Problemstatement
Itisnodoubttosayingthatrecently,regionaltradeagreements(RTA)havebecomea popularwid espreadtrendin th e internationaleconomicsystem,especially afterDoh a roundofGATT/
W T O AccordingtothedefinitionofWTO,regionaltradeagreement,includedfreetradeagreements( FTAs)andc u s t o m s u n i o n s (CUs),aret h e negotiationso f t wo o r mo re parties,i n whichthesepartic ipantsagreetoreducetheircurrentcustombarriers,suchastariffs,quotas.Sincee a r l y o f t h e 1 9 9 0 s , RTAshaveincreasedwidespread.A c c o r d i n g t o r e p o r t s o f W o r l d TradeOrganization(WT O),untilFebruary2016,thereare625notificationsofRTAsand419inwhichwereinforce.
Nations(ASEAN)isconsideredasasuccessfulm o d e l o f r e g i o n a l i s m andt h e c o m m u n i t y i s s t e p bys t e p g r e a t l y c o - o p e r a t i n g andintegratingtotheworldeconomy.Inaddition,Japan,aneconomywasgrowingrapidly,i nvolving1 7 % t o worldeconomici n 2 0 0 5 b u t reducedt o o n l y 6 % i n 2 0 1 5 (IMF,2015).H o w e v e r , h e r economicperformancehasamassiveinfluenceontheeconomyoftheentireregion.Forevidence,J a p a n i s o n e o ft op t h r e e tr ad in g partnersofAS EAN economies,e s p e c i a l l y Indonesiaandt h e P hilippines.
BeforeintegratingintoASEANregionaleconomies,Japanwasplayinganimportantroleintheregio naldevelopment.Inthe1970s,25%pertotalimportandexportvaluesofASEANweredoingw i t h Japan. Moreover,withlowercostinmaterialsandlabors,ASEANmarketswereattractivedestinationso f ca pitali n v e s t m e n t f l o w fromJ a p a n e s e companies.Itgeneratedworkj o b s andincreasedwo r k i n g wages, especially,w i t h hightechnologiesandh i g h - t r a i n e d employees,theyprovidedavaluableopportunityforlearningandtransferringinthisareaduringthe19 80sto1990s.T h e i n c r e a s i n g l y integrated businessn e e d a m a j o r o p p o r t u n i t y t o s t r e n g t h e n linkagesbetweenA S E A N andJapan.Thatis thereasonforraisinganeedfultalkaboutaregionalagreement.
Since2 0 0 3 , t h e governmento f J a p a n andt h e 1 0 countriesofA S E A N c o m p l e t e l y signedt h e generalf r a m e w o r k o f bilateralfreetradeagreementnamedA S E A N -
P.AttheendofDecember2 0 0 8 , t h e lastofficialroundwasfinalized,anagreementsignedamongAsiancountries,included:BruneiDar ussalam,C a m b o d i a , Indonesia,LaosP R D , Malaysia,Myanmar,Philippines,
Singapore,Thailand,VietnamandJapanhasbeenforced,supportmultilateraltradingbyreducingt h e tar iff.TheoriginobjectivesofthisFTAaretoencouragefreetradeacrosstheborderinintra- b l o c ASEANandJapan,strengthenAsiancountries,Japan economicintegration,enhancethe ireconomicintheworldmarket,aretransparentintradingprocedureandmaintainsustainabilityint h e economicarea.Itseemsamajoropportunityforhigh- techandmodernindustriesofJapansuchasautomobile,electronic,etct o enterA S E A N marketsaswellas encouragea s s e m b l y line i n regionsforJapanesefirms.
Statically,aftertradeagreementinforce,in2013,two- waytradingvolumeobtained$229billioncomparedwith$128billionin2000 Inthisyear,Japanre ported14%and15%forimportandexport valuetoASEAN,ThaiLand($22.5billion),Indonesia($3 2.2billion)andMalaysia($29.6b i l l i o n ) aret o p threeA s i a n biggestexporterst o J a p a n ( A S E A
N S t a t i s t i c s , 2014).T h e n o t a b l e productsmainlyexportedfromASEANtoJapanarefoods,ma nufacturedgoods,textiles,crudematerial.Conversely,machineryandequipmenttransportationtogat herwithchemicalandadvancedtechnologymanufacturingproductsareimportanttomajorexportfromJapantoASEANcountries.Forexample,accordingt o J a p a n automobileManufacturerA s s o c i a t i o n statistic sin2 0 1 4 , about4 7 % Japanesecars,8 0 % truckv e h i c l e s and8 5 % buseswereconsumedasfina lproductsin ASEANmarkets.
Researchobjectives
- Thesecondresearcho b j e c t i v e i s t o e x a m i n e t h e effecto f A J C E P o n sub- cataloguesi n particular:f o o d p r o d u c t s , agriculturalproducts,manufacturedproducts,M a c h i n e r y andequipmentof transportationandclothingandaccessoriesandtextile,fabric
Researchquestions
Accordingtonumerousstudiesbefore,theeffectofRTAshasnoguaranteepositiveeffecttohelpi t s mem bercountriesintegrating withthe globalmarket.Inmanycases,RTAsactually causeds o m e neg ativeeffects.Therefore,thisstudyaimstofindtheanswerstothesequestionsfollowing:
- Howthetrade creationandtrade diversioningeneraltotal exporthave beencausedbythefreetradeagreementwhichwassignedbyAJCEP toASEANmembercountries?
-Howthe tradecreationandtradediversionhavebeenaffectedbythefreetradeagreementwhichisbyAJCEP toASEANmembercountriesinthefivesub- catalogues:foodproducts,agriculturalproducts,manufacturedproducts,Machineryandequip mentoftransportationandclothingandaccessoriesandtextile,fabric?
Researchscope
2015withtotal5,920observationswithincluded09ASEANcountries:BruneiDarussalam,Cambodia,Ind onesia,LaoPDR,Malaysia,Myanmar,Philippines,Singapore,Thailand,Vietnamand1 5 biggestt r a d i n g partnersofJ a p a n 2 0 1 5 i n c l u d e : T h e U n i t e d State,C h i n a , S o u t h Korea(KoreaRep.),H o n g K o n g S A R C h i n a , Australia,SaudiArabia,T h e U n i t e d ArabEmirates,RussianFederation, Switzerland,NewZealand,UnitedKingdom,Germany,Mexico,NetherlandandJapan.
Toourknowledge,thereistherapiddevelopmentininvestingeffectofRTAsintheoreticalaswellasempirica laccesses.However,mostofthemareusuallyfocusongeneralquestions:whetherorn o t RTAshaveaff ectedtotradefloworcreatedtradecreation,tradediversion.Therearetwomainproblemsthatmanypreviousst udieshad.
Thefirstproblemisestimationchallengesofthegravitymodelwhichsolvearoundtheheteroscedas ticityandthefrequencyofzerotradeobservations.Theseproblems causechallengesi n concerningthemostsuitableestimationtechniquetoavoidbiasedandun- misinterpretedresult.T h e secondadvantageiswedoestimateregressionmodelbyusingtwosetso ftradeflowdata.T h e firstdatasetisaggregateddataisusedtoexamineforbilateraltotalexportflow.Th eseconddataseti s disaggregateddatai s o p t i m i z e d t o e s t i m a t e t h e A J C E P affectt o fives e p a r a t e s u b - categories:agriculture,manufacturing,chemicalindustry,machinery,transportationindustryandcloth ingandaccessoriesandtextile,fabric.Bytwodifferentapproaches,wecananalyzeimpactso f
Thesisstructure
This paper is structured as follows: Chapter 2 provides a literature review on trade theories related to international trade flows, including the theoretical and empirical support for the gravity model, and examines the impact of AJCEP on ASEAN members and Japan It also addresses the common issue of zero trade data frequency Chapter 3 outlines the methodology, model construction, estimation methods, and data scope employed in the study Chapter 4 discusses the results and findings derived from the regression model Finally, Chapter 5 summarizes the thesis results, offers recommendations, and acknowledges the study's limitations.
Int h i s chapter,w e w i l l summarys o m e relatedtradetheoriesw h i c h a r e p o p u l a r l y usedi n internationaltrade.Then,areviewoftheoreticalandempiricalsupportforgravity model ontradeareadded.Inaddition,weconsideraboutsomeliteraturereviewsaboutzerotradedataandthe developingofestimationtechniqueswhichsomepreviousstudiesused.
Tradetheories
International trade offers numerous benefits to countries, as outlined by various economic theories One prominent model is Adam Smith's Absolute Productivity Advantage, which posits that if labor and production factors are fixed and fully utilized, with constant technology and zero transportation costs, countries should export goods in which they have domestic effectiveness and import those where they are less effective (Howse and Trebicook, 1995) When a country holds an absolute advantage in two goods, David Ricardo's concept of comparative advantage comes into play By assessing the relative advantages or disadvantages in production, countries can produce goods with lower opportunity costs This specialization according to comparative advantage allows both nations to benefit from trade, enhancing their economic welfare.
O h l i n m o d e l e x p l a i n s internationaltradefocuso n t h e country’sresourcesabundancedifferenc esbetweencountriessuchaslabor,capitalandland.Itwillnotmeaningfulifo n l y mentionnumberof resources,forexample,labororlandresourcesthatcountryhas,sothedefinitiono f labor-andland- intensivehaveintroduced.Therefore,a c o u n t r y w i l l focusonmanufacturingproductsthatcountry hasanintensivefactor.
Tradecreationandtradediversion
Tradecreation
WhenanFTAisinforce,ingeneral,weexpectthatwithaneliminationoftariffaswellastradeincentivepo licies,FTAwillencouragetradeflowthatwouldnothaveexistedbefore.Itallowsmembercountriest oconcentrate andtradewiththeircomparativeadvantagesandgetthebenefitsfromeconomicofscale.All of thetradecreationcaseswillincreasecountry’snationalwelfare.
