richard bronk epistemological difficulties with...

23
Richard Bronk Epistemological difficulties with neoclassical economics Conference Paper Original citation: Originally presented at Southern Economic Association 2011, 20 th November 2011 This version available at: http://eprints.lse.ac.uk/39423/ Available in LSE Research Online: November 2011 © 2011 the Author LSE has developed LSE Research Online so that users may access research output of the School. Copyright © and Moral Rights for the papers on this site are retained by the individual authors and/or other copyright owners. Users may download and/or print one copy of any article(s) in LSE Research Online to facilitate their private study or for non-commercial research. You may not engage in further distribution of the material or use it for any profit-making activities or any commercial gain. You may freely distribute the URL (http://eprints.lse.ac.uk) of the LSE Research Online website.

Upload: duongnguyet

Post on 05-May-2018

222 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Richard Bronk Epistemological difficulties with ...eprints.lse.ac.uk/39423/1/Epistemological_difficulties_with... · Epistemological difficulties with neoclassical economics

Richard Bronk Epistemological difficulties with neoclassical economics Conference Paper Original citation: Originally presented at Southern Economic Association 2011, 20th November 2011 This version available at: http://eprints.lse.ac.uk/39423/ Available in LSE Research Online: November 2011 © 2011 the Author LSE has developed LSE Research Online so that users may access research output of the School. Copyright © and Moral Rights for the papers on this site are retained by the individual authors and/or other copyright owners. Users may download and/or print one copy of any article(s) in LSE Research Online to facilitate their private study or for non-commercial research. You may not engage in further distribution of the material or use it for any profit-making activities or any commercial gain. You may freely distribute the URL (http://eprints.lse.ac.uk) of the LSE Research Online website.

Page 2: Richard Bronk Epistemological difficulties with ...eprints.lse.ac.uk/39423/1/Epistemological_difficulties_with... · Epistemological difficulties with neoclassical economics

RichardBronkPresentedtoSEA,November2011

1

Epistemologicaldifficultieswithneo­classicaleconomics

Abstract

Thispaperaddressesseveralrelatedepistemologicaldifficultieswiththeneo‐classicalmicro‐foundationsofstandardeconomics,andexploreswhatthesedifficultiesimplyabouttherecentdominanceofthesemicro‐foundationsinbotheconomicresearchandappliedpolicyandfinancesettings.BuildingonthecrucialdistinctionmadebyKnightandKeynesbetweencalculableriskanduncertainty,thepapermakesafurtherdistinctionbetweentwotypesofuncertainty–epistemologicalandontological.Itarguesthatstandardeconomicsunderestimateshowproblematictheinterfacebetweentheoryandrealityisforbotheconomicagentsandsocialscientists,andoverlookstheimpossibilityofknowingthefutureyettobecreatedbyinnovationandthefreedomtochoose.Thepaperoutlinesalternativemicro‐foundationsthatcapturehowagentsuseimaginationandnarrativesalongsidemarketcoordinationtocopewithuncertaintyandconstructthefuture.Italsoexplainsseveralresearchandpolicyimplicationsoftakinguncertaintyseriously–thelimitsofprediction,theincoherenceofmanybolt‐onamendmentstoneo‐classicaltheory,thedangersofanalyticalmonocultures,andtheimportanceofdiscourseanalysisandpluralisminmodelling.Recognitionthatneo‐classicalmodelsareinappropriatefordealingwithproblemscharacterisedbywidespreadinnovationanduncertaintyshouldaugmentratherthanreducetheundoubtedvalueofthesesamemodelsinmanyotherareasofanalysis.Keywords:ontologicaluncertainty,epistemologicaluncertainty,rationalexpectationshypothesis,neo‐classicaleconomics,imagination,narrative,innovation,analyticalmonocultures,modellingpluralism,bolt‐onamendments

Contents

1.Introduction2.Theneo‐classicaltreatmentofinformationproblems3.KnightandKeynesoncalculableriskversusuncertainty4.Epistemologicaluncertainty:complexitynotcomputed5.Ontologicaluncertainty:thefutureisyettobecreated6.Romanticpost‐Kantianepistemology:implicationsforeconomists7.Thelimitsofreason8.Thelimitsofprediction9.Alternativemicro‐foundations:imagination,narrativeandemotion10.Whatdopricestellus?11.Monocultures,bubblesandtheneedfordiscourseanalysis12.Thedangersofbolt‐onamendmentsandthecaseforpluralisminmodelling

Page 3: Richard Bronk Epistemological difficulties with ...eprints.lse.ac.uk/39423/1/Epistemological_difficulties_with... · Epistemological difficulties with neoclassical economics

RichardBronkPresentedtoSEA,November2011

2

1.IntroductionItisfoolhardyatthebestoftimesforaphilosophertoapproacheconomiststocriticisetheirassumptionsandmethods.Worsestill,thispapermayappeartoignoretwopiecesofadvicemostofuswouldgivestudents:nottotackleatopictoobroadtobeansweredadequatelyinonepaperandtobewarecontesteddefinitionalterms.So,whatistheexcuseforsuchtemerity?First,nothingismoreimportanttoadisciplineaimingtoexplaineconomicactivitythananunderstandingofthefoundationandlimitsofknowledgefacingbotheconomicagentsandsocialscientists.Andonthesemattersofepistemology,itisatleastplausiblethatphilosophershavesomethingtooffereconomists.Secondly,whilethemodelling,researchandpolicyimplicationsdiscussedinthepapermaybefar‐reaching,thecentralargumentonwhichtheydependrestsonacoupleofrelativelysimple,ifwidelyignored,ideas:thattheinterfacebetweentheoryandsocialrealityisasproblematicasitisessential;andthatwecannotknowthefutureyettobecreatedbyinnovationandundeterminedchoice.Ourbeliefsabouttheworldwelivein,andtheactionswetakeonthebasisofthosebeliefs,areinevitablystructuredbythetheoreticalandconceptualframeworkweuse.Whatismore,therecanbenosinglerightwayofmodellingorthinkingaboutsocialreality,givenitsmulti‐facetedandpre‐interpretednatureandtheimpossibilityofcapturingthewholeinanyoneconceptualframework.Thecomplexitiesofreal‐worldproblemsoftencannotbesafelyreducedtoonetheoreticalframeworkand,evenwhentheycan,theyfrequentlydefyfullcomputationinpractice.Suchepistemologicaluncertaintyischallengingenoughfortheneo‐classicalprojectofreducingeconomicchoicetotheoptimisationbetweenknownpreferencesanddefinedoptionsunderwell‐understoodconstraints.Moredevastatingstillistheontologicaluncertaintyimpliedbytheimpossibilityofknowingeventhecategoriesandpossiblenatureofthefutureyettobecreatedbyinnovativeeconomicagentswhoarefreelychoosingbetweennewlyimaginedoptionsormakingidentity‐definingchoicesbetweenincommensurablevaluesorindeterminateoptions(i.e.,incaseswhere,bydefinition,thereisnosinglerationalchoice).Tohavecoherenceandavoidthechargeofcreatingastrawman,thispapermustalsoanswerattheoutsetthethornyquestionofhowtodefine‘neo‐classicaleconomics’.Colanderhasarguedthattheterm‘neo‐classical’isunhelpfulbecauseitisconfusinglyusedtodesignatebothanill‐definedhistoricalperiod(say,1870to1930),whichwasdominatedbycertainmarginalistassumptions,andthebroadswatheofstandardmoderneconomics;andthatinpracticethelatterisfarfrommonolithicorclearlydependentontheformerforitsunderlyingassumptions(Colander,2001,151‐61).LikeFontana(2010,585),however,thispaperarguesthatmostexamplesofmodernstandardeconomicsshareasufficientlycoherentkernelatthelevelofmicro‐foundations–andasufficientlyhomogenousmethodologicalbase–tomakeitreasonabletouseagenericlabelforthem.Atthesametime,thedominantequilibriummetaphorbehindmosteconomicmodelsentailsacrucialelementofcontinuitywiththeera

Page 4: Richard Bronk Epistemological difficulties with ...eprints.lse.ac.uk/39423/1/Epistemological_difficulties_with... · Epistemological difficulties with neoclassical economics

RichardBronkPresentedtoSEA,November2011

3

of‘socialphysics’(Mirowski,1989)usuallyassociatedwiththeterm‘neo‐classical’byhistoriansofeconomics.Asaresult,thispaperuses‘neo‐classical’torefertoanyeconomicmodelbasedonthemicro‐foundationsofindividualagentsrationallyoptimisingamongconsistentlyrankedpreferenceswithingivenconstraintssothatthesystemasawholetendstosomedeterminateandpredictablesocialequilibrium.Crucially,thisbroaddefinitionincludesnotonlytheobvious‘pure’examplessuchastheefficientmarketshypothesisandtherationalexpectationsmodelsofLucasandMuth,butalsothosemanyschoolswithinmodernstandardeconomicsthatusebolt‐onamendmentstomakethecentralmicro‐foundationsmoreplausibleasthebasisforpredictivemodels.Forexample,thisdefinitionincludesbyintentionmostofinformationandbehaviouraleconomics,wheretheymodelasymmetricinformation,framingeffectsetc.,aspredictableexceptionstothecorecasesofrationaloptimisationandefficientoutcomes.Inallcasesdesignatedhereas‘neo‐classical’,thereistheassumptionthatindividualsrationallyoptimisewithincertaineasilymodelledinformation,behaviouralandcognitiveconstraints,withtheresultthatoutcomesarestillpredictableoncethesepredictableconstraints,orexceptionstoperfectrationality,aretakenintoaccount.Bybeingusedinthisway,theterm‘neo‐classical’canalsoreferbyextensiontotwoimportantelementsthatenable(mostof)thedisparateschoolsofmoderneconomicstocoalesceintoastandard(or‘normalscience’)paradigmintheKuhnsense(Kuhn,1996):thefirstisthebeliefthattheeconomicsystemasawholecanandshouldbeexplainedbymodelsthatfurnishtestablepredictions(Friedman,1953).Centraltothemoderneconomicsparadigmistheassumptionthatthereisalwaysonecorrectanswertohowrationaleconomicagentswillactgiventhevariousconstraintstheyareunder–sothatoncetheconstraintsareknown,andanysystematicbehaviouralandcognitivebiaseshavebeentakenintoaccount,predictionsaboutbehaviourcanbemade(andtested).Asweshallsee,thiscentralassumptioniscalledintoquestionbytheprevalenceofepistemologicalandontologicaluncertainty.Thesecondrelated‘neo‐classical’featureofthemoderneconomicsparadigmisthewidespreadtendencytoexplainawayinconvenientcounter‐examplestocorerationalistandequilibriumassumptionsbytheorymendingthatleavesintactthebasicanalyticalmonocultureatthelevelofmicro‐foundations.Thetendencytobolt‐onbehaviouralandotheramendmentstostandardrationalistmicro‐foundationstocorrectforexplanatoryerrorsisreminiscentoftheattemptsbyPtolemaicastronomytocopewiththemanyexceptionstothepredictionsmadebyitscoremodelofanearth‐centreduniverse(Bronk,2009,10).ForKuhn,suchPtolemaicparadigmmendingledtoamonstroussystemof‘compoundedcircles’,whose‘complexitywasincreasingfarmorerapidlythanitsaccuracy’,andthiswasatell‐talesignofparadigm‘crisis’(Kuhn,1996,68f).Moreimportantly,suchastrongcommitmenttoonecoremethodologicalassumptioncausedcognitivemyopiaorbiasandafailuretospotanomalies.Thispaperarguesthatneo‐classicaleconomicssuffersthesameepistemologicalweaknessesgivenitsrelianceononesetofcoremicro‐foundationalassumptions.Especiallyinconditionsofuncertainty,aone‐paradigm‐fits‐allapproach–aninsistencethat

