De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical,...

26
Economics and Philosophy http://journals.cambridge.org/EAP Additional services for Economics and Philosophy: Email alerts: Click here Subscriptions: Click here Commercial reprints: Click here Terms of use : Click here Equilibrium and Disequilibrium in Economic Theory: A Confrontation of the Classical, Marshallian and WalrasHicksian Conceptions Michel De Vroey Economics and Philosophy / Volume 15 / Issue 02 / October 1999, pp 161 185 DOI: 10.1017/S0266267100003965, Published online: 05 December 2008 Link to this article: http://journals.cambridge.org/abstract_S0266267100003965 How to cite this article: Michel De Vroey (1999). Equilibrium and Disequilibrium in Economic Theory: A Confrontation of the Classical, Marshallian and WalrasHicksian Conceptions. Economics and Philosophy, 15, pp 161185 doi:10.1017/S0266267100003965 Request Permissions : Click here Downloaded from http://journals.cambridge.org/EAP, IP address: 144.173.6.37 on 25 Jul 2013

description

De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

Transcript of De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical,...

Page 1: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

Economics and Philosophyhttp://journals.cambridge.org/EAP

Additional services for Economics and Philosophy:

Email alerts: Click hereSubscriptions: Click hereCommercial reprints: Click hereTerms of use : Click here

Equilibrium and Disequilibrium in Economic Theory: A Confrontation of the Classical, Marshallian and Walras­Hicksian Conceptions

Michel De Vroey

Economics and Philosophy / Volume 15 / Issue 02 / October 1999, pp 161 ­ 185DOI: 10.1017/S0266267100003965, Published online: 05 December 2008

Link to this article: http://journals.cambridge.org/abstract_S0266267100003965

How to cite this article:Michel De Vroey (1999). Equilibrium and Disequilibrium in Economic Theory: A Confrontation of the Classical, Marshallian and Walras­Hicksian Conceptions. Economics and Philosophy, 15, pp 161­185 doi:10.1017/S0266267100003965

Request Permissions : Click here

Downloaded from http://journals.cambridge.org/EAP, IP address: 144.173.6.37 on 25 Jul 2013

Page 2: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

Economics and Philosophy, 15 (1999), 161-185 Copyright © Cambridge University Press

EQUILIBRIUM ANDDISEQUILIBRIUM IN ECONOMICTHEORY: A CONFRONTATION OFTHE CLASSICAL, MARSHALLIANAND WALRAS-HICKSIANCONCEPTIONS

MICHEL DE VROEY

UniversiteCatholique de Louvain

1 INTRODUCTION

When the economic theory of the last decades becomes a subject ofreflection for historians of economic theory, a striking feature which theywill have to explain is the demise of the disequilibrium concept.Previously, economists had no qualms concerning the view that diemarket or the economy was exhibiting disequilibria. Amongst manypossible quotations, the following, drawn from Viner's well-knownarticle on Marshall, illustrates that:

The ordinary economic situation is one of disequilibrium moving in thedirection of equilibrium rather than of realized equilibrium. (1953, p. 206)

Today, mainly under Lucas's impulse, such a statement is consideredunacceptable. To all intents and purpose, the term, disequilibrium, hasbeen banished from the vocabulary of economists. The basic aim of this

This research has been supported by a grant 'Actions de Recherches concertees' no 93/98-162 of the Ministry of Scientific Research of the Belgian French Speaking Community. Iwould like to thank Claude Wampach and Franco Donzelli for stimulating discussions onthe subject of this paper. Comments on an earlier draft by C. Benetti, J. Cartelier, P. DeVille, F. Magris and two anonymous referees are gratefully acknowledged.

161

Page 3: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

162 MICHEL DE VROEY

paper is to shed some light on the reasons underlying this evolution. Itsconcern is less the immediate episodes that led to its demise - forexample, the fall of disequilibrium theory a la Barro and Grossman - thansome prior antecedents. In particular, it purports to draw economists'attention to the fact that the Marshallian and the Walrasian theoreticaltraditions rest on different conceptions of equilibrium and disequili-brium. Put differently, my paper pleads for de-homogenizing Marshalland Walras on equilibrium.1 The main rationale for such a move from thetraditional interpretation has to do with the possibility of disequilibriumoutcomes. In the Marshallian approach, it makes sense to state that, at theclosure of a given market-day, the market can be in a state ofdisequilibrium. However, disequilibrium is not made synonymous withnon-market-clearance. On the contrary, market-clearing - that is, thematching of market-day supply and demand - is supposedly alwaysrealized, in equilibrium as well as in disequilibrium. In contrast, in theWalrasian approach disequilibrium states have only a virtual existence:they are eliminated before becoming effective. Moreover, market-clearingand equilibrium are considered to go hand in hand. As market-clearing isalways present, so is equilibrium. The lack of perception of this difference,it will be argued, is at the source of many confusions pervading present-day debates on topics such as, for example, unemployment theory.

Initially, my aim was to compare the Marshallian and Walrasianconceptions of equilibrium.2 However, in the course of writing thisarticle, I came to realize that I had to enlarge its scope in two respects.First, rather than just discussing the Walrasian conception of equilibrium,I felt that my attention should be focused on what I suggest should becalled the Walras-Hicksian conception. The latter is broader than theformer in that it brings to the fore the intertemporal perspective whichmay have been lacking in Walras's Elements of Pure Economics. TheHicksian modifier is then used to refer to Hicks's contribution to thisbroadening in Value and Capital.3 Secondly, it progressively turned outthat a comparison of the Marshallian and the Walrasian conceptionscould not validly be made without considering their predecessors'views. The rationale here is the lineage to be traced from the classics toMarshall as both approaches deserve the 'market-clearing disequilibriumtheory' label. Hence the view that the Marshallian stands in between theclassical and the Walras-Hicksian approaches. It shares the disequili-

1 In a companion paper (De Vroey 1999a), other differences between the Marshallian andthe Walrasian traditions related to the institutional set-up are analysed.

2 An interesting comparative study of Marshall and Walras on equilibrium is Dos SantosFerreira (1989). The emergence and development of the Walrasian research programme isdepicted in Ingrao and Israel (1990).

3 In view of Hicks's later defence of disequilibrium theory, his association with a viewpointexcluding the possibility of disequilibrium may seem odd. On this, see De Vroey (1999c).

Page 4: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

EQUILIBRIUM AND DISEQUILIBRIUM IN ECONOMIC THEORY 163

brium approach with the former and the subjective theory of value withthe latter. So, I ended up comparing three distinct approaches: theclassical, the Marshallian and the Walras-Hicksian.

The discussion is organized as follows. In Parts 2, 3 and 4, thedistinctive features of the classical, Marshallian and Hicks-Walrasianconceptions of equilibrium are brought to the fore and contrasted. InPart 5 the conceptual muddle arising from blurring them is illustratedwith a few examples. Concluding remarks are offered in Part 6.

Some supplementary remarks are in order before concluding thisintroduction. First, my exploration is not undertaken exclusively forhistorical reasons. Its main motivation lies in my strong belief that aretrospective exploration is a prerequisite for the understanding of thepresent meaning of equilibrium and the controversies and ambiguities itmay carry. Nonetheless, my paper considers only the canonical models,leaving aside their modern offspring wherein, for example, marketclearing may cease to be present. Second, each of the three approachesconsidered raises interpretative issues of their own. Hence some work ofreconstruction is needed, which necessarily involves making disputableinterpretative choices. In fact, as will be seen, I am less interested in whatthe classical political economists, Marshall, Walras and Hicks, meantthan in reconstructing what can be considered as the basic classical,Marshallian and Walras-Hicksian lines of thought, even though theymay possibly be different from these authors' original assertions. Third,my paper will adopt a rather relativistic viewpoint. Since my primaryaim is to bring to the fore differences in methodological perspectives, thequestion of their respective pros and cons will be dealt with briefly.Finally, before beginning my comparison of the three approaches, it isworth drawing a preliminary distinction between the issues of thedetermination and the formation of equilibrium. Determination pertains tothe assessment of the conditions of its logical existence as calculated bythe outside theorist, formation relates to the issue of the making ofequilibrium, that is, of the conditions, possibly institutional ones, neededfor bringing it about endogenously. At stake here is an outcome whoselogical possibility has been priorly asserted. Clearly, it does not sufficethat an equilibrium is logically possible to have it existing effectively.4

2 THE CLASSICAL CONCEPTION OF EQUILIBRIUM

The first equilibrium approach which will be considered is that ofclassical political economists such as Smith, Ricardo and Marx. It has

4 The issue of stability stands in between the two terms of my distinction. On the one hand,it addresses a problem of logical possibility, as does the issue of determination. On theother hand, as is the issue of formation, it is concerned with the process of attainingequilibrium.

