外文翻译能源消耗能源价格与经济增长之间的关系亚洲发展中国家的时间序列证据Word格式.docx

外文翻译能源消耗能源价格与经济增长之间的关系亚洲发展中国家的时间序列证据Word格式.docx

《外文翻译能源消耗能源价格与经济增长之间的关系亚洲发展中国家的时间序列证据Word格式.docx》由会员分享，可在线阅读，更多相关《外文翻译能源消耗能源价格与经济增长之间的关系亚洲发展中国家的时间序列证据Word格式.docx（9页珍藏版）》请在冰豆网上搜索。

EnergyEconomics,2000（22）：

615-625.

作者：

JohnAsafu-Adjaye

原文：

Abstract

ThispaperestimatesthecausalrelationshipsbetweenenergyconsumptionandincomeforIndia,Indonesia,thePhilippinesandThailand,usingcointegrationanderror-correctionmodelingtechniques.Theresultsindicatethat,intheshort-run,unidirectionalGrangercausalityrunsfromenergytoincomeforIndiaandIndonesia,whilebidirectionalGrangercausalityrunsfromenergytoincomeforThailandandthePhilippines.InthecaseofThailandandthePhilippines,energy,incomeandpricesaremutuallycausal.Thestudyresultsdonotsupporttheviewthatenergyandincomeareneutralwithrespecttoeachother,withtheexceptionofIndonesiaandIndiawhereneutralityisobservedintheshort-run.2000ElsevierScienceB.V.Allrightsreserved.

1.Introduction

Inthepasttwodecadesnumerousstudieshaveexaminedthecausalrelation-shipsbetweenenergyconsumptionandeconomicgrowth,witheitherincomeoremploymentusedasaproxyforthelatter.Todate,theempiricalfindingshavebeenmixedorconflicting.TheseminalarticleonthistopicwaspublishedinthelateseventiesbyKraftandKraft（1978）whofoundevidenceinfavorofcausalityrunningfromGNPtoenergyconsumptionintheUnitedStates,usingdatafortheperiod1947-1974.Theirfindingswerelatersupportedbyotherresearchers.Forexample,AkarcaandLong（1979）.foundunidirectionalGrangercausalityrunningfromenergyconsumptiontoemploymentwithnofeedback,usingUSmonthlydatafortheperiod1973-1978.Theyestimatedthelong-runelasticityoftotalemploy-mentwithrespecttoenergyconsumptiontobe-0.1356.

However,thesefindingshavebeensubjectedtoempiricalchallenge.AkarcaandLong（1980）,ErolandYu（1987a）,YuandChoi（1985）,andYuandHwang（1984）foundnocausalrelationshipsbetweenincome（peroxidebyGNP）andenergyconsumption.Onthecausalrelationshipbetweenenergyconsumptionandemploy-ment,ErolandYu（1987b,1989）,YuandJin（1992）,andYuetal.（1988）foundevidenceinfavorofneutralityofenergyconsumptionwithrespecttoemploy-ment,referredtoasthe‘neutralityhypothesis’.

Oneofthereasonsforthedisparateandoftenconflictingempiricalfindingsontherelationshipbetweenenergyconsumptionandeconomicgrowthliesinthevarietyofapproachesandtestingproceduresemployedintheanalyses.Manyoftheearlieranalysesemployedsimplelog-linearmodelsestimatedbyordinaryleastsquares（OLS）withoutanyregardforthenatureofthetimeseriespropertiesofthevariablesinvolved.However,ashasrecentlybeenproven,mosteconomictimeseriesarenon-stationaryinlevelsform（seeGrangerandNewbold,1974）.Thus,failuretoaccountforsuchpropertiescouldresultinmisleadingrelationshipsamongthevariables.

Followingadvancesintimeseriesanalysisinthelastdecade,recenttestsoftheenergyconsumptioneconomicgrowthrelationshiphaveemployedbivariatecausalityproceduresbasedonGranger（1969）andSims’（Sims,1972）tests.How-ever,thesetestsmayfailtodetectadditionalchannelsofcausalityandcanalsoleadtoconflictingresults.Forexample,recently,GlasureandLee（1997）testedforcausalitybetweenenergyconsumptionandGDPforSouthKoreaandSingaporeusingthestandardGrangertest,aswellascointegrationanderror-correctionmodeling.Theyfoundbidirectionalcausalitybetweenincomeandenergyforbothcountries,usingcointegrationanderror-correctionmodeling.However,usingthestandardGrangercausalitytests,theyfoundnocausalrelationshipsbetweenGDPandenergyforSouthKoreaandunidirectionalGrangercausalityfromenergytoGDPforSingapore.

Thedirectionofcausationbetweenenergyconsumptionandeconomicgrowthhassignificantpolicyimplications.If,forexample,thereexistsunidirectionalGrangercausalityrunningfromincometoenergy,itmaybeimpliedthatenergyconservationpoliciesmaybeimplementedwithlittleadverseornoeffectsoneconomicgrowth.Inthecaseofnegativecausalityrunningfromemploymenttoenergy（AkarcaandLong,1979）,totalemploymentcouldriseifenergyconservationpolicyweretobeimplemented.Ontheotherhand,ifunidirectionalcausalityrunsfromenergyconsumptiontoincome,reducingenergyconsumptioncouldleadtoafallinincomeoremployment.Thefindingofnocausalityineitherdirection,theso-called‘neutralityhypothesis’（YuandJin,1992）,wouldimplythatenergyconservationpoliciesdonotaffecteconomicgrowth.

