERS?

39

towards a critique of industrial location theeory

Documents