sign in sign up
<<
Home , Consumer choice, Marginal utility

A Post Keynesian Approach to Choice Author(s): Marc Lavoie Reviewed work(s): Source: Journal of Post Keynesian , Vol. 16, No. 4 (Summer, 1994), pp. 539-562 Published by: M.E. Sharpe, Inc. Stable URL: http://www.jstor.org/stable/4538409 . Accessed: 21/06/2012 18:26

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for- that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

M.E. Sharpe, Inc. is collaborating with JSTOR to digitize, preserve and extend access to Journal of Post .

http://www.jstor.org MARC LAVOIE

A Post Keynesian approach to

Therehave been few effortsmade by Post Keynesiansto explainhow consumersmake choices. Granted,there are severalstudies on the choices entrepreneurshave to make when theirfirms face uncertain prospects,A la Shackle.Also, neo-Ricardiansare known for their analysisof the choiceof techniquein .But, except for a few authors,such as Earl (1983, 1986),Baxter (1988), and Drakopoulos (1990,1992a, 1992b), Post Keynesians have been relatively silent about themicroeconomics of householdchoice.' As pointedout byEarl (1983, p. 2), thereis no chapteron consumerbehavior in theessays of thePost Keynesianstudy guide edited by Eichner(1979). Does thatmean that PostKeynesians, or all nonorthodoxeconomists for that matter, accept the neoclassicalway of detenniningthe compositionof output?Would Post Keynesiansendorse the axiomaticneoclassical presentationof consumer'schoice? The answerto both questionsis no. Althoughthere have been few contributions,even few comments,about consumer behavior by Post Keynesianauthors, there is a certaindegree of coherenceamong the few contributions.There are several pieces of workthat fit, like a puzzle,the major conceptionsadvanced in other fields of economicsby Post Keynesianauthors. The objectof this articleis to outlinethe views aroundwhich Post Keynesians agree, or around which they could agree. Threewanings are in order.First, my Post Keynesianapproach to consumerchoice does not arisesolely from contributions of well-iden- tifiedPost Keynesian authors. It hasmuch in commonwith the work of someInstitutionalists, Schumpeterian evolutionists, social , behavioralists,marketing specialists, famous but hard-to-labelnonor- thodoxeconomists, and even some dissidentmainstreamers. Second,

The authoris Professorof Economics at the Universityof Ottawa.He is gratefulfor the comments received from the referee. 1 There is also a chapteron consumerbehavior in Eichner(1987, ch. 9) and Lavoie (1992b, ch. 2), the main themes of which are taken afreshhere. Journal of Post KeynesianEconomics / Summer 1994, Vol. 16, No. 4 539 540 JOURNALOF POST KEYNESIANECONOMICS the principlesput forthto reflect a possible Post Keynesian approachto consumerchoice reflect my own views of tfie matter.Otlers interested in the field might find a differentset of principles.Finally, no axiomatic or formal model of choice behavior will be offered to the reader.The concepts ratherthan the mathematicswill be discussed.

1. Common pursuits

The neoclassical theory of consumer choice has been the target of criticism from its very inception;an example is Veblen's (1899) scath- ing assessment of the hedonistic and utilitarianneoclassical portraitof homo oeconomicus. Althoughthe focus of the crtique has now changed, some of the criticismsthat were leveled againstmarginal theory, or revealed preferencestheory even before its inception, are still rele- vant. Here are some extracts of such a critique, made by a Veblenian in 1910: Marginal-utilitytheory has usually been formulated in hedonisticterms ... .Hedonismis hopelesslydiscredited by modernpsychology .... Deliberation,reasoned choice plays but a minorpart in theaffairs of men. Habit,not calculation,governs the greaterpart of all ouracts ... .The habitsof thoughtwhich count for most in shapingchoice are not theresult of prevision,but are of the natureof conventionsuncritically accepted by virtueof membershipin a particulargroup... .Calculationis difficult work.It is mucheasier to acton a suggestionthan to weightalternatives

Adherentsof themarginal-utility school . . . denythat the marginal-util- ity doctrinesstand or fallwith hedonism ... .[They] assertthat econom- ics is concemedonly withthe fact of choicebetween or between alternativeactivities, and not withthe basis of choice.... If the marginalutility theory be interpretedhedonistically it is psycho- logicallyinvalid. If the theorybe deprivedof its hedonisticcontent it is reducedto theunobjectionable statements: that men will notbuy a thing unless they want it; thata commoditycannot be sold for more than somebodyis willingto payfor it. [Downey,1987, pp. 48, 49, 51, 53] Downey argues that marginalutility theory is based upon a deficient view of human conduct. He rejects neoclassical rationalitybecause it ignoreshabits and the difficulties of comparingchoices. He emphasizes the role of social conventions. He recognizes that mainstreamchoice 2 Ths anti-utilitarismsort of critiqueis still alive. See, for instance, the work of Caille (1988) and the paperspublished in the Revue du MAUSS. POST KEYNESIANAPPROACH TO CONSUMERCHOICE 541 theorydoes not requirethe conceptof utilityor hedonism-that is, some unique measure of pain and satisfaction; but then he suggests that, deprivedof it, the neoclassical theory of choice, or whathas come to be known as revealedpreference theory, is almost a tautology. The Post Keynesiantheory of householdchoice builds upon several of Downey's intuitions: habits and social conventions, a more proper psychological foundation,and a view of rationalitythat is better con- formed to our limited computationalabilities. The readermight think that he or she is being misled: that there is no more a Post Keynesian theoryof consumerchoice thanthere is a neo-Walrasiantheory of class .Perhaps one can talk of a behavioralistor a human- istic consumer theory, but not of a Post Keynesian consumer theory since Post Keynesians have made no concertedeffort to define such a theory.To this unbelieving reader,who has not read a recentattempt to demonstratethat behind Keynes's economic thoughtlies a nonorthodox consumertheory (Drakopoulos,1992b), the following quotes from five well-known Post Keynesian and neo-Ricardianauthors are offered:

Thereis a kindof competitionin consumption,induced by thedesire to impressthe Joneses, which makes each family strive to keepup at least an appearanceof beingas well off as thosethat they mix with,so that outlayby one inducesoutlay by others... .Generally speaking,wants standin a hierarchy(though with considerableoverlaps at each level) andan incrementin a family'sreal income is not devotedto buyinga little moreof everythingat the same level but to steppingdown the hierarchy.[Robinson, 1956, pp. 251, 354]. Althoughpossibilities of substitutionamong commodities are of course relevantat any given level of real income,there exists a hierarchyof needs.More precisely, there exists a verydefimite order of priorityin 'wants, and therefore among groups of goodsand services, whichmanifests itself as realincomes increase. [Pasinetti, 1981, p. 73]. Thereare hierarchies of needsfrom basic needs up to higherneeds such as the needfor self-fulfilment.The needsare taken as givenin a given environment.There are segments in thepopulation which correspond to income classes.... To differentsegments there correspond different patternsof consumptionto satisfythe hierarchiesof needs. [Schefold, 1985,p. 116] The consumptionof each class will be guidedby a conceptionof its appropriatelifestyle, given its place in thesocial pyramid.... Emulation effectsnormally follow the socialhierarchy; the consumptionstyles of 542 JOURNALOF POST KEYNESLANECONOMICS

therich and famous set standardsto whichthe rest aspire (or, sometimes, againstwhich they react). When can be reduced or raisedsufficiently, elements of moreprosperous lifestyles will be incor- poratedinto those of lowerlevels. [Nell, 1992,pp. 393, 3961. Post-Keynesiansgenerally assume that, in aneconomy that is expanding overtime, it is theincome effect that will predominateover the relative ,or substitution,effects.... Substitutioncan take place only within fairlynarrow subcategories. Consumer preferences are, in this sense, lexicographicailyordered.... A 'sconsumption pattem, at anygiven point in time,thus reflects the lifestyleof thehouseholds that constituteits socialreference group. [Eichner, 1986, pp. 159-160]