Tradediversion
When a trade agreement or customs union is established, trade flow often shifts from more efficient exporters to less efficient ones due to the removal of tariffs between member countries, while common tariffs remain for non-members In this scenario, countries typically import goods from the lowest-priced sources, which can lead to less efficient countries within the union exporting to member states, despite more cost-effective options outside the union The introduction of tariffs can increase costs for more efficient countries, diminishing their competitive advantage and raising prices Consequently, while trade flows may favor member countries, benefiting them from trade diversion effects, non-member countries suffer economically This dynamic results in a loss of overall welfare due to the presence of higher-cost producers gaining market access through reduced or eliminated tariffs.
Figure1showsthe welfareeffectsof joining afreetradeagreement.�and�aredenotedfordomes ticdemandanddomesticsupplyofaspecificproduct�ofacountryrespectively.
� �a n d � �are representedforexportingsupplyofproduct�fromintra-blocmemberc o u n t r i e s andextra-blocsnon-membercountries.
AccordingtoFigure1,beforeintegratingafreetradeagreement,� �+ �and� �+ �ared e n o t e d fors upplycurvesfromintra-blocandextra-blocrespectively.Assumingthatn o n - m e m b e r countriesprovideproduct�atalowerpricethanmembercountriesdo,� �lies under� �o r � �+
�liesunder� �+ �graphically.Thedifferencebetween� 0− � 0i s countryimportdemandfrom non-members.
Afterfreetradeagreementwasformed,thesupplycurve� �+ �i sunchangedbecausetariffi s stillappl iedtonon-membercountries.Meanwhile,thetariffisnolongercountedtosupplys o u r c e from� � Inthiscase,theequilibriumpriceofproduct�inthecountrywillbe� 1an d thedifferencebetween� 1−
� �,�while is, � the surplusofdomesticsmanufacturersfalls.Regardinggovernment,whenafreetradeagreementhasbeeni n forced,governmenti s n o longerco ll ec t ed t a x revenuebecausec u r r e n t l y alli m p o r t i n g valuescomesfrommembercountries,denotedbysumofarea�.Therefore,thetotaleffectoft r a d e creatio ncausedbyfreetradeagreementisthesumofareas� and�.
Regardingtradediversion,theswitchingtothehigher-costmanufacturersinintra-blocmembers insteadoflower-costfromextra-blocmembersisdenotedfortradediversiondenotedby�area.
Thetotaleffectsoffreetradeagreementinoverallwillbedeterminedbycomparingthe magnitudesoftradecreationandtradediversioneffects.Iftradecreationexceedstradediversioneffect,w elfareisenhancedduetofreetradeagreement.Oppositely,tradediversioneffectexceedstrade creation,itmeansthat countrywelfareis decreased duetothefreetradeagreement.
Thegravitymodelininternationaltrade
Theoriginofgravity model
Gravitymodelhasbeenusedasaworkhorsetoolstoanalyzetheinternationaltradeflow.Itwasdeveloped fromthelawofuniversalgravitationfoundbyIsaacNewtonin1967.Thelawstatethate v e r y two pointsattractanotheronewith aforcethat is indirectproportionto theproduct oftheirmassesandinverseproportion to thedistance betweenthem in square.
Thefirststudyusing gravitymodelwhichis derivedfromNewton’slaw ofgravitationtoanalyze internationaltradeflowsbyTinbergeni n 1 9 6 2 Fortradem o d e l , t h e bilateraltradev o l u m e betweentwocountries�,�hasbeenusedtoreplacefortheforceofgravityandeconomicsizes
� � ,� �h aveb eenu s edt o replacef ort h e masseso f� � ,� �r espectively.Generally,thegravity formulationhasbeenestablishedin thefollowingform:
Where�,� �maytakethevaluedifferentto1.Theydependontheelasticityofeconomics i z e s, ofexportingcountry,importingcountryanddistancerespectively.Incase,�=�nd�=2, ithasthesameformulationofNewton’sequation.Usually,economicsizesaredefinedasGDP,
GNP,realGDP,realGNP,incomepercapitaorpopulation.Theseessentialvariablesrepresentfors u p p l y anddemandforceofeachcountrythat determinecountry’stradevolume.
Regardingdistancevariable,itisdefinedbygeographydistancebetweentwoeconomicshubsorcapitals countedinlandmiles.Tinbergenstatedthatdistanceisnotonlyaproxyrepresentforrealdistancebutalsom aystandformanyothermarketfactorswhichinfluencetotradevolumesuchastransportationcost,transi tcost,communicationexchange costor even culturecost.
Usually,ineconomic regression,thesimplestgravity modelisestimatedunderOLSbytaking logarithmequation(2)andaddingerrorterm� �� Thecoefficientresultobtainedwillbeinterpretedaselastic itybecausetheregressiontook the doublelogform:
Iftheeconomyofcountry�/�increasesbyonepercent, thetradevolumebetweentwocountrieswillincrease� �percentrespectively whileotherfactor/ sareheldconstantly Similarly,t ra de volumewillreduce�percentifthedistancebetweentwoc ountriesincreasesbyonepercent.Allinthecases,errorterms� ��i s supposedthatitisindependentandnor mald i s t r i b u t i o n
Theoreticalframework
Theoreticalsupportandtheoreticalequation
Thefirstn o b l e w o r k i n applyinggravitym o d e l tointernationaltradebyTinbergen,1 9 6 2 Ho wever,itwasstillmissingpowerfultheoreticalapplicationbasicandstoodoutsideofmainstreamduet othepersistentperceptionofaphysicalgravitymodelmorethananeconomicmodel.Thefirstimporta ntcontributionwehavetomentionistheworkofAnderson(1979).Heb u i l t g r a v i t y m o d e l b a s e d o n Cobb-
The Douglass-Cramer-Smith-Evans (DCSE) preference function operates under several assumptions: each country specializes completely in producing differentiated goods based on their origin, known as the Armington assumption, while consumer preferences are homothetic and uniform across regions Additionally, there are no transport costs, tariffs, or trade barriers This aligns with the idea that gravity models depend on the share of national income allocated to international trade, which can be estimated through population and income functions To address the Armington assumption, researchers like Bergstrand (1985, 1989) developed a gravity model influenced by Paul Krugman's (1980) concept of monopolistic competition, indicating that countries specialize in production while consumers exhibit diverse preferences, leading to increased trade.
10 differentkindofcommoditiesfromtheidenticalcountry.Deardorff(1998)andKrugman(1985)contrib utedt h e newt h e o r y t o g r a v i t y m o d e l byappliedliteraturecomparativeadvantageofHecksch er-
Ohlintheory.EatonandKortum( 2 0 0 2 ) derivedg r a v i t y m o d e l byu s i n g Ricardianm o d e l , H elpmanetal (2008)addedfirmheterogeneitytoobtainthemodel.
Recently,manyresearchersd o t h e theoreticalcontributioni n g r a v i t y m o d e l byi m p o r t a n t l y concerningtheusageofvariablesandspecification.Inthissection,saythankstothecontributiono f A ndersonandvanWincoop(2003)whodevelopedmonopolisticcompetitionframeworkbasedo n theArmin gtonassumptionandconstantelasticityofsubstitution(CES).Assumethatcustomeru t i l i t y a m o n g c ountriesarei d e n t i c a l andh o m o t h e t i c , t r a d e g r a v i t y e q u a t i o n wasspecifiedasbelow:
(4) Where𝑉 ��i s bilateraltradevolume,� �, � � ,� 𝑤i s ��incomeof country�,�andglobalincomeres pectively,� �� ,� ��a r e bilateraltradebarrierbetweencountries�,�,denotesallbilateraltrade resistanceandassumedequally.Theyincludedistanceandsomebinaryvariablessuchascommon border,colony,tradeagreementetc.𝜎 istheelasticityofsubstitution,� � � �, ismultilateraltraderesis tanceand� � ,� �a reconsumerpriceindexofcountry�,�respectivelyandhaveafunctionas below:
� �i s theshareofcountry�incountry�’sconsumption,� � ,� �are exporterandimporterpricerespectiv ely.
Equation(1)and(2)showclearlythatanychanginginbilateraltraderesistance� ��in thenumerator
To estimate the border effect on international trade, the Non-Linear Least Squares (NLS) technique was employed The study consistently demonstrated that bilateral trade costs are influenced by factors such as distance, landlocked status, shared borders, and common language However, it failed to address several critical assumptions Firstly, the assumption of a two-way systematic trade cost between countries is flawed if bilateral or multilateral trade agreements exist Secondly, the study overlooked differences in customer preferences by assuming a balanced trade volume between the two countries Lastly, the analysis was constrained by a singular time period, neglecting the importance of a time-varying estimator in the regression model.
(2007)developedthemodelbyextensiontopaneldataandusedtime- varyingfixedeffecttoeliminateb i a s estimationcausedbytime- varyingtradecostvariables.BaldwinandTaglioni(2006)regressedt h e m o d e l w i t h t h e samem e t h o d whenchoosingcounty- pairf i x e d effectt o reducee n d o g e n e i t y biascausedbyFTAdummyvariable.
Fromthisstudy,manyauthorsuseitinthedifferenttypeofeconomicissuesastheworkhorseduet o i t s a b i l i t y toc o r r e c t l y estimatebilateralr e l a t i o n s h i p , f o r e xa mp le , immigrant,f o r e i g n directinvest mentasw e l l ast r a d e flow.T h i s m o d e l hasfirsttheoreticalclarificationpresentedbyAnderson(1 979)andtheoreticalbasiclaterprovedbyHelpmanandKrugman(1985),Bergstrand(1989),andDeardorf f( 1 9 9 8 ) Inadditional,t h i s m o d e l wasappliedt o m a n y studies,canbereferredto thecollectivepaperbySenandSmith's(1995).
