COMPLEXITY IN ECONOMICS.docx

COMPLEXITY IN ECONOMICS.docx

《COMPLEXITY IN ECONOMICS.docx》由会员分享，可在线阅读，更多相关《COMPLEXITY IN ECONOMICS.docx（24页珍藏版）》请在冰豆网上搜索。

COMPLEXITYINECONOMICS

COMPLEXITYINECONOMICS

EditedbyJ.BarkleyRosser,Jr.[bookforthcoming,EdwardElgar,2003]

Introduction

TheEmergenceofComplexityinEconomics

Increasinglyineconomicswhathadbeenconsideredtobeunusualandunacceptablehascometobeconsideredusualandacceptable,ifnotnecessarilydesirable.Whereasithadbeenwidelybelievedthateconomicrealitycouldbereasonablydescribedbysetsofpairsoflinearsupplyanddemandcurvesintersectinginsingleequilibriumpointstowhichmarketseasilyandautomaticallymoved,nowitisunderstoodthatmanymarketsandsituationsdonotbehavesowell.Economicrealityisrifewithnonlinearity,discontinuity,andavarietyofphenomenathatarenotsoeasilypredictedorunderstood.Atthesametimethebroadcoherenceofeconomicsystemsismoreimpressivethaneverinthefaceofsuchphenomena.Theorderoftheeconomyappearstoemergefromthecomplexinteractionsthatconstitutetheevolutionaryprocessoftheeconomy.

Thesephenomenahavecometobelabeledascomplexityineconomics.Evenwhatseemssimpleineconomicsgenerallyarisesfrombehaviornotreflectingrationalexpectations;weliveinaworldthatreflectstheenormousvarietyanddiversityofhumanityintheirknowledge,attitudes,andbehaviors,interactingwitheachotherinanenormousrangeofinstitutionalframeworks.Whatemergesintheaggregatemayhavelittletodowithwhathappensattheindividuallevel.Butthisaggregatecannotbesimplydescribedbysomesetofaggregateequations.Itemergesoutofthesoupoftheindividualandparticularwithallitsmultiforminteractionsandpeculiarities.

Thischangeofperspectivehasbroughtforthavarietyofnewapproachestoanalysis.Previouslyapremiumwasplacedondeductiveformalproofsoftheoremsthatsoughttoderivegeneralsolutionsbroadlyapplicable,aviewthatreacheditsculminationintheFrenchBourbakistschool.Nowweseeagreateremphasisoncomputersimulationsandexperimentalmethodstoinductivelydeterminepossibleoutcomesandrangesofsolutions.Emergentphenomenafromcomplexsystemsarenotusuallydiscoveredbytheoremsbutmorefrequentlybytheuseofincreasinglypowerfulcomputerstoexplorethelimitsandpossibilitiesthatcanarise.Thenewawarenessoftheubiquityofcomplexityistransformingthewaythatwethinkabouteconomics.

WhatisEconomicComplexity?

Theideathatrealityreflectscomplexitydidnotinitiallyariseineconomics.Ithascomefromnumerousdisciplinesofthoughtandhasreverberatedbackandforthacrossthem,frommathematicsandcomputersciencetophysicsandbiologytothesocialsciencesandbackandforth.Thatdifferentdisciplineshavecometoseeitasmeaningsomethingsomewhatdifferentfromwhatotheronesdoisnotsurprising.Indeed,thephysicist,SethMiller,hasgatheredatleast45differentdefinitionsof“complexity.”Manyofthesearenotappropriateforeconomics,oronlybarelysoforspecialoccasionsorcircumstances.Computerscienceistheoriginofmanyofthese,andarguablymanyoftheseareamongthemorerigorousofsuchdefinitions,suchasthelengthofanalgorithmforsolvingaspecificproblem.SuchideashadimportantinfluenceoneconomicthinkingthroughsuchfiguresasHerbertSimon,animportantfigureinthedevelopmentofartificialintelligenceincomputerscienceaswellasaNobelPrizewinnerineconomicsforhisconceptualizationofboundedrationality.Butthisisnotwhatmosteconomiststhinkofwhentheythinkofcomplexeconomicdynamics,althoughrealityofcomplexityineconomicshascomefullcircletoSimoninimplyingtheinevitabilityofboundedrationality.Economiccomplexityisnotsimplyamatterofthingsbeing“complicated,”althoughsomehavevieweditthisway,andcertainlyeconomicrealityisboth“complicated”aswellascomplexinthesensethathasincreasinglycometobeacceptedasmostrelevantforeconomics.

Attheriskofoversimplifyingcomplexity,weshallfollowRichardDay（1994）indefiningcomplexityineconomicsintermsofdynamicoutcomes.Aneconomicsystemisdynamicallycomplexifitsdeterministicendogenousprocessesdonotleaditasymptoticallytoafixedpoint,alimitcycle,oranexplosion.Somemightarguethatendogenouslimitcyclesalsoconstitutecomplexity,butweshallviewthemasmerelynearlycomplex.Wenotethatallsystemsthatfitthisdefinitionhavesomedegreeofnonlinearitywithinthem,howeverlimitedorarbitrary.Atthesametimetherearenonlinearsystemsthatarenotcomplex,suchasastandardexponentialgrowthmodel.