Page 5: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

164 MICHEL DE VROEY

been revived by modern Ricardian-Marxian economists, such as Gar-egnani (1976, 1987), Dumenil and Levy (1993) and Kurz and Salvadori(1995), who refer to it as the 'surplus approach' or the 'analysis of long-period equilibrium positions'. Donzelli (1989) calls it the 'stationaryequilibrium perspective'. Here I will however use the more neutralappellation of the classical equilibrium conception in order to avoidsome interpretative ambiguities born of the other labels.

In the classical equilibrium conception a distinction is drawnbetween two price concepts - the natural and the market price. Intheoretical terms the natural price is deemed to be more important thanthe market price. Theirs is a relationship of hierarchy: the market price issupposed to be subordinate to the natural price, in that any situation ofdeviation from the latter is supposed to trigger off some feed-back effect.Among the different possible definitions of the classical conception ofequilibrium, the following one drawn from Caminati (1990) is worthquoting:

A classical long-period position of the economy is, broadly speaking, astate of the system where the driving forces of the classical competition areat rest. More precisely, a classical long-period position is defined in thispaper as a given productive technique, a given state of distribution (e.g. agiven real wage rate), and an associated set of relative commodity prices(production prices) such that the rate of profit is uniform across industries.(Caminati, 1990, pp. 12-13)

Crudely stated, the natural prices are the prices allowing for theuniformity of the profit rate in all branches of the economy. Theireffective existence marks the realization of equilibrium. Market pricesand natural prices then coincide. On the contrary, disequilibriumprevails as soon as profitability differs across branches.

As is well known, unanimity among classical economists on valuetheory and hence on the determination of equilibrium did not exist.Whereas Smith defended an 'adding-up' or 'cost-of-production' determi-nation of natural prices, Ricardo and Marx gave it a labor theory of valuefoundation. Modern authors usually assume that natural prices coincidewith Sraffian prices of production. These differences are however lesscentral for my purpose. More important is the common claim that somefundamental price category determined otherwise than by market forces- the natural prices - exists and that this category plays the role of centerof gravitation for a less fundamental price category, market prices.

Two distinct formation processes should be separated as soon as adistinction is drawn between market and natural prices. As far as theformation of natural prices is concerned, the hallmark of the classicalconception is that whenever market prices deviate from natural prices,competitive forces intervene to bring the economy back to a state of

Page 6: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

EQUILIBRIUM AND DISEQUILIBRIUM IN ECONOMIC THEORY 165

equilibrium. Disequilibria arise because of mistaken production deci-sions (a lack of perception by firms of the size of the 'effectual demand'confronting them), though their mistaken character surfaces only expost. Over-investment in some branches and under-investment in othersleads the market price to respectively be below or exceed the naturalprice. Unequal profit rates across branches ensue, which in turn spurson the reallocation of capital from the low to the high profitabilitybranches. As a result, supply decreases (increases) in the former (latter)branches. Inequalities in profitability diminish and market prices comecloser to natural prices; all this occurs over a range of market-periods.Noteworthy, classical authors do not claim that equilibrium will ever beattained but rather that it acts as an attractor. The existence ofdisequilibrium states is in no way considered as an anomaly, on thecontrary: the resilience of the system is rather based on the fact thatthey elicit feed-back effects impeding their taking on a cumulativecharacter.

In the writings of classical economists this convergence process isdescribed in the crudest way, without any systematic examination ofeither its stability or institutional conditions. Nonetheless, it has a strongappeal for it points to a dynamic conception of competition wherein thelatter is depicted as an unceasing process rather than an end-state.5

Let me now turn to the issue of market price. In Chapter VII of BookOne of the Wealth of Nations, Smith defines it as 'the actual price at whichany commodity is commonly sold . . .' (1976, p. 73). This view, consistingof seeing market prices as day-to-day real world data, is defended bymost authors of the surplus approach.6 In this line of thinking, marketprices are seen as theoretically founded as well as accidental at one andthe same time. On the one hand, the amplitude of their variations islimited by the fact that any departure from natural prices elicits feed-back effects, on the other hand, it is also admitted that they are under thespell of casual factors. Hence the view that their actual size is of littletheoretical importance.7

At stake is the issue of whether a specific equilibrium concept exists

5 Cf. Kurz and Salvadori (1995, p. 20) and Caminati (1990, p. 15). Unfortunately, theirmodem followers have discovered the difficulty of recasting the classical insights inmodels that conform to the present-day canons of rigor. Hence the viewpoint taken byseveral of them that gravitation should be seen as an axiom.

6 See, for example, Boggio (1987, p. 392), Milgate (1987, p. 179), Panico and Petri (1987,p. 392).

7 In Kurz's and Salvadori's terms: 'And on the further premise that a general analysis ofmarket prices would be impossible anyway, it appears to be perfectly sensible to set asidealtogether the "temporary effect" produced by "accidental causes" and focus on the"laws which regulate natural prices, natural wages and natural profits, effects totallyindependent of these accidental causes'" (1995, p. 8). See also Eatwell (1987, p. 599).

Page 7: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

166 MICHEL DE VROEY

that is distinct from the natural price equilibrium, but associated withmarket prices instead. In other words, does a competitive process existwhich elicits that market prices are such as to ensure a matchingbetween supply and demand or market-clearing? Modern defenders ofthe classical approach suggest that there is no such mechanism andhence no further equilibrium concept. This view is however unconvin-cing to me. First, stating that the term 'market price', as used in thetheoretical discourse, refers to the real-world price phenomenon is far-fetched.8 Second, these authors are sitting on the fence. They accept thatsome adjustment process is at work as they endorse Smith's statementthat when the quantity supplied and the 'effectual demand' (the demandof those who are willing to pay the natural price of the commodity)diverge at the natural price, the market price changes (1976, pp. 73-4).However, they seem to think that this process stops short of its normalresult, that is, making effectual demand and supply match, for otherwisethey should accept the market-clearing conclusion and the existence of asecond equilibrium concept. Yet, if an adjustment process is present, is itnot better to assume that it works its way through until some equilibriumstate is reached?

According to Smith, the market price will either rise above or fallbelow the natural price whenever the quantity of any good supplieddiffers from effectual demand.

The market price of every commodity is regulated by the proportionbetween the quantity which is actually brought to the market, and thedemand of those who are willing to pay the natural price of the commodity,or the whole value of the rent, labor and profit, which must be paid inorder to bring it thither. (1976, p. 73)

For Smith the market price is thus determined by the ratio between twoquantities rather than by equality between market supply and demand.As is well known, he did not reason in terms of demand functions butmerely stated that agents express their effectual demand. This impliesthat they know the natural price, which is an unacceptable assumptionas soon as one admits that the natural price is logically anterior tomarket prices (the latter being a deviation from the former). A possiblealternative is to assume that agents pre-assign their expenditures acrossmarkets before the opening of the economy: hence, to borrow Benetti'sterminology (1981), the possibility of a 'natural expenditure curve'taking the form of a rectangular hyperbola. One may well not want to

8 This ambiguity is partially admitted in the first part of Roncaglia's following(contradictory) statement: 'Clearly, although the market price is a concept and as suchimplies a certain degree of abstraction, we are not confronted here with a theoreticalvariable' (1990, p. 104).