Inparticular,therearevariousempiricalstudieswhenappliedthismodeltointernationaltradeflow.Sol oagaandWinters(2001),AntonucciandManzocchi(2006)usedthismodeltoexaminet h e influenceo
12 fcountry’scharacteristicssuchasborder,geography,distancecombinedwithtradeagreementtotrade flow.
EmpiricalsupportforeffectofFTAtoASEAN
EmpiricalsupportforeffectofAFTAtointra-bloctradeflow
ThefirstFTAsignedbetweenASEANcountriesisAFTAin1992.Theoriginmembersincludes i x of tenASEANcountries:BruneiDarussalam,Malaysia,Philippines,Singapore,andThailand.T h e restof fourcountrieshavejoinedinVietnam(1995),LaoPRD(1997),Myanmar(1997)andCambodia(1999).U nderthisagreement,thetariffratewasreducedupto99percentforsixorigincountriesand95percentforre stoffourcountriesby2010.Atthepresent,eliminationoftariffunderAFTA hasbeen completed.
During the initial phase of AFTA implementation, various studies indicated that the impact on trade creation would be minimal DeRosa (1995) utilized a CGE model to estimate AFTA's effect on intra-bloc trade, revealing that the Most Favored Nation (MFN) status contributed more to trade liberalization than AFTA itself Similarly, Frankel and Wei (1995) employed a gravity model for ex-ante analysis, concluding that trade flows among ASEAN countries were still influenced by external factors beyond ASEAN relations Furthermore, Endoh (1999) was the first to introduce and apply the concepts of trade creation and trade diversion in assessing the effects of an FTA, providing insights into the trade dynamics during the sample period from 1960 onwards.
Sincethe2000s,adevelopmentofthemethodologytoestimategravitymodelhasbeenraised.Forevidence,Sol oagaandWinter(2001)hasappliedTobitmodeltoevaluatetheeffectofsomemajorPTAsonbilateraltradeto ASEANcountries.Accordingtohisresult,thecoefficientoftradeintra- b l o c wasinsignificantandnegative.H o w e v e r , o u t s i d e A S E A N tradeweres i g n i f i c a n t l y e ncouraged.Followingn e x t movemento n methodology,C a r r è r e (2006)appliedHausmanandTaylor m e t h o d byusing instrumentvariableandp a n e l data,s h e s h o w e d a p o s i t i v e tradef l o w w i t h i n A S E A N andreducingimportfromrestof theworld.
The growing interest in the development of methodology related to the ASEAN Free Trade Area (AFTA) has led to a moderate increase in studies examining its impact on the economies of ASEAN countries Elliot and Ikemoto (2004) utilized a gravity model to analyze the effects of AFTA on trade creation and diversion, discovering that ASEAN nations benefited from both aspects, with positive and significant coefficients Similarly, Bun et al (2009) implemented two types of FTA dummy variables to assess AFTA's impact, finding that it positively influenced trade over the collected data period They emphasized the importance of incorporating unobserved explanatory variables into estimation models to effectively capture trade trends influenced by tariff elimination under AFTA.
Research on the effects of the ASEAN Free Trade Area (AFTA) has examined tariff reduction progress under the Common Effective Preferential Tariff (CEPT) scheme Pelkmans-Balaoing (2007) analyzed trade data from 2001 to 2003 and found that AFTA had minimal or no impact on intra-bloc trade, while positively affecting certain products with a preference margin exceeding 25% However, the study suggested that the costs of implementing AFTA may outweigh the benefits when the tariff difference between the Most Favored Nation (MFN) and AFTA is marginal Additionally, Okabe and Urata (2013) explored the effects of tariff elimination under CEPT, focusing on 52 products from ASEAN member countries since 1980.
2010.T h e y f o u n d t h a t AFTAh a s createdtradecreationeffect.However,t h e magnitudeofeffectton ewmemberssuchasCambodia,Lao,Vietnamwaspuresmall.Thisresultcouldbeexplainedbythesmalls hareofnewmembersaswellasasubsequentscheduleoftariffreduction.Moreover,theimpacttotradeflo wisnotextremelystrong.Fromtheseargumentscanconcludethattariffreductionisnotanessentialmeas urementtopromoteregionaltradeflow.Top r o m o t e ASEANtradeflowaswellasincreasewelfar eforallmembercountries,otherfactorssuchastradefacilitation,eliminatingnon- tariffbarriers(NTMs),equalizingrulesoforigin(RoO)aswellasenhancingAFTAutilization should beconsidered.
EmpiricalsupportofeffectofASEAN+ 1FTAs
Recent studies have raised a critical definition regarding free trade agreements (FTAs) and regional trade agreements (RTAs), questioning whether they act as a "noodle bowl effect: stumbling or building block?" This definition accurately reflects the status of trade agreements within ASEAN Since the mid-2000s, ASEAN has established multilateral FTAs with six major economies: Australia, New Zealand, China, Japan, Korea, and India, alongside numerous bilateral FTAs Notably, Japan has signed bilateral FTAs with Indonesia, Malaysia, the Philippines, Singapore, Thailand, and Vietnam, while India has agreements with Malaysia and Singapore, and China has formed bilateral FTAs with Singapore and Thailand.
The integration trend in international trade, particularly with the enforcement of Free Trade Agreements (FTAs) related to ASEAN, has prompted various ex-ante studies to assess their impact Estrada et al (2011) utilized a simulation analysis through a Computable General Equilibrium (CGE) model to evaluate the welfare effects on member countries from FTAs with China, Japan, and Korea, revealing positive expectations and interest from these nations Similarly, Shenge et al (2012) employed the CGE model to predict the impact of ASEAN+1 FTAs, using a gravity model analysis covering the period from 1980 to 2008.
China( A C F T A ) hascausedm a s s i v e i m p a c t t o t r a d i n g flow.Especially,itaffectedtoadjacent productionlinkageinternationallyandpositiveimpactspreadingamongASEANcountries.Banoetal. (2013)analyzedtradeeffectcausedbyASEAN-Australia-
ASEANFTA(AIFTA)focusingonfisherydivision,hequotedthattheFTAhasimprovedtradebytari ffreduction,particularly low- developedcountries.Okabe(2015)usedCGEmodeltoforecasttheeffectofcurrentASEAN+1FT AsincludeACFTA,AKFTAandAJCEP.ShefoundthattradecreationhasbeencausedbyACFTA,AK FTAintwosub- categories:industrialsuppliesandcapitalgoods.Misa(2015)usedthesampledatafrom2002to2012a ndgravitymodeltoestimatetheeffectofASEAN+1FTAs.TheyfoundthatASEAN-
KoreaFTAhascausepositivetradecreationinindustrialsuppliesandcapitalgoodsamongmembercountr ies.Inaddition,ASEAN-Chinaalsofacilitatesconsumptiongoods.Incontrary,ASEAN-
Ingeneral,theimpactofAJCEPappearedlimitedlyatthemomentofex- postanalysis.Meanwhile,m o s t ofthestudiesaboutAJCEPshowanegativeoruncleareffects.Thereaso ncouldbeusedto explainthisresultwastariffreductionscheduleandRoOcertificationwhileotherformerFTAshaveb eenimplicatedinthelongertime.Thatisoneofthereasonencourageusre-estimatetheeffectofAJCEP toASEAN membersbyusingbytheex-postanalyst.
Zerotradedataproblem
Zerotradeflowbetweenagivenpairofcountriesisaproblemhasbeenwidelydiscussed.Becauset h e tradit ionalwayusuallyusedtoestimategravitymodelistakinglogarithmsleadingtodropoutz e r o valuet o t h e d a t a s e t However,ze ro da ta isn o t c o m p l e t e l y mean n o n - t r a d e b et we en t w o countries.Itmayimply thet r a d e v ol um ew it h v e r y smallflowso r evenm i s s i n g o r lo ss whenr e p o r t i n g processinmanycases.
Inreality,severalalternativezeroapproacheshavebeendiscussed.Thefirstmethodistruncatez e r o s dataandstillestimatelog- linearbyOLS.Bythismethod,zerotradedatawillbecompletelydeletedfromt h e m a t r i x T h e seconds o l u t i o n i s t h e c e n s o r i n g m e t h o d bys u b s t i t u t i n g a s m a l l constantvolume,forexample,on edollarbeforetakinglogarithmstradevaluewhenestimating g r a v i t y modelinlevel.Howeve r,thesemethodscanleadtoinconsistentestimatesanddistortthesignificantresultBurgeretal.,2009,G omez-
Herrera,2013.Moreover,accordingtoFlowerdewandAitkin(1982)indicatethattheresultissensitiv ewhenreplacingadhoczerodataforsmallvalueisnotguaranteedforanexpectedregressionresult,can- notbeavoidedinconsistentestimation.EichengreenandIrwin(1998),Heckman(1979),Helpmanetal., (2008)debatedthati n casezerodatadonotfollowtherandomdistribution,deletingzerodatafromtr adematrixcanleadto lossvaluableinformationandcreatebiasresults.
Witht h e sameresult,LinderandGrootp o s i t t h a t applyingtruncationo r censorm e t h o d whendealing withzerodatamayleadtomisunderstandingbilateraltradepatterns.Becausemaybeduet o fardistance ,lowlevelinGDPornon-linkageincultureorhistorical,non- profitableintrade,firmsm a k e d e c i s i o n reducingt r a d e o r n o n - t r a d i n g , t h e n w e eliminatez e r o dataw i l l causeunderestimatingcoefficients.Therefore,the reisanattentiononfindinganappropriatetechniquet o dealwith zerodatarecently.
Early studies, such as those by Rose (2004) and Andersen and Marcouiller (2002), have employed the Tobit model to analyze trade data, particularly when observations are rounded to zero or when trade flows are below positive values However, Linder and Groot (2006) challenged the appropriateness of using the Tobit model for zero trade flows, arguing that desired bilateral trade cannot be negative unless one or both countries have a GDP of zero, which is unrealistic They also contended that censoring trade flows at any positive value, even as low as $1, is inappropriate since zero trade may result from actual economic decisions rather than measurement errors Consequently, reliance on this method could significantly distort regression results, as noted by Frankel (1997).
