吴恩超（上海师范大学，上海浦东200235)
中图分类号：F822.0文献标识码：A文章编号：1003-2738（2012）02-0159-03
Abstract:Inviewoftheco-integrationtestandGrangercausalitytest,thispaperhasrespectivelyconductedthepositiveanalysisoftherelationshipbetweentheGDPinChinaandthemoneysupplyatalllevels.Theempiricalresearchhasdemonstratedthatinthelongtermthereisthebalancedco-integrationrelationshipandthecausalrelationshipbetweentheGDPandthemoneysupplywhileintheshorttermthecharacterofmoneysupply’sinfluenceontheGDPisessentiallythesameasthatofthelongterm.However,theinfluenceofM2ontheGDPisreversed,namelywhenM2risestheGDPwilldropinstead.Asaresult,duringmakingthepolicyofmoneysupplythestateshouldaimatdrivingthegrowthofGDPwhileduringmakingthepolicyofinterestratethestateshouldconsiderthebalancedinterestrate.Atthesametime,theyalsoshouldsyntheticallyapplythefiscalpolicysoastostrengthentheflexibilityandsustainabilityofthemonetarypolicy.
KeyWords:monetarypolicy;GDP;co-integrationtest;Grangercausalitytest
Thereisacloserelationshipbetweenthemoneysupplyandtheeconomicgrowthinthemodernmarketeconomy.Consequently,theanalysisoftherelationshipbetweenthechangesofmoneysupplyandtheeconomicgrowthwillbeimportanttoformulatethecorrectmacro-economiccontrolpolicy.
I.DataSourcesandStatisticalMethods:
DataExplanations.ThispaperhasaimedatexaminingtheinfluenceofthemonetarypolicyinChinaontheeconomicgrowth.Asaresult,intherespectofmonetarypolicywehaveselectedthedifferentmoneysupplyM0,M1,andM2astheresearchobjectswhileintherespectofreflectingthenationaleconomicgrowththeGDPmaybepidedintothenominalGDPandtheactualGDPaccordingtothedifferentaccountingpricelevels.ThispaperhasselectedthenominalGDPastheresearchobjectasthechangesofthemoneysupplywillleadtothechangesofthepricelevelandthenfurtherinfluencethatofthenominalGDP.Thereinto,thestatisticalapproachofthemoneysupplyatalllevelsisasfollows:
M0:thecurrencynoteincirculation;
M1:M0+currentdeposit
M2:M1+fixeddeposit+savingsdeposit+otherdeposit
AllthedatainthispaperisfromtheStatisticalYearbookof2009andthesampleintervalis1990-2008.TheEviews5.1softwarehasbeenusedforthedataprocessing.
Asthetransformationofthedata’snaturallogarithmwillnotchangetheoriginalco-integrationrelationshipbutinsteadcanbringthelinearizationintothetendencyandeliminatetheGARCHphenomenonexistedinthetimeseries,wehaveconductedthetransformationsofthenaturallogarithmofthenominalGDPandthethreetypesofmoneysupplyandrespectivelyusedlnGDPandlnMtoexpressthenaturallogarithmofthenominalGDPandthemoneysupply.
StatisticalMethods.Thispaperhasappliedthemethodsofco-integrationtestandGrangercausalitytesttoanalyzetherelationshipbetweentheGDPinChinaandthemoneysupplyatalllevels.Theanalysismaybeconcretelypidedintothefollowingfoursteps:
1.UnitRootTest.Asthetimeseriesofeconomyisinstableforthemostpart,itwillbelikelytoleadtothefallaciousregressionandthewrongconclusionifweadopttheinstabletimeseriestostudytheinterrelationshipbetweenthevariables.Consequently,whenweanalyzethetimeseriesweshouldfirstofallconductthestabilitytestoftheseriessoastoavoidthefallaciousregression.Theunitroottest,includingDFtestandADFtest,isfrequentlyusedforthestabilitytest.WegenerallyadopttheADFtestsoastoeliminatetheinfluencerelatedtotheerrorterms.
2.Co-integrationTest.Althoughsometimeseriesitselfisinstable,sometypeoflinearcombinationofthetimeseriesisstable.Thelinearcombinationreflectsthelong-termandsteadyrelationshipamongthevariableswhilesuchrelationshipiscalledtheco-integrationrelationship.Thetimeseriesthatowntheco-integrationrelationshipwillnotleadtothefallaciousregression.Thetwo-stageanalyticalmethodofEngelandGrangerisgenerallyusedfortheco-integrationtestofthedoublevariables.