Page 5: Richard Bronk Epistemological difficulties with ...eprints.lse.ac.uk/39423/1/Epistemological_difficulties_with... · Epistemological difficulties with neoclassical economics

RichardBronkPresentedtoSEA,November2011

4

thereisultimatelyonecoresetofmicro‐foundationsrequiredtomodelandanalyseeconomicreality–entailsseriousepistemologicaldifficulties.Inthisregard,though,thepaperdepartsfromastrictlyKuhn‐basedapproach:itdoesnotenvisageorrecommendashifttoasinglenewparadigmdesignedtoresolveboththePtolemaic‐stylecomplexityofrecentneo‐classicalmodelsandthefailureofthoserelyingonthemtospotanomaliesthrownupbyinnovativeanduncertainsituations.Instead,itrecommendsanewpluralisminmodellingandabetterunderstandingoftheboundariesofapplicabilityofneo‐classicalmodels.2.Theneo­classicaltreatmentofinformationproblemsWhilstearly‘neo‐classical’economistsdidoftenassumeperfectinformationaswellasperfectrationality,themodernneo‐classicalpositionismorenuanced.Atoneextremeistherationalexpectationshypothesis(REH)anditscorollary,theefficientmarketshypothesis(EMH).TheREHholdsthattheexpectationsformedbyeconomicactorswillonaveragebecorrectsincecompetitivemarketpressureswillensurethatsystematicforecastingerrorsareeliminatedsothatallusefulinformationistakenintoaccount.Asaresult,expectationswillonlybefalsified‘throughrandom,unpredictableerrors’(Backhouse,2010,131).FollowingMuth(1961),thestrongerassumptionisoftenmadethatagents’expectationswillreflectthepredictionsgeneratedbythemodeloftheeconomy(assumedtobecorrect)implicitintheparticulartheorybeingtested–‘anassumptionthatthreatenstointroduceameasureofcircularityintothemicro‐foundationsofthoseverytheories’(Bronk,2009,220).TheREHimpliesinturntheEMH,whichpositsthatmarketpricesarereliableindicatorsoffundamentalvalueandavailableinformationandareonlysubjecttoessentiallyrandomvariation.Implicitinbothhypothesesistheassumptionthattheeconomicworldis‘ergodic’,meaningthat(pacerandomevents)‘thefutureismerelythestatisticalshadowofthepast’(Davidson,2010,17):asaresult,expectationsaboutthefuturecanusuallytaketheformofcalculatedprobabilities.TheREHandEMHlaybehindmuchoffinanceandappliedeconomictheoryintherunupto2007,butthefinancialcrisishasnotbeenkindtoeitherhypothesis.Marketexpectationsandpricingarenowshowntohavebeenmisguidedandevendelusionalforalongtime,whilstthemodelsproducedbyeconomistsandinternalisedoveralongperiodbymostriskmanagersandothereconomicactorsarerevealedtohavebeensystematicallywrong,causinghighlycorrelated‘errors’thatwereanythingbutrandom.Hereareanomaliesthateventhemostdie‐hardofneo‐classicaleconomistsstruggletoignore.Manyeconomistsincludedundermylabelof‘neo‐classical’haveformanyyearsacknowledgedthattheREHandEMHarenotplausibleasgeneralmodels.Instead,theyhaveassumedthattheoptimisationbyindividualsofconsistentlyrankedpreferences,andthetendencyofthemarkettoanefficientequilibrium,areoftenvitiatedbycertaininformationproblems–notablytheproblemofasymmetricinformation.Akerlof(1970),forexample,showedthat,whendealingwithuncertaintyaboutthequalityofproducts,anyasymmetryofinformationbetweenbuyersandsellerswouldleadtoamarketfor‘lemons’orthinmarkets,withthemarketpricetendingtobeunattractivelylowforsellersofreliable

Page 6: Richard Bronk Epistemological difficulties with ...eprints.lse.ac.uk/39423/1/Epistemological_difficulties_with... · Epistemological difficulties with neoclassical economics

RichardBronkPresentedtoSEA,November2011

5

products.AndStiglitz(2011,168)amongothershaslaidtheblamefortherecentpredictiveshortcomingsofeconomictheoryonafailureofmicro‐foundationstotakeaccountof‘informationasymmetriesandmarketimperfections’.AnotherlineofattackonREHandEMHfromwithintheeconomicsparadigmhasbeentofocusoncertainsystematiccognitiveorbehaviouralbiases–suchastheframingorformulationeffectsisolatedbyKahnemanandTversky(2000,9f),wherethewordingorreferencepointsusedtoframeoptionsbiasdecision‐making.Suchbiasesandheuristicsaremodelledasexceptionstorationalexpectationsandoptimisation.Thereisclearlymuchtothesearguments,buttheystillbeganumberofquestions:arethemostimportantinformationproblemsfacingmarketparticipantsthoseofasymmetricinformationratherthan‘symmetricignorance’?(Skidelsky,2009,45;Bronk,2011,8).Isitalwaysatleasttheoreticallypossibletohaveperfectmarketssolongaswehavetransparencyofinformationandmorecompleteriskmarkets?Andshouldrelianceonheuristicsandframingeffectsalwaysbeseenasadeparturefromfullyrationalchoice,andonethatcouldbecorrectedbyattentiontoovercomingthesebiases?Onlyiftheanswertothesequestionsis‘yes’isthedominant‘neo‐classical’strategyofbolting‐onamendmentstothecentralassumptionsofrationalchoiceandoptimisationvalid.Butasthispaperwillshow,theanswertoeachofthesequestionsmayneedtobe‘no’whenweconsidertheimpactofgenuineuncertainty.3.KnightandKeynesoncalculableriskversusuncertaintyOneofthestrikingfeaturesofmostmodernstandardeconomicsishowfaritignoresacentraldistinctionmadebyFrankKnight(1921,233)betweenmeasurable‘risk’(wherepossibleoutcomescanbeclassifiedingroupsandprobabilitiescalculated)andimmeasurable‘uncertainty’(wherenoprobabilitycanbecomputedbecause,forexample,thesituationisunique).ForKnight,uncertaintywasnotsomeregrettablemarketfailure(touseananachronisticexpression)butaninevitablepartofentrepreneurialactivity(ibid.,232);indeed,withoutuncertaintytherecouldbenoprofitsinacompetitivesystem,sinceprofitsthatarepredictable(orprobabilisticallyforecastable)wouldbequicklycompetedaway.1Knightunderscoreshisepistemologicalpositionasfollows:‘Profitarisesoutoftheinherent,absoluteunpredictabilityofthings,outofthesheerbrutefactthattheresultsofhumanactivitycannotbeanticipatedandthenonlyinsofarasevenaprobabilitycalculationinregardtothemisimpossibleandmeaningless.’(Knight,1921,311)Keynes,likewise,focusedontheproblemthatmarketvaluationsrelatingtoinvestmentsinthefuture‘cannotbeuniquelycorrect,sinceourexisting

1NotethatKnightishereusingtheassumptionofotherwiseperfectcompetitionconjoinedwiththefactofentrepreneurialprofitstoarguefortheexistenceofuncertainty.Aswemightputittoday,theexistenceofprofitsinafreemarketcanonlybeexplained–atleastintheabsenceofmonopolyrentsorasymmetriesofinformation–bythepresenceofuncertaintythatcannotbeturnedintomeasurablerisk.