In these quotes the notion of hierarchy appears: each household defines a patternof needs that have to be fulfilled in a priorityorder. This patternmay be influenced by the consumption structureof the that are of equal or higher standingin the socioeconomic ladder.There is a class structureor at least a socioeconomic structureto the composition of consumption.The consumptionof an individualis not independentof his or her environment.Finally, some emphasizethe importanceof income effects over substitutioneffects. The latterwould only be relevantwithin goods fulfilling the same need. Nonorthodox authors who have grappled mostly with production theory issues seem to agree about the constituents of an alternative approachtoconsumertheory. The elementsthat they mentionare among the six principlesthat define what a Post Keynesiantheory of consumer choice could look like. These views are in agreementwith each other, and areconsistent wi'th the views thathave been espousedmore recently by Post Keynesian specialists in consumer choice theory such as Earl and Drakopoulos. Some coherent theory of consumer choice is emerging from Post Keynesian circles, although the latter have paid little attentionto the question.This Post Keynesianview of household's choice is compatible with the rest of the Post Keynesian researchprogram. It can be related to the views of the classical authorsand to those of Keynes, thus reassertingthe linkages between neo-Ricardianand the Post Keynesian authors.3

3In addition,some of the ideas involved in a nonorthodoxconsuner theorymay have feedback effects on otherparts of economics, such as labor theory or the , as can be seen from the works of Baxter (1988) andThurow (1991). POST KEYNESIANAPPROACH TO CONSUMERCHOICE 543

2. Commonground Thecommon ground of PostKeynesian consumer choice theory can be summarizedunder the formof the followingsix principles: (i) Theprinciple of proceduralrationality; (ii) Theprinciple of satiableneeds; (iii) Theprinciple of separabilityof needs; (iv) Theprinciple of subordinationof needs; (v) Theprinciple of the growthof needs; (vi) Theprinciple of nonindependence.

Eachof theseprinciples will be discussedin turn.Two points should be made.First, the conjunctionof the principleof separabilityand that of subordinationof needsleads to whatGeorgescu-Roegen (1954, p. 515) hascalled the principle of theirreducibility of wants.Second, many Post Keynesianauthors who have discussed the issue of consumerbehavior have linked some of the above principleswith the notion of character- istics, as they can be found in the works of Lancaster(1972) and Ironmonger(1972).4 In the analysesof Anrous(1978, pp. 277-9), Pasinetti(1981, p. 75), Leaet al. (1987,pp. 496-501), andNell (1992, p. 394), needs areassociated with characteristics or groupsof charac- teristics.Preferences are not directlyassociated with goods, but rather withtheir characteristics.

Procedural rationality Proceduralrationality does not appearin the quotes from the five nonorthodoxauthors presented above, but proceduralrationality is a principlethat should guide the construction of a PostKeynesian theory of consumerchoice. It has been arguedelsewhere that procedural rationality,in contrastto neoclassicalsubstantive or instrumentalratio- nality, is one of the presuppositionsof the Post Keynesian research program (Lavoie, 1992a), the others being realism, organicism, and production. In the realworld, in contrastto neoclassicalmodels, agentslack perfect

4 lronmongerspeaks of wants, ratherthan characteristics,but he has a similarcon- sumptiontechnology. Lancaster'sfirst paperon the subject was publishedin 1966, while Ironmongerdeveloped his version of characteristicsin his 1961 dissertation (Lancaster,1991, p. 180). 544 JOURNALOF POST KEYNESIANECONOMICS knowledge and the ability to process a large amount of information. They generally lack confidence in their informationand their assess- ment of it. The substantiverationality of neoclassical models can be neithera guide nor a descriptionof decision making.Hence, an approach to rationalitydifferent from the neoclassical one is required.Such an approachto rationality,in cases of uncertaintyor of insufficient capa- bilities to process existing information, consists of means to avoid complex calculations and considerations,and of proceduresenabling decisions to be takendespite incompleteinformation. These means and proceduresmight comprise rules of thumb, the acceptance of social conventions,reliance on the hopefullybetter informed opinion of others. Agents follow proceduresthat are sensible given theirbounded knowl- edge and computationalcapabilities. Their rationality,as Simon (1976) would have it, is of a proceduralsort. Proceduralrationality is also sometimes called bounded rationality.5 Agents take decisions, but within a very constrained framework, bounded on the one hand by the set of informationto which they have access and which they are able to process, and on the otherhand by the rules and conventions that they have imposed on themselves to handle this infonnation. One may add, following Shackle (1966, p. 101), that the decisional frameworkis also bounded by the imaginationof the decision maker.6As a consequence,procedural rationality goes beyond maximization subject to constrainedcomputational abilities. In most situationsoccurring in the real world, the solutions sought can only be good or satisfying solutions,because it is impossible to know what the optimalsolution is, or whethersuch an optimalsolution exists. As many scholars have recently pointed out, this is consistent with Keynes's views of rationality, as he "may be viewed as basing the whole of economictheory on a single, broad,non-Neoclassical conception of agentrationality" (O'Donnell, 1989, p. 272). In the case of consumerbehavior, it has long been establishedby

S Simon (1979) himself uses this term.Some authors,such as O'Donnell (1991, p. 15) speak of weak rationality,contrasting it to the neoclassical strong rationality (which looks like Simon's global neoclassical rationality).He also refers to the latter by the terms of hyper-rationality,ultra-rationality, and pseudo-rationality (1991, p. 45). I fnd these threeexpressions illuminating.On the other hand, proceduralor boundedrationality could also be called reasonable or sensible rationality,if such couplings are not pleonasms. 6 Hence the tidtlesThe EconomicImagination and The CorporateImagination of the books writtenby PeterEarl (1983, 1984). POST KEYNESIANAPPROACH TO CONSUMERCHOICE 545 marketingspecialists that consumer choice usually involves very simple procedures(Earl, 1986, p. 58). Veryoften there is no decisionprocess to speakof: purchasesare made on recommendations,in confonnity to socialnorms, with the consideration of few altemativesand on thebasis of few criteria.Some of thefollowed procedures are conscious, we may thenspeak of rules,while others are unconscious, we mayrefer to them as habits(Hodgson, 1988, p. 106). For instance,a large partof our repetitivespending on consumptiongoods is based on routinesand habits. Furthermore,short-cuts are used to arriveat quick decisions. As shown by Earl (1986, p. 183), variousforms of compensatoryand noncompensatoryfiltering procedures are used to processalternatives: formsof lexicographicscreening rules as well as multistagedecision processes.8Very often, the chosenrule or short-cutwill resultfrom conventions,based on the choices of ourrelatives, our friends, our social class,our culture. Procedural rationality is a socialphenomenon, linked to institutions. In short,in most of our decisions,optimizing is impossibleeither becauseof a lackof informationand fundamental or because of anoverload of informationand deficient computational capabilities. Evenwhen optimizing is possible,it maynot be theobjective pursued.9 Proceduralrationality seen fromthe perspectiveof neoclassicalsub- stantive rationalitymay seem to be "ad hocery,"but procedural responses are the only sensible answer to the sort of economic environmentsurrounding most of our decisions in the real world (HargreavesHeap, 1989, p. 164).The individualdescribed by proce- duralrationality is no less rationalthan the homo oeconomicusof neoclassicaltheory. The environmentfaced by the Post Keynesian economic and the rationalitycompelled by this environment are simplydifferent.