3.ErrorCorrectionModel.AccordingtotheGrangerTheorem,amongthevariablesthatowntheco-integrationrelationshipthereshouldbetheerrorcorrectionmodelthatreflectstheshort-termdynamicinfluencerelationshipamongthevariables.Whenwetransformtheinstableseriesintothestableseriesthroughthedifferential,thelong-terminformationoftheeconomicvariablerelationshipwilllosewhileatthesametimetheregressionmodelserieswillownthereciprocityandtheregressionanalysiswillbecomeinvalid.Beingabletoovercometheseproblems,theerrorcorrectionmodelnotonlycanretainthelong-termdynamicinformationofthevariablerelationshipbutalsocanensurethevalidityoftheregressionanalysis.
4.GrangerCausalityTest.GrangerhaspointedoutthatthereshouldbetheGrangercausalrelationshipatleastinonedirectionamongthosevariablesifthereistheco-integrationrelationshipamongthevariables.GrangercausalrelationshipisakindofstatisticalrelationshipthatdescribestheinterrelationshipandinfluencebetweentwovariablesanditisrealizedinviewofthedoublevariableVAR.
II.TestingResultsandAnalysis
(1)ADFTestingResults.Table1istheADFtestresultsoftheGDPinChinaandthemoneysupplyatdifferentlevels.FromTable1wecanfindthattheADFteststatisticoftheoriginalserieslnGDPis-0.622529anditismorethanthecriticalvalue-2.660551ofthe10%significancelevel.Asaresult,theoriginalseriesisinstablebutafterthefirst-orderdifferentialtheADFstatisticis-2.7129thatislessthanthecriticalvalue-2.673459anditisstable.Namely,afterthefirst-orderdifferentialtheinstableserieslnGDPwillbecomestableanditshouldbethefirst-ordersingleintegratedseries.Inthesameway,thehorizontalseriesoflnM0,lnM1,andlnM2allhavenotpassedthehypothesisofdatastabilityandtheyareinstable.However,thefirst-orderdifferentialserieshaspassedthehypothesisanditisstable.Consequently,thetimeseriesofthoseeconomicvariablesallbelongtothesingleandintegratedfirstorderandtheycanbeconductedtheco-integrationtestbetweenthevariables.
Table1TestofStability
VariablesTestValuesCriticalValuesTestType(C,T,K)AICSCConclusions
lnGDP0.62259-2.660551（C，N，0）-4.48297-4.38404Instable
lnM0-1.128203-2.660551（C，N，0）-2.2888-2.18987Instable
lnM1-1.092922-2.660551（C，N，0）-2.52631-2.42738Instable
lnM2-2.14018-2.660551（C，N，0）-3.90357-3.80464Instable
d(lnGDP)-2.7129-2.673459（C，N，
2）-3.7961-3.6469Stable
d(lnM0)-4.645083-2.673459（C，N，
1）-2.9222-2.77734Stable
d(lnM1)-4.385061-2.673459（C，N，1）-3.00807-2.86321Stable
d(lnM2)-2.831042-2.666593（C，N，1）-3.94112-3.8431Stable
Note:Thetesttype(C,T,K)respectivelyexpresstheconstanttermoftheunitroottestequation,thetimetendency,andthelaggingexponentnumberwhileNmeansthatdonotincludethetimetendency.
(2)Co-integrationTestResults.ThispaperhasadoptedtheEngle-Grangertwo-steptestmethodtotesttheco-integrationbetweenlnGDPandlnM.Firsttheco-integrationaggressionshouldbeconductedtolnGDPandlnMbytheleastsquaremethodandthentheunitroottestshouldbeconductedtotheresidualerrorfromtheco-integrationaggression.Iftheresidualerrorseriesisstable,thereshouldbetheco-integrationrelationship,otherwisethereshouldnotbe.ThetestresultscanbereferredtoinTable2.Fromthatwecanfindthattheteststatisticvaluesofseriese0,e1,ande2allarelessthanthecriticalvaluesandthenwecanestimatethattheresidualerrorserieseshouldbestable.Therefore,thereshouldbethelong-termandstablebalancerelationshipbetweenlnGDPandlnM0,lnM1,andlnM2,namelytheco-integrationrelationship.
