The sustainability of trade accounts of the ASEAN-5 countries

Transcription

1 Olaolu Rchard Olayen, and Avral Kumar Twar The susanably of rade accouns of he ASEAN-5 counres. Journal of Chnese Economc and Foregn Trade Sudes (forhcomng) Absrac The susanably of rade accouns of he ASEAN-5 counres Purpose- The presen sudy s nended o analyze he susanably of he rade defcs n he ASEAN-5 counres usng panel framework durng he perod from 1965 o Desgn/mehodology/approach- We appled a baery of frs and second generaon panel un roo ess and Pedron s (1999), Kao and Chang s (1998), Weserlund (2007), and D-Ioro and Fachn (2011) conegraon ess o acheve our objecve. Fndngs- We found he evdence of susanable rade defc n ASEAN-5 counres whle ulzng panel un roo ess as well as panel conegraon eses. Research mplcaons- Our fndngs have mporan macroeconomc polces mplcaon for ASEAN-5 counres ha hese polces had been effecve n leadng expors and mpors o long run seady sae equlbrum relaonshp among he ASEAN-5 counres. Orgnaly/value- The man conrbuon of he paper s o show ha he macroeconomc polces of ASEAN-5 counres had been effecve n leadng expors and mpors o long run seady sae equlbrum relaonshp. To our bes knowledge, n hs area hs s he frs sudy n he panel framework for ASEAN counres. Paper ype Research Noe Keywords: Expors, Impors, rade accouns susanably, ASEAN-5 JEL Code: C22, F13, F14, F35 1. Inroducon The behavour of he rade (curren accoun) defc, parcularly n an open economy, s an ndcaor of macroeconomc sably. Hence, for such an economy, one mporan aspec of neremporal plans s he me pah of he curren accouns, whch measures changes n naonal ne ndebedness. Problems ha arse n an open economy are no caused by rsng mpors bu are due o he msmach beween expor and mpor growh. Therefore, he saonary of curren accouns s mporan for wo reasons. Frs, a saonary curren accoun s conssen wh he accumulaon and susanably of exernal debs (whch ndcaes ha here s no ncenve for a counry o defaul on s nernaonal debs) as well as an ndcaor of poenal exchange rae

2 Olaolu Rchard Olayen, and Avral Kumar Twar The susanably of rade accouns of he ASEAN-5 counres. Journal of Chnese Economc and Foregn Trade Sudes (forhcomng) realgnmen. Second, he saonary of curren accouns agrees wh he mplcaon of he modern ner-emporal model of curren accoun, and hence suppors s valdy (Obsfeld and Rogoff, 1996, p. 90). However, Bohn (2007) argues ha a saonary curren accoun balance s suffcen bu no necessary for he susanably of exernal debs. The me-seres properes of he curren accoun and he long-run relaonshp beween mpors and expors are noneheless nformave. In he face of shor-run urbulence, governmens have neres n wheher or no, and n wha way, a curren accoun defc s lkely o correc owards a more sable level. In hs regard, Hused (1992) shows ha under he null hypohess ha he economy sasfes s ner-emporal budge consran, we expec ha expors and mpors have a conegrang relaon wh conegrang vecor (1,-1), graned ha hey are I(1). Thus, he fac ha he rade balances are saonary s he necessary condon o sasfy he economy s neremporal budge consrans. There are wo ses of sudes based on he applcaon of he ess eher n me seres or panel framework. One se apples convenonal un roo ess such as ADF PP and KPSS on he curren/rade accoun whereas second se apples conegraon es beween expors and mpors and ess he resrcon of he conegrang vecor (1, -1). The focus of he second approach s o es for he long-run relaonshp beween expors and mpors. Such a relaonshp would mply ha he wo seres would never drf oo far apar. Noeworhy sudes n hs area are Hused s (1992), Bahman-Oskooee and Rhee (1997), Founas and Wu (1999), Arze (2002), Irandous and Ercsson (2004), Narayan and Narayan (2005), Holmes (2006), Upender (2007), Erbaykal and Karaca (2008), Konya and Sngh (2008), Hamor (2009), Konya (2009), Twar (2011), Twar and Pandey (2011), Yn and Hamor (2011), Greendge e al. (2011), Twar (2012), and Twar (2013). For he Assocaon of Souheas Asan Naons (ASEAN), he queson of curren accoun susanably s unambguously of prmary neres. Formed from s predecessor, he Assocaon of Souheas Asa (ASA) n 1967, he ASEAN was orgnally an assocaon of fve counres he ASEAN-5 namely, Indonesa, Malaysa, Phlppnes, Sngapore, and Thaland. They are he major economes of he exsng ASEAN-10. Alhough he group encourages cooperaon on many frons, economc cooperaon s one of he prncpal facors ha brough hem ogeher. Trade flows n he group are of neres especally n recen me when he regon s collecvely and ndvdually forgng rade es wh he res of he world parcularly he US