Recently,HeadandMayer(2013)hasproposednewapproachwhendealingwithasetofdatawith2 5 percen tbygravitymodel,basedonEatonandKortum(2001),namedEKTobitmodel.Bythis method,theywillreplaceallzerotradedatafromcountry�toalldestinationcountry�byminimum leveloftradedatarecorded.Thismethodhastwoadvantages.First,withoutanycriteria,wewill easilycollectminimumtradevaluewhichusedtoreplace.Second,easilyestimatethemodelby usingcommand�����𝑔inStata.
The Pseudo Maximum Likelihood (PPML) method addresses issues related to logarithm transformations and zero trade flow data, particularly in datasets where 62% of the entries reflect zero trade By regressing the model at the level rather than applying logarithmic transformations, PPML effectively manages observed heterogeneity and provides a straightforward approach for estimating models with zero trade, minimizing bias compared to other estimators However, critics such as Martin and Pham (2008) and Burger et al (2009) have highlighted limitations of the PPML method, particularly its failure to account for unobserved heterogeneity In response to these concerns, Head and Mayer (2014) introduced a new approach called the Multinomial Pseudo Maximum Likelihood method.
MaximumLikelihood(MPML).Inthismethod,dependentvariablewillbevaria bleandestimatebyStatacommand������alongwithfixedeffect.
(2009)t o t a k e careunobservedh e t e r o g e n e i t y a r e NegativeBinomialPseudoMaximumLikelih ood(NBPML)andZero- inflatedPseudoMaximumLikelihood(ZIPML).However,theyp o s i t t h a t t h i s m e t h o d i s n o t we ll-suitedf o r t h e casesa numberofzerotradedatapredictedbythemodelislowerthananumberofzerotradedataflowobserved.
Insum,accordingtolistedreview,eachmethodhasownprosandconsandthebestmethodhasn o t yett obedefined,remainindebateandarenotanuncleardecision.However,inthispaper,dataisupdated fortimeperiodrecentlyfrom2000-
2015,thetradezerodatawererecordedat15percentageandconcentrateondataofCambodiaandBrunei In2015,thetotalexportvalueofCambodiaandBruneitotherestoftheworldwerereportedat0.42per centand0.34percentpertotale x p o r t valueo f th e countriesi n t h i s sample.Therefore,w e choset h e si m pl es t estimationm e t h o d incaseofpresentzerotradeis droppingthem out thedata.
Chapterremark
Theapplicationofgravitymodeltoexplainthetraderelationsintheinternationaltradebecomev e r y popular.Itcouldbeprovedbyarigorousdevelopmentinrelatedtheoreticalsupportaswellasempiricalco ntribution.However,inspiteofthepopularity,itremainssomequestionsaboutthea d e q u a c y ofm o d e l specificationandproperestimationtechniquewhichs h o u l d b e usedf o r a consistentestimationt odealwiththe frequencyofzerotradedata,occurringwhenabout50%tradedatah a v e f o u n d i nz e r o Eachm e t h o d h a s o w n p r o s andconsa n d hadn o t b e defined.Therefore,withinthisstud ywithabout15%zerotradedata,wewillusethesimpletechniquetotacklewithit.However,ouranalyst willbewiderbycomprisingsomeregressiontechniquesinorderto obtainsuitableestimatorforourdatabase.
Int h i s chapter,w ewi ll diagnoseth e g r a v i t y modelappliedt o internationaltrade frommultiplica tivetologarithmicform.Wedid itbyaddingvariableswhicheffecttobilateraltradetog r a v i t y modeltoinvestigatetheimpactofAJC EPtoASEANaswellasJapan.Then,weprocessbyusingav ar ie ty ofes ti ma t io n techniquessuchas OLS,Fixedeffectm o d e l (FEM),Randomeffectmodel(REM),Hausman-
Taylorestimator.Inaddition,anoverviewofdatascopeanddatasourcesarealsomentioned in thisChapter.
Modelspecificationandvaliditytesting
Model specification
Followingtothegravitymodelofinternationaltrade,thefunctionofthetotalbilateralvolumeof exportofacoupleofcountries� ��i s includedtheirGDPs,population,distanceand� ��denotes asetofdu mmyelementswhichencouragingordiscouragingbilateraltradeflowincludedBorder,
Language,Colony,Land-lockedness,Freetradeagreement.
Tradeflow(� ��𝑡 ):inthegravitymodel,weemployvariabletradeflowbyexportvolumefromexport ingcountry�toimportingcountry�attimetatcurrentUS$.
GPDatcurrentUS$isthefirstindependentvariableswhichcollectedfrom thedatabaseofWorld tradeorganizationWTO.Thisvariabledenotedfortotalvaluewithinacertaintime�atcurrentUS$ ofallfinalgoodsandservicesproducedinacountry.Ingravitymodel, itincludes� � � � 𝑡 a nd
��� �𝑡whi chconsider�isexportingcountryand�isimportingcountry.Accordingtoutilitytheory, whentheincomeandoutputofacountryincrease,itwillincreaseconsumerdemandforgoods andservice,leadingtoincreasingproductionandexport.NellisandParker(2004),���presents forcountry’sincomeandpurchasing power,therefore,GDPwillhavethepositivesignwithtotal importp l u s exportvolume.However,Basat( 2 0 0 2 ) indicatedt h a t positiverelationo n l y w i t h middledevelopmentcountries,thereis noevidenceforlowandhighdevelopment countries.
Thepopulation is t he secondindependentvariable, dividedto t w o p o p u l a t i o n variables,
The population of the exporting country, denoted as \( t_d \), and the importing country, denoted as \( t_w \), are measured in millions It is anticipated that countries with larger populations will exhibit greater demand for both imports and exports However, Aiken (1973) suggested that in populous countries, the ratio of domestic market demand to foreign market demand exceeds one, implying that countries with smaller populations may achieve higher export volumes Additionally, Oguledo and Macphee (1994) indicated that the expected impact on trade can vary, showing either negative or positive signs depending on the level of integration of each country.
WeighteddistanceisthethirdindependentvariableiscalculatedbyMayerandZignago(2005), basedoninspiredideaofHeadandMayer(2002)iscalculatinggeographicdistancebetweentwo countries�and�bybiggestcitiesdistance,innercitiesdistancebeingweightedbypopulationr a t i o of thecityto thetotalcountry’spopulation.
Thereasonforusingthismethodinsteadofsimplegeodesicdistancewhichcalculatedbyusinglongitu desandl a t i t u d e s i s t o avoidovero r underestimatet h e effecto f theborder.T a k i n g ane x a m p l e oftradingvolumebetweenVietnamandChina,oneofthefactorincludetotradevolumei s t h e compari sono f domesticstransportationcosti n s i d e C h i n a i n t e r n a l l y andinternationaltransportatio ncostbetweenChina-Vietnamandpopulationinthesecities.Thegeneralfunction wasdevelopedbyHeadandMayer(2002)tocalculatetheweighted- distancefromcountry�t ocountry�is:
��� � ,��� �are populationagglomerationofcities�,�belongstocountry�and� respectively
AccordingtoBakman,Garretsen and Marrewijk(2001),thefurther distance,thehigherin thecostoftransportation,culturedifference,thehighercostoftrade.Inotherwords,therelationbe tweendistanceandtradeflowareanegativecorrelation.
Languagei s t h e firstd u m m y variable,usedasa measurementt o o l t o compareculturefacture differences.Accordingtothetheoreticalsupport ofLinnermann(1966),HackerandJohansson(2 001),thevolumeoftradebetweentwocountrieswasinfluencedbythecommonlanguageasthecommunicat ionbarrier.Iftwocountriesspeakthesamelanguage,theycaneasytocommunicate andreducetransactioncost.Therefore,therearetwobinaryvaluesof������variable,equal1 if twocountrieshavesamelanguage andzero ifotherwise,thecoefficientisexpectedpositivesign.
Apairofcountriessharingthecommonborderwitheachother,theywillhavealowercostof transportation,leadinghighertradeflow,Bakman,GarretsenandMarrewijk,
(2001).Therefore,t h i s dummyvariabletakesthebinaryvalue,equal1ifapaircountriessharingaco mmonborderand0 ifotherwise.Thecoefficientisexpectedpositivesign.
Iftwocountrieshavebeencoloniesofeachotheroracommoncolonizer,thisdummyvariable takesvalue1 i f t w o c o u n t r i e s sharea commonborderand0 ifo t h e r w i s e T h u s , apriori,t h e co efficientisexpectedpositivesign.
Ifatleastoneofcountriesinapairistheland- lockedcountries,thisdummyvariablewilltakethevalueat1and0ifotherwise.Variousstudiestriedtohig hlightthecostofacountrywhenbeinglandlocked.Mackellar,WorgotterandWorz(2002)indicatethatb einglandlockedconnectedwithincreaseimport priceaswellasreduceinexport revenueduet o higher inintermediateexport services,Stone (2001).Therefore,thecoefficientisexpectednegativesign.
1ifbothcountries�and�arebelongedt o AJCEPafter2008andzeroifotherwise.���_2 ��𝑡= 1ifexp orter�isthememberofA J C E P officiallyinyeartwhiledestinationimportingc o u n t r y�isnotbelongt o A
J C E P andequalzero otherwise.���_3 ��𝑡= 1i ncaseexporter�i sfromextra- blocAJCEPmembersinyeartandexporting todestinationintra- blocAJCEPcountry�andgetzerovaluei f otherwise.