7 Davidson (1972, p. 19), for instance, writes that"much of consumptionactivity in- volves repetitive acts in similar circumstanceseach day." 8 It is often shown thatunder these rules the transitivityassumption of choices may not hold. See Earl (1986, pp. 244-247) for instance. 9 An enormousamount of literatureabout choice in situations(such as lotteries) has developed following Allais's (1953) critiqueof neoclassical rationalitybased on expected utility theory (see Machina, 1987). For some authors,a properlygeneral- ized expected utility theorymight eventuallytake care of Allais's paradoxand all other observedchoice behavioranomalies, such as Ellsberg's paradox, re- versals, and framing.For other authors,including presumably most Post Keynesian authors,those so-called anomaliesdemonstrate that "no optimizationprinciples of any sort lie behind the simplest of humanchoices" (Gretherand Plott, 1979, p. 623). 546 JOURNALOF POST KEYNESIANECONOMICS

Satiablewants

The second principle,that of satiablewants, can be likenedto the neoclassicalprinciple of diminishingmarginal utility, but it takes a particularmeaning in the Post Keynesiantheory of the consumer.In standardneoclassical consumer theory, more of one goodor moreof a characteristicwill bringeventually less additionalutility, however that last tenn is defined.For neoclassicaleconomists, however, satiation arisesonly if pricesare null or if incomesare infinite. InPost Keynesian economics, satiation arises with positive prices and finiteincome. There are threshold levels of consumptionbeyond which a good,or its characteristics,brings no satisfactionto its consumer.This view of the principleof satiationis morein line withwhat the earlier neoclassicaleconomists had inmind. Whereas mostpresent neoclassical thinkersassociate consumption activity with the maximizationof a uniquesort of need,utility, classical and earlier neoclassical economists associatedconsumption with the fulfilmentof a varietyof needs,each standingin hierarchy.To be ableto go fromone needto the next,the principleof satiationof needsthus had to be assumed(see Lancaster, 1972,p. 147).When neoclassical economists took away the notion of a multiplicityof needs,the principleof satiationhad to be modifiedto accommodatethe new theoryof utility maximization(Ironmonger, 1972,pp. 10-12). As a consequenceof the above,one has to carefullydistinguish wants fromneeds, as do Lutz and Lux (1979). In neoclassical standard versions of consumertheory, diminishing is explainedby the varietyof wants.It is possibleto rankthe various wants, but this ranking solely dependson the abilityof eachwant to createutility, and in that sense all wantsare equal. A Post Keynesianview wouldfocus on the hierarchyof needs,where some needs are more basic than others, which impliesthat they must be fulfilledin priority.In that sense all needsare notequal. Some needs are bound to be satiatedmuch earlier than others, as recognizedinitially by Mengerand Marshall. We shouldthus differ- entiatebetween needs and wants. Needs are liable to a hierarchic classificationand are the motorof consumerbehavior. Wants evolve fromneeds. They can be substitutedfor eachother and constitute "the variouspreferences within a commoncategory or level of need"(Lutz andLux, 1979,p. 21). Thisbrings forth the next two principles of a Post Keynesianconsumer theory. POST KEYNESIANAPPROACH TO CONSUMERCHOICE 547

Separateneeds

Categoriesof needs can be distinguishedfrom each other.Two rela- tively recentcontributions to mainstreamconsumer theory can give an aperfu of how this canbe formalized.First, a needmay be associated with a characteristicpossessed by a good. Goods,however, possess manycharacteristics, and it maybe thata singleneed exhibits several features.If thematrix of consumptiontechnology featuring character- isticsand goods can be decomposed,goods of a givensubmatrix are then saidto constitutean intrinsicgroup (Lancaster, 1972, p. 126).The group of characteristicscorresponding to thatgroup of goodswill constitutea multifacetedneed. A needwill thus be represented by a setof characteristics defininga submatrixof the decomposablematrix of the consumption technology.Each need consistsof a distinctset of characteristics.The principleof separabilityof needsthus gives some substance to thenotion of categoriesof needsthat can be distinguishedfrom each other. The principleof the separabilityof needshas beenused in a slightly differentcontext by mainstreamauthors. The most widely used models in econometricconsumer demand studies (the "Rotterdam" and "Linear Expenditure"models) assume the separability of theutility function. In so doing,they assume the principle of theseparability of needsdescribed above,in so faras one is readyto acceptthat utility is derivedfrom the satisfactionof needs and their sets of characteristics.By assuming separabilityon theutility of broadcategories of goods,these empirical modelsassume the existence of intrinsicgroups and their corresponding distinctneeds. The assumptionof the separabilityof needsis pushedone stepfurther in the so-calledutility-tree approach (Strotz, 1957). There, each broad categoryof expendituresmay be furthersubdivided in severalsub- branches,each subject to its own separableutility function in the goods space.°1The implicationof this,in the characteristicsand goods space, is thatthe submatricesof the matrixof consumptiontechnology can themselvesbe decomposedto formsubgroups (Arrous, 1978, p. 259), each subgroupcorresponding to a second-orderelement of the utility tree.We couldcall subneedsthose second-order elements. Theutility-tree approach is clearlylinked to procedural rationality. The separabilityof needs (the existenceof intrinsicgroups in Lancaster's terminology)allows the consumerto dividethe decision-makingpro- cess into a seriesof smallermultistage decisions. The consumerfirst

10Lancaster (1991, pp. 70-71) discusses the relationshipbetween utility-treeanaly- sis and his analysis of characteristics. 548 JOURNALOF POSTKEYNESIAN ECONOMICS makesan allocationof his or herbudget among needs, then spends that allocationamong the various wants or subgroups of eachneed, indepen- dentlyof whathappens for the other needs. The principleof separabilityof needs,by itself,offers only a limited challengeto neoclassicalconsumer theory. It still involveswhat Lan- castercalls preferencesubstitution between groups of characteristics, thatis, betweenthe various needs. There is, in theutility-tree approach, marginalsubstitution between rival budget categories (Earl, 1983, pp. 154-155). Changesin the relativeprices of goods within a given categoryof wantswill haveno effecton thebudget allocation between variousneeds, while a fall in the overallprice of a groupof goods correspondingto a given need will have repercussionson the budget allocationof all needs.The all-importantneoclassical principle of substi- tutionis stillpreserved. The principle of theseparability of needs,however, setssubstantial restrictions to theprinciple of substitution.As Deatonand Muellbauer(1980, p. 128)note, "separability places severe restrictions on the degreeof substitutabilitybetween goods in differentgroups." Further restraintsmay be addedif onegoes beyond the principle of separabilityof needs,by introducingthe principle of thesubordination of needs.