Page 8: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

EQUILIBRIUM AND DISEQUILIBRIUM IN ECONOMIC THEORY 167

call this a demand function, yet to all intents and purposes, it amounts tothe same; market prices are then those prices which ensure a matchingbetween this fixed pre-assigned expenditure and supply. With such areconstruction it can be stated that market prices are market-clearingprices.9

The above is about the determination of market prices. On the topicof their formation, classical theory is mute. Which assumptions aboutagents or the organization of trade are necessary to ensure marketclearing? It may be surmised that the only way to get such a result is toassume the market organization is centralized: a market secretary, thestory would then run, announces prices at the natural price and changesthem until supply and demand match. Trading at a false price or bilateraltrade is forbidden. Competition among the dealers, referred to by Smith(1976, p. 73), would then be a kind of auction process. Obviously, thisassumption is hardly satisfactory. Yet, it seems that there are noalternatives for ensuring unicity of price and the matching betweenmarket supply and demand. Once this assumption is made, the resultthat market prices are obtained instantaneously or in logical time cannotbe avoided.

If my interpretation is accepted, the classical approach buttresses theco-existence of two equilibrium concepts, natural price equilibrium andmarket price equilibrium. However, they should not be put on the samefooting. Theirs is a relationship of hierarchy, with market priceequilibrium (natural price equilibrium) as the 'lower' ('higher') equili-brium concept. The term, 'full equilibrium', can be used to designate asituation where the two equilibria are jointly present or, in other words,where the market price coincides with the natural price. 'Disequilibrium'is then used as a short-hand for 'natural disequilibrium', that is, anystate where the market deviates from the natural price. It refers to a lackof 'full equilibrium' rather than 'full disequilibrium', the (non-consid-ered) case where neither of the two equilibrium criteria are matched. Asfull equilibrium is considered to be rarely obtained, states of effectivedisequilibrium are seen as normal phenomena.

Before leaving the classical viewpoint, let me say a few words ontheir treatment of time, an issue closely related to their conception ofequilibrium. In the classical approach, an economy is analysed over agiven time span, the length of which is not made precise, except for twofeatures. First, an implicit yet crucial distinction is drawn betweenpermanent and transitory features of reality. Put differently, the economicreality is supposed to be formed by two layers: a deep or fundamentaland a superficial layer, the main item to be put under 'fundamentals'9 The view that the classical conception comprises two equilibrium concepts is defended,

amongst others, by Arena, Froeschle and Torre (1990), Benetti (1981), Hollander (1987, pp.64-9) and Kubin (1990).

Page 9: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

168 MICHEL DE VROEY

being the technological conditions of production.10 By definition, thefundamental layer is seen as immutable during the time span of thetheoretical investigation, whereas the superficial one is viewed as subjectto transitory changes. This assumption can be called the 'constancy ofbasic data' assumption.11 The second feature concerns the length of theperiod under analysis. The convergence process towards natural pricesrequires changes in the quantities produced. Hence it is a time-takingprocess. However, nothing precise is stated about its duration, exceptthat it ought to be more lengthy than the adjustment towards marketprices. In as far as it is accepted that the latter adjustment takes placeinstantaneously, this is a very weak requirement.12

3 THE MARSHALLIAN CONCEPTION OF EQUILIBRIUM

3.1 Marshall's price and equilibrium concepts

A good point of entry for studying Marshall's conception of equilibriumis his famous fishing industry example (1920, p. 369). Herein, Marshallreflects on how firms react to changes in demand conditions.13 If thechange is expected to be of a very short duration, supply will remainunchanged. If it is expected to be of a moderate length (i.e., to last for a'short period', to which Marshall attributes the length of a year or two),only their variable capital will be modified. Finally, if it is expected tolast for a 'long-period', fixed capital will vary as well, which will entailbigger changes in production. The ensuing picture is, as Hicks put it(1946, p. 122), a tripartite equilibrium classification: it is composed of

10 The term fundamental is used by classical economists to draw a contrast between 'deep'and accidental features of reality. In Walrasian theory this term is used differently anddesignates the 'givens' of the economy, without any contrast being drawn between'fundamental' and 'non-fundamental' givens.

11 In Petri's terms: 'And to this end the equilibrium's data must be sufficiently persistent, sothat the equilibrium does not change over time and the deviations of the economy from ithave time to be corrected, or to compensate one another' (1991, p. 273). See also Eatwell(1987, p. 599).

12 In this light, the judiciousness of Garegnani's rechristening of the classical approachunder the label 'long-period equilibrium positions analysis' (1976) is open to question.This terminology suggests that classical analysis bears on long time spans. Yet, as stated,this interpretation is hardly compelling. I, for one, would be more inclined to see thelong-period as formed of a succession of non-overlapping natural equilibrium time-periods, each of which is associated with a specific state of techniques of no definite time-length. The long-period would be the time span with which an issue such as the declinein the rate of profit could be associated. It could then be stated that classical economistshave a deep interest in the long-period yet not that their value theory is a long-periodtheory.

13 As regards very quick changes, Marshall refers to changes in supply conditions, but forlonger ones demand is considered the triggering off factor.

Page 10: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

EQUILIBRIUM AND DISEQUILIBRIUM IN ECONOMIC THEORY 169

three equilibrium concepts - the ultra-short, the short- and the long-period equilibrium - each of which is associated with a specific priceconcept, the market-day price, the short-period normal price and thelong-period normal price respectively.

The standard interpretation of the interrelationship of these conceptsruns as follows: first, the three price concepts are not given the sameimportance, the market-day price being considered the least important.Second, the three equilibrium concepts are seen as all designating amultiple of some basic unit of time, for example, the hour. The market-day can then be defined as comprising a certain number of hours, theshort-period as comprising a certain number of such days, each of thembeing furthermore considered as separated by a certain time span, and,finally, the long-period as comprising a certain number of short-periods.Third, the relationship of the three price concepts is described asconsisting of a twofold gravitation-like process, from the market-periodtowards the short-period equilibrium and from the short- towards thelong-period equilibrium (Frisch, 1950).

To me, this standard view raises several interpretative problems. InDe Vroey (1999b) I have suggested an alternative interpretation whosethrust can be summarized under three headings.

(a) The basic divide is between market-day and normal equilibriumTo me the main divide of the Marshallian conception of equilibrium isbetween market-day equilibrium and normal equilibrium, which oughtto be associated with two price concepts, market-day and normal price.14

They stand in the same relationship of hierarchy as the classicaleconomists' two equilibrium concepts, with normal equilibrium beingconsidered as more fundamental than market-day equilibrium. Thecriterion for the latter is market-clearing. That is, a state where agents'trading plans prove to be compatible on a specific market-day con-sidering and taking into account the specific constraints they face, inparticular the fact that changes in supply cannot be implementedinstantaneously. Put differently, market-clearing refers to states wheremarket-day demand and supply functions match. In the context ofMarshall's analysis, Book V of the Principles, with its focus on firms'decision problems, normal equilibrium is defined as a situation wherefirms have no incentive to change their production decisions, it beingmoreover assumed that the effects of such past decisions have had timeto be fully worked out. Put differently, normal equilibrium is character-ized by the matching of normal supply and demand functions. Thenotions of 'full equilibrium' might be used in the same sense as in the

14 The market-day equilibrium is also called the temporary equilibrium, yet this term willnot be used in order to avoid confusion with Hicksian terminology.

Page 11: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

170 MICHEL DE VROEY

classical approach, that is, to designate any situation where the twocriteria are fulfilled or, in other words, where the market-day pricecoincides with the normal price.15 In turn, the notion of disequilibriumcharacterizes any situation where the higher equilibrium criterion is notfulfilled, that is, where market-values differ from their normal values.Whereas the possibility of disequilibrium against market equilibrium isexcluded, disequilibria against normal equilibrium can exist effectively.Thus, we have the same 'market-clearing disequilibrium' pattern as inthe classical approach.