Page 7: Richard Bronk Epistemological difficulties with ...eprints.lse.ac.uk/39423/1/Epistemological_difficulties_with... · Epistemological difficulties with neoclassical economics

RichardBronkPresentedtoSEA,November2011

6

knowledgedoesnotprovideasufficientbasisforacalculatedmathematicalexpectation’(Keynes,1936,152).ForKeynes,uncertainty(asopposedtocalculablerisk)wasthecentralproblemineconomicsandbusiness:‘Theoutstandingfactistheextremeprecariousnessofthebasisofknowledgeonwhichourestimatesofprospectiveyieldhavetobemade.Ourknowledgeofthefactorsthatwillgoverntheyieldofaninvestmentsomeyearshenceisusuallyveryslightandoftennegligible.’(Keynes,1936,149)Asaresult,Keynesarguedthatinvestorshavenochoicebuttorelyonconventions,theemotionsofconfidenceandfear,and‘animalspirits’.Thiswasnotbecauseofanyirrationalityoranydisinclinationtocalculatewherepossible,butbecauseuncertaintygivesinvestorsnochoicebuttofallbackon‘whimorsentiment’(ibid.,163).NotwithstandingKeynes’hugeinfluenceoneconomics,theKnight‐Keynesconceptofuncertaintyhasbeenmarginalisedwithinrecentneo‐classicaleconomicsandlargelyabsentfrommainstreamjournalarticlesforalongperiodafterthe1980s(Hodgson,2011a,161).Uncertaintyhasincreasinglybecomeconflatedwithprobabilisticandstandarddeviationmeasuresofrisk(Davidson,2003,231)–aconflationthatimpairsnotonlythemicro‐foundationsofstandardeconomicmodelsbutalsotheriskmodelsofbanks.Indeed,thewholediscourseof‘riskmanagement’frequentlyrepresentsconfusionbetweenthemanagementofactuarialriskandthemanagementofKnighteanuncertainty(Power,2006,26).Buttogaugehowdamagingthisconfusionandelisionbetweenriskanduncertaintyhasbeen,weneedtoexaminefurtherthenature,sourcesandprevalenceofimmeasurableuncertainty.AndtodothisitishelpfultoexploreadistinctionmadebySkidelsky(2009,88)andBronk(2011,8‐12)between‘epistemological’uncertainty,whererelevantprobabilitiesareinpracticeunknown,and‘ontological’uncertainty,wheretheyarelogicallyunknowable.4.Epistemologicaluncertainty:complexitynotcomputedTheepistemologicaluncertaintyfacingeconomicagentsrelatestothecomplexityoftheirpredicament,inanon‐technicalandtechnicalsenseoftheterm‘complexity’.Inthefirstinstance,itrelatestotheenormousvolumeofinformationbombardingeconomicagentsandthehugecostsintermsoftimeormoneythatwouldnecessarilybeinvolvedincollectingandassimilatingrelevantdataadequatelyforanoptimaldecisiontobemade.Modernfinancialmarketsprovideanumberofgoodexamplesofthis,notleasttheenormouscomplexityofcollateraliseddebtobligations(CDOs),wheretypicallyaninvestorwouldneedtomasterthecontentsofabillionpagesofprospectusinformationtounderstandasingleproduct’singredientsfully(Haldane,2009b,17).Suchvolumesofrelevantinformationhaverenderedthenotionofmarkettransparencylargelyacademic.Atthesametime,‘thestandardideathat(atworst)onepartytoanexchangemaybeataninformationdisadvantagetoanotherseemsfarlesspertinentthanthedangerofsymmetricalignoranceamongallplayers’(Bronk,2011,10).This

Page 8: Richard Bronk Epistemological difficulties with ...eprints.lse.ac.uk/39423/1/Epistemological_difficulties_with... · Epistemological difficulties with neoclassical economics

RichardBronkPresentedtoSEA,November2011

7

isespeciallysowhenthedifficultyoffathomingfinancialnetworkinterlinkagesmushroomsastheircomplexityreachesthresholdsofcomputationalintractability(Mirowski,2010,427f).Complexityeconomics,andcomplexitysciencemoregenerally,highlightsanotheraspectofepistemologicaluncertainty–thedifficultyofmakingexantespotpredictionsindynamicnon‐linearsystems.Using‘complexity’inamoretechnicalsense,economistsbasedattheSantaFeInstituteandelsewherehavepointedtotheprevalenceofincreasingreturnsandcriticalthresholdeffects,whichentailthatverysmalldifferencesininitialconditionscansnowballintoradicallydifferentlong‐termoutcomes(Arthur,1990;Waldrop1994;Fontana,2010).Whentheresultofsmallfirst‐moveradvantageintrade,orofinitiallyminorfinancialmishapsincomplexmarketnetworks,canchangethecourseofhistorythankstoincreasingreturnsandpositivefeedbackeffects,theattempttounderstandeconomicrealityresemblestheuncertainstudyofweathersystemsorfishstocksratherthanthemechanicaluseofpredictiveequilibriummodelsasfurnishedbystandardeconomics.Inthesecases,wehavetoacceptthepracticalimpossibilityofpreciseexanteprediction,evenifwecangetbetteratspottingemergingpatternsbyusingsimulationsofthresholdeffects.Epistemologicaluncertaintyisalsoafunctionofotherontologicalissues:2socialrealityismulti‐faceted–madeupofinterlockinglayersofphysicalconstraints,institutionalframeworks,individualcognitionandaction,socialmeaningandmacro‐levelemergence.Thisontologicallayeringposesconsiderableepistemologicalchallengesforeconomicactorsandsocialscientistsalike,especiallyasitiscompoundedbytheimpossibilityofencompassingallfacetsinoneconceptualormodellingframeworkwithoutdrainingthemofmostoftheirtruesignificance.Asweshallexplorelaterinthepaper,economicagentsandeconomistsbothhavetocontendwiththeneedtoexercisejudgementaboutwhichofmanypossiblemodelsortheoriestousewhentheonlycertaintyisthatnosingletheorycanencapsulateeverythingsignificantandrelevanttotheproblemathand.Butfirstweneedtoconsideranotherorderofuncertaintyentirely.5.Ontologicaluncertainty:thefutureisyettobecreatedOntologicaluncertainty‘impliestheimpossibilityofknowingeventhecategoriesandpossiblenatureofwhathasyettobecreatedoryettoevolve’(Bronk,2011,9),andisendemicinfast‐changinginnovativesituations,(Lane&Maxfield,2005,11).Bybreakingpredictablelinksbetweenthepastandthefutureanddisturbingpreviouslysystematicregularities,innovationsensurethattheworldisnon‐ergodic–morethanamerestatisticalshadowofthepast(North,2005,2Forclarity,itisimportanttonotethattherecanbeissuesofontology–likethemulti‐facetednatureofrealityandtheprevalenceofnon‐linearfunctions–thatimpactonthedegreeof‘epistemologicaluncertainty’,inthesenseofweakeningthebasisforprecisepredictionandthepossibilityinpracticeoffirmknowledge.Forthepurposesofthispaper,theterm‘ontologicaluncertainty’referstoindeterminacyatthelevelofrealityitselfandtothelogicalunknowabilityofthebasiccategoriesandentitiesoffuturerealityaheadofitscreation.

Page 9: Richard Bronk Epistemological difficulties with ...eprints.lse.ac.uk/39423/1/Epistemological_difficulties_with... · Epistemological difficulties with neoclassical economics

RichardBronkPresentedtoSEA,November2011

8

19f).Indeed,radicalinnovationsmayinjectsomethingentirelynovelintotheequationsandboundariesoflife,andsubvertthebest‐laidschemesofclassificationandprediction.Shacklewroteofour‘ownoriginal,ungovernednoveltiesofimagination…injecting,insomerespectexnihilo,theunforeknowablearrangementofelements’(Shackle,1979,52);andhespeltoutclearlytheontologicalimplicationsofthis:‘Thehistory­to­comewhichwillflowfrommen’sdecisionsisnon­existentuntilthosedecisionsthemselvesaremade.Whatdoesnotyetexistcannotnowbeknown.Thefutureisimaginedbyeachmanforhimselfandthisprocessoftheimaginationisavitalpartoftheprocessofdecision.Butitdoesnotmakethefutureknown.Theabsoluteandeternaldifferencebetweentherecordedpastandtheunformedfuture,despiteitsoverwhelmingsignificancefortheverystuffofhumanexistence,hasbeenoftenoverlookedinoureconomictheories’(Shackle,1992,3f).Shackleexaggeratesthelevelofuncertaintyweface.Thefutureispartlypredictablegivenourobservationofsociallyconstructedregularitiesinbehaviour,anditisinsomerespectsthedeterminateoutcomeofrationalchoiceamongstwell‐understoodconstraintsandoptions.Nevertheless,ontologicaluncertaintyremainsimportant.Whereverinnovationexistsandnovelpossibilitiesareimagined,expectationsandchoicecannotbepurelytheoutcomeofcalculation;andconsequentvagariesinthewayweimaginethefuture,anddegreesoffreedominthechoicesetsweface,introducefurtherindeterminacy.Crucially,thefirst‐orderuncertaintyimpliedbyinnovationandtheimaginativecreationofnovelpossibilitiesis‘compoundedbyuncertaintyaboutthesecond‐ordercreativereactionsofothers’(Bronk,2011,9).Atallpointsinthiscreativechainstretchingintothefuture,thedecisionsofactorscouldbedifferentwithoutanyirrationalitybeinginvolved,byvirtueoftheirimaginingtheoptionsdifferently.Furthermore,indeterminacyandontologicaluncertaintyisnotonlyaproductofthehumanimaginationandthefreedomtochoosebetweennewlyimaginedoptions.Itisalsoafeatureofmanycomplexself‐organisingsystems,whereovertimenon‐linearreactionstosmall(oftenrandom)mutationscanleadtotheemergenceofgenuinelynovelforms.Theimplicationsofontologicaluncertaintyforeconomicsareverysignificant.AsBuchananandVanberg(1991:inHausman,1994,323)note,onceweseethemarketasa‘creativeprocess’,itmakesnosensetoseethefutureasinsomesense‘outthere’waitingtobediscovered,nortousethelanguageoferrorforforecaststhatgetoverturnedasif‘correct’forecastswerepossible.Indeed,anappreciationofthewidespreadandrevolutionaryimpactofinnovationandimaginedfuturesis‘corrosiveofthestandardnotionthatforward‐lookingmarketvaluationscanbestableandefficientlypriced–thatthereisastaticreality‘outthere’onwhichrationalexpectationswillconvergeinresponsetocompetitivepressures’(Bronk,2011,9).TheideacentraltotheREH–thatsystematicerrorswillloseoutincompetitivemarketsandbeeliminatedsothatforecastsareonaveragecorrect–hasnovaliditywhentherealitytowhichtheexpectationspertainisstillasetofimaginedideasinpeople’sminds,andwhenwearelivinginanon‐ergodicworldwithfewsystematicregularities.Inother

Page 10: Richard Bronk Epistemological difficulties with ...eprints.lse.ac.uk/39423/1/Epistemological_difficulties_with... · Epistemological difficulties with neoclassical economics