3 Olaolu Rchard Olayen, and Avral Kumar Twar The susanably of rade accouns of he ASEAN-5 counres. Journal of Chnese Economc and Foregn Trade Sudes (forhcomng) and he European Unon (EU). Recenly, for nsance, he US-ASEAN Expanded Economc Engagemen (E3) nave was launched n November 2012 o promoe closer economc es beween he ASEAN counres and he Uned Saes. The E3 nave s poenally amed a promong rade flows beween he group and he US. Whle he formaon of he E3 nave s glamorously welcome, he EU-ASEAN es have been meeng wh a srng of bolenecks especally n he servce expors ha s he core of he EU s expor secor (Robles, 2013). The ensung polcy ncoherence wll ulmaely ell on rade balance n he group, hence warranng a close examnaon of he curren accoun susanably. However, he fndngs are far from conclusve on he subjec of curren accoun susanably n ASEAN markes and ndeed he dearh of relevan leraure n hs area s of concern. Ths s wha we plan o brdge n he leraure. We acheve hs n a panel framework gven he proxmy and he negraon effecs (Pesaran, 2004) n he ASEAN counres. Unl now such a framework has been mssng n he leraure on he ASEAN counres generally and on he ASEAN-5 consdered n hs sudy n parcular. The srucure of he paper s as follows. The followng secon presens n bref revew of leraure. The hrd secon presens a basc model of rade susanably and fourh secon presens he daa and mehodology. Ffh secon presens resuls of our emprcal analyss and he fnal secon concludes and gves polcy mplcaons. 2. Relaed leraure Hused (1992) n hs poneerng work usng quarerly US rade daa ( ) shows ha expors and mpors are conegraed, and subms ha he exsence of conegraon beween expors and mpors mples ha counres do no volae her IBC. Ths fndng lends suppor o he effecveness of macroeconomc polces n resorng long-run equlbrum. However, for he US economy Founas and Wu (1999) fnd no evdence of conegraon beween expors and mpors durng and conclude ha US curren accoun defc s no susanable. Bahman-Oskooee and Rhee (1997) lend suppor o conegraon wh posve coeffcen on expors for he Korean economy usng quarerly daa on expors and mpors. Arze (2002) examnes he queson of curren accoun susanably usng quarerly daa from

4 Olaolu Rchard Olayen, and Avral Kumar Twar The susanably of rade accouns of he ASEAN-5 counres. Journal of Chnese Economc and Foregn Trade Sudes (forhcomng) 1973 o 1998 on 50 counres ncludng OCED and developng ones. He fnds conegraon for 35 ou of 50 counres; and for 31 ou of 35 counres, he fnds expor coeffcens o be posve. Irandous and Ercsson (2004) expand on Hused s (1992) analyss by focusng on sx counres (namely he US, Germany, he UK, France, Sweden and Ialy) for he perod They fnd evdence of conegrang relaonshp beween expors and mpors n Germany, he US and Sweden. Hence, auhors conclude ha hese counres do no volae her neremporal budge consrans and ha rade mbalances are shor-run phenomena susanable n he long run. Narayan and Narayan (2005) apply he bounds esng approach o conegraon o nvesgae a long-run relaonshp beween expors and mpors for 22 leas developed counres (LDCs). They fnd ha expors and mpors are conegraed only for sx ou of 22 counres, wh a coeffcen of expors less han one. In examnng susanably of curren accoun defc of 11 OECD counres from 1980 o 2002, Holmes (2006) fnds ha sx of 11 counres (Ausrala, Belgum, Canada, Japan, he UK and he USA) exhb susanably n curren accoun, whle France, Germany, Ialy, Norway and Span do no. Upender (2007) shows ha Inda's nomnal expors and mpors are conegraed by employng daa for he perod o Erbaykal and Karaca (2008) esablsh he exsence of a conegraed relaonshp beween expors and mpors suggesng ha rade defcs are only shor-run phenomena and, herefore, susanable n he long run. Konya and Sngh (2008) apply one-me srucural break ( ) o Inda's expors and mpors seres from o , bu do no fnd evdence of conegraon. They choose he srucural break exogenously assumng ha he swch from a fxed o floang exchange rae regme n March 1993 would have some mpac. Hamor (2009) examne he susanably of he rade accouns of he G-7 counres by analysng he long-run equlbrum beween expors and mpors. He fnd ha rade accouns are no susanable for he G-7 counres; hus, he recommends ha some approprae polces mus be mplemened o preven rade mbalances from ncreasng. Holmes and Panagods (2009) conduc an nvesgaon no he exen of conegraon beween mpors and expors and asymmeres n he adjusmen of he US curren accoun over he sudy perod 1960Q4-2007Q2. They fnd evdence n favour of conegraon hrough he applcaon of he sandard Johansen s mehodology. However, ulsng he race es procedure recursvely, hey denfy wo dsnc regmes accordng o wheher or no mpors and expors are conegraed. Wh her non-lnear conegraon and