Subordinateneeds

Withinthe sole context of theseparability of needs,although every need couldbe madecharacteristically distinct, each need couldbe brought backunder a uniqueutility label. With the additionof the principleof the subordinationof needs,utility can only be representedby a vector. Hedonismis at stake. Thenotion of subordinateneeds can be relatedto thepyramid of needs suggestedby AbrahamMaslow and his humanisticschool of psychol- ogy. This pyramidis saidto be constitutedby five sets of needs,from themore basic ones to thehighest needs: physiological needs (air, water, food, sex, sleep); safety needs (health,education, shelter, stability, protection);social needs, subdivided in two sets: belongingness and love needson theone hand,and self-respect and the esteem of otherson the otherhand; finally the moralneeds, which Maslow called self-actual- ization-that is, the searchfor truth, justice, aesthetics, the meaning of life, achievement(Lutz and Lux, 1979,p. 11;Lea et al., 1987,p. 499). This rankinglist is very similarto the one proposedby Marshall(see Haines,1982, p. 111).This does notmean that all individuals have the same mnldng,or that there is nointeIpenetration of the needs. But we cancertainly POST KEYNESIANAPPROACH TO CONSUMERCHOICE 549 presume that every individual entertainsa rankingof the above sort, especially at the level of the more basic needs, which may be construed as physical health and mental health (Doyal and Gough, 1991). Variousnonorthodox authors have presentedanalyses based on some version of the principleof subordinationof needs, in particularLutz and Lux (1979), Canterbery(1979), andMiller (1983). The latter's analysis is particularlyinteresting in view of whathas beenpresented in the above subsection. Miller (1983) makes use of the separabilityand the subor- dinationprinciples by setting up a utility-treeanalysis with separability of the utility function, where the first-stage budgeting problem is re- solved by assumingthat money is first allocatedto necessities and then to discretionaryneeds. There is no substitutionany more between the budget categories allocatedto necessaryneeds and discretionaryones. In thatspirit, most notableare the studiesof Roy (1943) andIronmonger (1972), who both call the principleof subordinationby the name of the principle of hierarchy,while Georgescu-Roegen(1954, p. 515) chris- tens as the principleof the irreducibilityof needs the conjunctionof the principles of separationand of subordination.For all those authors, needs are orderedaccording to lexicographicpreference ordering, with satiation.11All the principles previously invoked culminate in this hierarchy:needs are separable,the most basic needs are first taken care of in theirpnrority order, until they are satiatedat some thresholdlevel. The above variantof lexicographic orderingis called "an order of a lexicographic nature"by Arrous (1978, p. 230). Earl (1986, p. 233) refers to "behaviourallexicographic" rules, while Drakopoulos(1990, p. 154) speaks of "targetlexicographic" ordering. That sort of lexico- graphicordering is not only reasonablebut also compatiblewith proce- duralrationality, since a complete utility map is not required.Decisions about the most basic needs can be taken quite independentlyof the infornational requirementsof the higherneeds. Consumersneed know nothing whatsoever about the prices of the goods that are part of the higherneeds, and they need not rankalternatives that they cannotattain or that are beyond their satiationlevels.12 Orderingof a lexicographicnature plays a role at several stages. First, if the principleof subordinateneeds is correct,it plays a role at this highly

11 Some authorshave formalizedthe lexicographicelements of choice. See for in- stance Encarnaci6n(1964) and Fishbum (1974). 12 It can furtherbe arguedthat procedural rationality and rules akin to it are a neces- sity in a world where consumerslook at characteristicsrather than goods as such; it 550 JOURNALOF POST KEYNESIANECONOMICS abstractMaslowian level. Second, the hierarchyprinciple may be pre- dominantwith respect to the choice of materialneeds. Several authors have arguedthat there exists an orderof prioritiesamong needs strictly relatedto physical consumptiongoods (Clarkeand Soutar, 1982). This is clearly the case of the five Post Keynesianand neo-Ricardian authors who were cited earlier.Wants stand in a hierarchy,although, as pointed out by Robinson(1956, p. 354), "thereare considerable overlaps at each level.",13Third, as was arguedin the section on proceduralrationality, choices between similar goods (i.e., those relatedto wants ratherthan needs)may implythe lexicographic ordering of the featuresof thosegoods. While some neoclassical authors, such as Lancaster (1991, p. 77), recognizethe realismof lexicographicchoices in this thirdcontext, they usually deny that needs or materialneeds are subjectto the principleof subordination.This, it must be presumed,is mainly due to the devasta- ting consequencesof the irreducibilityof needs for neoclassical theory and its substitutionprinciple. Irreducibleneeds imply that they are incommensurable and therefore that "everything does not have a price.','4Neoclassical authors usually assume that all needs can be subsumedinto wants, so thatin the end one can always offer enough of a good to compensatefor the loss of another.The axiom of Archimedes, so popularwith choice theorists,does not hold any more (Earl, 1986, p. 249), nor does the axiom of gross substitution(Eichner, 1987, p. 632), so often invoked among generalequilibrium theorists.

Growingneeds Having assumed that indeed there exists a hierarchyof needs, or of subneeds, how do consumers move up the steps of the pyramid?The cannot be otherwise. The reason is thatin general the consumptiontechnology ma- trix is not squared,the numberof goods being differentfrom the numberof character- istics. This implies thatthe inverse of the matrixcannot be exactly calculatedto take care of the ,which is in the goods space. It follows that the "com- modity composition of the optimal consumptionbundle is indeterminate"(Roth, 1989, p. 76). Substantiveneoclassical rationalityis powerless. 13 Some authorsassociate "behaviorallexicographic" ordering with goods insteadof needs. See Drakopoulos(1990) for such an analysis, which is, however, in the spirit of the principleof the irreducibilityof needs. 14 This is a point made by Lane (1978, pp. 811-815), and more recently acknowl- edged by Hirschman(1982, p. 20). Lane arguesthat the satisfactionderived from friendship,family, worship,self-esteem, justice, and othersimilar needs-ecorre- spondingby the way to the upperechelons of the Maslowian pyramid-are to a large extent incommensurablewith the satisfactionderived from marketablegoods. POSTKEYNESIAN APPROACH TO CONSUMERCHOICE 551 basic answer is that individualsmove upwardsin the hierarchydue to income effects. Of course, as was alreadypointed out, differentindivid- uals have different scales and it is likely that widely different income levels are needed to reach the upper grades of the hierarchyof needs. Differentindividuals may have differentthreshold levels. The principle, however, remains. Unless there is some Freudianfixation with some type of need-that is, a neuroticobsession with some set of goods-the needs of individuals will grow as their lower needs are gradually fulfilled. Beyond the principle of satiation, discussed above, lies the principleof the growth of needs. When a need has been fulfilled, or more precisely when a threshold level for thatneed has been attained,individuals start attending the needs that are situatedon a higher plane. There are always new needs to be fulfilled. The needs of the lower levels, however, requireincome to be satisfied. To go from one level of need to anotherdictates an increase in the real income level of the individual.The fulfilment of new needs andtherefore the purchaseof new goods or new services arethus related to income effects. This is the microeconomic counterpartof the Post Keynesian focus on -that is, on macroeconomic income effects. What is being assertedis that income effects are much more importantto explain the evolution of expenditureson goods than are substitutioneffects. The latter only play a minor role in a static analysis of consumerbehavior, when similar goods or goods fulfilling the same wants arebeing considered.Indeed, changes in relativeprices have an impacton budgetallocation between needs only in so far as they have an impacton real income. The emphasisof traditionaltheory on substitutioneffects has led to the neglect of the study of income effects, but it has also led to a neglect of the learning process in consumptiontheory. How do consumers react when their incomes increase? How do they rank their new spending opportunities?How do they learn to spend their additionalspending power. A key element to these questionsis the impactof life-styles.