Thispaperexaminestheenergyincomerelationshipforfourenergy-dependentAsiandevelopingcountries:

India,Indonesia,thePhilippinesandThailand.Thesecountrieswerechosenbecausetheyrepresentenergy-dependentLDCswhicharepoisedfortake-offintoaphaseofindustrialization.Wedepartfrompreviousstudiesbyconsideringatrivariatemodel（energy,incomeandprices）ratherthantheusualbivariateapproach.Thisapproachofferstheopportunitytoinvestigateotherchannelsinthecausallinksbetweenenergyconsumptionandeconomicgrowth.

2.Economicandenergyuseprofiles

Thefourcountriesareheavilypopulatedandhaveacombinedtotalof1.3billionpeople（Table1）.Ofthefour,Indiaistheleastwealthyonapercapitaincomebasisofcomparison,withapercapitaGDPofUS$380（1996dollars）whichistheaveragefortheSouthAsiaregion.TheothershavepercapitaincomesofoverUS$1000（seeTable1）.Allfourcountriesrecordedhighannualgrowthratesintheirmanufacturingsectorsin1996,rangingfrom10.5%forIndonesiato5.6%forthePhilippines.Ofcourse,theseimpressivegrowthrateswouldhavedeclinedin1997andbeyondinviewoftheAsianfinancialcrisis.Tomaintainthehighlevelsofeconomicoutputthesecountriesmakehighdemandsonenergyresources.

Table1reportsfiguresforpercapitaenergyuseandcarbondioxideemissionsforthefourcountriesinthesample.EnergyusepercapitaishighestforThailandin1995with878kg,followedbyIndonesiawith442kgpercapita.Indiahasthelowestpercapitaenergyusewith260kg.Carbondioxideemissionspercapitaarealsorelativelyhigh,rangingfrom2.9metrictonsforThailandto0.9metrictonsforthePhilippines.Mostofthecountrieshavetorelyonimportsfortheirenergyneeds,exceptIndonesiawhichisanetexporteroffuel.Indiaisamongthelargestconsumersofenergyintheregion.India’senergysourcescomprisemainlycoal,andwereestimatedtobe244millionmetrictonsin1991（OECD,1993）.

Table1

Percapitaenergyuseandcarbondioxideemissions（1995）a

Indicator

India

Indonesia

Thailand

Philippines

Populationmid-1996（millions）

945.1

197.1

60.0

71.0

GNPpercapita1996（US$）

380

1080

2960

1160

Manufacturing（averagegrowthrate%p.a.）

8.1

10.5

7.7

5.6

Energyusepercapita（kg）

260

442

878

307

CO2emissionspercapita（mton）

1.0

1.5

2.9

0.9

aSource:

WorldBank（1998）.

TheabovefiguresshowthatAsianLDCsaccountforasignificantproportionofworldenergyconsumption.Giventherecentphenomenalgrowthinawarenessofandconcernforglobalwarming,anexaminationoftheenergyincomerelationshiphasimplicationsforenergypolicyinthesecountries.Itisimportanttoaddthatmostofthestudiesreferredtoabovehavedealtwithadvancedornewlyindustrializedcountries（NICs）anditmaybearguedthattheresultsarenotapplicabletocountriesatadifferentstageofdevelopment.

3.Methodologyanddata

ThemodelingstrategyadoptedinthisstudywasbasedonthenowwidelyusedEngleGrangermethodology（seeGrangerandNewbold,1974;

EngleandGranger,1981）.TheaugmentedDickeyFuller（ADF）andPhillipsPerron（PP）testsofstationaritywereused（DickeyandFuller,1981;

PhillipsandPerron,1988）.Followingtheunitrootandcointegrationtests,weestimatedthefollowingerrorcorrectionmodel:

Dyt=A21（L）Dyt-1+A22（L）Dent-1+A23（L）Dpt-1+λyECTt-1+u2t

（1）

Dent=A11（L）Dyt-1+A12（L）Dent-1+A13（L）Dpt-1+λenECTt-1+u1t

（2）

Dpt=A31（L）Dyt-1+A32（L）Dent-1+A32（L）Dpt-1+λpECTt-1+u3t（3）

where[yt,ent,pt]arerealincome,energyconsumptionandprices,respectively;

Disadifferenceoperator;

Aij（L）arepolynomialsinthelagoperatorL;

ECTisthelaggederror-correctionterm（s）derivedfromthelong-runcointegratingrelation-ship;

andtheuitsareerror-correctiontermsassumedtobeuncorrelatedandrandomwithmeanzero.Thecoefficients,λi（i=en,y,p）,oftheECTsrepresentthedeviationofthedependentvariablesfromthelong-runequilibrium.

Throughtheerror-correctionmechanism,theECMopensupanadditionalcausalitychannelwhichisoverlookedbythestandardGranger（1969）andSims（1972）testingprocedures.IntheGrangersenseavariableXcausesanothervariableYifthecurrentvalueofYcanbetterbepredictedbyusingpastvaluesofXthanbynotdoingso.TheGrangercausalitytestingprocedureinvolvestestingthesignificanceoftheAijsconditionalontheoptimumlags1.ThroughtheECT,anerrorcorrectionmodeloffersanalternativetestofcausality（orweakexogeneityofthedependentvariable）.If,forexample,λeniszero,thenitcanbeimpliedthatthechangeinentdoesnotrespondtodeviationinlong-runequilibriuminperiodt-1.Also,ifλeniszeroandbothA11andA13arezero,itcanbeimpliedthatincomeandpricesdonotGranger-causeenergyconsumption.Thenon-significanceofboththetandWaldF-statisticsintheECMwillimplythatthedependentvariableisweaklyexogenous.2

Ifthevariables,yt,entandptarecointegratedthenitisexpectedt