RichardBronkPresentedtoSEA,November2011

9

words,theassumptionimplicitintheREHthatthereis‘anobjectiverealityindependentoftheobserverthatservesastheanchorofthecommonknowledgesharedbyalleconomicactors’isfalse(Nelson&Katzenstein,2010,51)whenevertherealityconcernedisafutureyettobeinvented.Norcantheproblemofinnovationbereducedtooneofasymmetricinformationbetweeninventorandotherplayers.Forinnovationsoncelaunchedtakeonformsthatexceedtheimaginedconceptionofeventheinventor,andthisontologicaluncertaintyiscompoundedforinventorandothersalikebytheunpredictableinnovativecountermovesofothermarketparticipants.3Ininnovativeenvironmentseveryoneinvolvedisgropinginthedark:thefutureissimplyunknowableandwemustadapttoitasbestwecanasitunfolds.6.Romanticpost­Kantianepistemology:implicationsforeconomistsTounderstandfullythelimitsofknowledgeineconomicsandthewaysinwhichsomeoftheselimitscanbeovercome,itishelpfultoconsidertheimplicationsofRomanticpost‐Kantianphilosophy.TheRomanticsunderstoodthatweneverhavesomeunmediatedaccesstofacts,andthatourbeliefsarenotsimplyreflectionsofareality‘outthere’.Rather,theworldweseeispartlyacreationoftheconceptualstructuresthatourmindssupply.IntheterminologyofWordsworth(1798),wehalf‐createtheworldwesee.Themindplaysanessentialandcreativeroleinallempiricalobservationbysupplyingaframeworkofinterpretation(Bronk,2009,258)–apointmadeclearbyColeridgeinthesewordsspokentoanaïveempiricistofhisacquaintance:‘Youmusthavealanterninyourhandtogivelight,otherwiseallthematerialsintheworldareuseless,foryoucannotfindthem,andifyoucould,youcouldnotarrangethem’(Coleridge,1830).ThisRomanticconceptionoftheconstructiveroleoftheoryandlanguageinourobservationandunderstandingoftheworldhasanumberofimportantimplications.Inparticular,whileatheoryormodelhelpsmakesenseofthechaosaroundusinthesamewaythatalamphelpsusseeinthedark,itisalsotruethatanytheoryormodel(likeanylamp)hasitslimitations.Ifweuseonlyonemodellingortheoreticalapproach–onesourceoflight–wewill‘keepstumblingonaspectsofrealitythatareoutsidetheareailluminated’byourtheoryormodel(Bronk,2010,103).Herethenisasourceofepistemologicaluncertainty:nosingletheoryormodelcanexplaineverything.AsMurdochwroteinherdescriptionofSartre’sphilosophy:‘Whatdoesexistisbruteandnameless,itescapesfromtheschemeofrelationsinwhichweimagineittoberigidlyenclosed,itescapesfromlanguageandscience,itismorethanandotherthanourdescriptionsofit’(Murdoch,1999,42).3Whencell‐phonemanufacturersfirstincludedthefunctionoftextingintheirrevolutionaryproducts,theycouldnothaveforeseenmanyofthenovelusesofthismediumthatwouldbemadebyconsumers,northewaythatitwouldhelptransformthenatureofsocialinteraction.IamgratefultoNicholasBarrforthisexampleofhowtheontologicaluncertaintyimpliedbyinnovationisdeepenedbythecreativecountermovesofotherplayers.

Page 11: Richard Bronk Epistemological difficulties with ...eprints.lse.ac.uk/39423/1/Epistemological_difficulties_with... · Epistemological difficulties with neoclassical economics

RichardBronkPresentedtoSEA,November2011

10

Ourhopeofcapturingthemultifacetednatureofrealityisespeciallypoorifweremainweddedtoonetheoreticalframework,oneconceptuallanguage–if,thatis,weeschewmodellingpluralism.Itisforthisreasonthateventhehardnaturalscienceofmodernphysicsnolongerreliessolelyononetheoreticalframework:itstillmakesuseofNewtonianmechanicswhererelevantalongsidemoremoderntheories;andthereistodatenounifiedtheorythatcanencompasstheadditionalrespectiveinsightsofrelativitytheory,quantummechanicsandchaosorcomplexitytheory.Andifthereisnounifiedtheoryinphysics,howmuchmoreforlornistheprojectoffindingsuchaunifiedmodellingframeworktodescribeandadequatelyexplainthemulti‐facetedsocio‐economicworldinwhichwelive?AsDiamondargues:‘Thecomplexityoftheeconomycallsfortheuseofmultiplemodelsthataddressdifferentaspects…[T]akingamodelliterallyisnottakingamodelseriously.Itisworthrememberingthatmodelsareincomplete–indeed,thatiswhatitmeanstobeamodel.’(Diamond,2011,1045f)Theinappropriatenessineconomicsofrelyingononesetofmodels(andtreatingthemasencapsulationsofthetruthratherthanasmerelythrowingapartiallightonaproblem)isnotonlytheresultofepistemologicalandontologicaluncertaintyandthecomplexityofsocialreality;itisalsoafunctionoftheneedforeconomiststointerpretapre‐interpretedworld.HereinliesthesecondimportantimplicationofRomanticphilosophyforeconomists:ourtheoreticalandconceptualframeworksstructureouractionsaswellasourbeliefs,andsohavethepowertotransformsocialreality(Bronk,2009,260).Inotherwords,thebehaviourthateconomistsstudyisalreadypartlyconstructedbythesociallyformednarrativesandeconomictheoriesthatindividualactorshaveinternalised.Thismeansthateconomistscannotfullyexplainorpredicteconomicandmarketbehaviourunlesstheyhavelearnedtoempathisewith(thebettertointerpret)thevariousmindsetsandconceptualstructuresthatinfluenceactors’beliefsandreasonsforaction.Atfirstsight,recognitionofthetransformativeimpactoftheoryonactionmightnotseemtopresenttoomuchdifficultyforneo‐classicaleconomists:indeed,theyaregenerallyonlytoohappytoadmitthattheirtheoriesworkbestwheretheyare‘performative’(Backhouse,2010,17)andhavebeenusedtoconstructmarketsthatresemblemoreclosely‘theidealworldoftheory’(Ibid.,76).Atthesametime,theREHmakesavirtueofassumingthatmarketparticipantsinternalisethe(onetrue)modeloftheeconomythathappenstobethesameasthatproposedbytheeconomistsmodellingthesituation(Cassidy,2009,99;Frydman&Goldberg,2011,57f).Itdoesnotrequiremuchanalysistounderstandthetheoreticalconstructionofagents’beliefsandactionsiftheyhavenochoiceincompetitivemarketsbuttointernalisethecorrecttheorytheeconomistconcernedhappenstobeworkingon.ThereisnoneedtounderstandwhytheFrenchortheSwedesmayinfactseetheworlddifferently,becausetheywillsoonhavetolearntheerroroftheirwaysorbeeliminatedincompetitionwiththosewhohaveinternalisedthecorrecteconomicmodel.

Page 12: Richard Bronk Epistemological difficulties with ...eprints.lse.ac.uk/39423/1/Epistemological_difficulties_with... · Epistemological difficulties with neoclassical economics

RichardBronkPresentedtoSEA,November2011

11

Itisonlynecessary,ofcourse,tospelloutthese(generallyimplicit)stepsinREH‐basedreasoningtoseethattheyarefallacious.Firstofall,aswehaveseen,thereis(inaworldofinnovationandontologicaluncertainty)nosingleefficientoutcomeoutthereinthefutureyettobecreatedonwhichrationalexpectationsmustconverge.And,secondly,notheorycanshinealightonmorethancertainaspectsofwhatisgoingon.Moreover,theperformativenatureoftheorypresentshugepitfallsfortheneo‐classicalscientificproject.Forwhileitistruethatawholesetofeconomicagentsinternalisingatheorysuchastheefficientmarketshypothesiswillmakemarketsconformmorecloselytothetheoryforatime(if,forexample,theyallstructureandpricefinancialproductsinaccordancewiththetheory),thisdoesnotpreventthemodelfailingwhenthesesameagentsbumpupagainstdynamicsnotcapturedinthemodel.Furthermore,earlyapparentconfirmationorplausibilityofatheoryresultingfromactorshavinginternalisedit,andsobehavingaccordingly,mayonlyservetoblindeconomiststothefailureofthemodeltocaptureothereffects.Itisnothealthyfortherigoroustestingofatheorytohaveittestedinconditionswherethesample’sbehaviourisstructuredinaccordancewiththemodel.Suchwasthefateuntilrecentlyofmostfinancetheorybasedontheefficientmarketshypothesis.7.ThelimitsofreasonItshouldnowbeclearhowdamagingtheprevalenceofuncertaintyisforthecorerationalistassumptionsofneo‐classicaleconomics.Lucasisoftenquotedforhisviewthat‘incasesofuncertainty,economicreasoningwillbeofnovalue’(Lucas,1977,15:quotedinFontana,2010,590;Hodgson,2011b,193).And,asdiscussedabove,therationalexpectationshypothesisisrenderedproblematiconceweacceptthateconomicagentscannot,exceptinrestrictedcircumstances,haveaccesstoonecorrectmodelthatcaptureseverythingsignificantabouttheirpredicament.Atthesametime,theREHnotionthatatleastsystematicerrorsinagents’forecastswillbeeradicatedbycompetitionlosesitsforcewhenweacceptthat,inconditionsofwidespreadinnovationandemergentnovelty,muchoffuturerealityisneitherpredeterminedatthemomentofdecisionnorsubjecttosystematicregularities.Moredamagingstill,inconditionsofepistemologicaluncertainty,themostbasicneo‐classicalassumption–thatrationaleconomicagentsseektooptimisetheirpreferenceswithingivenconstraints–becomesdubious.Forinmanycomplexsituations,wherethecostsofcollectingtheinformationnecessarytomakeavalidpredictionoranoptimalchoicearetoohigh,optimisationmaysimplynotbeafeasibleoreffectivestrategy.Thisisespeciallytruewhenthereissignificantlegitimatedoubtastowhichofseveralmodellingframeworksbestcapturestheessenceofthesituation–necessitatingtheneedtokeepseveralinterpretiveframeworksinplayandexercisejudgementbetweenthem.ThedictatesofeconomicRationalChoicemakesuperhumandemandsonthecomputationalandimaginativecapacitiesofindividualeconomicagents(Bronk,2009,216).Moreover,inmanycases,theseinformationproblemsaffectallmarketparticipantssymmetrically–makingitlessthanobvioushowevenintheorythe‘marketfailure’or‘imperfection’couldbesolved.Finally,whereitexists,

Page 13: Richard Bronk Epistemological difficulties with ...eprints.lse.ac.uk/39423/1/Epistemological_difficulties_with... · Epistemological difficulties with neoclassical economics