Nonindependence

The notion that consumers care about their relative position in the hierarchyof consumersis nothing new in the literature.Veblen (1899) is preciselyknown for having developed this point.Keynes (1972, p. 326) himself arguedthat there are two sortsof needs:the basic materialneeds and those needs "which are relativein the sense that we feel them only 552 JOURNALOF POST KEYNESIANECONOMICS if theirsatisfaction lifts us above,makes us superiorto, our fellows." ForKeynes, in contrastto themore basic needs, the needs of thesecond sortare insatiable because "the higher the generallevel, thehigher still are they."Mainstream economists have paid little attentionto this aspect,with few exceptions,such as Duesenberry's(1949) relative income hypothesisand Leibenstein's(1950) Bandwagonand Snob effects. The principleof nonindependencefocuses on the factthat decisions andpreferences are not made independently of thoseof otheragents, an assumptionthat runs in oppositionto the standardneoclassical models of choice.In mainstream theory, households with the same income level aregenerally presumed to havewidely different tastes. In Post Keynes- ian tlieory,households with similarincomes are presumedto fulfill needs in the same orderand to face similarthresholds. Beyond the physiologicalneeds, conventionis the main reasonfor which it is believedthat a hierarchyof needs or subneedswill be more or less identicalfor all individualsof a similarculture. A household'spattem of consumption"reflects the lifestyle of otherhouseholdsthat constitute its socialreference group" (Eichner, 1986, p. 160).The thresholds that eachindividual consumer sets as a targetare defined by normsarising fromthe referencegroup to whichthe consumerwishes to conforn (Miller,1983, p. 48). Theconsumption pattern of individualhouseholds is thusinfluenced by the demandstructure of householdswith similar incomesor similartypesofjobs, as hasbeen empirically shown recently (Alessieand Kapteyn, 1991). Althoughthe orderingof needsmay be to some extentthe resultof innatepreferences, it is clear,primarily with respect to the moresocial needs,that the publicitybeing exerted by the producingsector, as well as existinghabits and conventions, will havea substantialimpact on the compositionof consumption.This is where the principleof non- independenceplays its part.What is beingasserted is thatthe behavior of theconsumer, as well as the satisfactionderived from consumption, is not independentof thebehavior of otherconsumers. To thehierarchy of needscorresponds a hierarchyof consumers,which marketing offi- cersof largecorporations take advantage of, somethingGalbraith (195 8, ch. 11) calls thedependence effect. Individualsfeel relativelydeprived when they want a certaingood that a referencegroup possesses, and when they believe that it is feasibleto obtainthe good (Baxter, 1988, p. 52). Thegoal of publicity,besides the creationof purchasinghabits, is to provokeenvious feelings. Publicity POST KEYNESIANAPPROACH TO CONSUMERCHOICE 553 makesone realizewhat the Jonesesare up to, and indicateshow to suppressthe unhappy feeling of envy:the consumerneed only buythe good thatthe Joneseshave already incorporated into their structure of consumption. Thereare otherreasons for whichone shouldexpect individual de- mandsfor a given productto be a positivefunction of the demandof societyfor thatproduct. Consumers watch and copy otherconsumers becausein so doingthey learn how to spendtheir increased purchasing power.Consumers need to discovertheir preferences (Pasinetti, 198 1, p. 75). They arenot innate.15They areacquired by experienceand by imitationof the consumptionpattern of friendsor of peopleof higher ranksin the consumers'hierarchy. Purchases of specificproducts in chainreaction are thus explained by the infornationalcontent of con- sumptionby neighbors,relatives, friends, or acquaintances(Marris, 1964, p. 146). The impacton purchasesof socioeconomiccontact reinforcesthe belief that the composition of demanddepends on socio- economicclasses. The normsof consumption,while dependenton past standards,will evolve and be set to a large extenteither by imitationor by envy.16 Whateverthe springof action,the resultis the same:the hierarchyof needsor subneedsbecomes the samefor all, sinceall consumerstry to emulatethose that belongto the upperechelons of the hierarchyof society.These consumer elites set the trends,as underlinedby Nell in the quotes of the first section. It is ratherironic to note that some of the efforts made by neoclassical general equilibriumtheorists to evade the destructiveconsequences of the Mantel-Debreu-SonnenscheinImpos- sibility theoremhave reliedupon the assumptionof income classes with identicalpreferences (Kirman, 1989).17 These theoristsare thus foresak- ing one of the four presuppositionsof neoclassical analysis, tfiat of methodological individualism,to escape instability results in general 15 Or as Hirschman(1982, p. 21) would put it "We never operatein terns of a com- prehensivehierarchy of wants establishedby some psychologist surveying the multi- farious pursuitsand needs of mankind,but at any one point of our existence ... we pursuesome goals which then get replacedby others." 6 See Hirsch (1976) among many sociologists arguingalong those lines. 17 The various versions of the Impossibilitytheorem show that,starting from the usual assumptionsof maximizingbehavior which arerequired to demonstratethe ex- istence of equilibriumin neo-Walrasiangeneral equilibrium models, excess demand functions satisfying Walras'slaw can take almost any form. This result has devasta- ting consequences because it precludes the possibility of ever demonstratingthe uniqueness and stabilityof a neo-Walrasianequilibrium. 554 JOURNALOF POST KEYNESIANECONOMICS

equilibriummodels. In so doing, they have reinstateda form of organ- icism, which, in Post Keynesian consumer theory, is given substance throughthe principleof nonindependence.