5 Olaolu Rchard Olayen, and Avral Kumar Twar The susanably of rade accouns of he ASEAN-5 counres. Journal of Chnese Economc and Foregn Trade Sudes (forhcomng) asymmerc adjusmen analyss, hey conclude ha adjusmen owards long-run equlbrum s prmarly drven by US expors respondng o curren accoun defcs. Greendge e al. (2011) nvesgae he susanably of he curren accoun defc n Barbados over he perod The auhors fnd ha he curren accoun of Barbados s susanable and ha devaons from long-run equlbrum beween real expors and mpors are correced n he shor-run wh mpors makng he adjusmen. Twar (2011) examnes he long-run relaonshp beween expors and mpors for Chna and Inda usng monhly daa from January 1992 o February He endogenously deermnes srucural breaks n un roo and conegraon and fnds ha rade defc s susanable n Inda bu no n Chna. Yn and Hamor (2011) use nonsaonary me seres approach o es for he susanably of he curren accoun defcs n Chna over he perod They fnd ha, despe he conegrang relaonshp beween mpors and expors n Chna, he ner-emporal exernal consrans may be volaed. Thus, hey conclude ha he rade balance surplus experenced over he pas several years canno be susanable n he fuure. Twar and Pandey (2011) examne Indan daa from Aprl 1984 o March 2009 and fnd long-run relaonshp beween expors and mpors n all he cases (.e. by usng dfferen daes of srucural breaks). In anoher sudy, Twar (2012) uses dsaggregaed Indan daa a level from 1970 o 2007 and allows for endogenous deermnaon of srucural breaks n he un roo and exogenous srucural breaks for conegraon. In hs novel approach, he specfcally ess for long-run susanably n he curren accoun defc n Inda usng ol and non-ol expors and mpors. He fnds srong evdence of long-run relaonshp beween non-ol expors and mpors, bu no for ol expors and mpors. He also fnds long-run unsusanable rade defcs for ol and long-run susanable rade defcs for non-ol commodes. Twar (2013) examnes he long-run susanably of rade defcs for he ASEAN-5 economes, vz., Indonesa, Malaysa, he Phlppnes, Myanmar and Thaland, n he presence of srucural breaks. He fnds ha here s a long run relaon beween expors and mpors for Indonesa, Myanmar and Thaland, and fnds susanable long-run rade defc only for Myanmar. However, n hs paper, we apply a baery of panel un roo ess as well as panel conegraon ess. Ths s because by applyng he panel un roo and conegraon ess, we can perform boh me seres and cross-seconal analyses whou lmng her power due o a small sample sze. Whle a large number of counres have been analyzed, o he bes of our

