recursive macroeconomic theory time series

《recursive macroeconomic theory time series》由會(huì)員上傳分享，免費(fèi)在線閱讀，更多相關(guān)內(nèi)容在工程資料-天天文庫。

1、Chapter2TimeSeries2.1.TwoworkhorsesThischapterdescribestwotractablemodelsoftimeseries:Markovchainsand?rst-orderstochasticlineardi?erenceequations.Thesemodelsareorganizingdevicesthatputrestrictionsonasequenceofrandomvectors.Theyareusefulbecausetheydescribeatimeserieswithparsimony.Inlaterchapters

2、,weshallmaketwouseseachofMarkovchainsandstochasticlineardi?erenceequations:(1)torepresenttheexogenousinformation?owsimpingingonanagentoraneconomy,and(2)torepresentanoptimumorequilibriumoutcomeofagents’decisionmaking.TheMarkovchainandthe?rst-orderstochasticlineardi?er-encebothuseasharpnotionofas

3、tatevector.Astatevectorsummarizestheinformationaboutthecurrentpositionofasystemthatisrelevantfordetermin-ingitsfuture.TheMarkovchainandthestochasticlineardi?erenceequationwillbeusefultoolsforstudyingdynamicoptimizationproblems.2.2.MarkovchainsAstochasticprocessisasequenceofrandomvectors.Forus,t

4、hesequencewillbeorderedbyatimeindex,takentobetheintegersinthisbook.Sowestudydiscretetimemodels.Westudyadiscrete-statestochasticprocesswiththefollowingproperty:MarkovProperty:Astochasticprocess{xt}issaidtohavetheMarkovpropertyifforallk≥1andallt,Prob(xt+1

5、xt,xt?1,...,xt?k)=Prob(xt+1

6、xt).Weassumet

7、heMarkovpropertyandcharacterizetheprocessbyaMarkovchain.Atime-invariantMarkovchainisde?nedbyatripleofobjects,namely,–29–30TimeSeriesann-dimensionalstatespaceconsistingofvectorsei,i=1,...,n,whereeiisann×1unitvectorwhoseithentryis1andallotherentriesarezero;ann×ntransitionmatrixP,whichrecordsthepr

8、obabilitiesofmovingfromonevalueofthestatetoanotherinoneperiod;andan(n×1)vectorπ0whoseithelementistheprobabilityofbeinginstateiattime0:π0i=Prob(x0=ei).TheelementsofmatrixParePij=Prob(xt+1=ej

9、xt=ei).Fortheseinterpretationstobevalid,thematrixPandthevectorπ0mustsatisfythefollowingassumption:Assumpt

10、ionM:a.Fori=1,...,n,thematrixPsatis?esnPij=1.(2.2.1)j=1b.Thevectorπ0satis?esnπ0i=1.i=1AmatrixPthatsatis?esproperty(2.2.1)iscalledastochasticmatrix.Astochasticmatrixde?nestheprobabilitiesofmovingfromonevalueofthestatetoanotherino