3. Consequences for price theory

The emphasis of orthodox theory on static behavior has led to an excessive amountof researchon substitutioneffects. On the otherhand, income effects have been eitherneglected or assumedaway. Those who have attemptedto estimate the importanceof pure substitutioneffects on the general categories of consumption expenditures,after having taken into considerationthe income effects throughtime, have discov- eredthat these substitutioneffects, own-priceelasticities andcross-elas- ticities, arequite negligible (Deatonand Muellbauer, 1980, p. 71). While the own-priceelasticities of food, fuel, drinks,travel, entertainment, and otherservices turnedout to be negative, as expected, the absolutevalue of these elasticities were found to be no greaterthan 0.05. The price of clothing and housing was not statistically different from zero. These findings seem to correspondto the picture of consumer behavior drawn in the preceding section. The cause of these small substitutioneffects, within the Post Keynesian framework,is that the large categories of consumerexpenditures fulfill importantneeds that cannot be compensated one for another. Variations in their relative prices induce no change in consumptionbehavior, or very small ones. Only within each one of these large spending categories could one possibly observe more substantialsubstitution effects. One can thus presumethat the more disaggregatedthe analysis is, the more likely we are to find high absolute values of price elasticities. The findings of Houthakkerand Taylor (1970), however, show thateven at a muchmore disaggregatedlevel-that is, with over eighty categories of consumer goods-consumption expendituresare mainly detennined by habitsand income effects, while price substitutioneffects play a fairlymodest role. The crucialissue here is thatfluctuations in the price of a good, unless they are really substantial,will not have much impact on the quantities sold.'8 The majorexception to this predictionwould be new goods being

18 Thispoint was clearly made by Roy (1943,p. 16),in his exceptionallyclear- sightedpaper: "Variations in priceshave no impact,in ouropinion, on theoverall ranking,for those price changes only induce substitution effects within each group, withoutmodifying the ranking position of eachgroup. The notion of thehierarchy of needs,and hence that of commodities,is a reflectionof physiological,moral and psy- POST KEYNESIANAPPROACH TO CONSUMERCHOICE 555 introducedto consumers. Innovationson the consumer either create new needs or fulfill existing needs that were previously poorly attended.Besides these innovativecommodities, the reasonthat fluctu- ations of relative prices would have little impact on demandis that all goods respondto a need (or to a set of needs). Providedthat these needs, or subneeds, are arrangedin a preset order,the decrease in the price of a good will only make it more attractiveto consumerswho have already attainedthat partof the pyramid.All those who are still trying to attain their thresholdlevels with respectto lower needs will not be concerned by this price decrease. Furthermore,since a substantial amount of expendituresare done on the basis of habits and customs, the decrease in prices may go unnoticedunless it is heavily publicized.The decrease in the price of a good will only have an impactto the extent that it can replace other goods fulfilling the same needs, or more precisely what we have called the same wants.This is the traditionalsubstitution effect, limited, however, to the goods that have similar characteristics.In classical theory, these goods were for practical purposes treated as identical (Schefold, 1985, p. 112). The symmetricconsequence of the above is that changes in the prices of goods fulfilling needs of the higher levels of the hierarchywill have no impact whatsoever on the consumptionof the goods of the lower levels of the hierarchy(Roy, 1943). The reason is that these goods are not part of the basket or of the hypotheticalbasket of consumptionof all consumersthat have not yet fulfilled the requiredthresholds of their lower needs. These poorerconsumers just do not care aboutthe prices of the goods thathelp to fulfill the higherneeds since these goods cannot be acquireddue to theirbudget constraint.On the otherhand, if thereis an increase (or a decrease)in the relativeprice of the goods thathelp to fulfill the lower needs, this will have repercussionson the quantitiessold of all goods belonging to the higher part of the hierarchy. A lower relativeprice will increasethe real income of all households,leading to an increase in the consumptionof all goods fulfilling the higher needs of the hierarchy.19

chological motives, which, to a large extent, are indifferentto the variationsof prices that can be observed in practice."Later Roy adds:"Only within each group do the relative prices, combined with individualtastes, have any effect for such or such com- modity, throughthe mechanismof substitution"(p. 22, my translation). 19Again, Roy (1943, p. 20) showed clear-sightedness:"As a result of the hypothesis made about the allocationof income, any price change related to a group of rank i 556 JOURNALOF POT KEYNESIANECONOMICS

What we have here is an asymmetric relation similar to the one established by the classical authorswhen they were discussing neces- sary goods and luxury ones (Roncaglia, 1978, p. 52). According to the classical classification, were nonnecessary goods that were not consumedby the workers.Only the rentiersand the capitalists could spend their income on luxury goods. Necessary goods, on the other hand, were consumedby both the workersand the upperclasses. The consequence of this, accordingto Ricardo,was that changes in the productionconditions of luxury goods or on their prices did not have repercussionson the overall or on the cost of producing necessary goods. On the other hand, changes in the prices or the production conditions of necessary goods had repercussions on the overall rateof profit andon the cost of productionof luxury goods. The neo-Ricardianshave drawnsimilar conclusions from Sraffa's analysis of basic and nonbasic commodities, the formerplaying the role of the necessary goods, so to speak, while the latter replacedthe role played by luxury goods. Steedman(1980) has integratedthe classical theory with the neo-Ricardianapproach by showing that,in a world of hetero- geneous labor,the rateof profitdoes not dependon the overall average real rate. Rather,only the real wage rate of workers producing goods consumed by workersis a detenninantof the rate of profit. There is thus a strong relationshipbetween the classical and neo- Ricardianasymmetric conception of the economy andthe Post Keynes- ian theory of consumer behavior, based on a hierarchyof needs and hence presumablyon a hierarchyof goods, from the more basic neces- saries to the productsof high luxury. Whereasthe neo-Ricardianfocus is on the consequencesof this asymmetricrelation for the detenninants of relative prices, the impact of the asymmetryfor the Post Keynesian theory is on the quantitiesconsumed. Using the same conceptualframe- work that makes a distinctionbetween goods fulfilling lower needs and those responding to higher needs, the neo-Ricardians show how a change in the conditionsof productionor in the compositionof demand by the workerscould affect relative prices or the purchasingpower of consumers through the cost side. On the other hand, a proper Post will have no impacton the demandfor the goods belonging to a group of inferior rank, whereas it will influence not only the demandfor the goods of group i but also the demandfor all the other groupsof superiorrank. Another way to put it is to say that price changes of luxurygoods will have no influence on the demandfor goods of first necessity or on the demandfor goods thatprovide some degree of comfort withoutbeing luxury"(my translation). POSTKEYNESIAN APPROACH TO CONSUMERCHOICE 557

Keynesiantheory of consumerchoice would show how thesechanges in purchasingpower are translatedinto increasesin the quantities consumedof thevarious products, and how little substitution effects are inducedby changesin relativeprices, owing to the existenceof hierar- chicalneeds and subneeds. Onemay thus conclude that when changes in relativeprices are small, the substitutioneffects that they induce can be ignored,either because they are negligible,as in the case of goods thatare part of different groups,or becausethey concerngoods that for all practicalpurposes maybe consideredidentical. On the otherhand, substantial changes in the relativeprice of a good areusually associated with novel products, which createnew needs.Pure substitution effects in this case do not arise,since this consumption innovation needs to be incorporatedwithin thehierarchy of needsand wants. This picture of thelack of importance of the substitutioneffect in consumerbehavior is certainlycompatible withthe views of neo-Ricardianauthors: If theeffect of theprice on the quantitybought is not appreciable,then the effect can be ignoredwithout great error. Alternatively, when the effectis importantenough to needgeneral consideration, it seems it will oftenbe thecase that the effect constitutes an irreversible change, which is incompatiblewith its treatmentin termsof a demandfunction. That is the effect will entaila permanentchange in the habitsof consumers, which even marginalistauthors would have to treatas a change in "tastes."[Garegnani, 1990, p. 131] The Post Keynesian consumer theory presented above offers some justificationto the dichotomythat some authorswish to introducein the analysis of prices on the one hand and of quantities on the other (Roncaglia, 1978, p. 15). Whereasin neoclassical analysis, everything simultaneouslydepends on everything,asymmetric relationships are oftenunderlined by PostKeynesian and neo-Ricardian authors. If prices have little effect on the structureof consumption,then it becomes possible to claim, as does Pasinetti(1981, p. 141), that "costs of production, i.e., technology, determine relative prices," while "4consumers'sdecisions, i.e., theirdemand, determine relative quanti- ties.",20 Leavingmicroeconomics for ,it is obviousthat by acceptingthe building blocks of thePost Keynesian theory of consumer theory, macroeconomicconstructions which focus on classes and