(b) The short-flong-period distinctionIn my view, the short-/long-period distinction should be made only in asecond stage of reasoning. Properly speaking, it characterizes the shockin normal demand and the type of adjustment it triggers off rather thanthe new equilibrium arising after the adjustment (whatever the shock,the new equilibrium qualifies as a 'normal equilibrium'). In other words,a short-period adjustment will be initiated if the expected duration ofsome change in demand is longer than the time-span required to make achange in variable capital worth its salt, yet not long enough forjustifying a change in total capital. The technical characteristics of theproduction process are therefore involved and there is no reason tobelieve that the underlying thresholds will be the same across branches.

(c) No adjustment front short- to long-period equilibriumIt has been argued above that the natural equilibrium is obtainedthrough successive adjustments of market equilibria in the classicalapproach. The same applies for the relationship between market andnormal equilibria in my interpretation of the Marshallian approach. Inboth cases, the formation of the higher equilibrium arises over timethrough successive changes in the lower concept. It therefore makessense to state that an adjustment process from the lower to the higherequilibrium concept is at work. Contrary to what is often argued, forexample, by Frisch, the same is not true, however, for the relationshipbetween short-period and long-period normal equilibrium. The reasonlies in the fact that short- and long-period changes in demand arealternative occurrences. The firm has to make a decision as to whether anobserved change in demand will last long enough to justify a modifica-tion of its total capital or, if not, as to whether a change limited tovariable capital is justified. As soon, as it is assumed that the firm hasperfect information, as is implicitly the case in Marshall's reasoning, norisk of confusion arises and the choice is clear-cut. Hence, it makes little15 Marshall himself used the notion of full equilibrium in another sense to designate a state

where all firms of a branch are in equilibrium, in contrast to his alternative 'statisticalequilibrium' category. Cf. Newman (1960).

Page 12: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

EQUILIBRIUM AND DISEQUILIBRIUM IN ECONOMIC THEORY 171

sense to state that the short-period normal equilibrium will adjusttowards the long-period normal equilibrium since only one or the othertype of adjustment will be elicited.

3.2 The formation of equilibrium prices

As in the classical perspective, in as far as two equilibrium concepts areseparated, a distinction ought to be drawn between two specificadjustment processes geared respectively to the formation of market andnormal equilibrium. The formation of market equilibrium is addressedby Marshall in his corn-market analysis, in Book V Chapter 2 of thePrinciples. Two different cases are considered.16 First, it is assumed thatagents hold perfect information about the market. Market-clearing thenfollows automatically. Marshall's references to higgling and bargainingto the contrary notwithstanding, the formation of equilibrium shouldthen be considered as occurring in logical time. In a second stage,Marshall makes the constant marginal utility of money assumption,amounting to considering a quasi-linear utility function. As a result, thepossibility of income effects is discarded. In this case, false trading isallowed to occur. Since it is not accompanied by path-dependency, themarket will end up with the same quantity traded as in the perfectinformation case with the last exchange occurring at Marshall's 'trueequilibrium' price. Whereas most interpreters tend to give precedence tothe second stage of Marshall's analysis, I, for one, see it rather as anamendment of the first argument, geared towards making the point thatthe perfect information assumption is less heroic than it may appear atfirst sight, since the same result can be obtained without it.

Some reconstruction is needed as regards the issue of the formationof normal equilibrium since nothing on this topic is to found either inMarshall's Principles or in the writings of authors such as Frisch, Viner orHicks. The point to elucidate is whether the market-clearing disequili-brium result noticed apropos classical theory is likely to occur in aMarshallian context. Let me refer again to the fishing industry example.It is implicit in Marshall's reasoning that changing the quantitiesproduced is a time-taking process. This is the reason why, for example,new boats are not constructed if the change in demand is supposed tolast only two years. This may not allow enough time for them to be built.Moreover, they could not be amortized over such a time-span. Here, it isthe first of these factors - the 'time-to-build' element, to use present-dayterminology - which needs to be considered. Once it is brought into thepicture, disequilibrium states become a likely occurrence. Suppose that ashock arises and the owners of the representative fishing firm correctlydiagnose its short-period nature. Suppose also that the changes decided

16 For a more in-depth analysis and assessment, see De Vroey (1999b).

Page 13: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

172 MICHEL DE VROEY

on in variable capital cannot be implemented in the time-span separatingone market-day from the next. As a result, market outcomes can becharacterized as disequilibrium states for the whole period separatingthe market-day where the shock arose and some subsequent market-daywhere the decision to change variable capital will turn out to haveexerted its full effects. During the different market-days occurring inbetween, market prices and quantities will diverge from their normalequilibrium, all this, however, going along with market clearing. Clearlyenough, the same reasoning is even more valid when shocks are of along-period nature.

The conclusion can then be drawn that, as far as the adjustmenttowards normal equilibrium is concerned, the Marshallian resembles theclassical account on the basic point that its realization occurs throughsuccessive displacements of the (market-clearing) market results, as theyemerge at every period of trading. This is a point which was obvious toearly interpreters, whereas its obviousness may have been lost todaybecause of the prevalence of the Walrasian meaning of equilibrium/disequilibrium.

Finally, the question may be raised of whether disequilibrium statescould still arise without the time-to-build assumption. The answer is yes.An enlightening example is Friedman's discussion of the expectations-augmented Phillips Curve (1968). Here, the linchpin of the disequili-brium result is the asymmetry in expectations across firms and workers.Whereas firms hold rational expectations and hence correctly expect therate of inflation, workers suffer from being endowed with adaptiveexpectations. Expansionary monetary policies then result in over-employment states. Clearly, against my definitions, this is a state ofmarket-clearing disequilibrium. On the one hand, both the quantitytraded and the real wage arising at the end of the market-day followingthe shock differ from their normal magnitudes. On the other, market-daysupply of and demand for labor still match.

3.3 The treatment of time

As with the classical approach, the Marshallian approach is underpinnedby the constancy of data assumption. Yet the distinction betweenpermanent and transitory features is now dropped. Economic data areassumed to stay constant during the time-span of the adjustment processin order to ensure the success of the adjustment towards equilibrium(either of the short- or the long-period type). Yet irreversible moves orchanges in equilibrium positions are accepted. This process-limitedconstancy requirement is the price to be paid for the introduction ofduration without forgoing the view that equilibrium states are effectivelyachieved. In view of the fact that, according to Marshall, adjustment may

Page 14: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

EQUILIBRIUM AND DISEQUILIBRIUM IN ECONOMIC THEORY 173

be the matter of a decade or more (when the long-period is concerned),this requirement still proves to be formidable and runs counter toMarshall's characterization of economic systems as being subject togradual and continuous transformations.

4 THE WALRAS-HICKSIAN CONCEPT OF EQUILIBRIUM

The Walras-Hicksian label refers to Walras's equilibrium model devel-oped in his Elements of Pure Economics (1954), and extended by Lindahl,Hayek and Hicks in order to encompass the intertemporal perspective.17

Its subject matter is a sequence of instantaneous equilibria, each referringto a given point in time. Owing to a lack of space, I will not enter into theissue of whether this perspective is already present in Walras's work, asargued by Donzelli (1989), or whether the Lausanne economist shouldrather be interpreted as having stuck to the classical equilibriumconception, as claimed by his first interpreters and by the surplus schooleconomists nowadays. According to the stance taken in this respect, itcan be asserted either that Lindahl, Hayek and Hicks rediscovered aninsight already present in Walras's and Pareto's work, yet neglected bytheir immediate followers, or that they were responsible for an importantshift in the Walrasian research program, away from what Walras himselfhad spelled out. Be this as it may, the Lindahl-Hayek-Hicks interpreta-tion has become the backbone of the neo-Walrasian research programtoday.