Ifintheregressionresult,thecoefficientof���_1��𝑡o b t a i n sapositiveandstatistically,itimplies thattradecreationeffectisgenerated.Inotherwords,intra-regionaltradeflowhasbeenpromoted morewhenthe freetradeagreementisin forceintime tthan normal.Similarly,astatistically significantandpositivecoefficientof���_2 ��𝑡i s impliedthatFTAhascreatedtradecreationeffec tintermofexports.Expressinginthedifferentway,exportactivitieshasbeenswitchedfrom
AJCEPmembercountriest o e x t r a - b l o c A J C E P countriesbyr e g i o n a l integrationagreement. Contrarily,a s t a t i s t i c a l l y significantandn e g a t i v e signo f coefficiento f���_2��𝑡expressesa dec reaseinexportsvaluefrommemberAJCEPtonon-memberAJCEPcountriesandimpliesan exporttradediversioneffect.Aregressionresultshowsapositivesignandstatisticallysignificant thecoefficientof���_3 ��𝑡 ,itmeansthatatradecreationeffectintermsofimportfrome x t r a - b l o c tointra-blocAJCEP.Inthiscase,weconcludethatAJCEPhasencouragedexportflowfromnon- membercountriestomembercountries.Contrarily,wewillconcludethatatradediversioneffect inrelationtoimportshasbeencreatedifthecoefficientof���_3 ��𝑡is negativeands t a t i s t i c a l l y significant.Themodel should be:
Modelvaliditytesting
Therearemanytechniquestoestimatetogravitymodelwhicharementioned.Ifignoringpanel dataandassumingt h a t errorterm𝜇 n o tcorrelatedw i t h t h e dependentvariableandidenticaldistribu tionandconstantvariance�~�(0,𝜎 2 ),wecanregressmodelbyPoolOLStechnique.
However,poolOLSestimator obtainedthe biasedresultdueto omittedvariables.
To achieve a more accurate estimation of export direction and time effects, it is essential to consider unobserved time-invariant variables such as history, culture, and education, which influence export trends These unobserved variables may correlate with explanatory variables Additionally, time effects indicate that factors like export potential, trends, or business cycles can impact the direction of exports Therefore, incorporating these effects into regression models is necessary We can employ two models: the fixed effect model and the random effect model, and use the Hausman test to determine which model is more appropriate.
The model under consideration incorporates both time-varying elements, such as GDP and population, as well as time-invariant factors like bilateral distance, language, borders, colonies, and landlocked status Econometrically, while the Fixed Effect Model (FEM) cannot estimate time-invariant variables, the Random Effect Model (REM) can provide regression results for both time-varying and time-invariant variables However, REM assumes that unobserved individual effects are included in the error terms and are not correlated with the model's variables Consequently, this assumption can lead to inconsistent results in many cases due to the potential correlation between unobserved effects and explanatory variables.
ThealternativemodelsuggestedbyHausmanandTaylor(1981)tofixthedisadvantagesofFEMandREM namedHausman-Taylormethod.Itisarandomeffectincludinginstrumenttechniqueto eliminatethedisadvantageofREMwhenfixingthecorrelationbetweenincludedvariablesinthem o d e l anderrorterm.Basically,ahybrid ofFEMandREMtakesthe followingequation:
Inwhich:� ��𝑡i s fordependentvariableforcountry�tocountries�intime�.X- matrixincludestwotypesofvariables:�′ 1��𝑡are thetime- varyingindependentvariablesanduncorrelatedw i t h
� ��𝑡 ,�′ 2��𝑡are thetime-varyingindependentvariablesandcorrelatedwith� ��𝑡 Z- matrixincludes�′ 1��are time-invariantindependentvariablesanduncorrelatedw i t h � ��𝑡 �
Hausman-Taylori s t h e m e t h o d i n whichu s i n g variablesfromX - m a t r i x correlatedw i t h�t oproduceunbiasedestimatesof�’sbydeviationfromindividualmean saswellasinstrumentsf o r forZwithcorrelatedwith�byusingindividual.Inaddition,accordingt oBaierandBergstrand
(2002),FTAdummyvariablesperhapsendogenousduetothecorrelationofunobservedvariableso r om ittedvariablessuchasculture,politic,history.Therefore,itwillcausebiasedestimation.Wooldridge (2000)suggested to usingFTAdummyvariablesasinstruments to fixendogeneity.
Tocheckt h e c o n s i s t e n c y ofHausman-Taylorestimatort o u s e H a u s m a n t e s t f o r o v e r - identification.ItwillcomparetheestimatorbetweenHausman-
Taylormethod,wewillregressmodelinmanymethods: Pool OLS,FEM,REMandHausman-Taylor.
Dataanddatasources
2015withtotal5,920observationswithincluded09ASEANcountries:BruneiDarussalam,Cambodi a,LaoPDR,Malaysia,Myanmar,Philippines,Singapore,Thailand,Vietnamand15biggesttradingpart nersofJapan2015include:T h e U n i t e d State,C h i n a , S o u t h Korea( K o r e a R e p ) , H o n g K o n g S A R C h i n a , Australia,SaudiArabia,TheUnitedArabEmirates,RussianFederation,Switzerland, NewZealand,UnitedKingdom,Germany,M e x i c o , NetherlandandJ a p a n RegardingIndonesia,u n t i l 2 0 1 6 r e c o r d ,
IndonesiahasnotyettofinalizedocumenttojoinintoAJCEP,therefore,Indonesiaisnotincludedi n selecte ddataofASEAN.
ComtradedatabaseataggregatedanddisaggregatedlevelandareextractontheStandardinternationaltrad eclassification(SITC),revision3andstayo n nominaldatatoeliminateerrorinmeasurement(Baldwin
&Taglioni,2006).Disaggregateddatawillbespliti n t o fivesub- sectionsseparately whichabsorb thehighesttariffreduction onA J C E P agreement:agricultur alproducts(sumexportvalueofSITCNo.0,1,2,4,minus 27,28), manufacturedp r o d u c t s ( s u m o f e x p o r t valueo f SITCN o 5 , 6 , 7 , 8 e x c l u d i n g 6 6 7 p l u s 6
Machineryandequipmentoftransportation(SITC7)andclothingandaccessoriesandtextile,fabric(SIT C84 plus 65).
DataonGDPandpopulationareobtainedfromWorldBankdevelopmentindicatorsdatabase.Theuniverseof RTAsiswithdrawnfromthe“RegionalTradeAgreementInformationSystem”(RTA-
IS)databaseoftheWorldTradeOrganization(WTO).Geographicandculturaldataondistance,commo nlanguage andadjacencyarefromCEPIIdatabase.
Thissectionwillbedividedintotwosub- section.Thefirstwillbestatisticsofvariablesandtheireconomicsmeanings.Thesecondonewill beeconometricsresultsanditsmeaning.Inthesecondsub- sectionw i l l b e a d d e d s o m e t e s t i n g t o f i n d o u t t h e bestm o d e l e s t i m a t e d andreje ctt h e u n s u i t a b l e models.
Descriptivestatisticsofvariables
The analysis of full panel data from 24 countries, encompassing 5,920 observations from 2000 to 2015, reveals that total export values account for 0.34% of the total gross domestic product (GDP) The minimum export value is reported as zero, which may stem from missing data, rounding errors, or decisions by firms not to engage in exports with each other Within the total export trading value, manufactured goods dominate, comprising approximately 77.80% This category is further divided into subcategories, with chemical products making up 8.64% and machinery and transportation equipment at 48.72% Agricultural products represent 6.94% of total exports.
Dummyvariable���_1takesthevalue1ifbothimportingandexportingcountryarebelongto freetradeagreementA J C E P from2 0 0 8 T h e effectivet i m e o f eachc o u n t r y w i l l b e adjusted accordingtosummaryinTable10.Thedescriptivestatisticsofothervariablesincase���_1=1 is shown onTable3
Dummyvariable� � �_ 2takest h e value1 i f b o t h exportingc o u n t r y i s belongtoAJCEPandimpo rtingc o u n t r y i s n o t belongt o freet r a d e agreementA J C E P from2 0 0 8 T h e d e s c r i p t i v e statisticsofothervariablesincase���_2=1isshownonTable4
Table4:Descriptivestatisticsofvariablesifexportingcountryis belong toAJCEPand importingcountriesisnotbelongs toAJCEPfrom2008
Dummyvariable���_3takesthevalue1ifbothexportingcountryisnotbelongtoAJCEPandimport ingcountryisbelongtofreetradeagreementAJCEPfrom2008.Thedescriptivestatistics ofothervariablesincase���_3=1isshownonTable5
Table5:Descriptivestatisticsofvariablesifexportingcountryis notbelongtoAJCEPand importingcountriesis belongs toAJCEPfrom2008
Testingmulticollinearity
Thepresentofmulticollinearitycouldhavetheimpactonregressionresult.Therefore,weneedtodetectmu lticollinearitybetweenthevariablesbyusingPearsonandthecalculationofVarianceInflationFactor (VIFs).Table6 shows thecorrelationbetweenvariablesinthisstudy.
Testing results indicate a positive correlation between exports and several key variables: GDP of exporting countries (0.3068), GDP of importing countries (0.346), population of exporting countries (0.2532), population of importing countries (0.2251), common border (0.0008), and common language (0.1272) The positive relationship between exports and the GDP of importing countries suggests that wealthier nations tend to import more goods Furthermore, a common language enhances communication, thereby reducing transaction costs, while sharing a common border fosters cultural proximity and lowers transportation costs, ultimately predicting a higher trading flow in exports.
Contrary,thecolony,distanceandland-lockedpresent thenegativerelationwith export(- 0.0218,0 0 6 1 7 , -
0.086).Theyimplythatiftwocountriesarecolony,fardistanceoroneoftwocountriesi s land- lockedintendtoreducein export.
RegardingFTAvariables,���_2and���_3showthepositiverelationwithexportwhile� � �_ 1 showsnegativerelation.ItimpliesthatAJCEPencouragestradingbetweenAJCEP‘smemberto partnersofJapanandviceversaandhavenotyettoboosttradingeffectwithinAJCEP’smembers.