20 A similarclaim is made by Earl (1983, p. 50). 558 JOURNALOF POST KEYNESIANECONOMICS groupsof agentsbecome more appealing.Post Keynesiananalysis leads to the belief that agents or households belonging to the same income class have a similar structureof consumption.We know, however, that agents do not have identical incomes. There is thus amplejustification for utilizing income classes rather than the representativeagent in macroeconomicstudies. Furthermore, since income changes ratherthan price changes appearto be the main influence on changes in consump- tion, the importanceof the study of income effects in macroeconomics is reasserted. A furthereffect of a Post Keynesiantheory of consumerbehavior based on lexicographic choice with satiationlevels has recently been under- lined by Drakopoulos(1992b). Ironmonger(1972) had alreadydemon- stratedthat such behaviorled to kinked individualdemand curves, but thatthe kinks would disappearwhen individualshave differentincome levels and different threshold levels or priority order. However, if individualswith identicalincome entertainthe same conventions-that is, the same hierarchy of needs and similar threshold levels-even aggregatedemand curves will exhibit a multitudeof kinks. It has been pointed out by Drakapoulosthat this could offer an additionalexplana- tion of price stickiness in the face of variationsin demandof consumer goods, even when marginalcosts are not constant.This may be a valid lesson for those who do not wantto rejectorthodox theory completely.2'

4. Conclusion

It has been claimed from the startthat, althoughPost Keynesianshave paid little attentionto issues relatedto consumerchoice, thereis agree- ment on what shouldconstitute the basic principlesof a Post Keynesian theory of consumerbehavior. It was first shown that well-known Post Keynesians-not, however, renowned for their in consumer theory-have put forthstatements that show substantialoverlaps about the aspects of consumerbehavior that should be emphasized. The common groundupon which a Post Keynesiantheory of consumer choice could be constructedwas then presented. Six principles were identified:the principleof proceduralrationality, the principleof satia- ble wants, the principleof separabilityof needs, the principleof subor-

21 It has also been shown by Watts and Gaston (1982-83) that, if the consumption technology is nonlinear,Lancaster's characteristics will not yield smooth demand curves, even though lexicographicchoices and satiationdtresholds are being omitted. POST KEYNESIANAPPROACH TO CONSUMERCHOICE 559 dinationof needs,the principle of thegrowth of needs,and the principle of nonindependence.It was shown that these principles entail the use of noncompensatoryprocesses of choice, based for instanceon lexico- graphicordering with satiation thresholds. This is theapproach put forth notablyby Georgescu-Roegen,who arguedthat needs areirreducible andstand in a hierarchy.It was furthershown that the Post Keynesian authorswho have dealt with the issue of consumerchoice have favored ananalysis based on Ironmonger'sor Lancaster's characterisfics, where a needis identifiedby a groupof characteristics. Two of the main consequencesof this Post Keynesiananalysis of consumerchoice may now be recapitulated.First, macroeconomic modelsbased on the analysisof incomeclasses and on incomeeffects area legitimateoutgrowth of a PostKeynesian theory of the consumer where price substitutioneffects are not importantor severelycon- strainedto goods that respondto similarcharacteristics, and where increasesor changesin demandare mostly detennined by increasesin real incomes or changes in consumer preferences. Second, the Post Keynesian theory of consumption,based on the hierarchicalnature of needs, is also reminiscentof the classical distinctionbetween necessar- ies and luxury goods and of the neo-Ricardiandivision between basic and nonbasic commodities. Under these circumstances,to ignore sub- stitution effects, based on relative prices, appears to be much less disastrousthan to ignore income effects and thresholdlevels. Fromthe empiricalstandpoint, the Post Keynesiantheory of consumer choice points in the direction of studies that would try to identify the groups of goods that would correspondto the various needs and sub- needs, and that would trace the order of acquisition of these various groupsof goods. Also, ratherthan trying to estimateimaginary demand curves, empiricalstudies shouldfocus on estimatesof income andprice elasticities, taking into account preference interdependenceand the imbedded and hierarchicalnature of needs or groups of goods within categoriesof expenditures.

REFERENCES Alessie,Rob, and Kapteyn, Arie. "Habit Formation, Interdependent Preferences and DemographicEffects in theAlmost Ideal Demand System." Economic Journal, May 1991,101, 404-419. Allais,Maurice. "Le comportement de l'hommerationnel devant le risque:critique des postulatset axiomesde l'ecole americaine."Econometrica, October 1953, 21, 503-546. 560 JOURNALOF POST KEYNESlANECONOMICS

Arrous,Jean. Imperfection de l'information,incertitude et concurrence.Doctoral dis- sertation,Universite de Strasbourget Universitedes Sciences Sociales de Grenoble, 1978. Baxter, John L. Social and Psychological Foundationsof EconomicAnalysis. Lon- don: Harvester-Wheatsheaf,1988. Caille, Alain. Critiquede la raison utilitaire.Paris: La Decouverte, 1988. Canterbery,E. Ray. ",Necessities and DistributiveEfficiency." In James H. Gapinsky and CharlesE. Rockwood (eds.), Essays in Post-KeynesianInflation. Cam- bridge, MA: Ballinger, 1979, pp. 79-103. Clarke,Y., and Soutar,G.N. "ConsumerAcquisition Patternsfor DurableGoods: AustralianEvidence." Journal of ConsumerResearch, March 1982, 8, 456-460. Davidson, Paul. Money and the Real World.London: Macmillan, 1972. Deaton, Angus, and Muellbauer,John. Economics and ConsumerBehavior. Cam- bridge:Cambridge University Press, 1980. Doyal, Len, and Gough, Ian. A Theoryof HumanNeed. London:Macmillan, 1991. Drakopoulos,Stavros A. "The Implicit Psychology of the Theory of the Rational Consumer."Australian Economic Papers, December 1990, 29, 182-198. . "PsychologicalThresholds, Demand and Price Rigidity."Manchester School of Economics and Social Studies, June 1992a, 40, 152-168. . "Keynes' Economic Thoughtand the Theory of ConsumerBehaviour." ScottishJournal of ,August 1992b, 39, 318-336. Downey, Ezekial H. 'The Futility of MarginalUtility." In R. Albelda, C. Gunn, and W. Waller (eds.), Alternativesto Economic Orthodoxy.Armonk, NY: M.E. Sharpe, 1987, pp. 48-59 (firstpublished in Journal of Political Economy, 1910). Duesenberry,James S. Income, and the Theoryof Conswner Behavior. Cam- bridge, MA: HarvardUniversity Press, 1949. Earl,Peter E. The EconomicImagination: Towards a BehaviouralAnalysis of Choice. Armonk,NY: M.E. Sharpe, 1983. . The CorporateImagination: How Big CompaniesMake Mistakes.Brighton, UK: Wheatsheaf, 1984. . Lifestyle Economics: ConsumerBehaviour in a TurbulentWorld. Brighton, UK: Wheatsheaf, 1986. Eichner,Alfred S., ed. A Guide to Post-KeynesianEconomics. White Plains, NY: M.E. Sharpe, 1979. Eichner,Alfred S. Towarda New Economics: Essays in Post-Keynesianand Institu- tionalist Theory.London: Macmillan, M.E. Sharpe, 1986. . The Macrodynamicsof AdvancedMarket Economies. Armonk,NY: M.E. Sharpe, 1987. Encamacion,Jose. "A Note on LexicographicPreferences." Econometrica, January- April 1964, 32,215-217. Fishburr,Peter C. "LexicographicOrders, and Decision Rules: A Survey." ManagementScience Theory, July 1974, 20, 1442-1471. Galbraith,John K. The AffluentSociety. London:Hamish Hamilton, 1958. Garegnani,Pierangelo. "Sraffa: Classical versus MarginalistAnalysis." In K. Bharadwajand B. Schefold (eds.), Essays on : Critical Perspectives on the Revival of Classical Theory.London: Unwin Hyman, 1990, pp. 112-158. Georgescu-Roegen,Nicholas. "Choice,Expectations and Measurability."Quarterly POST KEYNESIANAPPROACH TO CONSUMERCHOICE 561