6 Olaolu Rchard Olayen, and Avral Kumar Twar The susanably of rade accouns of he ASEAN-5 counres. Journal of Chnese Economc and Foregn Trade Sudes (forhcomng) knowledge, here has been no emprcal nvesgaon of he behavor of rade accoun balance of ASEAN-5 counres n he panel framework. Based on he panel un roo and conegraon ess, we fnd ha expors and mpors are conegraed and he conegrang coeffcen s sgnfcanly dfferen from zero, bu no sgnfcanly dfferen from one. These fndngs mply ha he curren accoun defcs among ASEAN-5 counres are susanable. 3. Basc Model Based on Hused (1992), Arze (2002) and Irandous and Ercsson (2004), we examne he neremporal budge consran and analyze he dynamcs of he rade balance. Followng Hakko and Rush (1991), Hused (1992) provdes a smple framework ha mples a longrun relaonshp beween expors and mpors. The ndvdual curren-perod budge consran s: where, C, I, B ζ (1) = C + I Y + ( 1+ r ) B 1 = + (1 + r ) B 1 Y, B and r are curren consumpon, nvesmen, oupu, nernaonal borrowng, and a one-perod neres rae respecvely. ζ = Y C I EX M ) and ( 1+ ) B 1 ( r s he nal deb sze. We can rewre Eq. (1) recursvely o oban he neremporal budge consran. Frs, noe ha Eq. (1) can be wren as B ζ = ( 1+ r) B 1 + ( r r) B 1 where r = Er = Er + 1. The hrd erm η = ( r r) B 1 represens he devaon of sochasc neres rae valued deb repaymen from he consan neres rae valued deb repaymen and can be reaed as random error. Nex, we recursvely updae Eq. (1), arrvng a he followng neremporal budge consran. Takng expecaon we oban he followng: (1 + r) B ζ [(1 r B ] + 1 = E + lm E + ) = 0 (1 + r) Followng Ahmed and Rogers (1995), we dfference he neremporal budge consran above usng he fac ha Eq. (1) mples oban: B B + + rb 1 ζ 1 E ζ = lm E[(1 + r) B ] +. (1 + r) = 0 1 = 1+ r ) B 1 (1 + r 1 ) B 2 = r B 1 ( ζ o

7 Olaolu Rchard Olayen, and Avral Kumar Twar The susanably of rade accouns of he ASEAN-5 counres. Journal of Chnese Economc and Foregn Trade Sudes (forhcomng) In hs case, he ransversaly condon guaranees ha lm E [(1 + r) B ] = 0. The erm + = 0 E ζ r) wll be saonary f mpors and expors are boh I(1) varables. Thus, + (1 + rb 1 ζ 1 wll be saonary. Defnng IM M + rb esable model: where EX, =, we arrve a he Hused (1992) EX β + IM + u (2) = 0 β1 IM and u are he expors of goods and servces, he mpors of goods and servces plus ne neres paymens and ne ransfer paymens and dsurbance a me, respecvely. Under he null hypohess, for an economy ha sasfes s neremporal budge consran (.e., for a susanable curren accoun defc), s expeced ha β = 1 1 and u s a saonary process. In oher words, f expors and mpors are negraed of order one,.e., I(1), hen under he null hypohess, hey are conegraed wh a conegrang vecor (1, -1). 4. Daa and Mehodology 4.1 Daa We nclude fve counres n our sample: Indonesa, Malaysa, Phlppnes, Sngapore, and Thaland. The annual daa from 1965 o 2011 are used n our emprcal analyss. Our measure of expors ncludes expors of goods and servces and our measure of mpors ncludes mpors of goods and servces and expressed n mllon Uned Saes (US) dollar values. Boh expors and mpors are measured as a percenage of Gross Domesc Produc (GDP). All daa, obaned from offcal webse of World Bank, were exraced on 27 June, Fgure 1 shows he movemens of he expors, mpors and rade balances of he ASEAN- 5 counres. Ths fgure ndcaes ha rade mbalances generally end o hover around he zero lne, suggesng rade balance, for all he counres. However, owards he end of he sample perod, we observe ha rade mbalances have shown up and have been expandng for boh Malaysa and Sngapore.

8 Olaolu Rchard Olayen, and Avral Kumar Twar The susanably of rade accouns of he ASEAN-5 counres. Journal of Chnese Economc and Foregn Trade Sudes (forhcomng) Fgure 1: Expors, Impors and Trade balance Indonesa Malaysa Phlppnes 250 Sngapore Thaland IMPORTS EXPORTS TRADE_BALANCE 4.2 Mehodology Panel un roo ess The frs hng we do n hs secon s o perform un roo ess on expors and mpors. We dvde hese ess no wo groups, namely, frs generaon un roo es and second generaon un roo ess. The frs generaon of panel un roo ess employed n hs sudy ncludes he LLC es (Levn, Ln and Chu, 2002), he IPS es (Im, Pesaran and Shn, 2003) and he MW es (Maddala and Wu, 1999). The second generaon ess are he MP es (Moon and Perron, 2004), Pesaran es (Pesaran, 2007) and Cho es (Cho, 2006). Whle he frs generaon assumes