lnGDP＝1.298366＋1.072621lnM0（1）
（5.2646）（40.85478）
lnGDP＝2.42845＋0.837883lnM1（2）
（11.66）（42.9697）
lnGDP＝2.37＋0.7757lnM2（3）
（14.6576）（55.7063）
Table2ADFTestResultsofResidualErrorSeries
VariablesADFStatisticCriticalValuesTestFormConclusions
e0-4.871730-2.681330(C,N,1)Stable
e1-4.401906-2.690439(C,N,1)Stable
e2-5.478759-2.701103(C,N,0)Stable
Table3TestResultsofCausalRelationshipbetweenlnM0andlnGDP
OriginalHypothesisFStatisticPValueConclusions
lnM0isnottheGrangercauseoflnGDP7.056280.00942Refuse
lnGDPisnottheGrangercauseoflnM01.033700.38531Accept
Table4TestResultsofCausalRelationshipbetweenlnM1andlnGDP
OriginalHypothesisFStatisticPValueConclusions
lnM1isnottheGrangercauseoflnGDP17.66990.00027Refuse
LnGDPisnottheGrangercauseoflnM11.005870.39458Accept
Table5TestResultsofCausalRelationshipbetweenlnM2andlnGDP
OriginalHypothesisFStatisticPValueConclusions
lnM2isnottheGrangercauseoflnGDP6.715330.01104Refuse
lnGDPisnottheGrangercauseoflnM26.571730.01182Accept
Equations(1)-(3)aretheco-integrationequationsbetweentwovariables,namelythelong-termbalancerelationshipbetweenthevariables.Theco-integrationtestresultsindicatethatthereistheco-integrationrelationshipbetweenthemoneysupplyandtheGDP.Moreover,thereisthepositivecorrelationbetweenthemoneysupplyandtheGDP,namelytheexpandingmonetarypolicywillbeabletodrivetheincreaseofGDPandtoacceleratetheeconomicgrowthwhiletheretrenchingmonetarypolicywillbeabletoslowdowntheeconomicgrowth.Inconclusion,themoneysupplywillexertimportanceinfluenceontheGDP.
(3)ErrorCorrectionModel.Accordingtothetheorem,theerrorcorrectionmodelcanbeestablishedaslongasthereistheco-integrationrelationshipamongthesingleandintegratedvariables.ThroughtheEviews5.1software,wecanfind:
lnGDP＝0.075＋0.557△lnM0＋0.6277et-1
lnGDP＝0.055＋0.5514△lnM1-0.2754et-1
lnGDP＝-0.0298＋0.96△lnM2-0.1575et-1
Theco-integrationequationsdescribethelong-termrelationshipamongthevariableswhiletheerrorcorrectionmodeldescribestheshort-termrelationship.Besides,theerrorcorrectionmodelcandeterminethemutualadjustingspeedandtheshort-terminteractiveinfluencebetweenthevariables.
FromthemodelwecanfindthatifM0changesby1%,theGDPwillchangeby57.7%andtheerrorcorrectioncoefficientwillbe0.6277whileifM1changesby1%,theGDPwillchangeby55.14%andtheerrorcorrectioncoefficientwillbe-0.2754,whichconformstothereverseadjustmentmechanismand27.54%ofthedeviationbetweenthecurrentshort-termfluctuationandthelong-termbalancewillbeadjustedinthenextperiod.However,ifM2changesby1%,theGDPwillchangeby96%andtheerrorcorrectioncoefficientwillbe-0.1575,whichconformstothereverseadjustmentmechanismand15.75%ofthedeviationbetweenthecurrentshort-termfluctuationandthelong-termbalancewillbeadjustedinthenextperiod.Asaresult,thechangesofmoneysupplyinChinahaveobviouslyacceleratedtheeconomicgrowth.
(4)GrangerTest.TheGrangertestbetweenthepairwiseeconomicvariablesshouldbeconductedwhiletheresultscanbereferredtoinTable3,4,and5.(Table3,4,5)
Fromthecausalitytest,wecanfindthatlnM0,lnM1,andlnM2allshouldbetheGrangerfactorsthatleadtothechangesoflnGDPwhilelnGDPshouldbetheGrangerfactorthatleadstothechangesoflnM2.Consequently,inthesenseofGrangerthereshouldbethecausalityrelationshipbetweenthemoneysupplyandtheeconomicgrowthandtheyinfluenceeachothersoastoformacomplexcirculation.Namelyononehandthechangesofmoneysupplywillleadtothechangesofeconomicgrowthwhileontheotherhandthechangesofeconomicgrowthwillalsoleadtothechangesofmoneysupply,whichreflectsthatthereshouldbecertaininteractiverelationshipbetweenthemonetarypolicyandtheeconomicgrowth.