Populationi mportcountr y Export Colony Distance Border
Regressionresult
Comparisonofestimatorproperties
Accordingtopreviousstudies,OLSusuallyincludedheterogeneitybias.Iftheheterogeneityi s no taddressedin themodelands ome - ho w variableswhichincludedin themodelarecorrelatedwith,theestimationwillbebiased.This biascanbedetectedonFigure2bytestingresidualagainst totalexportinlevel.Theresultpresents thatatthelowleveloftotalexport,theresidualisnegative,meanthatexportisoverestimatedinothe rwords.
Todeterminewhichi s thebetterm od el betweenFixedeffectandRandomeffect,w e w il l r u n Hausman testtocheckitwiththenullhypothesisthatthecoefficientsofrandomeffectestimatoraresameastheones
31 offixedeffectestimated.Acceptingthenullhypothesismeansrandomeffects h o u l d b e used.Theresult ofHausmantestis shown in followingTable7
Accordingtotestingresult,significanceP- value,Prob>chi2smallerthan0.05,weshouldstronglyrejectthenullhypothesisandconcludethatfixedeffecti sthemorecorrectmodelshouldbeappliedoverrandomeffect.
Fixedb =consistentunderHoandHa;obtainedfromxtreg Random(B)=inconsistent underHa,efficientunderHo;obtainedfromxtregTest:Ho:differenc eincoefficientsnotsystematic chi2(7)=(b-B)'[(V_b-V_B)^(-1)](b-B)
= 143.76Prob>chi2= 0.0000(V_b-V_Bis not positivedefinite)
Regressionresults
Table8showstheoverviewaboutnumericregressionresult.Eachcolumnpresentseachtypeofestimator:OLS,fixedeffectmodel,randomeffectmodel,Hausman-Taylorrespectively.
Table8:OLS,FEM,REM andHausman-Taylor regressionresultfortotalexport
(4) Hausman- Taylor LogGDP exportingcountryLogGDP importingcountry
The regression analysis indicates a strong positive relationship between GDP and export value, confirming findings from previous studies by Diks and Panchenko (2005) and Kim and Link (2009) Specifically, a 1% increase in the GDP of an exporting country is associated with an approximate 0.623% rise in export value Additionally, the export value is influenced not only by the exporting country's GDP but also by the GDP of its trading partners For instance, a 1% increase in the GDP of the importing country is expected to lead to a 0.696% increase in the exporting country's export value This suggests that larger economies tend to produce more goods for export and have a higher demand for imports, highlighting that bilateral trade among countries is significantly driven by economic growth.
However,therelationbetweenpopulationandexportiscomplexanddifferentamongestimators.Forexp ortingpopulation,whileOLSshowthenegativerelationandinsignificant,randomeffectandHausman- Taylorestimatespositiveandsignificantresults.Mostof thepopulationvariableofi m p o r t i n g countryareinsignificantinfixedeffectandrandomeffectes timatorswhileOLSandHausman-
The relationship between population size and economic growth is complex and remains a topic of debate among scholars Thirlwall (1994) and Becker et al (1999) suggest that a larger population can yield both positive and negative effects on exports and economic growth On one hand, a substantial population can create a robust labor force, enhance domestic consumption, and foster competition that leads to improved skills and technological advancements Conversely, it can also result in challenges such as food scarcity, savings development issues, and unemployment (Meier, 1995) Additionally, studies by Simon (1992), Kelley, and Schmidt (1996) as well as Ahlburg (1996) indicate a lack of statistical evidence linking population growth directly to exports and economic growth Regression analysis in this paper reveals both positive and negative correlations between population and export activity, highlighting the nuanced nature of this relationship.
Taylorestimator,ifincreasep o p u l a t i o n o f e x p o r t i n g c o u n t r y byo n e percent,t h e e x p o r t w i l l b e expectedt o increase0 4 3 percent,ifincrease populationofimporting countrybyo nepercent,exportwillincrease0.403percent.
Regardingtherelationbetweenweighteddistanceandexportisconfirmedastrongnegativeimpactandsignific antinallestimators,statingthatdistanceis atrade- barrierfactor.Ifbilateralweighteddistancesincreaseonepercent,bilateralexportisdecreasedapproxi mately0.701percentinHausman-
The proximity of countries significantly influences the intensity of bilateral trade, as noted by Leamer and Levinsohn (1995) This phenomenon can be attributed to transportation costs, which increase with greater distances between origin and destination countries Additionally, transaction costs rise due to unfamiliarity with product demand, market characteristics, cultural differences, and institutional variations, as highlighted by Melitz (2003) and Andersson (2007) Consequently, new firms face challenges entering unfamiliar markets due to uncertainties related to fixed pre-period costs and the necessity of advertising to expand their market shares, as discussed by Villarrubia (2008) and Arkolakis (2008).
Concerningaboutcommonlanguageandcolonyeffect,mostoftheestimatorspresentasignificantandpositi veeffectwhichcomplieswithprediction.Languageisconsideredasatradebarrierint h e globaltrade FrankelandRose(2002)foundoutthatapproximately1.8timeshigherinbilateraltradeiftwopartnercountr iesspeakanofficiallanguagethanotherwisecountries.AndersonandVanWincoop(2004)countedat 7percentforcostoflanguagemeanwhileHelpmanatal.
(2009)predicted10percentincreaseinbilateraltradeincasecommonlanguage.ForcurrentinternationalE nglishlanguage,K u a n d Zussman( 2 0 1 0 ) predictedt h a t t h e a b i l i t y ofu s i n g Englishw o u l d s t r o n g l y promoteinternationaltrade,however,thelimitationofthisstudyisauthorignoringthe a b i l i t y tospeakanotherlanguage,soitmayover- estimatetheroleofEnglishintheglobaltrade.Similarly,thecoloniallinkispresentedpositiveandsignifi cantinOLS,randomeffectmodelandinsignificantintherestofestimators.Theeffectofcolonylinkwiththe tradeisstillvagueandhasn o t yettobeconsideredinstudiesrecently.Regardingtheoreticalaspect,thisl inkcanbecausedbythesimilarinstitutionalframeworkthatmayleadtoreducingtransactioncost.Inad ditional,colonialtiescancreatethetradingnetworkbetweenthecountryandtheirformercolonies,Disdiera ndMayer(2007).
Taylor's analysis reveals that insignificant coefficients indicate that bilateral trade flow between neighboring countries may not be mutually beneficial This paradox can be attributed to the weak cross-border transportation infrastructures among ASEAN nations In contrast, the landlocked variable is significant across all estimators, showing a negative correlation with bilateral trade flow Previous studies, such as Radelet and Sachs (1998), highlight that landlocked countries face transport and insurance costs that are twice as high as those of coastal nations Gallup and Mellinger (1999) suggest that these countries struggle with cross-border migration control, infrastructure development, and political issues with neighbors Overall, trading connectivity for landlocked countries is severely limited and heavily reliant on political relations with transit nations Any sensitive political situation or diplomatic conflict can lead to border blockages, underscoring the need for strategies to support the development of landlocked developing countries (LLDCs) as a global concern.
The analysis of the ASEAN-Japan Comprehensive Economic Partnership (AJCEP) reveals that the coefficients for intra-bloc trade are significantly negative, indicating that the agreement has not yet fostered trade creation among member countries or with extra-bloc nations Instead, AJCEP appears to have resulted in trade diversion in terms of exports Although the positive coefficient for trade flow from extra-bloc countries to AJCEP members is noted, it is insignificant, suggesting limited impact Overall, the effectiveness of AJCEP remains constrained, despite Japan being one of ASEAN's largest trading partners, highlighting the need for enhanced commitment to realize the agreement's potential benefits.
Taylorwithpaneldataofvarioustypesofproducts:agriculturalgoods,manufacturedproducts,chemica lproducts,Machineryandequipmentoftransportationandclothing,accessoriesandtextile,fabric.The main resultsareshowninTable9
Table9:OLS,FEM,REMandHausman-Taylorregressionresultforsub- catalogueproducts
(5) Clothing,acces soriesandtexti le,fabric
Accordingtoregressionresults,AJCEPhascausednegativeeffectsignificantlyinmostofkindso f pro ductsincollectedsample,exceptingFTA_2andFTA_3ofagriculturalgoods.FTA_2hassignificantan dpositivesigns,theymeanthatAJCEPhasencouragedtradecreationinagriculturaltradingfromASEAN countries andJapantobiggestpartnersofJapan.Regardingpositiveandsignificantofcoefficient
FTA_3,itpresentsthatFTAhascreatedtradecreationintermofimportagriculturalgoodsfromextra-bloc to ASEANandJapan.
First,theleveloftariffeliminationofAJCEPincomparewithotherASEAN+1FTAs.Atthepresen t,ASEANhasfivedifferentFTAswithbigcountriesintheregion,include:ASEAN-
ASEAN-India(AIFTA)shown in thefollowingsummaryTable10.