9 Olaolu Rchard Olayen, and Avral Kumar Twar The susanably of rade accouns of he ASEAN-5 counres. Journal of Chnese Economc and Foregn Trade Sudes (forhcomng) cross-seconal ndependence, second generaon ess are robus o cross-seconal dependence and are more powerful han he frs generaon ess (Hurln, 2004)[1] Panel conegraon ess Provded boh expors and mpors are nonsaonary n level and saonary n frs dfference.e., hey are I(1) varables, our nex sep s o perform conegraon ess beween expors and mpors. The rade accoun (or rade balance) s defned by he followng equaon: where EX TB = = EX IM ( 1 1) (3) IM TB, EX, and IM represen he rade balance, expors and mpors a me, respecvely. Consequenly, f expors and mpors have a conegrang relaonshp wh conegrang vecor (1,-1), hen he lnear combnaon of EX and IM n Eq. (3) and hence rade balance becomes a saonary. Thus, follows ha esng for a conegrang relaonshp beween expors and mpors under hs consran s effecvely he same as esng for un roos n rade accouns. To furher asceran he robusness of he emprcal resuls, we also apply alernave conegraon ess developed by Pedron (1999), Kao and Chang (1998), Weserlund (2007), and D-Ioro and Fachn (2011). Consder he followng fxed-effec panel regresson, = α β, = 1K,, N ; = 1K,, T (4) EX, + IM, + u, where IM = IM + e { EX, IM } are ndependen across cross-seconal uns and,, 1, ;,, ε, = (,,, ) e u s a lnear process ha sasfes he assumpons n Pedron (1999), and Kao and Chang (1998). Furhermore, s he ndex of cross-secon dmenson and s ndex of me dmenson, and α and β are respecvely me nvaran counry-specfc nercep and he slope coeffcens ha vary across counres. Pedron (1999) derves he asympoc dsrbuon and explores he small sample performances of seven dfferen sascs. Of hese seven sascs, four are based on poolng along wha s commonly referred o as he whn-

10 Olaolu Rchard Olayen, and Avral Kumar Twar The susanably of rade accouns of he ASEAN-5 counres. Journal of Chnese Economc and Foregn Trade Sudes (forhcomng) dmenson, and hree on poolng along wha s commonly referred o as he beweendmenson. Pedron (1999) descrbes he former and laer as panel conegraon sascs and group mean panel conegraon sascs, respecvely. Kao (1999) presens DF and ADF ypes of conegraon ess n he panel daa. Kao and Chang (1998) derve lmng dsrbuons for he ordnary leas square (OLS), fully modfed (FM) and dynamc ordnary leas square (DOLS) esmaors n a conegraed regresson and hen show ha hey are asympocally normal. However, mos of he panel and me seres ess are based on he assumpon ha he long-run parameers for he varables n her levels are equal o he shor-run parameers for he varables n her dfferences referred o as a problem of a common-facor resrcon (see Banerjee e al., 1998; Kremers e al., 1992) and whch can cause a sgnfcan loss of power for resdual-based conegraon ess. In order o overcome hs problem, Weserlund (2007) develops four new panel conegraon ess ha are based on srucural raher han resdual dynamcs and, herefore, do no mpose any common-facor resrcon. The dea s o es he null hypohess of no conegraon by nferrng wheher he error-correcon erm n a condonal panel error-correcon model s equal o zero. The new ess are all normally dsrbued and are general enough o accommodae un-specfc shor-run dynamcs, unspecfc rend and slope parameers, and cross-seconal dependence. Two sascs are desgned o es he alernave hypohess ha he panel s conegraed as a whole known as panel ess (denoed as P and Pa), whle he oher wo sascs, known as group-mean ess and denoed by G and Ga, es he alernave ha a leas one un s conegraed[2] Panel conegraon esng va resdual based boosrap The basc problem we address n hs secon s ha he error erms for counry, u,, n Eq. (4) 2 hough may be dencally and ndependenly dsrbued for all, ha s, u ~ d(0, σ ), hey, u could be cross-seconally correlaed. Boosrap panel conegraon esng hrough resdual resamplng has become a sandard approach o assessng counry-specfc crcal values whle dealng wh a panel daa srucure and can be adaped o address he cross-seconal dependence problem. Earler novel applcaons can be found n Paparods and Pols (2001, 2003), D Ioro and Fachn (2011), and Weserlund and Edgeron (2007). D-Ioro and Fachn (2011) show ha he Resdual-based Saonary Boosrap (RSB) es proposed by Parker e al. (2006) s helpful