RichardBronkPresentedtoSEA,November2011

12

ontologicaluncertaintymakestheplightofthewould‐beoptimiseressentiallyhopeless:forthenshehasnowayofknowingforsureeventhenatureoftheoptionsshefaces,letalonethelikelyemergentoutcomesofthecreativeinteractionofherchoiceswiththoseofotherinnovativeagents.Eventheaxiomsofrationalitythatformthebedrockofneo‐classicaleconomicoptimisationmodelsmaybechallengedbyontologicaluncertainty:whyshouldtheconsistencyandtransitivityofpreferences(especiallyovertime)beprivilegedwhenourpreferencesrelatetoanuncertainandindeterminatefuture.Itmaynotbeworthplacingourpreferencesinrankorderatanymoment,letaloneassumingtheirconsistencyovertime,whentheoptionsopentousareinaconstantstateofflux.‘Thisrenderstheverynotionofoptimalchoiceinfuture‐directedactionsuspect;anditrendersforlornattemptstopredictbehaviourinuncertainconditionsonthebasisofindifferencecurvesconstructedfrompreferencesrevealedatonetime’(Bronk,2009,238).Uncertaintydoesnotimplythatthereisnoroleforreason,ofcourse.But,asweshallsee,itdoesimplythatreasonhastoworkhandinhandwiththeimagination–forexample,stress‐testingtheoptionsweimagine.Italsoimpliesthatwewilloftenhavetobecontenttomakejudgementsforreasonsthatareultimatelyinconclusivebecause,giventheindeterminacyofoptionsortheincommensurabilityofmodellingframeworks,thereisnoonerightanswertobehad.AsCassidy(2009,176)putsit:‘Rationalityisn’tabinaryvariable;itexistsonacontinuum’.Thereismuchreasonedthoughttobehadinthespacebetweenpuredeductivelogicandcompleteirrationality.Reasondoesnotimplycertainty.8.ThelimitsofpredictionAsaparadigm,neo‐classicaleconomicsisstronglycommittedto‘model‐basedprediction’and‘therhetoricofprediction’(Hodgson,2011a,164f).Oneofthediscipline’scanonicaltextsisFriedman’sfamousessayinwhichhearguedthatthetaskof‘positive’economicsis‘toprovideasystemofgeneralisationsthatcanbeusedtomakecorrectpredictions’,andthatitsstatusasasciencedependsonitsabilitytotestthesepredictionsfortheir‘precision,scope,andconformitywithexperience’(Friedman,1953:inHausman,1994,181).Itispartlythisconceptionofwhatbeingascienceentailsfortheirdisciplinethatleadsneo‐classicaleconomiststobesoweddedtomicro‐foundationsassumingthatactorsengageinrationaloptimisationwithingivenconstraintsonthebasisofrationalexpectations.Forthesemicro‐foundationsallowthemtomodelmarketsastendingtowardssomedeterminateandpredictableequilibriumoutcome(Bronk,2009,131).Thisprojectofhavingtestablepredictionsremainsfirmlyintactifeconomistsboltonamendmentstothesecoremicro‐foundationstomodelpredictableexceptionstotheruleasaresultofinformationasymmetriesorsystematicbehaviouralandcognitivebiases.Indeed,itremainsintactsolongastheyrefusetoadmittheimportanceofanyuncertaintythatcannotbeturnedintocalculableriskortreatedasanisolatedmarketimperfection;andsolongasanydeparturefromrationalchoiceisassumedtobeasystematicbiasthatcanbebuiltintothepredictivemodel.

Page 14: Richard Bronk Epistemological difficulties with ...eprints.lse.ac.uk/39423/1/Epistemological_difficulties_with... · Epistemological difficulties with neoclassical economics

RichardBronkPresentedtoSEA,November2011

13

Asthispapermakesclear,however,uncertaintyshouldnotbeignoredbyeconomicsunlessitwantstolimititselftothestudyofnarrowlyfocusedergodicsystems,wherethefutureisashadowofthepast.Ifeconomicsistobeusedtoexplainwhatisgoingonincomplexsystemsandinnovativemarkets,itwillhavetoacceptdifferentstandardsofproofthantherigoroustestingofexantespotpredictions(Bronk,2009,27).Indeed,theveryfailureofmostpredictionsintheareaofmacroeconomicsandfinancetheoryoverthelastdecadeshoulditselfpromptquestionsandrenderattractivethisdowngradingofthestatusofapparentlyprecisepredictivemodels.AsHodgson(2011b,191)putsit,the‘underlyingerrorliesinoverestimatingtheimportanceandpossibilityofprediction.’Nordoeseconomicsneedtobeembarrassedaboutanylossofscientificcredentialsifitreliesonmodelsthatexplainandsimulateratherthanpredictpreciseoutcomes.Afterallthequeenofmodernnaturalsciences–biology–rarelyattemptstopredictthefuturewithanyprecision,recognisingthecentralimportanceofrandommutations,thresholdeffectsandincreasingreturns.Andtherearemanywhowouldarguethatthetimehascomeforeconomicstotakemoreinspirationfrombiologicalmetaphors,and‘followbiologyinembracingcomplexityanddowngradingpredictioninfavouroftheprimarygoalofcausalexplanation’(Ibid.,192).Downgradingpredictiondoesnot,ofcourse,meangivingupanyhopeofusingeconomicsmodelstohelpuspeerintotheunknowablefuture.Justasthereisnobinarydistinctionbetweenreasonandirrationality,sothereisnobinarydistinctionbetweenpredictionandbeingcluelessaboutthefuture.Wecanandshouldusemodelsas‘diagnostic’tools(Sutton,2000,16)forteasingoutsuchsystematictendenciesasthereareintheapparentchaosofmarkets;andwecanusemodelstohelpusspotemergingpatternsandsimulatethepossibleimpactsofcrossingcertainthresholds.Butifwewanttouseeconomicsmodelsinthiswaytoenableustoreadthefutureasitunfolds(aswellasrefineourexplanationsofthepast)weneedaccesstonewmodels(alongsidethoseofneo‐classicaleconomics)–modelsthatmovebeyondtheparadigmof‘fullypredeterminedmodels’involvinga‘mechanisticnotionofrationality’(Frydman&Goldberg,2011,22,51).Complexityeconomicsisoneobvioussource,modellingmarketsasspontaneousself‐organisingsystemsattheedgeofchaos(Waldrop,1994),characterisedbypositivefeedbackeffectsandthresholdeffects.Butwealsoneedtodevelopanalternativesetofmicro‐foundationsthatisthecounterpartofseeingmarketsinthiswayandcanhelpexplainattheleveloftheindividualhowdecisionsaremadeinconditionsofuncertainty.AsBeckertputsit:‘Thegoalisnottodevelopamoreaccuratetheoryofpredictionbutratheratheoryoftheunpredictabilityoftheworldandofhowintentionalityunfoldsdespitethisunpredictabilityofoutcomes.’(Beckert,2011,24)9.Alternativemicro­foundations:imagination,narrativeandemotionInconditionsofuncertainty,theexpectationsonwhichdecisionsarebaseddependuponimaginationaswellasreason,aremediatedthroughnarrativesand

Page 15: Richard Bronk Epistemological difficulties with ...eprints.lse.ac.uk/39423/1/Epistemological_difficulties_with... · Epistemological difficulties with neoclassical economics

RichardBronkPresentedtoSEA,November2011

14

stories,andinvolvefeelingsandsentiments.Formicro‐foundationstoreflectthis,weneedtoreplacehomoeconomicuswithhomoromanticus(Bronk,2009).ItwasShacklewhofirstpointedoutthevitalroleplayedbytheimaginationinconditionsofeconomicuncertainty,arolehesummedupintheaphorism:‘Valuationisexpectationandexpectationisimagination’(Shackle,1992,8).Becausewehaveonly‘fragmentaryandconfusingevidence’ofwhattomorrowwillbring(ibid.,),wemustbuildupapictureofwhatiscomingwiththehelpoftheimagination.Inaneconomicsetting,choiceisrarelyamatterofchoosingbetweenwell‐definedoptions,butinsteadinvolvescreating,‘byconjectureandreasonedimaginationonthebasisofmeresuggestionsofferedbyvisibleorrecordedcircumstance,’apictureoftheoptionswefaceandthegoalsweaspireto(ibid.,96).Imaginationandcreativity,thatis,arenotonlyamajorcauseofontologicaluncertainty;theyarealsoourmaintoolsforcopingwiththatuncertainty.Imaginationfillsthevoidleftbytheindeterminacycreatedbyinnovationandthefreedomtochoosebetweennoveloptions;itsketchesoutvisionsofhowtheworldmightbeandhowwewouldlikeittobe.Itprovideswarningsandgoals;itconstructspossibleoptions.Andnoneofthisisantitheticaltoreason.Indeed,imaginationmustrelyonreasontostresstestitsvisionsofthepossiblefortheirlikelyfeasibilityanddesirabilityinthelightofpastexperience(Bronk,2009,221).Imaginationalsoplaysaroleinkeepingeconomicagentsopen‐mindedaboutthenatureoftheirpredicamentandcreativeinswitchingbetweendifferentcognitivespectacles.ForthepoetKeats(1817),theessenceoftheimaginativegeniuswasthequalityof‘negativecapability:thatis,whenmaniscapableofbeinginuncertainties,mysteries,doubts,withoutanyirritablereachingafterfactandreason,’of‘remainingcontentwithhalf‐knowledge’(Keats,1817).Andforentrepreneursandothereconomicactorstryingtocharttheirwaythroughtheunknownfuture,itislikewisehelpfultoremainalertandimaginativelyreceptivetopointersastheyemergeratherthanstraintopredicttheunknowablefutureorimposeastandardinterpretationontheincompleteevidencebeforethem.Butsuchpassiveimaginationisnotenough.Sincenosinglemodelormetaphorcancaptureallimportantfacetsofasituation,thesuccessfuleconomicagentisalsolikelytobetheoneadeptatexperimentingwithdifferentmetaphorsandperspectives,withdifferenttheoreticalframes.Itisfromthe‘generativefriction’ofusingmultipleperspectivesandevaluativeprinciplessidebysidethatwederivebothourcapacityforinnovativethinkingandusefulhintsabouthow‘tonavigatethroughunchartedterritory’(Stark,2009,xvi,16‐19).Afullunderstandingofthefrequentlyignoredroleoftheimaginationineconomicdecision‐makingandcognitionisahelpfulantidotetonarrowanddeterministicmodelsofrationalityandrationalexpectations.But,initself,itmaynothelpmuchinthecreationofnewanalyticaltechniquesandmodelsforeconomistsnorexplainwhyeconomicactorsmakeaparticulardecision.Forthesepurposes,itishelpfultofocusonthewayinwhichimaginationfindsmoredeterminateexpression,namelythroughthecreationofnewnarrativesandthroughtheuseandrecombinationofexistingnarratives.