ACFTA AKFTA AJCEP AANZFTA AIFTA
China Jul-05 Korea Jul-05 Japan Dec-08 Australia Jan-10 India Jan-10
Darussalam Jul-05 Darussalam Jul-05 Darussalam Jan-09 NewZealand Jan-10 Darussalam Jan-10
Cambodia Jul-05 Cambodia Jul-05 Cambodia Jan-10 Darussalam Jan-10 Cambodia Jan-11
Indonesia Jul-05 Indonesia Jul-05 Indonesia Stillpending Cambodia Jan-11 Indonesia Jan-10
LaoPDR Jul-05 LaoPDR Jul-05 LaoPDR Dec-08 Indonesia Jan-12 LaoPDR Jan-10
Malaysia Jul-05 Malaysia Jul-05 Malaysia Feb-09 LaoPDR Jan-11 Malaysia Jan-10
Myanmar Jul-05 Myanmar Jul-05 Myanmar Dec-08 Malaysia Jan-10 Myanmar Jan-10
Philippines Jul-05 Philippines Jul-05 Philippines Jul-10 Myanmar Jan-10 Philippines May-11
Singapore Jul-05 Singapore Jul-05 Singapore Dec-08 Philippines Jan-10 Singapore Jan-10
Thailand Jul-05 Thailand Jan-10 Thailand Jun-09 Singapore Jan-10 Thailand Jan-10
Vietnam Jul-05 Vietnam Jul-05 Vietnam Dec-08 Thailand Mar-10 Vietnam Jan-10
TherearenodoubtthattariffeliminationisoneifthekeyelementofanFTA.Inaddition,tariffeliminatio ntargetyearcommitmentistheend- yearwhenthetransitionperiodmustbedone.Thetargettariffeliminationsummaryis shownin Table11
AANZFTA ACFTA AIFTA AJCEP AKFTA Average
AccordingtoTable11werecognizethataftertransitionperiod,thecommitmenttariffeliminationo f Japa naswellasASEANmembercountriesarestillhigherthanothercurrentASEAN+1FTAssuchasAANZFTA,ACFTA,AKFTA.RegardingtotimetargetispresentedinTable12wefoundt h a t while China andKo reaandsixA S E A N countrieshasreacheliminat ing tariffrateon two FTAs:ACFTA,AKFTA,Ja panandallASEANcountriesneedmoretimetofinalizetheirtargeto n A J C E P
*1includes6ASEANcountries:BruneiDarussalam,Indonesia,Malaysia,Philippines,Thailand,andSi ngapore
2* includerestof fourASEANcountries:Cambodia,Lao,MyanmarandVietnam
ThesecondreasonmaycomefromeachASEANmemberbilateralfreetradeagreementwithJapanindividua lly.AccordingtoTable13,wecanrecognizethatcurrently,Japanhasmanybilateralagreements withIndonesia,Malaysia,Philippines,Singapore,ThailandandVietnamwhicheffectiveindifferenttime
.Thus,withtheseindividualagreements,firmsoftwooptionstoestimatetariffwhentheyexportsothattheycan obtainthehighestbenefitfromFTAs.Thisconsiderationdependson somefactors:
However,accordingtoMedalla(2011,2015),currently,bilateralfreetradeagreementshasafasterreducti onontariffrateyeartargetaswellasmoreconsiderableonpreferencemargin.Moreover,R o O r u l e o f bilateralfreet r a d e agreementsi s e a s i e r andmoreliberalt h a n i t s presentationo n A J C E P Ther efore,thus,AJCEPisbecomenormalatleastwithASEANcountriesperspectiveinwhichalreadyhavee xistedbilateralfreetradeagreementwithJapan.ThetotalbilateralFTAsofA S E A N countriesaresho wn inTable13
China Korea Japan Australia NewZealand India
Indonesia Jul-08 Jul-12 Aug-07 Sep-12 x Oct-11
Malaysia x x Jul-06 Jan-13 Oct-09 Jul-11
Singapore Oct-08 Mar-06 Nov-02 Jul-03 Jan-01 Aug-05
Thailand Oct-03 x Nov-07 Jan-05 Jul-05 Jan-04
Ingeneral,t o p r o m o t e tradinga m o n g countriesi n FTAs,eliminatingtariffi s n o t c o m p l e t e l y n e c e s s a r y elementinfluencetoresultandAJCEPisnotstandoutthisrule.Therefore,thethir dreasoncomesfromnon- tariffbarriers.However,exceptRuleofOrigin(RoO),itisverycomplicatedt o q u a n t i t a t e andco mpareo t h e r n o n - t r a d e barrierss u c h asc u s t o m s procedures,technicalregulations,s a f e t y s t a n d a r d s , administ rativef e e s , transparentl a w s andregulations…
Therefore,i n t h i s paper,w e o n l y considereffecto f R o O whencomparingw i t h o t h e r currentA S
WTOestablished.Followingtheserule,eachcountryhasleftt h e rightt o c e r t i f y p r o d u c t origin byt h e m s e l v e s andorigincertificatesw i l l b e acceptedbymembers.Intheperfectprovision,ifa nFTAwhichgettheharmonization andconfirmationtounrestrictedRoOs regulation, itwillfaci litatetradingamong membercountries Therearetwo m a i n m e t h o d s t o calculateR o O : W h o l l y O b t a i n e d criteria( W O ) andN o n - W h o l l y O b t a i n e d criteria.
RegardingWhollyObtained(WO)criteriaorsubstantialtransformationcriteria,inotherwords,i t i s t h e basiscriteriaappliedwhengoodsorproducts areproduceddomestically100 percent. ForN o n -
W h o l l y Obtained,thedeterminationoforiginwillbaseonminimumchanginginoperationo f man ufacturingaproduct.Therearethreedifferentapproachesinsubstantial transformation:
Thefirstisthevalue- addedcriteria(VA)orRegionalValueContent(RVC)suchasRVC(40),R V C ( 3 5 ) meanthatto becertifiedorigin,theproductmustmeettheregionalvalueisnotlowert h a n 4 0 percentand 35percentrespectively.
Thesecondapproachistariff- headingcriteriaorchangeintariffclassification(CTC).Itmeanst h a t thelevelofclassificationofp roductininternationalharmonizationsystemmustbechanged,suchaschangeinchaptercriteria(CC),c hangeintariffheading(CTH)orchangeintariffsub- h e a d i n g (CTSH)whenproductsareconsumedacrossmembercountries.
43Thel a s t approachi s SpecificProcessR u l e ( S P R ) T h i s approachc o n s i d e r s aboutchangingi n pr oductionmethodacrossmembercountries.
+1FTAs.A c c o r d i n g tot h i s result,wecanrecognizethatACFTAmo st ly usesRVC(40) asa gen eralapproachtheRoOwhileAJCEPusesmanykindsofapproaches.Moreover,themainCTCapproachhas manyexceptionalc a s e s t o c o n t r o l originsource.T h a t s e e m s l e a d i n g t o v e r y complicatedandincreasingcostforfirmsandexporterstochoosewhichRoOshouldbeused.Int h i s case,acompanymusthavemorecalculationmethod,databaseandevensourceinformationmean whiletheauthoritywhoissueorigincertificatemusthavecompleteprocessandexaminationt o work.Inco nclusion,themoreprohibitivetheRoOsare,the higher the non-tradebarriers.
Table14:Frequencyoftype of RoO inASEANFTAs+ 1
ROO type ATIGA AANZFTA ACFTA AIFTA AJCEP AKFTA
Inconclusion,byusingmanyregressionmethods:OLS,FEM,REM,Hausman-
Taylortoestimatet h e effectofAJCEPtotradecreationandtradediversiontoASEANcountries,result showsthatfromthein- forcedatetotheendof2015,AJCEPhasnotyettocreatepositiveeffectonASEANcountriesaswellasJap an.Evenwhenweestimategravity modelwitheachspecific products,regressionshows the similarresultswith totaldata.
Conclusion
ThispaperanalyzestheeffectofAJCEP,focusingontradecreationandtradediversione ff ects toASEANcountriesandJapan.Aggregatedanddisaggregateddatasetsforfive d i f f e r e n t productsinclude:Agriculturalgoods,manufacturedproducts,chemicalproducts,Machi neryandequipmentoftransportationandclothing,accessoriesandtextile,fabrictradei n p e r i o d t i m e from2 0 0 0 -
2 0 1 5 b e t w e e n n i n e A S E A N c o u n t r i e s a n d 1 4 b i g g e s t p a r t n e r s wereemplo yed.Byusinggravitymodelandvarious kindofestimators:OLS,FEM,REMandHaus man-Taylor,however,totakeintoaccountnotonlycountry- specificu no bser ved effectsbutalsoyieldingaconsistentestimatorfortime- invariantvariables,H a u s m a n - T a y l o r isourpreferredestimator.
TheestimationresultshownsthatGDPofimportingand exportingcountries, com monb o r d e r arepositivefactorsthathelptoencouragethebilateraltrade.Similarly,cu lturald i s t a n c e factorssuchaslanguage,colonyhavethepositiveeffectsandfosterbilateraltradeaswel l.Meanwhile,theempiricalfindingstatesthatgeographicfactorssuchasgeographicd i s t a n c e , land- lockedasthetrade- restrictivefactorwhichhavestrongtheegativeeffecttot r a d e Inaddition,anegativeeffectofgeo graphicdistanceontradeisconsistentwiththeinternationaltradegravitymodel‘shypothesis.
Using aggregated total export data, regression results indicate that despite the elimination of tariff commitments, the ASEAN-Japan Comprehensive Economic Partnership (AJCEP) has not been successful in promoting trade among member countries or with external partners Additionally, the disaggregated data shows that the trade agreement between ASEAN and Japan yields either negative or negligible effects on various sub-categories, including agricultural goods, manufactured products, chemical products, machinery, transportation equipment, clothing, accessories, and textiles This suggests that the removal of tariffs under AJCEP has not effectively encouraged intra-bloc trade among AJCEP members or trade flows from AJCEP members to Japan Furthermore, testing for inward AJCEP trading from Japan reveals insignificant results, indicating that AJCEP has yet to enhance the integration of its members into international trade.
Policyimplication
Inthepastdecade,wehavewitnessedarisingupofbilateralandmultilateralfreetradeagr eementsingeneralandFTAsinASEANregioninparticular.Itprovestheimportantr o l e o fASEANmembersininternationaltrade.However,mostimportantly,thecurrentF T
Thefirstreasonisthel i b e r a l i z a t i o n le v e l o f cu r r e n t FT As d o e s no tconvince F o r t hi s r e a s o n , i n t e r m o f t h e t a r i f f , n e w F T A s s h o u l d a i m t o b r i n g a h i g h e r l e v e l i n t a r i f f eliminationthancurrentASEANFTAs+1.