Nowweobservethatthisisa“broadtent”definitionofcomplexitythatincludessystemsthatsomewouldargueshouldnotbeincluded.Thereisa“narrowtent”viewofeconomiccomplexity,butitlacksasimpleorcleardefinition.Itcanonlybecharacterizedbyfittingwithasetofcharacteristics.Arthur,Durlauf,andLane（1997）providesuchalistofcharacteristics.Theyinclude1）dispersedinteractionamongheterogeneousagentsactinglocallyoneachotherinsomespace;2）noglobalcontrollerthatcanexploitallopportunitiesorinteractionsintheeconomyeventhoughtheremightbesomeweakglobalinteractions;3）cross-cuttinghierarchicalorganizationwithmanytangledinteractions;4）continualadaptationbylearningandevolvingagents;5）perpetualnoveltyasnewmarkets,technologies,behaviors,andinstitutionscreatenewnichesintheecologyofthesystem,and6）out-of-equilibriumdynamicswitheitherzeroormanyequilibriaexistingandthesystemunlikelytobenearaglobaloptimum.

Howdothesetwoapparentlycompetingperspectivesrelatetoeachotherandhowarewegoingtodealwiththeminthiswork?

Thesecondwillgenerallybeasubsetofthefirst（althoughtherearecertainlysystemsthatcouldfulfillallsixofthesecharacteristicsandyetviolatethefirstdefinition）.Moreaccuratelyitcanbearguedthatthesecondisthehistoricaloutcomeofthefirst.InhisessayreproducedinSectionI,JohnHorgan（1995）criticizestheconceptof“complexity”moregenerally,notjustineconomics,asjustthelatestinastringoffads,“thefourC’s.”InhisviewthesefourC’sarecybernetics,catastrophe,chaos,andcomplexity.Thisreflectstheapproximateorderoftheirappearanceasbroad,transdisciplinaryfads,withcyberneticshotinthe1960s,catastrophetheoryhotinthe1970s,chaostheoryhotinthe1980s,andcomplexitycominginduringthe1990s.InHorgan’svieweachofthesewasoverhyped,wasasortofintellectualbubblethatblewupandthencrashed.Forhimcomplexityisjustthelatestoftheseandwillcrashalso.

TheviewofthisintroductionistotakeHorgan’sargumentseriouslybuttoturnitonitshead.First,heiscorrecttoidentifyintellectuallinkagesbetweentheseintellectualmovements.Theyallareapartofcomplexityinour“broadtent”sense.Notonlythat,buttheyhavetosomedegreearisenoutofeachother,eventhougheachhasspecificsharpdifferencesfromtheothers.Thisisevenreflectedinthatcertainindividualshaveplayedrolesinvariousofthesestages,withafewevenarguablyhavingbeeninvolvedinallfour.Thereisanintellectualunityto“broadtent”complexity.

Second,heiswrongtohaveviewedthemashavingsimplydisappearedwhenthey“crashed”aftertheiradmittedlyrespectiveoverhypings（althoughitisnotclearthatthelast,“narrowtent”complexity,hasactually“crashed,”despitevariousdoubtsandcaveatshavingbeenraisedaboutit）.Cyberneticslivesoninthenewercomplexitythroughtheuseofsimulationandtheideathat“counterintuitive”resultscanarisefromsimulatingsystemsofnonlineardifferentialequations.Catastrophetheory,perhapsthemostcriticizedand“academicallyunacceptable”ofthefour,persistsinthewidespreadperceptionoftherealityofprofounddiscontinuity,whetheritisthecrashofastockmarketoracurrency,orthecollapseofanecologic-economicsystem.Thenewercomplexitytheoryofdispersedagentshasdiscovereditsownwaysofgeneratingthisoutcomeoftheoldertheory.Thesituationwithchaostheoryismorecomplicatedasitsprogenitorsandpractitionersareevenmorecloselylinkedtothenewerviews.ThisfactisrecognizedbyHorganinhislabelofchaoplexology,whichmaybeanon-category,butwhoseveryexpressiongivescredencetotheideaoftheunityofthestreamofideasthathavebeenbuildingoverseveraldecadestoreachacriticalintellectualmassthatisnowmoreprofoundlyalteringtheviewsofeconomistsmorewidely.

Ourapproachwillbetopresentpapersthattendtoemphasizethemorerecentviewpoint,whichisclearlywhatisgenerallydrivingmorerecentresearch.However,weshallalsoincludecertainpapersthatrepresenttheearlierperspectivesofthebroaderdefinitionofeconomiccomplexitythatwehavepresented.Thiswillallowforbothhistoricalaccuracyandalsotoshowtheessentialunityoftheseapproachesasmanypaperswillmixvariousofthesesupposedlydistinctviewpoints.Howeverthebookwillbeorganizedbyareasofeconomics.Thuswithineachareatherewillbepapersofvariousoftheperspectives,althoughgenerallywithagreateremphasisonthemorerecentviewofwhateconomiccomplexityis.WenotehoweverthatweshallnotincludeanypapersthathavealreadyappearedinthevolumeonchaostheoryeditedbyDechert（1996）,althoughsomeofthosewouldhavebeencandidatesforthisvolume.

Weshallnowproceedwithasummaryofthepaperspresentedintheirtencategories.

PhilosophicalandMethodologicalOverviews

Thefirstpaperdatesfrom1962,longbeforethecurrentdiscussionsofthemeaningornatureofcomplexityineconomics.Indeeditisnotex