III.PolicySuggestions:
Fromtheabovepositiveanalysis,wecandrawtheconclusion:inthelongtermthemoneysupplyshouldbeprincipalfactorthatdrivestheeconomicgrowth.ThemonetarypolicycaninfluencetheGDPthroughthemoneysupplysothatthemacro-economiccontroloftheeconomicgrowthcanbeconductedthroughthesuitablemonetarypolicy.Asthemoneysupplywillexertthepositiveinfluenceontheeconomicgrowthinthelongterm,theretrenchingmonetarypolicywillbeabletorestraintheeconomicovergrowthwhilethestablemoneysupplywillbeabletoavoidtheovergrowthofconsumptionandinvestment,toeffectivelystabilizethemarketeconomy,andtoavoidtheinflation.InChina,besidesthemacro-economiccontrolthroughthemonetarypolicythestatealsoshouldadoptthesuitableancillarymeasuresasaresultoftheinfluenceofthespecialnationalconditions.Concretelyspeaking,themeasuresshouldincludethefollowing:
Firstly,thepolicyformulatingofmoneysupplyshouldaimatpromotingtheGDPgrowth.Intheshorttermtheexpansionarymonetarypolicythatcounterstheeconomiccycleshouldbeemphasizedwhileinthelongtermtheamountofacurrencyincirculationshouldbeemphasizedsoastoavoidtheinflation,sincetheinflationwillincreasetherunningcostwhiletheexcessiveinflationwillalsoleadtothesocialunrestandallthesewillreducethegrowthofthelong-termGDP.
Secondly,thebalancedinterestrateshouldbeconsideredwhenweformulatingthepolicyofinterestrate.Theactualinterestrateshouldinfinitelyapproachthebalancedinterestrateandshouldbeadjustedintimesothatwecangetthemaximumcapitalaccumulationandensurethelong-termeconomicgrowthinChina.InviewofthepracticeinChina,thepolicyofinterestrateformulatedbythecentralbankhaspaidmuchattentiontotheconsumerpriceindexbutseldomemphasizesthenegativeactualinterestrate.However,thenegativeinterestratewillinevitablyinfluencethecapitalaccumulationandthelong-termeconomicgrowth.Therefore,weshouldformulatethebalancedinterestratethatconformstotheactualconditions,reflectsthecapitalcost,andownstheequaldemandandsupply.
Thirdly,thecentralbankinChinashouldproperlydealwiththerelationshipbetweenthemaintenanceofrelativelygoodandrapideconomicdevelopmentandtheadjustmentofeconomicindustrialstructure,maintainthechronicity,continuity,andstabilityofthemacro-economicpolicy,continuetoimplementthereasonablyloosemonetarypolicy,properlycontrolthestrengthandrhythmofthepolicy,andsyntheticallyadoptthefiscalpolicysoastostrengthentheflexibilityandsustainabilityofthemonetarypolicy.
Reference:
[1]SpadaroJV,RablA:Pathwayanalysisforpopulation-totalhealthimpactsoftoxicmetalemissions.RiskAnalysis2004,24:1121-1141.
[2]EuropeanCommission:ExternE-ExternalitiesofEnergy-Methodology2005Update.Luxemburg:OfficeforOfficialPublicationsoftheEuropeanCommunities2004.
[3]EuropeanCommission:ESPREME-Estimationofwillingness-to-paytoreducerisksofexposuretoheavymetalsandcost-benefitsanalysisforreducingheavymetalsoccurrenceinEurope-Publishablefinalactivityreport,502527edn.2007.
[4]RiberC,FredriksenGS,ChristensenTH:HeavymetalcontentofcombustiblemunicipalsolidwasteinDenmark.WasteManagementResearch2005,23:126-132.
[5]WHO:EnvironmentalHealthCriteria165-InorganicLead.WHO-WorldHealthOrganization1995.
[6]Alleyne,K.,Lewis-Bynoe,D.andMoore,W.2004:Anassessmentofthegrowth-enhancingsizeofgovernmentintheCaribbean.AppliedEconometricsandInternationalDevelopment4,77–94.
[7]Barro,R.1990:Governmentspendinginasimplemodelofendogenousgrowth.JournalofPoliticalEconomy98,S103–S125.
[8]Carr,J.1989:Governmentsizeandeconomicgrowth:Anewframeworkandsomeevidencefromcross-sectionandtime-seriesdata:Comment.AmericanEconomicReview79,267–71.