4.1 The two equilibrium concepts

The starting point of Hicks's reasoning is Walras's static equilibriumconcept. The latter is denned in the following way:

A market is in equilibrium, statically considered, if every person is actingin such a way as to reach his most preferred position, subject to theopportunities open to him. This implies that the actions of the differentpersons trading must be consistent. (1946, p. 58)

Hicks's contribution is to have extended this concept to an intertemporaldimension. To this purpose, recourse is made to two distinct equilibrium17 The impulse for the intertemporal extension of Walras's static model extension is mainly

due to Lindahl and Hayek. Lindahl's seminal paper, published in Swedish in 1929, wastranslated into English under the title "The Place of Capital in the Theory of Price' andpublished as Part Three of Lindahl's book, Studies in the Theory of Money and Capital(1939). Hayek's paper is 'Intertemporal Price Equilibrium and Movements in the Value ofMoney', published in German in 1928 and reprinted in Hayek (1984). Hicks, however,popularised it in his Value and Capital and is responsible for what was to become thestandard terminology. On the issue of the origin of the intertemporal approach see Currieand Steedman (1990), Hansson (1982) and Ingrao (1989).

Page 15: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

174 MICHEL DE VROEY

concepts which may be called 'equilibrium at one point in time' (or'temporary equilibrium) and 'equilibrium over time' (or 'intertemporalequilibrium'). 'Equilibrium over time' designates an equilibrium path. InValue and Capital it is defined by the 'condition that the prices realized onthe second Monday are the same as those which were previously expectedto rule at that date' (1946, p. 132, author's emphasis). It exists wheneverthe succession of point-in-time equilibria happen to belong to theequilibrium path. The temporary equilibrium notion is borrowed fromMarshall where it designates the market-day result. To Hicks, itconstitutes a new, expectations-augmented, characterization of Walras'sstatic equilibrium. It has the same conditions of existence, namely thatthe excess demand for every good or service traded is equal to zero. Assoon as the tatonnement hypothesis is made, equilibrium at one point oftime should be considered as arising instantaneously.18 Its introductionserves the purpose of emphasizing that what appears as an equilibriumagainst the backdrop of the conditions for static equilibrium may ceaseto do so when set against an intertemporal perspective. Hence the viewthat it is a provisional reality, exactly like Marshall's market-day result.In the temporary equilibrium context, markets exist only for theexchange of commodities at the date under consideration as well as fortrading of the numeraire commodity from the present to the next futuredate. As far as these commodities are concerned, agents' equilibriumplans are made mutually consistent through tatonnement. However,agents' expectations about future prices are not made compatible. As aresult, agents may regret their choices ex post. Thus, any point-in-timeequilibrium is a temporary equilibrium when set against the intertem-poral perspective, irrespectively of whether it proves to belong to theintertemporal equilibrium trajectory or not. This is how disequilibriummight enter the picture.

In this (analytically important) sense the economic system can be taken tobe always in equilibrium; but there is another wider sense in which it isusually out of equilibrium, to a greater or less extent. . . The wider sense ofEquilibrium - Equilibrium over Time, as we may call it, to distinguish itfrom the Temporary Equilibrium which must rule within any current week- suggests itself when we start to compare the price-situation at any twodates. (Hicks, 1946, pp. 131-2)19

18 Hicks, himself, did not want to adopt the tatonnement hypothesis. However, without it,income effects are bound to arise. In Hicks's story, time is divided into 'Weeks' andtrading is supposed to occur on 'Mondays'. Contrary to what he states explicitly, hisMonday concept is better interpreted as comprising no duration. Whatever happens onMondays occurs in logical time. Cf. De Vroey (1999c).

19 Or, as stated in Capital and Growth (Hicks, 1965, p. 26): 'And it is similarly true that anequilibrium at a point of time which is not an equilibrium over the period in which thatpoint of time occurs, is a disequilibrium position, from the point of view of the period. (It

Page 16: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

EQUILIBRIUM AND DISEQUILIBRIUM IN ECONOMIC THEORY 175

4.2 The treatment of time

Contrary to the other two approaches, the Walras-Hicksian viewpointrequires that all events, be they concerned with trading or delivery, aretime-indexed or dated and that each date is considered as an indivisiblepoint in time. A rather subtle treatment of time, based on a distinctionbetween 'logical' and 'real' time, ensues. In Donzelli's terms:

The set of 'real' time instants is the time set over which the economy undertheoretical investigation is supposed to evolve; the set of 'logical' timeinstants is instead the time set over which the equilibration process isassumed to take place . . . by virtue of that distinction the process ofadjustment towards equilibrium, though being interpreted as an enduringprocess with respect to 'logical time', can nevertheless be viewed as adurationless phenomenon with respect to the 'real' time set through whichthe evolution of the economy is supposed to take place. But, beingdurationless with respect to 'real' time, the adjustment process cannot giverise to any observable disequilibrium phenomenon; as a consequence, ithas the character of a purely virtual process that is structurally unable toaffect the data constellation characterizing the economy at the instant (of'real' time) to which the analysis is meant to refer. By the same token, theequilibrium state associated with that data constellation, though it can beconceived as a rest point of an adjustment process unfolding itself in'logical' time, can at the same time be supposed to be instantaneouslyreached by the economy at the relevant instant of 'real' time. (1989,pp. 27-8)

Another issue is whether the constancy of economic data assump-tion, found in one way or another in the classical and the Marshallianapproaches, is also present in the Walras-Hicksian program. In oneobvious sense the answer is in the negative: when time unfolds, the setof future states of the world become partitioned in the set of effectivelyrealized states on the one hand, and the set of beforehand possible yetunrealized states on the other. This move should be interpreted as theemergence of 'novelty over time' and hence as running on a collisioncourse with the immutability assumption. The point is then to seewhether the latter assumption should be overruled completely orpartially maintained. Usually, neo-Walrasian authors seem to think thatthe fundamentals (tastes, endowments, technology) remain the same.However, there is no intrinsic necessity to stick to this viewpoint. TheWalras-Hicksian approach can perfectly accommodate cases where thefundamentals are changing over time. Adopting such a radical disconti-nuity of economic data perspective actually amounts to a Heraklitean

is better to say that the path, on which the disequilibrium occurs, is not an equilibriumpath, over the period.)'

Page 17: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

176 MICHEL DE VROEY

depiction of the Walras-Hicksian program. In this line of thought, everytatonnement round should be seen as concerned with a radicallydifferent economy having possibly nothing to do with the economieswhich existed at earlier dates. This assumption is obviously too strong,yet it clearly points out that the Walrasian approach does not require theconstancy of data assumption as the classical and Marshallianapproaches do.

4.3 The contrast between the Walras-Hicksian and the classical andMarshallian conceptions of equilibrium

At first sight, the Walras-Hicksian and the classical and Marshallianconceptions of equilibrium seem to be rather similar. First, temporaryequilibrium seems to be on the same footing as the market clearing resultwhich is obtained in the last two approaches. Second, a hierarchy similarto that which is established in the other two approaches is implicitlyestablished between Hicks's two concepts. Market-clearing, as theequilibrium principle obtained by tatonnement, seems to be supersededby a higher equilibrium principle bearing on the fulfillment of priceexpectations.20 Third, the impression prevails that the two equilibriumconcepts are linked by some convergence process in the same way as theattainment of equilibrium in the classical and Marshallian approachesproceeds through successive displacements of market values. Theremust be, it is suggested, some forces at work that result in bringing theeconomy back on its equilibrium path whenever it happens to deviatefrom it.

Were this impression confirmed by a deeper probing, the basic claimof this paper - that no unique equilibrium conception pervades economictheory - would not hold. Some significant differences break the surfaceupon closer scrutiny however, suggesting an important differencebetween the Walras-Hicksian approach on the one hand and the classicaland Marshallian one on the other. They arise when the following twoquestions are raised about the Walras-Hicksian approach: first, is therereally a convergence process towards equilibrium over time at work?Second, is there a hierarchy between the temporary equilibrium and theintertemporal equilibrium concepts?

4.3.1 Convergence towards equilibrium over time?

Market prices are deemed to oscillate around natural prices in theclassical approach. In the Marshallian approach, whenever market

20 As stated by Leijonhufvud when commenting on Value and Capital: 'Market-clearing,however, was equilibrium in a "limited sense"; in the more fundamental sense of"Equilibrium over time", Hicks emphasised, the economic system was "usually out ofequilibrium'" (1984, p. 31).