ThesecondreasonisanunexpectednoodlebowlFTAssituationleadingtocomplexi typ r a c t i c es E s p e c i a l l y , A S E A N i s c o n s i d e r e d a s a h i g h l y c o m p l e x t r a d e a r e a i n w h i c h overabundance ofRulesofOrigin(RoO)areused.ThisRulesnetworktendtocha ngefromo n etradeagreementtoanotherone,difficulttomanageandmaycreatedifficultiesintrade.T h e mo stt r o u b l e s o m e o f F T A s i n t e r m o f R o O i s h o w tod e f i n e t h e e l i g i b i l i t y o f preferentialtreatmentforeachkindofproduct.MultiplechoiceinRoOisanightmarenoton lyforexporters,manufacturersbutalsoforcustomsauthoritieswhoneedtoverify thecomplianceofrules.Therefore,toreducesignificantlytheadministrationcostander asen o o d l e bowlcircumstance,thenewFTAsshouldavoidsettingcomplexrulesasmanyasp ossible.O t h e r w i s e , R o O s s h o u l d b e t r a n s p a r e n t , straightforward,andconsistent.Itis n e c e s s a r y toallowfullliberalizationintradeamongFTAs.
Thelast isfocusingon tradefacilitationandNon-tariff tradebarriers.Recently,thespeedo f bilateralandmultilateralFTAsnegotiationroundwouldbethekeyt ocreatingnewmoreattractivepackages.However,interestedineliminating tariffcan nolongerbethecorei s s u e s discussed.Inparticular,non- tarifftradebarriers(NTBs)suchasfoodandsafety standards,technicalrequirements,customs procedures,sensitivelistproducts… areaddedtotackleinthenegotiationroundgradually.
Limitationsofthestudy
Beforeendingthispaper,wewouldliketoshowoutthelimitationofthisstudyforfuturer e s ea r ch ontheimpactofAJCEPtointernationaltrade.
Thefirstlimitationisthein- forcetimeofAJCEP.ComparingwithACFTA,AKFTAwerei n force2002and2007respectively ,AJCEPwasinforceinDecember2008.Therefore,withthelimitedtimeduringtheperiodfro mbeginning2008to2015,wecan-notestimateth e fulllong- termeffectsofAJCEPtoASEANandJapan.Inaddition,thetargeteliminationtimeofAJCEP hasbeenstilllastedto2018,wedonothaveanoverviewofA J C E P effect.
Thes e c o n d l i m i t a t i o n i s t h e c o n s t r u c t i o n o f thed a t a b a s e Wec o u l d n o t i n c l u d e s o m e variablesthatwouldeffecttobilateraltradesignificantlysuchascontrolvariablesfornon- tradebarriers,administrationcost,RoOs,foodsandsafetystandards…
AJCEPcontent.FormanyFTAsingeneralandAJCEPinparticular,theyhavecommitmentschedul eofelimination/ remo val o f F T A s t a r i f f s r a t h e r t h a n i m m e d i a t e l y a f t e r i n f o r c e F T A s T h e s e p h a s i n g schedulesshouldbeincorporatedtotheanalysisprocessingofimpactofAJCEPasw ellasFTAstohavedeeplyunderstoodtheeffectoftariffeliminationgradually.
Aiken,D.E.1973.Proecdysis,setaldevelopment, andmoltpredictionintheAmericanlobster(Hom arusamericanus).J.Fish.Res.BoardCan.30: 1337–1344.
(2002).Insecurityandt h e p a t t e r n oftrade:A n empiricalinvestigation.ReviewofEconomicsand statistics,84(2),342-352.
Anderson,J E andE VanW i n c o o p , 2 0 0 3 , G r a v i t y withGravitas:A S o l u t i o n T o T h e BorderP u z z l e , AmericanEconomicReview93, pp 170–192.
Antonucci,D.,&Manzocchi,S.(2006).DoesTurkeyhaveaspecialtraderelationwiththeEU?:A g r a v i t y model approach.Economic Systems,30(2), 157-169.
Baier,S.L.andJ H Bergstrand,2 0 0 9 , E s t i m a t i n g t h e Effectso f FreeTradeAgreementso n Int ernationalTrade
Baier,ScottL.&JeffreyH.Bergstrand(2004).‘Economicdeterminantsoffreetradeagreements,J o u r n a l o f InternationalEconomics64(1),29–63(October).
(2006).Gravityfordummiesanddummiesforgravityequations(No.w12516).NationalBureauofEco nomic Research.
Bergstrand,J.H.(1985).The gravitye q u a t i o n ininternationaltrade:somemicroeconomicfoundationsandempiricalevidenc e.Thereviewofeconomicsandstatistics,474-481.
(2001).A n introductiont o geographicaleconomics:Trade,locationandgrowth.Cambridgeuniver sitypress.
Diks,C.,&Panchenko,V.(2005).AnoteontheHiemstra-JonestestforGrangernon-causality.
(2007).Jet'aime,moinonplus:Bilateralopinionsandinternationaltrade.EuropeanJournal ofPoliticalEconomy,23(4), 1140-1159.
Eaton,J.,&Kortum,S.(2002).Technology,geography,andtrade.Econometrica,70(5),1741-
Egger,Hartmut,Egger,Peter,Greenaway,David,2008.Thetradestructureeffectsofendogenousregionaltr adeagreements.Journal ofInternationalEconomics74, 278–298
Egger,P.,(2005).‘Alternativetechniquesforestimationofcross- sectiongravitymodels,’Reviewo f InternationalEconomics,13(5),881-891.
(1998).T h e r o l e o f h i s t o r y i n bilateraltradeflows.InT h e regionalizationofthe world economy(pp 33-62).UniversityofChicagoPress.
Elliott,R.J.,&Ikemoto,K.(2004).AFTAandtheAsianCrisis:HelporHindrancetoASEANIntra- RegionalTrade?.AsianEconomicJournal,18(1),1-23.
(1995).Tradingb l o c s andt h e Americas:T h e natural,t h e unnatural,and thesuper- natural.Journal ofdevelopmenteconomics,47(1),61-95.
Frankel,JeffreyA.,1997.RegionalTradingBlocs.InstituteforInternationalEconomics,Washingto n,DC.
Fukunaga,Y.,&Isono,I.(2013).TakingASEAN+1FTAstowards theRCEP: amappingstudy.
Hausman,A.&E Taylor(1981) ‘Paneldata andunobservableindividualeffects,’Econometrica4 9 , 1377–1398.
Heckman,James,Lochner,Lance,Taber,Christopher,1998.General- equilibriumtreatmenteffects:astudyoftuitionpolicy.AmericanEconomicReview88(2),381– 386May.
(1985).Marketstructureandforeigntrade:Increasingreturns,imperfectcompetition,and theinternationaleconomy.MITpress.
( 2 0 0 8 ) Wages,unemploymentandi n e q u a l i t y w i t h heterogeneousfirmsandworkers(No.w 14122).NationalBureauofEconomicResearch.
Krugman,P (1980).S c a l e economies,p r o d u c t differentiation,andt h e patterno f trade.T h e AmericanEconomicReview,70(5),950-959.
Linnemann,Hans,1 9 6 6 A n EconometricStudyo f InternationalTradeF l o w s NorthHolland,Amst erdam.
(2002).Economicgrowthoflandlockedcountries.Inệ k o n o m i e i n Theorieund Praxis (pp. 213-226).SpringerBerlinHeidelberg.
Manchin,M.,&Pelkmans-Balaoing,A.O.(2007).RulesoforiginandthewebofEastAsianfreetrade agreements.WorldBankPolicyResearchWorkingPaper,(4273).
(2008).Estimatingtheg r a v i t y equationwhenz e r o tradeflowsarefrequent.MPRAPaper. Mayer,T.,&Zignago,S.
(2011).NotesonCEPII’sdistancesmeasures:TheGeoDistdatabase.McCombie,J.S.,Thirlwall,A.P.,
&Thompson,P.(1994).Economicgrowthandthebalance-of- paymentsconstraint.NewYork:St.Martin'spress.
Melitz,M.J.(2003).Theimpactoftradeonintra‐ industryreallocationsandaggregateindustryproductivity.Econometrica,71(6),1695-1725. Nellis,J.G., &Parker,D.(2004).Principlesofmacroeconomics.Pearsoneducation.
(1994).Gravitymodels:areformulationandanapplicationtod i s c r i m i n a t o r y tradearr angements.AppliedEconomics,26(2),107-120.
Silva,J.S.,&Tenreyro,S.(2011).FurthersimulationevidenceontheperformanceofthePoissonpseudo- maximumlikelihoodestimator.EconomicsLetters,112(2),220-222.
Soloaga,I.andA.Winters,2001,HowHasRegionalisminThe 1990sAffectedTrade?,NorthAme ricanJournal ofEconomicsandFinance12,pp 1–29.
(1962)‘ShapingtheWorldEconomy:SuggestionsforanInternationalTradePolicy’,NewYork:Twentie thCenturyFundPress.
Fixed-effects(within)regression Obspergroup:min = 1
R-sq: within =0.4742 F(7,324) = 93.90 between=0.4601 Prob >F = 0.0000 overall=0.4847
R-sq: within =0.4721 F(7,324) = 93.90 between=0.7068 Prob >F = 0.0000 overall=0.6977 corr(u_i,X)= 0(assumed) Waldchi2(12) = 985.13
Table 19:Hausman-Taylorestimator (Totalagricultural goodsexport)
Hausman-Taylor estimation Obspergroup:min = 1
Hausman-Taylor estimation Obspergroup:min = 1
Table 23:Hausman-Taylorestimator(TotalCloth,accessoriesandtextilesandfabricproductsexport)
Hausman-Taylor estimation Obspergroup:min = 1
LogtotalCloth,accessoriesandt e x t i l e andfabricproducts export Coefficient Robust Std. t P>|t|