Page 16: Richard Bronk Epistemological difficulties with ...eprints.lse.ac.uk/39423/1/Epistemological_difficulties_with... · Epistemological difficulties with neoclassical economics

RichardBronkPresentedtoSEA,November2011

15

Ithaslongbeenrecognisedoutsideeconomicsthatnarrativeis‘oneoftheimportantdeviceshumansusetogivemeaningtolife’sactivities,’and‘tocreatethecommitmenttoact’(Tuckett,2011,xvii).Weareespeciallyrelianton‘narrativestructure’,whenweneedto‘keepontologicaluncertaintyatbay’(Lane&Maxfield,2005,4).Facedwithanuncertainfuture,wegenerallychoosetocharacteriseitaccordingtooneofanumberofpossiblenarratives;andthenarrativeorscriptchosenprovidesuswithanunderstandingoftheroleweshouldplay(ibid.,).Theseaction‐guidingnarrativesareoftensuppleinthesensethatweengagein‘narrativeshifting’(ibid.,16f)whenwebumpupagainstaspectsofemergentrealitysignificantlyatoddswiththestorywehavetoldourselves.Thenarrativesusedbyeconomicagentsareinessenceimaginative‘fictions’,butthey‘provideparametersfordecision‐makingandtherebyprovideorientationdespitetheuncertaintyinherentinthesituation’(Beckert,2011,5).Andcruciallyitis‘throughtheirstory‐structurethatimaginingsoffuturestatesbecomedeterminate’(ibid.,7).Narrativesarenot,ofcourse,merelyaproductoftheindividualimagination.Mostofthenarrativesweusetostructureourbeliefsaboutthefutureandgiveusreasonsforactionaresociallyconstructed–theproductofconventionalnarratives.Andinconditionsofuncertainty,‘conventionalexpectationsareperfectlyrational’(Skidelsky,2009,93).Conventionalnarrativesorscriptsarealsooftenconsciouslycraftedbyeconomicactorssuchascentralbankswishingtochannelexpectationsinacertaindirection(Beckert,2011,8),orbyadvertisingcompanieswishingtocreatetheneedforaproductbyembeddingitinameaningfulnarrative.Beforeexploringmorefullytheimplicationsoftheroleofnarrativeforbotheconomictheoryandthewayeconomicanalysisiscarriedout(seesections10and11),itishelpfultoexaminealsotheroleofemotions.Amongeconomists,itwasKeyneswhomostfamouslyfocusedattentionontheimportanceinconditionsofuncertaintyofemotionsandthesentimentsofconfidenceandfear‘wherenosolidbasisexistsforareasonablecalculation’(Keynes,1936,154).Inrecentyearstherehasbeenarenewedupsurgeofinterestinhowemotionscolourourjudgmentsandevenhelpusmakechoicesinconditionsofuncertaintybyactingasan‘EmotionalPositioningSystem’(Brooks,2011,21).Tuckett(2011,18f)advocatesan‘emotionalfinanceapproach’thatanalyseshowemotionalstatesalterthewaysinwhichweprocess‘ambiguousinformationaboutanuncertainfuture’;andheisatpainstopointoutthatemotionsdonotalwaysmislead.Instead,emotionsandgutfeelingsmay‘quicklyandefficientlyfacilitategooddecision‐makinginuncertainbuturgentsituations’,makingitinappropriatetoseetheroleofemotionsasabiasordeparturefromrationalityasitisnormallyportrayedbybehaviouraleconomics(ibid.,13f).Perhapsthemostimportantaspectofemotionsfromthepointofviewofmodellingmicro‐foundationsishowtheyinteractwithnarrativesandtheproductsofeconomicagents’imaginations.Emotionsareboththevectorsandproductsofnarratives:confidenceandfearcanmakeeconomicagentsmorereceptivetoparticularnarratives,andthenarrativesmayinturninstilinagentshigherorlowerlevelsofconfidenceorfear.Moreover,emotions–andthestorieswithwhichtheyareintertwined(Tuckett,2011,1)–aresociallycontagious;theyspread‘likeviruses’andcanbemodelledwiththehelpof

Page 17: Richard Bronk Epistemological difficulties with ...eprints.lse.ac.uk/39423/1/Epistemological_difficulties_with... · Epistemological difficulties with neoclassical economics

RichardBronkPresentedtoSEA,November2011

16

modelsborrowedfromepidemiology(Akerlof&Shiller,2009,56).Narrativescanalsobeacceptedandretainedpasttheirsell‐bydateforemotionalratherthanlogicalreasons.The‘grandnarrativeofriskmanagement’,forexample,hasbeensobeguilingbecauseofthe‘illusionofcontrol’itfostersformanagement(Power,2007,viii,98,);andmanyapparentlyscientificnarrativescanactasatalismanagainstinsecurity,atranquilizeragainstanxiety(Beckert,2011,22).Emotionsalsoattachtotheidealisedproductsoftheimagination:thepleasurewederivefromday‐dreamingaboutlifewithanewcaroranotherholidaymotivatesustopurchasethemnow–adynamicwellunderstoodbytheadvertisingindustry(Campbell,1987;Bronk,2009,213,250).Indeed,emotionandimaginationtogetherareessentialeventotheindividualpursuitofself‐interest.AsHazlittarguedtwocenturiesago,weimaginetheinterestthatwewouldfeelintheimaginedoutcomeofactionwearecontemplatingtodayandthisimaginedinterestexcitesinusanimmediate‘emotionofinterest’sufficienttopropelustoact(Hazlitt,1805;Bronk,2009,200).Lookedatthisway,eventhequintessentiallyeconomicactofpursuingourownself‐interestbecomesanimaginativeandemotionalenterprise.Itisemotionsattachingtobothpublicnarrativesandprivatevisionsthatmotivateusquiteasmuchasrationalcalculationoflikelyconsequences.10.Whatdopricestellus?Theefficientmarketshypothesisassumesawayepistemicuncertaintyandpositsthatmarketpricesreflectallavailableinformationandfundamentalvalues(Cassidy,2009,86f;Frydman&Goldberg,2011,81f).Itignoresthefactthatinacomplexworldagentsindividuallyandcollectivelycometorelycognitivelyandemotionallyoncertainheuristics,modelsandnarrativesthatmakesomesenseoftheirpredicament,butnecessarilytellonlypartofthestory.Morecrucially,itignoresthecentraltruththatinaworldofontologicaluncertainty,themarketcannotform‘the“best”possibleestimate’ofvalue,‘becausethis“best’estimatedoesnotexist’(Orlean,2008,15):wecannotrelyonmarketcompetitiontoensurethatpricesareonaveragecorrectbyeliminatingsystematic‘errors’inforecasting,becausethereisnoobjectiverealityinthefutureyettobecreatedagainstwhichtojudgeestimatesatthetimetheyaremade.Butthisstillbegsthequestionofwhatmarketpricesdotellusandhowmuchtheycanhelpinsolvingtheepistemologicalproblemsfacedbyeconomicactors.Toanswerthisquestion,itisusefultostartbyrecallingtheinsightsofHayek.Hayekdidnotbelievethatthemarketeverreachedanoptimalequilibrium(Gamble,1996,69),andhewouldhavebeenimpatientwiththeclaimsoftheefficientmarketshypothesis.Buthedidbelievethatthepricesysteminafreemarketisanunrivalled‘marvel’initsfunctionasacoordinationmechanismforcommunicatingdecentralizedinformation(Hayek,1945,526f).Hayekarguedthatitwasthisroleofmarketpricesinhelpingeachofusdiscovermoreaboutouruncertainpredicamentthatmadeitsosuperiortosocialistattemptstocalculateourneeds.Forhim,thecentralproblemwasoneofwhatIcallepistemologicaluncertainty.Ashewrote:

Page 18: Richard Bronk Epistemological difficulties with ...eprints.lse.ac.uk/39423/1/Epistemological_difficulties_with... · Epistemological difficulties with neoclassical economics

RichardBronkPresentedtoSEA,November2011

17

‘…theknowledgeofthecircumstancesofwhichwemustmakeuseneverexistsinconcentratedorintegratedform,butsolelyasthedispersedbitsofincompleteandfrequentlycontradictoryknowledgewhichalltheseparateindividualspossess.’(Hayek,1945,519)Hayek’sargumentwasthatthesignalsgivenbymarketpricescommunicatesummaryinformationbasedontheinnumerabledecisionsofindividualsandsoreflectkeyattributesofthedispersedandcontextualizedknowledgefrequentlyavailableonlytothoseindividuals.ForHayekthemarketwasa‘discoveryprocedure’(Hayek,1978)forsolvingtheproblemofknowledge;anditwastheverypluralismofindividualperspectivesthathelpedgivethesystemepistemologicalrobustnessandimportance.FewdoubtthatHayekwasarticulatingaveryimportanttruthhere,onethatwasunderlinedbytherepeatedfailureofcommunisteconomiestofindsolutionstoconsumerandcorporateneedswithoutthehelpofafunctioningpricesystem.Butrecenteventsinworldmarketshavealsoconfirmedthatitisonlyapartialtruth.Since2007,wehavecometoacceptthatmarketpricescanoftenbeprofoundlymisleadingandtheirmovementspositivelydestabilising.Thispaperhelpstounderstandwhy.ItbuildsonKeynes’centralinsightthat,inconditionsofuncertainty(wherenobasisexistsfor‘calculatedmathematicalexpectation’),investorsfrequentlyfallbackonaconventionalapproachtovaluations,orupon‘whimorsentiment’(Keynes,1936,152,163).Marketpricesdonotsimplyreflectfundamentalsasperceivedbydispersedagents;theyalsoreflectdominantnarratives,groupemotionsofconfidenceandfear,andstoriesofnewpossibility.Andhereinliesthepotentialforpricestobeprofoundlymisleading.Asthispaperhasargued,thereisnocorrectversionofthefuture,sinceitisstilltobedeterminedbyinnovationsnotyetconceivedandchoicesnotyetmade;and,inthisspaceofpossibilities,currentpricescanonlyreflectourbestguesses,preferrednarratives,andfleetingsenseofoptimismorpessimism.Solongasmarketpricesarethesummationofheterogeneousperspectivesandnarratives,variedemotionsanddifferentdreams,thentheymayatleasthelpusworkoutemergingpatternsintheunderlyingeconomy,sincethemultipleperspectivesandintuitionsreflectedarebetterthanone.ButalltoooftenmarketpricesdonotreflectthedecentralizedknowledgethatHayekassumedbutahomogeneousnarrativeandasharedemotionalstate.Marketsarenotinhabitedbythelonerangersassumedbythemethodologicalindividualismthatiscentraltoneo‐classicaleconomics,butratherbysocialanimalswhoprefertoshareadominantscript,acollectiveimageofthefuture.Andthiscollectivecognition,homogeneityofbelieforconvention,andtendencyto‘groupfeel’,canthreatenthesmoothworkingsofmarkets(Tuckett,2011,19)bydrivingmarketpricessofarinonedirectionthattheyfailtoreflectignoredelementsofemergingfundamentals.Since,inconditionsofontologicaluncertainty,wecannotassume–astheREHdoes–convergenceononetruemodel(becausethereisasyetnotruthaboutthefuturetoactasananchorforexpectations),convergenceonadominantmodelmaysimplyimplyagatheringcollectivedelusion,andtherebysowtheseedsforabubbleandsubsequentcrash.Thisisoneofseveralwaysinwhichtheepistemologicalargumentsinthispapercanaddtoourunderstandingofwhy