Journal of Economics, November 1954,48,503-534. Grether,David M., and Plott, CharlesR. "EconomicTheory of Choice and the Prefer- ence Reversal Phenomenon."American Economic Review, September1979, 69, 623- 638. Haines, WalterM. 'The Psychoeconomicsof HumanNeeds: Maslow's Hierarchy and Marshall's OrganicGrowth." Journal of , Winter 1982,9, 97-121. HargreavesHeap, Shaun.Rationality in Economics. Oxford:Basil Blackwell, 1989. Hirsch, Fred. Social Limits to Growth.London: Routledge and Kegan Paul, 1976. Hirschman,Albert 0. ShiftingInvolvements. Princeton, NJ: PrincetonUniversity Press, 1982. Hodgson, Geoffrey M. Economics and Institution:A Manifestofor a ModernInstitu- tional Economics, Cambridge:Polity Press, 1988. Houthakker,H.S., and Taylor, L.D. ConsumerDemand in the UnitedStates: Analy- ses and Projections, 2d ed. Cambridge,MA: HarvardUniversity Press, 1970. Ironmonger,D.S. New Commoditiesand ConsumerBehaviour. Cambridge: Cam- bridge University Press, 1972. Keynes, JohnMaynard. Essays in Persuasion, The Collected Writings,vol. 9. Lon- don: Macmillan, 1972. Kirman,Alan. 'The IntrinsicLinmts of Modem Economic Theory:The EmperorHas no Clothes."Economic Journal, Supplement1989, 99, 126-139. Lancaster,Kelvin. ConsumerDemand: A New Approach.New York: ColumbiaUni- versity Press, 1972. . "Hierarchiesin Goods-CharacteristicsAnalysis." In Modern ConsumerThe- ory. Aldershot,UK: EdwardElgar, 1991, pp. 69-80. Lane, RobertE. "Marketsand the Satisfactionof HurnanWants." Journal of Eco- nomic Issues, December 1978, 12, 799-827. Lavoie, Marc. 'Towards a New ResearchProgramme for Post-Keynesianismand Neo-Ricardianism."Review of Political Economy, January1992a, 4, 37-78. . Foundationsof Post-KeynesianEconomic Analysis. Aldershot,UK: Edward Elgar, 1992b. Lea, Stephen E.G.; Tarpy,Roger M.; and Webley, Paul. The Individualin the Econ- omy:A Survey of EconomicPsychology. Cambridge:Cambridge University Press, 1987. Leibenstein,Harvey. "Bandwagon,Snob, and Veblen Effects in the Theory of Con- sumer Demand."Quarterly Journal of Economics, February1950, 64, 183-207. Lutz, MarkA., and Lux, Kenneth.The Challenge of HumanisticEconomics. Menlo Park,CA: Benjamin/Cummings,1979. Machina,Mark J. "Choiceunder Uncertainty: Problems Solved and Unsolved."Jour- nal of EconomicPerspectives, Summer 1987, 1, 121-154. Marris,Robin. The EconomicTheory of Managerial .New York: Free Press of Glencoe, 1964. Miller, Michael S. "Methodologyand the Theory of ConsumerBehavior." Review of Social Economy,April 1983, 41, 39-51. Nell, EdwardJ. "Demand,Pricing and ."In TransformationalGrowth and EffectiveDemand: Economics after the Critique.London: Macmillan, 1992,pp.381-451. 562 JOURNALOF POS KEYNESIANECONOMICS

O'Donnell, Rod M. Keynes:Philosophy, Economics and Politics. London:Macmil- lan, 1989. . "Keynes on Probability,Expectations and Uncertainty."In Rod M. O'Donnell (ed.), Keynes as Philosopher-Economist.New York: St. Martin'sPress, 1991,pp.3-60. Pasinetti,Luigi L. StructuralChange and .Cambridge: Cambridge University Press, 1981. Robinson, Joan. The Accumulationof Capital. London:Macmillan, 1956. Roncaglia, Alessandro.Sraffa and the Theoryof Prices. New York: Johl Wiley, 1978. Roth, Timothy P. The Present State of ConsumerTheory, 2d ed.. New York: Univer- sity Press of America, 1989. Roy, Rene. "Lahierarchie des besoins et la notion de groupesdans l'economie de choix." Econometrica,January 1943, 11, 13-24. Schefold, Bertram."On Changes in the Compositionof Output."Political Economy, 1985, 1 (2), 105-142. Shackle, G.L.S. The Nature of Economic Thought.Cambridge: Cambridge Univer- sity Press, 1966. Simon, HerbertA. "FromSubstantive to ProceduralRationality." In S.J. Latsis (ed.), Method and Appraisal in Economics. Cambridge:Cambridge University Press, 1976, pp. 129-148. . "RationalDecision Making in Business Organizations."American Eco- nomic Review, September1979,69,493-513. Steedman,Ian. "HeterogeneousLabour and Classical Theory."Metroeconomica, February1980,32, 39-50. Strotz, RobertH. "The EmpiricalImplications of a Utility Tree."Econometrica, April 1957, 25, 269-280. Thurow, LesterC. "ProducerEconomics." In Lloyd G. Reynolds, Stanley H. Mas- ters, and Coletta H. Moser (eds.), Readings in Labor Economics and Labor Rela- tions. Englewood Cliffs, NJ: Prentice-Hall,1991, pp. 41-49. Veblen, Thorstein.The Theoryof the Leisure Class. New York:Modem Library, 1931 (first publishedin 1899). Watts, MartinJ., and Gaston,Noel G. "The 'Reswitching' of ConsumptionBundles: A Parallelto the CapitalControversies?" Journal of Post KeynesianEconomics, Win- ter 1982-83,5, 281-288.