11 Olaolu Rchard Olayen, and Avral Kumar Twar The susanably of rade accouns of he ASEAN-5 counres. Journal of Chnese Economc and Foregn Trade Sudes (forhcomng) n generang saonary pseudo seres, whch, for our purpose, has a clear advanage over he block boosrap panel un roo es also shown o be asympocally vald (Palm e al., 2011). I should be noed ha he RSB self s a block boosrap and so preserves auocorrelaon srucure embedded n he daa se. Ths allows us n our applcaon o re-sample by channg blocks of observaons sarng a random locaons (D-Ioro and Fachn, 2011), he approach ha conrass wh he block boosrap ypcally of fxed block lengh. To proceed, we f he Engle-Granger conegraon model gven n Eq. (5).We es for he null of no cross-seconal dependency gven by H cov( u, u ) 0 for all and j agans he alernave hypohess gven by 0 :, j, = H 1 : cov( u,, u j, ) 0 for some pars j. We model he error erm, u, n (4) for counry a perod as an AR(1) process gven by u, ρ u, 1 + ε, = (5) Tesng for conegraon beween our varables herefore amouns o esablshng ha he esmaed auoregressve coeffcen varables are sad o be conegraed for un f here exss ρ n Eq. (5) s less han un. In oher words, hese β such ha u, s saonary. I should be observed ha saonary of ε, does no mply he exsence of conegraon as ε, s always saonary (D-Ioro and Fachn, 2011). Nex, we employ he resdual-based boosrap o accoun for cross-seconal dependence. We acheve hs by applyng he re-samplng procedure dscussed above o he enre cross-secons a he same me and hus preservng he crossseconal correlaons. More precsely n sep 2 of he algorhm n Appendx I, we apply he resamplng echnque on he N T marx of resduals, ˆε o produce ε = [ ˆ ε 1, K, ˆ ε N ], where ε = [ ˆ ε, K, ˆ ε ].The procedure for re-samplng s gven n Appendx. ˆ 1 N 5. Emprcal Analyss Resul of panel un roo ess for expors, mpors and rade balance are presened n he Table 1 below. Table 1: Second Generaon Panel Un Roo Tess Types of es sasc Tes sasc expors Second-generaon panel un roo es on expors Tes sasc mpors Tes sasc rade balance

12 Olaolu Rchard Olayen, and Avral Kumar Twar The susanably of rade accouns of he ASEAN-5 counres. Journal of Chnese Economc and Foregn Trade Sudes (forhcomng) Moon Perron s ( a ) es (2004) *** Moon Perron s ( b ) es (2004) *** Pesaran s ( CIPS ) es (2007) ** Cho s ( P M ) es sasc (2006) *** Cho s ( Z ) es sasc (2006) *** Cho s ( L * ) es sasc (2006) *** Noe: (a) Crcal values for Moon Perron s ( a ) and ( b ) (2004) ess and Cho s ( Z ) es sasc (2006) and Cho s ( L * ) es sasc (2006) a 1%, 5% and 10% respecvely are , , (b) Crcal values for Pesaran s ( CIPS ) es (2007) a 1%, 5% and 10% respecvely are , , (c) Crcal values for Cho s ( P M ) es sasc (2006) a 1%, 5% and 10% respecvely are , , Source: Auhors calculaon Second column of Table 1 shows he resuls of he panel un roo ess performed on expors. The Akake nformaon creron (AIC) was used o selec he number of lags. The frs and second-generaon panel un roo ess provde smlar resuls.e., expors has a un roo. Resuls of panel un roo ess on mpors are presened n he hrd column of Table 1, and n hs case oo we fnd ha all ess provde he smlar resuls. Hence, we conclude ha expors and mpors are nonsaonary varables n he level form. Nex, we denfy he order of negraon of he varables and fnd ha boh varables are frs dfference saonary. Ths mples ha boh varables are I(1)[3]. Havng found ha boh expors and mpors are nonsaonary varables n level and saonary n frs dfference.e., I(1), we now perform conegraon ess beween expors and mpors. Frs we presen resuls of conegraon analyss. Followng on Eq. (3), we presen resuls of panel un roo ess on rade balance n he fourh column of Table 1. As s clear from he fourh column of Table 1, rade balance does no have a un roo n any of he cases (excep for he LLC es). Ths suggess ha here s srong evdence of conegrang relaon beween expors and mpors wh conegrang vecor (1,-1). In he nex sep we adop conegraon es proposed by Pedron (1999), Kao and Chang (1998), and Weserlund (2007) and presen resuls n Table-2 below. Table 2: Conegraon es Mehod Tes sasc P-Value Pedron Resdual Conegraon Tes: Trend assumpon:- No deermnsc rend Panel v-sasc