Page 19: Richard Bronk Epistemological difficulties with ...eprints.lse.ac.uk/39423/1/Epistemological_difficulties_with... · Epistemological difficulties with neoclassical economics

RichardBronkPresentedtoSEA,November2011

18

marketsarepronetobubblesandsubsequentcrashes–aphenomenonlargelyignoredbyneo‐classicaltheory(Stiglitz,2011).11.Monocultures,bubblesandtheneedfordiscourseanalysisPriortoabubble,thereisusuallyaspateofinnovationthatintroducesontologicalaswellasepistemologicaluncertainty.Intheresultingperiodofuncertainty,investorsgenerallyclingtoamixtureofconventionalexpectationsandconfidence‐inspiring‘newerastories’(Akerlof&Shiller,2009,55).Theseconventionsandnewerastorieshavetwoeffects.Thefirstisaphenomenonreferredtobythelegendaryinvestor,Soros,as‘reflexivity’:thepartialanddistortednarrative(ornewerastory)isreflectedininvestorbehaviourandthereforeinmarketprices,andthesepricemovementsinturnservetoreconfirmtheincreasinglydominantnarrativein‘reflexivefeedbackloops’thatcanleadmarketswayoutoflinewithfundamentals(Soros,2010,14‐16).Thesecondeffectcomplementsthefirst:ifaconventionorstorybecomesdominantenough,itleadstoananalyticalmonoculture,whichensuresthattheinevitablepartialdistortionimpliedbyrelianceonthisonenarrativeormodelleadstoageneralizedcognitivemyopiaandadestablisinghomogeneityofbehaviour(Bronk,2011,15).Theresultisnotonlyblindnesstotheunexpectedbutveryhighcorrelationsinbehaviour.Pricesaredriveninonedirectionforalongtime,untilelementsofbruterealityignoredbythedominantnarrativeormodelbecomesosalientthattheycallthenarrativeintoquestion.Atthatpoint,acrashisonlyamatteroftime.Therecentfinancialcrisisprovidesagoodexampleofalltheseaspects.Inthedecadeleadinguptoit,therewasawholespateoffinancialproductinnovations,relatingtothesecuritizationandrepackagingofmortgagesinevermoreexoticderivativeinstruments.Theseinnovationschangedmarketdynamicsforeverandupendedpreviouslysystematicregularitiesofbehaviour.Marketparticipants,though,initiallyhadnoproblemcopingwiththeresultantuncertainty,thankstorelianceonconventionandnewerastories,whichinthiscasewerecloselylinked:thedominantmarketandregulatoryconventionbecamethatuncertaintycould(contrarytotheforgottenteachingsofKnightandKeynes)beturnedintocalculablerisk(Skidelsky,2009,87),eveninthefaceoframpantinnovation.Indeed,itwasassumedthatfinancialmarketshadentered‘anewera’of‘simultaneouslyhigherreturnandlowerrisk’,resultingfrom‘ashiftinthetechnologicalfrontierofriskmanagement’(Haldane,2009a,4).Inthegoodyearsbeforethis‘newera’conventionwasproventobecatastrophicallymisguided,ithadtworelatedeffects.First,itinfluencedmarketbehaviourandcausedfinancialgroupstotakepositionsthattheygenuinelybelievedtobe‘riskcontrolled’butinfactrepresentedhugespeculativeuncertainty.And,secondly,thegrowingmonocultureofriskmanagementcausedblindnesstotheproblemsunfoldinganddangerouslyhighcorrelationsinmarketbehaviour.Witheveryoneinthemarketpursuingasimilarstrategyandusingessentiallythesameriskmodels,pricesrespondedaccordingly,apparentlyconfirmingthestrategyanddominantneweranarrative.Buttheseedsofdisasterwerethere,notleastbecauseonefactormissingfromtheriskmodelswasthedestabilisingriseincorrelationscausedbytheinternalisationbyallplayersofthesamebusiness

Page 20: Richard Bronk Epistemological difficulties with ...eprints.lse.ac.uk/39423/1/Epistemological_difficulties_with... · Epistemological difficulties with neoclassical economics

RichardBronkPresentedtoSEA,November2011

19

strategiesandthesamenew‐erariskmodels(Bronk,2011,15).Therestishistory.Theimplicationforthedisciplineofeconomicsisclear:economists(likeriskofficers)needtopaycloseattentiontothenarrativesthatmarketparticipantsareusing.Toexplainthepast,orspotwhatiscurrentlyunfolding,economistsshouldusediscourseanalysistocapturethemodelingframeworksandnarrativesthatstructurethebeliefsandactionsofeconomicagents.Economistshavenochoicebuttointerpretapre‐interpretedworld.Butthisisnottheendoftheirtask.Itisalsotheircivicaswellasacademicdutytobewillingtoemploydifferentmodelsfromthoseinternalisedbytheactorsthemselvesandsobeabletospotanomaliestheactorshavemissed.Economistsshouldconstantlywarnagainstthelazydogmaof‘bestpractice’(or‘bestmodel’)thatimpliesonerightwayoflookingattheworld.Theymustremember,andremindtherestofus,thatinaworldofKnightianuncertainty,wecannotknowexantewhatbestpracticeorthebestmodelwillbe(ibid.,17).12.Thedangersofbolt­onamendmentsandthecaseforpluralisminmodellingManyeconomistshavegrappledpiecemealwithsomeoftheissuesraisedinthispaper.Thedominantresearchstrategy,however,hasbeentoboltonamendmentstothecorerationalistmicro‐foundationsofneo‐classicaleconomics–inthehopeofcorrectingformanageableinformationproblemsandsystematiccognitivebiaseswithoutlosingthepredictivepowerofthemodelsinquestion.Butasthispapersuggests,thisbolt‐onstrategyhasanumberofproblemsifintendedasabidtocountertheuncertaintyfacingbotheconomicagentsandeconomists.Bywayofconclusion,itishelpfultorehearsethemostimportantoftheseproblems.Inthefirstplace,itisfarfromclearthatthecentralassumptionsofneo‐classicaleconomics–individualsrationallyoptimisingconsistentlyrankedpreferenceswithingivenconstraintsonthebasisofrationalexpectations–haveanyexplanatoryroleinconditionsofeitherontologicalordeepandsymmetricalepistemologicaluncertainty.Asaresult,whendealingwithproblemscharacterisedbyuncertainty(asopposedtocalculablerisk),itmakeslittlesensetotreatdeparturesfromrationalchoiceassumptionsasbiasesorerrorsthatcanbesystematicallymodelledandboltedontorationalistmicro‐foundations.For,aneconomicagentfacinguncertaintyasopposedtoriskhasnochoicebuttorelyonimagination,emotionsandnarrativestomakesenseofherpredicament.Anyattempttooptimisehersituationisdoomedtofailure.Asecondandrelatedpointisthatbolt‐onamendmentscanleadtologicalincoherenceattheheartofmodels.AgoodexampleofthisisEndogenousGrowthTheory,whichinastylisedfashionbuildsintoitsmodelscreativedestructionandincreasingreturnstoresearchanddevelopment,butstillreliesonmicro‐foundationsthatassumethateconomicagentsaredrivenbyrationalprobability‐calculatingexpectationsandthegoalofmaximisingtheirconsumptionorprofitpotential(Bronk,2009,80).Giventheontological

Page 21: Richard Bronk Epistemological difficulties with ...eprints.lse.ac.uk/39423/1/Epistemological_difficulties_with... · Epistemological difficulties with neoclassical economics

