INTRODUCTION TO THE MARKET MICROSTRUCTURE LITERATU

IntroductionINTRODUCTIONTOTHEMARKETMICROSTRUCTURELITERATUREAPPLIEDTOTHEFOREIGNEXCHANGEMARKET1IntheirforthcomingchapterforTheHandbookofInternationalEconomics,FrankelandRose(1994)reviewempiricalandtheoreticalstudiesofexchangeratemovements.Themodelsusedtoexplaintheexchangeratesbetweenindustrializedcountriesrelyonmacroeconomicfundamentalssuchasprices,moneysupplies,andtrade-relatednationalaccounts.Thesefactorsarebelievedtoaccountfortheleveloftheexchangerate.AnearlypaperwhichdescribestheserelationsisMussa(1976).Itwasbelievedthatmacroforcesdrivetheexchangerateprocess.Examplesofthesemacroforcesincludethedegreeofpriceflexibility,changingriskfactorsduetotheaccumulationofchangestoaneconomy,ortheactionsofthegovernmenttotheinfluencemacroeconomicfactors.ModelsincludingdynamicfactorsbegantoappearafterworkbyDornbush(1976).Theintelligentagentsinthesemodelsaretheforeignexchangetraders,usuallybanks,whounderstandthedynamicsoftheexchangerateprocessandinterpretthemacroeconomicfundamentals.Inmostearlymodels,theseagentsareoftenpassiveparticipantsandarenotmodeledexplicitly.Astraightforwardinterpretationofthefundamentalsusingtime-seriesmodelingoftheexchangeratehasfailedtoadequatelyexplaintheleveloftheexchangeratewithinsampleortoforecastfuturemovements.Addingdynamicelementshasledtomorecomplexmodelsofpricestickinessandovershootingorundershootingbuttheseenhancements,whileinterestingastheoreticalmodels,havenotsignificantlyimproveduponthebasiceconometricmodels.Asimplemodelusingarandomwalkprocessasanexplanationofthemovementoftheexchangeratehasperformednearlyaswellasthemorecomplexmodels.ThisisdocumentedintheworkofMeeseandRogoff(1983)aswellasothers.Aftertheirreviewoftheexistingempiricalresultsfromnearlyalloftheexistingmodels,FrankelandRose(1994)concludethattheexistingstructuralmodelshavelittleintheirfavorbeyondtheoreticalcoherence.Thefailureoftheexistingtheoreticalstructuralmodelshasturnedthefocusofcurrentresearchtotheintelligentagentasparticipant,andoftenoverridinginfluenceonthepathoftheexchangerate.Agentscaninfluencetheexchangeratebyforming2expectationsofthefutureleveloftheexchangerate,andbyactingontheseexpectationsasactiveparticipantsintheforeignexchangemarket.SincethemodelofMussa(1976),thepresenceofexpectationshasbeenrecognized.Thiscanbeseenbytheexpressionfortheleveloftheexchangerateinthesimplemonetarymodel:et=ft+bEt[det]/dt.whereetistheleveloftheexchangerate,ftistheunderlyingfundamental,Etistheexpectationoperator,anddet/dtisthechangeintheleveloftheexchangerateovertime.Theinitialinterpretationoftheexpectationtermwasthatitwasduetothedynamicsofthemodel.Ifpricesmoveslowlyandtheequilibriumrelationsareknown,thenthereisanexpectationofreachinganequilibriumsometimeinthefuture.Alaterinterpretation,attributedtoDornbush(1978)andFrenkel(1981),stressedtheroleofexpectationsderivedfromnewinformationavailabletoagents.Thisrepresentedasubtleshiftintheinterpretationofexpectations.Expectationswerepresentinthemodelsbutnowtheseexpectationsareduetoinformationrelatedtothefundamentals.Earliermodelsassumedthefundamentalwasknown,buttheselatermodelsassumeonlythatinformationaboutthefundamentalisknown.Thisinterpretationofexpectationsledtoempiricalworktomeasurefactorswhichcouldaffecttheformationofexpectations.EmpiricalstudiesreviewedbyFrankelandRose(1994)useinnovations(shocks)ineconometrictime-seriesmodelstostudyhowexpectationsmightbeaffectedbyinnovationstointerestratedifferentials,themoneysupplyratio,orchangesinpolicy.Afterthedisappointingresultsfromthesestudies,theexpectationsthemselveswereestimated.Surveydataorofficialpolicychangeannouncementswereusedtotesttheeffectofchangesinexpectations.FrankelandFroot(1987)andFrankelandChinn(1993)areexamplesofthislineofresearch.Researchofthiskindisnotuniquetothestudyofforeignexchange.Muchoftheworkonexpectationshasbeendirectedtotheunderstandingofmarketefficiency.Thisworkalsobringsintofocustheimportanceof3theassumptionofrationalexpectationsunderlyingthemodelsofforeignexchangedetermination.Whileexpectationshavebeenincorporatedintothemonetarymodelsofexchangeratedeterminationsincethemid-1970s,ithasbeendifficulttoexplainhowtheseexpectationsareformedortoobservetheseexpectations.Perhapsmoreimportantly,howtheseexpectationsareusedbyagentsinthemarkethasnotbeenadequatelyaddressed.Theissueofhowexpectationsareusedhasledtothedevelopmentofthemarketmicrostructureliterature.Inthisliterature,attentionisplacedonthemarketinstitutionandtheactivitiesofagentswithinthemarketinstitution.Theformationofexpectationsbyagentsismodeledasadynamicprocesswhereinformationisprocessedregardingthefundamentalsofthemarket,andstrategicinteractiontakesplaceamongtheagentsinthemarket.Themicrostructureliteratureaddressestheroleofinformationinthemarketinrelationtotheformationofexpectations.Theprimaryquestionishowinformationisincorporatedintomarketprices.Informationisassumedtobedispersedamongmanyagentsinthemarket.Towhatextentdoesasinglemarketclearingpriceincorporatethedispersedinformation?Therearemanytheoriesastohowthisisaccomplishedandmanydoubtthatitispossibleatall.Rational