13 Olaolu Rchard Olayen, and Avral Kumar Twar The susanably of rade accouns of he ASEAN-5 counres. Journal of Chnese Economc and Foregn Trade Sudes (forhcomng) Panel rho-sasc Panel PP-Sasc Panel ADF-Sasc Group rho-sasc Group PP-Sasc Group ADF-Sasc Kao Resdual Conegraon Tes: Trend assumpon:- No deermnsc rend ADF Weserlund (2007) conegraon es: lags(1 2) leads(0 1) lrwndow(2) consan rend Sasc Value Z-Value P-Value G Ga P Pa Source: Auhors calculaon The frs of he smple panel conegraon sascs, he panel v-sasc, s a nonparamerc varance rao sasc. The second, he panel rho-sasc, s a panel verson of a nonparamerc sasc analogous o he famlar (Phllps and Perron, 1988) rho-sasc. The hrd, he panel -sasc (nonparamerc), s a nonparamerc sasc analogous o he Phllps and Perron -sasc. The fourh, he panel -sasc (paramerc), s a paramerc sasc analogous o he famlar ADF -sasc[4]. The oher hree panel conegraon sascs are based on a group mean approach. The frs, he group rho-sasc, s analogous o he Phllps and Perron rho-sasc. The las wo, he group -sasc (nonparamerc) and group - sasc (paramerc), are analogous o he Phllps and Perron -sasc and he ADF -sasc, respecvely. Table 2 shows he resuls of panel conegraon ess performed on expors and mpors. I ndcaes ha he null hypohess of no conegraon s rejeced n all seven cases (excep one) a he 5% level of sgnfcance. Smlar fndngs are repored by ADF-ype Kao (1999) es. Thus, s lkely ha expors and mpors are conegraed for he ASEAN-5 counres. Ths resul s conssen wh one obaned n Table 3. Our resuls based on Weserlund s (2007) wo (parcularly Pa and Ga) ou of four panel conegraon ess provde evdence o rejec he null hypohess ha expors and mpors are no conegraed. In he fnal sep we presen resuls of conegraon es proposed by D-Ioro and Fachn (2011) wh resdual

14 Olaolu Rchard Olayen, and Avral Kumar Twar The susanably of rade accouns of he ASEAN-5 counres. Journal of Chnese Economc and Foregn Trade Sudes (forhcomng) based boosrap approach[5]. In prncple, hese ess should rejec he null even f here s conegraon n only one of he 5 ASEAN counres. Hence, we adoped a counry specfc boosrap approach o valdae our fndngs. Our resuls of resdual based boosrap approach are presened n Table 3. Table 3: Resdual based boosrap conegraon es sascs (a) Cross-seconal dependence es LM (Breusch-Pagan, 1980) CD1 (Pesaran, 2004) CD2 (Pesaran, 2004) [0.000] [0.000] 4.819[0.000] (b) Indvdual counry s cross-seconal-dependence conegraon ess Counry Sascs 10% 5% 1% Dckey-Fuller GLS Indonesa ** Malaysa Phlppnes *** Sngapore *** Thaland *** Phllps-Perron Indonesa ** Malaysa Phlppnes *** Sngapore *** Thaland *** Modfed Phllps-Perron Indonesa * Malaysa * Phlppnes *** Sngapore *** Thaland *** (c) Group cross-seconal-dependence conegraon ess Mean Medan Max Dckey-Fuller GLS [0.000] [0.000] [0.039] Phllps-Perron [0.000] [0.0002] [0.000] Modfed Phllps-Perron [0.000] [0.0019] [0.000]