RichardBronkPresentedtoSEA,November2011

20

uncertaintyengenderedbyinnovation,itishardtoseehowtheagentsconcernedaresupposedtogoaboutmaximisingconsumptionorprofitsonthebasisofprobabilitycalculations.Likewise,asFontana(2010,585,593f)arguesmoregenerally,theneo‐classicalparadigmcannotstretchtoaccountingforcomplexeconomicphenomena,andtheuncertaintycentraltothem,without‘losingitsanalyticalandtheoreticalcoherence.’Thebolt‐onprojectalsobetraysaconvictionthatitispossibletohaveonebestmodellingframework–theproductofcontinuouslinearprogressinknowledgeandtheslowaccretionofwisdom.But,aswehaveseen,itisimpossibletoreduceallthemulti‐facetedaspectsofrealitytooneperspectivewithoutlosingmuchofitssignificanceandtexture.Moreover,inanindeterminateworldofconstantinnovationandtheemergenceofgenuinenovelty,therecanbenosystematiceliminationoferrorsandconvergenceontruth(Budzinsky,2007,315).Givenouruncertainpredicament,wecanneverknowexantewhatthebestmodelorparadigmwillbe.Thebolt‐onamendmentsoftoday–evenassumingtheyworkedtopatchupthepredictivepowerofourmodelsfornow–maystillleaveusunabletospotandexplaintheproblemsoftomorrow.Facedwithpermanentuncertainty,itmakessensetohaveanumberofanalyticaltoolkitsathand.Forthisreason,itmayactuallypaytoreverseengineerthedisciplineofeconomicsmoreoftenandengagewiththeresearchquestionsandmethodsofeconomistsinthepast.Wehavemuchtolearn,forexample,fromKnight,KeynesandHayekabouttheimplicationsofuncertainty–implicationsthatwereperhapsmoreobviousintheturbulent1920s,1930sand1940sthanintherelativeeconomiccalmofthelatetwentiethcentury.Thehistoryofeconomicsprovidesaready‐madesourceofsuchalternativeperspectivesandmodellingapproaches.Ratherthanconstanttheorymending,economistswouldbebetteradvisedtoembraceatleastameasureofpluralisminmodellingandtovarythecognitivespectacleswithwhichtheyanalysesocialreality.Thiswillnotonlyincreasetheirchancesofmakingsenseofthemulti‐facetednatureoftoday’ssocialreality,butalsohelpthemspotthenovelproblemsoftomorrow.Moreover,sucheclecticismcanandshouldbedisciplinedintwoimportantways:itshouldseektomatchtheoriestoproblemsforwhichtheyaresuited;anditshouldfocusonestablishingclearerboundariesofapplicabilityforthemodelsavailable(Bronk,2009,296).Neo‐classicalmodelswillgainanewleaseoflifeifweuselogicandexperiencetolearnwheretheyarelikelytobeanalyticallyfruitfulandwheretheyarenot.Inconditionsofuncertainty,innovation,andincreasingreturns,theyhavelittletooffer;butelsewheretheyhaveprovedthemselvestobeamongthegreatestintellectualachievementsofsocialscience.Epistemologicalandontologicaluncertaintydonotimplytheneedforaparadigmshiftawayfromneo‐classicalmicro‐foundationstoanothermodellingmonoculture,butratheranewrespectfordisciplinedeclecticismandtheuseofseveralparadigmssidebyside.

Page 22: Richard Bronk Epistemological difficulties with ...eprints.lse.ac.uk/39423/1/Epistemological_difficulties_with... · Epistemological difficulties with neoclassical economics

RichardBronkPresentedtoSEA,November2011

21

ReferencesAkerlof,GeorgeA.(1970),‘TheMarketfor“Lemons”:QualityUncertaintyandthe

MarketMechanism’,TheQuarterlyJournalofEconomics,84(3),488‐500Akerlof,GeorgeA.&Shiller,RobertJ.(2009),AnimalSpirits–HowHumanPsychology

DrivestheEconomyandWhyitMatters,PrincetonUniversityPressArthur,W.Brian(1990),‘PositiveFeedbacksintheEconomy’,ScientificAmerican,

February,80‐85.Backhouse,RogerE.(2010),ThePuzzleofModernEconomics–ScienceorIdeology,

CambridgeUniversityPressBeckert,Jens(2011),‘ImaginedFutures:FictionalityinEconomicAction’,Max‐Planck‐

InstitutWorkingPaper11/8Bronk,Richard(2009),TheRomanticEconomist–ImaginationinEconomics,

CambridgeUniversityPress– (2010),‘ModelsandMetaphors’,inTheEconomicCrisisandtheStateofEconomics,

ed.R.Skidelsky&C.WesterlindWigstrom,PalgraveMacmillan,101‐110– (2011),‘Uncertainty,ModellingMonoculturesandtheFinancialCrisis’,The

BusinessEconomist42(2),5‐18Brooks,David(2011),TheSocialAnimal,RandomHouseBudzinsky,Oliver(2008),‘Monocultureversusdiversityincompetitioneconomics’,

CambridgeJournalofEconomics,32,295‐324Buchanan,JamesM.andVanberg,ViktorJ.(1991),‘TheMarketasaCreativeProcess’,

EconomicsandPhilosophy,Vol.7;reproducedinThePhilosophyofEconomics:AnAnthology,2ndedition,ed.DanielM.Hausman,CambridgeUniversityPress,315‐335

Campbell,Colin(1987),TheRomanticEthicandtheSpiritofModernConsumerism,Blackwell

Cassidy,John(2009),HowMarketsFail,AllenLaneColander,David(2001),TheLostArtofEconomics–EssaysonEconomicsand

theEconomicsProfession,EdwardElgarColeridge,SamuelTaylor(1830),TableTalk,21September,reprintedinSamuelTaylor

Coleridge,ed.H.J.Jackson,OxfordUniversityPress,1985,596Davidson,Paul(2003),‘UncertaintiesandGovernmentPolicies’,inThephilosophyof

Keynes’seconomics–probability,uncertaintyandconvention,ed.JRunde&SMizuhara,Routledge

–(2010),‘RiskandUncertainty’,inTheEconomicCrisisandtheStateofEconomics,ed.R.Skidelsky&C.WesterlindWigstrom,PalgraveMacmillan,13f

Diamond,Peter(2011),‘Unemployment,Vacancies,Wages’,AmericanEconomicReview,101,1045‐1072

Fontana,Magda(2010),‘Canneo‐classicaleconomicshandlecomplexity?Thefallacyoftheoilspotdynamic’,JournalofEconomicBehavior&Organization76,584‐596

Friedman,Milton(1953),‘TheMethodologyofPositiveEconomics’,fromEssaysinPositiveEconomics,UniversityofChicagoPress;reproducedinThePhilosophyofEconomics–AnAnthology,2ndedition,ed.DanielM.Hausman,CambridgeUniversityPress,1994,180‐213

Frydman,Roman&Goldberg,Michael(2011),BeyondMechanicalMarkets,PrincetonUniversityPress

Gamble,Andrew(1996),Hayek–TheIronCageofLiberty,PolityPressHaldane,AndrewG.(2009a),‘WhyBanksFailedtheStressTest’,BankofEngland– (2009b),‘RethinkingtheFinancialNetwork’,BankofEngland

Hausman,DanielM.(1994),ThePhilosophyofEconomics–AnAnthology,2ndedition,CambridgeUniversityPress

Page 23: Richard Bronk Epistemological difficulties with ...eprints.lse.ac.uk/39423/1/Epistemological_difficulties_with... · Epistemological difficulties with neoclassical economics

RichardBronkPresentedtoSEA,November2011

22

Hayek,F.A(1945),‘TheUsesofKnowledgeinSociety’,AmericanEconomicReview,35(4)

– (1978)‘CompetitionasaDiscoveryProcedure’inNewStudiesinPhilosophy,Politics,EconomicsandtheHistoryofIdeas,byF.A.Hayek,UniversityofChicago

PressHazlitt,William(1805),‘AnEssayonthePrinciplesofHumanAction’inTheSelected

WritingsofWilliamHazlitt,ed.DuncanWu,Vol.1,Pickering&Chatto,3‐82Hodgson,Geoffrey(2011a),‘TheEclipseoftheUncertaintyConceptinMainstream

Economics’,JournalofEconomicIssues,45(1),159‐175– (2011b),‘ReformingEconomicsaftertheFinancialCrisis’,GlobalPolicy,2(2),190‐

195Kahneman,Daniel&Tversky,Amos(2000),Choices,Values,andFrames,Cambridge

UniversityPressKeats,John(1817),LettertoGeorgeandTomKeats,21December,extractreprintedin

Romanticism–AnAnthology,2ndedition,ed.DuncanWu,Blackwell,1998,1019Keynes,JohnMaynard(1936),TheGeneralTheoryofEmployment,InterestandMoney,

MacmillanKnight,FrankH.(1921),Risk,UncertaintyandProfit,HoughtonMifflinKuhn,ThomasS.(1996),TheStructureofScientificRevolutions,3rdedition,The

UniversityofChicagoPressLane,DavidA.andMaxfieldRobertR.(2005),‘Ontologicaluncertaintyandinnovation’,

JournalofEvolutionaryEconomics,15(1),3‐50Mirowski,Philip(1989),MoreHeatthanLight–EconomicsasSocialPhysics,Physicsas

Nature’sEconomics,CambridgeUniversityPress– (2010)‘InherentVice:Minsky,Markomata,andthetendencyofmarketsto

underminethemselves’,JournalofInstitutionalEconomics,6(4),415‐443Murdoch,Iris(1999),Sartre:RomanticRationalist,VintageMuth,JohnF.(1961),‘RationalExpectationsandtheTheoryofPriceMovements’,

Econometrica29(3),315‐35Nelson,Stephen&Katzenstein,PeterJ.(2010),‘UncertaintyandRiskandtheCrisisof

2008’,Paperpreparedforthe“PoliticsintheNewHardTimes”workshophonoringPeterGourevitch,UniversityofCalifornia‐SanDiego,April23‐24

North,DouglassC.(2005),UnderstandingtheProcessofEconomicChange,PrincetonUniversityPress

Orlean,Andre(2008),‘KnowledgeinFinance:ObjectiveValueversusConvention’,Unpublished–onlinepdfonly,Paris

Power,Michael(2007),OrganisedUncertainty–DesigningaWorldofRiskManagement,OxfordUniversityPress

Shackle,G.L.S.(1979),ImaginationandtheNatureofChoice,EdinburghUniversityPress– (1992),EpistemicsandEconomics–ACritiqueofEconomicDoctrines,Transaction

PublishersSkidelsky,Robert(2009),Keynes–TheReturnoftheMaster,AllenLaneSoros,George(2010),TheSorosLectures,PublicAffairsTMStark,David(2009),TheSenseofDissonance–AccountsofWorthinEconomicLife,

PrincetonUniversityPressStiglitz,Joseph(2011),‘Rethinkingeconomics’,GlobalPolicy,2(2),165‐175Sutton,John(2000)¸Marshall’sTendencies–WhatCanEconomistsKnow?,MITPressTuckett,David(2011),MindingtheMarkets–AnEmotionalFinanceViewofFinancial

Instability,PalgraveMacmillanWaldrop,Mitchell(1994),Complexity–TheEmergingScienceattheEdgeofOrderand Chaos,PenguinWordsworth,William,‘LineswrittenaFewMilesaboveTinternAbbey’(fromLyrical

Ballads,1798)reprintedinRomanticism–AnAnthology,2ndedition,ed.DuncanWu,Blackwell,1998,265‐269