Page 18: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

EQUILIBRIUM AND DISEQUILIBRIUM IN ECONOMIC THEORY 177

values diverge from their normal magnitude, they will change until theycoincide. Equilibrium plays the role of an 'attractor' in both cases. Theissue to be examined here is whether the same can be stated aboutequilibrium over time vis-a-vis temporary equilibrium.

As far as Hicks is concerned, the quotations given above suggest hisadhesion to the idea of the existence of a convergence process. However,in subsequent writings, in particular in Chapter 2 of Capital and Growth(1965), he seems to have changed his mind. In this respect the followingpassage is worth reflecting on:

. . . the kind of equilibrium concept (or concepts) that we require indynamics. We need (1) equilibrium at a point of time; the system is inequilibrium in this sense, if 'individuals' are reaching a preferred position,with respect to their expectations, as they are at that point. It is only to suchan equilibrium that there can be tendency. We also need (2) equilibrium overa period of time. . . . But for period equilibrium there is the additionalcondition that these expectations must be consistent with one another andwith what actually happens within the period. Period equilibrium isessential, in dynamic theory, as a standard of reference; but is hard to seehow there can, in general, be any 'tendency' to it. (1965, p. 24)

Hicks draws a crucial difference between his two equilibrium conceptsin this passage. On the one hand, he asserts the existence of anadjustment towards point-in-time equilibrium.21 On the other hand,however, he denies it in so far as equilibrium over time is concerned. Thelatter concept, he states, is a standard of reference, a yardstick andnothing more. In other words, intertemporal disequilibrium elicits nomechanism tending towards its disappearance. Although Hicks, unfortu-nately, did not elaborate further on his observation, I would surmise thatit is underpinned by the issue of whether changes in the data areaccepted. In a nutshell, if it is assumed that agents have to devise theiroptimizing behavior plan against a completely new set of data at eachnew point in time, the idea that they are able to correct their pastmistakes over time becomes irrelevant. The more radical the disconti-nuity of data over time, the less the idea of convergence makes sense.

Semantics prove to be troublesome at this juncture. Once Hicks'sobservation is admitted and one furthermore accepts that the equili-brium concept implies some underlying adjustment process, what he

21 His assertion that there is a tendency towards equilibrium at a point in time is somewhatmisleading. When taken literally, it means that if equilibrium prices are not realized,forces will be triggered off tending to realize them. This suggests the possibility of falseprice trading, a possibility which Hicks himself did not objet to, yet which cannot beaccepted in a tatonnement context. Thus, it should be emphasized that the adjustmenttowards equilibrium at a point in time occurs instantaneously and does not allow forfalse trading.

Page 19: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

178 MICHEL DE VROEY

calls 'equilibrium over time' should in no way be subsumed under theequilibrium label, for as he states earlier in the same chapter: 'It isnecessary, if the equilibrium assumption is to be justified, that we shouldbe able to assert the existence of a tendency to equilibrium' (1965, p. 17).However, Hicks did not abide by his own conclusions on this point. Towit, the quotation given above in note 19, according to which anequilibrium at a point in time can be a disequilibrium position from thepoint of view of the period, is to be found just two pages after thepassage where he makes the point about the lack of attraction.

As far as more recent works in the Walras-Hicks tradition areconcerned, no clear-cut position stands out. In Grandmont's interpreta-tion of temporary equilibrium, much emphasis is placed on agents'expectation functions (1977, p. 542; 1987). In his view, the formation ofexpectations in temporary equilibrium models can be formulated as theresult of the application of classical statistical techniques by the agentsrather than of following some simple rules-of-thumb. To Grandmont, theexpectation functions ought to be formulated in a general way, self-fulfilling expectations being considered as a special case. The conver-gence idea is part of the picture when a general formulation is adopted.22

However, as soon as Grandmont's general expectations function isreplaced by the rational expectations assumption, the prevailing ap-proach nowadays, the picture changes again and the idea of a randomwalk substitutes itself for that of convergence. The 'equilibrium overtime' term is then used to designate a sequence of temporary equilibria,without claiming that they are governed by some superior equilibriumprinciple and that a convergence process is at work.

To conclude, the Walras-Hicksian hypothetical convergence cannotbe put on the same footing as the classical convergence process. Whereasthe latter is considered as the embodiment of the competitive process, inthe Walras-Hicksian perspective this process is, on the contrary,embodied in the formation of temporary equilibrium. Nor can therelationship between point-in-time and intertemporal equilibrium beassimilated to the relationship between Marshall's market-day equili-brium and normal equilibrium categories. Notice also that if what is

22 As stated by Grandmont in his New Palgrave Dictionary entry on temporary equilibrium:'[The temporary equilibrium method] permits to incorporate in the analysis the fact thattraders usually learn the dynamic laws of their environment only gradually and thus tostudy in principle how convergence towards self-fulfilling expectations may or may notobtain in the long run' (1987, p. 622). The same viewpoint is found under Radner's pen:'In the evolution of a sequence of momentary equilibria, each agent's expectations will besuccessively revised in the light of new information about the environment and aboutcurrent prices. Therefore, the evolution of the economy will depend upon the rules orprocesses of expectations formation and revision used by the agents. In particular, theremight be interesting conditions under which such a sequence of momentary equilibriawould converge, in some sense, to a (stochastic) steady state' (1991, p. 437).

Page 20: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

EQUILIBRIUM AND DISEQUILIBRIUM IN ECONOMIC THEORY 179

called an intertemporal equilibrium path is used as a reference againstwhich the path of successive temporary equilibria can be gauged ratherthan as an attractor, it makes little sense to consider states where thelatter paths deviate from the former as disequilibria. Thus, the concept ofdisequilibrium which is so central in the two other approaches should beconsidered as irrelevant when it comes to the Walras-Hicksian approach.

4.3.2 A hierarchy between the temporary equilibrium and theintertemporal equilibrium concepts?

As seen above, in both the classical and the Marshallian approaches thetwo equilibrium concepts are viewed as organically connected and asbeing in a relationship of hierarchy. In the Walras-Hicksian approach,two equilibrium concepts may well also be present but they featureneither interconnectedness nor hierarchy. First of all, assessing theconditions for market-clearing is deemed to be a central theoreticalobjective in this approach, which is hardly the case in the other two. Inspite of Hicks's insistence on the intertemporal dimension, the staticversion of equilibrium was considered as able to capture sufficientinsights on its own to deserve most of the attention. After the advent ofequilibrium business cycle theory this is of course no longer true today.However, even if dynamic rather than static analysis is now beingaccorded more importance, it remains true that the two matters remainunconnected. Equilibrium at a point in time and equilibrium over timeare not part and parcel of the Marshallian approach as are the market-day result and normal equilibrium. Likewise, Leijonhjufvud's assertionto the contrary notwithstanding (see note 20), no foundation exists forclaiming that equilibrium over time is more fundamental than equili-brium at a point in time. The relationship of hierarchy characterizing theclassical and Marshallian approaches has no raison d'etre in the Walras-Hicksian approach.

5 SEMANTIC PITFALLS

It follows from my above analysis that the conceptions of equilibriumfound in the classical, the Marshallian and the Walrasian approachesshould not be confounded. Unfortunately, economists are hardly awareof such a need. Some examples of the confusion liable to arise when theyare blurred are given in this last section.

First, let me briefly reflect on the fate encountered by the short-/long-period divide. The latter, as may be seen, is a typical Marshalliandistinction. In Marshall's writings it refers to two particular cases of theadjustment towards normal equilibrium. Today, the divide is still usedin the Marshallian literature yet in a quite different sense. What is nowcalled short-period equilibrium is in fact Marshall's market-day

Page 21: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

180 MICHEL DE VROEY

equilibrium - to all intents and purposes, it ought to be understood asreferring to an instantaneous adjustment - whereas the long-periodequilibrium term ought to be interpreted as designating what wascalled normal equilibrium above.23 This is how, for example, Friedmanunderstands the short-/long-period divide. Here, the problem issemantic rather than substantial. More unfortunate, however, is theexportation of this divide to the Walras-Hicksian approaches, wherein,strictly speaking, it does not belong. There is no reason to redefine'point-in-time equilibrium' as 'short-period equilibrium' and 'equili-brium over time' as 'long-period equilibrium'. When this is done, it ishard to resist instilling into the Walras-Hicksian approaches connota-tions which are alien to them, such as the idea of a relationship ofsubordination of short- to long-period equilibrium. The muddle causedby such an import is further revealed by the fact that the Walrasianstatic model is seen as a case of short-period analysis by some authors,whereas it is viewed as a long-period analysis by others. The latterstance is taken up by the authors of the surplus approach, who therebyassign the same role of center of gravitation to the Walrasian staticequilibrium as to their own natural equilibrium concept. More oddlyhowever, this viewpoint is also taken up by some neoclassicalauthors.24

Another example of the semantic muddle resulting from con-founding the Marshallian and the Walrasian conceptions of equilibriumconcerns the introduction of the involuntary unemployment concept inneoclassical theory.25 Without entering into a substantive discussion, letme just note that the characterization of the involuntary unemploymentprogram differs according to whether it is embedded within theMarshallian or the Walrasian approach. In the Marshallian perspective,involuntary unemployment should be characterized as an equilibriumresult, whereas in the Walrasian perspective it can only be characterizedas a disequilibrium state. Thus, in reference to the Marshallian approach, atheory of involuntary unemployment aims at turning the Marshallianstandard market-clearing disequilibrium result upside down and repla-cing it with models of non-market clearing equilibrium. The 'Keynesian-Walrasian' involuntary unemployment program is quite distinct as itconsists of substituting a point-in-time market rationing result for a23 Thereby, the distinction between short-period and long-period as based on the difference

between changes in variable and total capital, vanishes from the scene.24 For example, in his Theory of Unemployment Reconsidered (1977), Malinvaud presents the

phenomenon of price rigidity as having a short run existence, whereas he assignsWalrasian equilibrium to the long run: 'The Walrasian equilibrium is appropriate forlong-run economic analysis, because in the long run prices are actually flexible and playthe role that was traditionally given to them' (1977, p . 34). A similar viewpoint is to befound in Bliss's entry on Hicks in the Neiv Palgrave Dictionary of Economics (1987, p. 643).

25 Cf. De Vroey (1998).

Page 22: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

EQUILIBRIUM AND DISEQUILIBRIUM IN ECONOMIC THEORY 181

point-in-time market-clearing result, thus replacing the temporaryequilibrium outcome by its converse - a 'temporary disequilibrium'.

In the same vein, the meaning of the notions of full employment andnatural rate of unemployment (or, to consider the opposite of thestandard appellation, the natural rate of employment) differs accordingto whether they are set against a Walrasian or a Marshallian equilibriumconception (here an exclusive concern for static Walrasian equilibriumsuffices). The concept of full employment makes sense in the former, butnot in its ordinary meaning. Full employment coincides with being onthe supply of labor curve.26 No increase in employment can arisewithout increases in the real wage, due to the voluntary nature ofexchange. In this context, the notion of natural rate of employment isirrelevant. The opposite is true however in a Marshallian perspective.Here it is the notion of full employment that makes little sense, whereasthat of a natural rate of employment, which should then be considered torefer to the quantity dimension of normal equilibrium, does. Unlikewhat happens in the static Walrasian context, this quantity is noinsuperable threshold: the market-day equilibrium level of employmentcan very well be greater or smaller than the natural level, in which casethere is either over- or under-employment. Labor-market disequilibrium,so defined, is thus a plausible effective result, going along, I repeat, withmarket-clearing. Neither the notion of over- or of under-employmentmakes sense in a static Walrasian framework.

All this, I hope, suffices to illustrate the need of having a firm graspof the differences between Marshall and Walras; not just for the sake ofgetting one's history of economics straight, but also in order to be able toavoid semantic pitfalls when discussing modern economic theory.

6 CONCLUDING REMARKS

This article has pursued a critical aim that consists in comparing theequilibrium conceptions found in three main economic paradigms: theclassical, the Marshallian and the Hicksian-Walrasian approaches.27

Stating that they have nothing in common would be a crude exaggera-tion. They are similar in at least two central points: the first one concernsthe role given to the equilibrium concept, the backbone of the wholetheoretical reasoning in all three approaches. The second featurecommon to all three paradigms is market-clearing. It thus turns out that,contrary to a wide-spread opinion, market-clearing is not a moderninvention. It has pervaded economic thinking since its inception, theKeynesian episode being the exception which proves the rule.

26 Cf. Pat inkin (1965, p . 315). Such a micro-founded definition of full employment is polesapar t from the common-sense mean ing of this te rm. Cf. De Vroey (1998).

27 Table 1 overleaf summar izes its main results.

Page 23: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

Tab

le 1

. The

dis

tinc

tive

fea

ture

s of

the

thr

ee e

quil

ibri

um a

ppro

ache

s00

The

cla

ssic

al a

ppro

ach

The

Mar

shal

lian

app

roac

hT

he W

alra

s-H

icks

ian

appr

oach

1. T

he e

quil

ibri

um c

once

pts

natu

ral

and

mar

ket

equi

libr

ium

no

rmal

and

mar

ket

equi

libr

ium

2. T

he r

elat

ions

hip

betw

een

the

two

conc

epts

3. T

he a

djus

tmen

t pro

cess

4. E

ffec

tive

ness

of m

arke

t-cl

eari

ng

5. D

iseq

uili

briu

m5.

1. m

eani

ng

5.2.

pos

sibi

lity

of e

xist

ence

rela

tion

ship

of

hier

arch

y

two

dist

inct

adj

ustm

ent

proc

esse

s:-

logi

cal

tim

e ad

just

men

tto

war

ds m

arke

teq

uili

briu

m-

diac

hron

ic a

djus

tmen

tto

war

ds n

atur

al p

rice

equi

libr

ium

yes

non-

coin

cide

nce

of m

arke

tan

d na

tura

l pr

ices

yes

rela

tion

ship

of

hier

arch

y

two

dist

inct

adj

ustm

ent

proc

esse

s:-

logi

cal t

ime

adju

stm

ent

tow

ards

mar

ket

equi

libr

ium

- di

achr

onic

adj

ustm

ent

tow

ards

nat

ural

pri

ceeq

uili

briu

m

yes

non-

coin

cide

nce

of m

arke

tan

d no

rmal

pri

ces

yes

equi

libr

ium

at o

ne p

oint

in

tim

ean

d eq

uili

briu

m o

ver

tim

e

no o

rgan

ic l

ink

betw

een

the

two

conc

epts

one

adju

stm

ent

proc

ess:

- lo

gica

l tim

e ad

just

men

tto

war

ds e

quil

ibri

um a

t on

epo

int

in t

ime

- no

adj

ustm

ent

tow

ards

equi

libr

ium

ove

r ti

me

yes

non-

mar

ket

clea

ranc

e

non m r* a m s

Page 24: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

EQUILIBRIUM AND DISEQUILIBRIUM IN ECONOMIC THEORY 183

Having acknowledged these two common features, the questionmay then be raised of whether the spontaneous ranking of the threeapproaches - that is, putting the classical approach on one side and theMarshallian and Walras-Hicksian ones together on the other, is con-firmed by my analysis. No clear-cut answer can be given to this question,the reason being that the Marshallian approach sits on the fence betweenthe other two. It ought not to be confused with either of them. True, theMarshallian and the classical traditions differ in several importantrespects: their value theory, the place given to micro-foundations, theiradoption of either a partial or general equilibrium perspective. On theother hand, however, they all give the disequilibrium concept a centralplace in their analysis. It is also true that the Marshallian and the Walras-Hicksian approaches share important similarities. In particular, they arealike as far as the micro-foundation aspect is concerned. In view of thefact that this topic forms the bulk of microeconomic theory, smallwonder that most economists feel no need to draw a distinction betweenthese two approaches. Yet, they stand in sharp contrast as regards thepossibility of disequilibrium results. As seen, in the Walras-Hicksian,contrary to the Marshallian conception, no discrepancy between marketclearing and equilibrium exists and states of disequilibrium have noeffective existence (as, rightly enough, deviations from steady-stategrowth are scarcely labeled as disequilibria).

So, it turns out that the three approaches are not poles apart. It israther that their differences are subtle, similarities being interwoven withdissimilarities. This is precisely where the difficulty lies. The fact that, forexample, they are similarly underpinned by two distinct equilibriumconcepts may prompt the conclusion that they could be subsumed underthe same general divide such as the short-/long-term distinction. Thecontribution of my paper is to show that such an unifying interpretationelicits more heat than light.

Finally, to turn back to the observation with which I began thispaper, namely the Lucasian view that no room exists for disequilibriumin modern economic theory, the preliminary lesson which can be drawnfrom my reflection is as follows: this view can be true only to the extentthat the Walras-Hicksian approach has effectively become the exclusiveway of practising economic theory.

REFERENCES

Arena, Richard., Claude Froeschle and Dominique Torre. 1990. 'Gravitation Theory: TwoIllustrative Models'. In Caminati and Petri (eds.). pp. 287-308

Benetti, Carlo. 1981. 'La question de la gravitation des prix de marche dans la Richesse desNations'. Cahiers d'Economie Politique, 6:9-31

Bliss, Christopher. 1987. 'Hicks, John Richard'. The New Palgrave. A Dictionary of Economics,Vol. 2, pp. 641-46. Macmillan

Page 25: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

184 MICHEL DE VROEY

Boggio, Luigi. 1987. 'Center of Gravitation'. Political Economy, Studies in the Surplus Approach,1:392-4

Caminati, Mauricio. 1990. 'Gravitation: an Introduction'. In Convergence to Long-PeriodPositions. Mauricio Caminati and Fabio Petri (eds.). Special Issue of Political Economy,Studies in the Surplus Approach, Vol. 6, pp. 11-44

Currie, Martin and Ian Steedman. 1990. Wrestling with Time. Problems in Economic Theory.Manchester University Press

De Vroey, Michel. 1998. 'Accounting for Involuntary Unemployment in NeoclassicalTheory. Some Lessons from Sixty Years of Uphill Struggle'. In Economics andMethodology: Crossing Boundaries, pp. 177-224. Roger Backhouse, Daniel Hausman,Uskali Maki and Andrea Salanti (eds.). Macmillan

De Vroey, Michel. 199a. 'The Marshallian market and the Walrasian economy. Twoincompatible bedfellows'. The Scottish Journal of Political Economy, forthcoming

De Vroey, Michel. 1999b. 'Marshall on Equilibrium and Time. A Reconstruction'. TheEuropean Journal of the History of Economic Thought, forthcoming

De Vroey, Michel. 1999c. 'Hicks on Equilibrium and Disequilibrium'. History of EconomicsReview, 29:31-44

Donzelli Franco. 1989. The Concept of Equilibrium in Neoclassical Economic Theory. An Inquiryinto the Evolution of General Competitive Analysis from Walras to the 'Neo-Walrasian'Research Programme. Ph.D. Dissertation. University of Cambridge

Dos Santos Ferreira, Rodolphe. 1989. 'Equilibre marshalien et equilibre walrasien'.Recherches economiques de Louvain, 55:399-424

Dumenil, Gerard and Dominique Levy. 1993. The Economics of the Profit Rate. Competition,Crises and Historical Tendencies in Capitalism. Edward Elgar

Eatwell, John. 1987. 'Natural and Normal Conditions', The New Palgrave. A Dictionary ofEconomics, Vol. 3. Macmillan

Friedman, Milton. 1968. 'The role of monetary policy'. American Economic Review,58:1-17

Frisch, Ragnar. 1950. 'Alfred Marshall's Theory of Value'. The Quarterly Journal of Economics,64:495-524

Garegnani, Pieranjelo. 1976. 'On a Change in the Notion of Equilibrium in Recent Work onValue and Distribution'. In Essays in Modern Capital Theory, pp. 25-45. M. Brown,K. Sato and P. Zarembka (eds.). North-Holland

Garegnani, Pieranjelo. 1987. "The surplus approach to value and distribution'. The NewPalgrave Dictionary of Economics. Vol. 4, pp. 560-74

Grandmont, Jean Michel. 1977. 'Temporary general equilibrium theory'. Econometrica,45:535-72

Grandmont, Jean Michel. 1987. 'Temporary Equilibrium'. The New Palgrave. A Dictionary ofEconomics, Vol. 3, pp. 620-23. Macmillan

Hansson, B. A. 1982. The Stockhom School and the Development of Dynamic Method. Croom HelmHayek, Friedrich A. 1984. Money, Capital and Flucutations. Early Essays. RoutledgeHicks, John R. 1946. Value and Capital. 2nd. edn. Clarendon PressHicks, John R. 1965. Capital and Growth. Clarendon PressHollander, Samuel. 1987. Classical Economics. Blackwell.Ingrao, Bruna. 1989. 'From Walras' General Equilibrium to Hicks's Temporary Equilibrium'.

Recherches economiques de Louvain, 55:365-97Ingrao, Bruna and Giorgo Israel. 1990. The Invisible Hand. Economic Equilibrium in the History

of Science. The MIT PressKubin, I. 1990. 'Market prices and natural prices: a model with a value effectual demand'.

Political Economy, Studies in the Surplus Approach, 6:175-92Kurz, Heinz and Neiri Salvadori. 1995. Theory of Production. A Long-Period Analysis.

Cambridge University Press

Page 26: De Vroey Equilibrium and Disequilibrium in Economic Theory a Confrontation of the Classical, Marshallian and Walras-Hicksian Conceptions

http://journals.cambridge.org Downloaded: 25 Jul 2013 IP address: 144.173.6.37

EQUILIBRIUM AND DISEQUILIBRIUM IN ECONOMIC THEORY 185

Leijonhufvud, Axel. 1984. 'Hicks on time and money'. Oxford Economic Papers, November,Supplement, pp. 26-46

Lindahl, Erik. 1939. Studies in the Theory of Money and Capital. Reprint. Augustus M. Kelley,1970

Malinvaud, Edmond. 1977. The Theory of Unemployment Reconsidered. Basil BlackwellMarshall, Alfred. 1920. Principles of Economics. 8th edn. MacmillanMilgate, Murray. 1987. 'Equilibrium: Development of the Concept'. The New Palgrave. A

Dictionary of Economics, Vol. 2, pp. 179-82Newman, Peter. 1960. 'The erosion of Marshall's Theory of Value'. Quarterly Journal of

Economics, 74:587-601Panico, Carlo and Fabio Petri. 1987. 'Long-run and Short-run'. The New Palgrave. A

Dictionary of Economics, Vol. 3, pp. 238-40. MacmillanPatinkin, Don. 1965. Money, Interest and Prices. 2nd. edn. Harper and RowPetri, Fabio. 1991. 'Hicks's recantation of the Temporary Equilibrium Method'. Review of

Political Economy, 3:268-88Radner, Roy. 1991. 'Intertemporal General Equilibrium'. In Lionel W. McKenzie and Stephano

Zamagni (eds.). Value and Capital Fitfy Years Later, pp. 423-60. MacmillanRoncaglia Alessandro. 1990. 'Is the Notion of Long-Period Position Compatible with

Classical Political Economy?' In Caminati and Petri (eds.), pp. 103-11Smith, Adam. 1976. An Inquiry into the Nature and Causes of the Wealth of Nations. Oxford

University PressViner, Jacob. [1931] 1953. 'Cost Curves and Supply Curves'. In Readings in Price Theory, pp.

198-232. George Stigler and Kenneth Boulding (eds.). Allen and UnwinWalras Leon. 1954. Elements of Pure Economics. Trans. W. Jaffe. Allen and Unwin