15 Olaolu Rchard Olayen, and Avral Kumar Twar The susanably of rade accouns of he ASEAN-5 counres. Journal of Chnese Economc and Foregn Trade Sudes (forhcomng) Noes: In Panel (a), LM, CD1 and CD2 sand for he Breusch-Pagan (1980) lkelhood mulpler es, he frs and he second cross-seconal dependence ess of Pesaran (2004) respecvely. In Panel (b), ***,** and * refer o 1, 5 and 10% sgnfcance levels respecvely. The resuls are based on re-drawngs wh replacemen. In Panel (c), he p-values are repored n parenheses. The p-values n all cases are based on he re-drawngs compued as he proporon of re-drawngs for whch he compued sascs (mean, medan and maxmum of he sascs for he cross-secons) s greaer han he boosrap sascs. The frs se of ess we carry ou n hs secon s cross-seconal dependence ess. We use hree (3) ess o confrm f here s cross-seconal dependence. They are he Breusch-Pagan (1980) LM es, he frs and he second Pesaran (2004) cross-seconal dependence ess. The resuls n Panel (a) of Table 3 show ha a 1 percen all he hree ess ndcae he presence of crossseconal dependence. Based on hese resuls, we proceed o accommodae cross-seconal dependence n our conegraon ess. Wha we do here s o esablsh he saonary propery of he daa by facorng n he cross-seconal sylzed fac common o panel-daa srucure. Panel (b) of Table 3 repors he counry specfc conegraon ess beween expors and mpors. The based on he hree (3) sascs repored n hs paper, namely, he Dckey-Fuller GLS, he Phllps-Perron (PP) and he Modfed Phllps-Perron conegraon ess, rade balance for Phlppnes, Sngapore and Thaland s conssenly found o be saonary (ha s, mpors and expors are conegraed) a 1 percen sgnfcance level, whle for Indonesa rade balance s saonary a 5 percen level of sgnfcance based on he Dckey-Fuller GLS and he Phllps- Perron (PP) and a 10 percen based on he Modfed Phllps-Perron. For Malaysa, rade balance s no saonary a he sandard levels of sgnfcance usng he Dckey-Fuller GLS and he Phllps-Perron (PP) and only margnally saonary a 10 percen. In Panel (c) we repor he group sascs (mean, medan and maxmum) based on he Dckey-Fuller GLS, he Phllps- Perron (PP) and he Modfed Phllps-Perron conegraon ess. The mean sasc mples ha he rejecon of he null hypohess of no-conegraon agans he alernave of conegraon mples ha n mos of he cross-secons he mass of he dsrbuon should be sgnfcanly far from 1. Ths same nerpreaon can be gven o he medan sasc. The max sasc on he oher hand requres ha he rejecon of he null hypohess of no-conegraon depend on fndng conegraon for all he cross-secons, meanng ha ρ < 1 n all cases. Based on hese sascs, we fnd ha conegraon s found for ASEAN-5 as a group, he assocaed p values beng suffcenly less han 5 percen for all he sascs.

16 Olaolu Rchard Olayen, and Avral Kumar Twar The susanably of rade accouns of he ASEAN-5 counres. Journal of Chnese Economc and Foregn Trade Sudes (forhcomng) 6. Dscusson, conclusons and polcy mplcaons The presen sudy examnes he naure of long-run relaonshp beween expors and mpors for he ASEAN-5 counres namely: Indonesa, Phlppnes, Thaland, Malaysa, and Sngapore. Perod suded n he analyss s from 1965 o As demonsraed earler by Hused (1992), he rade balance of an economy s a saonary varable when he economy sasfes neremporal budge consrans. Hused (1992), Arze (2002) and Irandous and Ercsson (2004) show ha rade defc s saonary and ha n many counres, rade accoun mbalances wll converge o a ceran equlbrum over he long run. Our sudy provdes suppors o he fndng ha rade defc s saonary as we fnd ha expors and mpors are conegraed and rade defc s saonary. Economc heory suggess ha n a well-funconng economy where here are neher permanen producvy shocks nor polcy dsorons, here s a naural endency owards conegraon n he rade balance. A volaon of hese mananed hypoheses wll lead o non-conegraon n rade balance. Our resuls suppor he vew ha here s a endency of rade defc o converge o zero n he long run n he ASEAN-5 counres. Ths mples ha he economes ASEAN-5 counres are well-funconng n such a way ha hese economes are able o adjus for echnologcal and polcy shocks. From our analyss, we are able o come-up wh few polcy mplcaons for ASEAN-5 counres. Frs, our fndngs of conegraon ndcae ha hese counres do no volae her neremporal budge consrans. Second, he resuls reveal ha, n hese counres, he mbalances are emporary and susanable n he long run. Thrd, macroeconomc polces (such as fscal, moneary, and exernal polces) n ASEAN-5 had been effecve n brngng rade balance o converge oward zero n he long run. Fourh, snce rade balance follows meanreverng process, polcy analyss can use o predc he fuure behavour and can sugges polcy decsons accordngly. Ffh, n he framework of BOP-consraned growh model, ASEAN-5 counres are followng BOP equlbrum growh rae for long me. Acknowledgemen The auhors are graeful o wo anonymous referees for her very consrucve commens and suggesons on an earler draf of he manuscrp whou whch he paper would have remaned flawed and ncomplee. The auhors are solely responsble for any remanng errors.

17 Olaolu Rchard Olayen, and Avral Kumar Twar The susanably of rade accouns of he ASEAN-5 counres. Journal of Chnese Economc and Foregn Trade Sudes (forhcomng) Noes 1. For a dealed revew on he dfference beween dfferen ess of frs and second generaon please refer o Hurln (2010). 2. For elaborae dscusson on hese es sascs please refer o orgnal work of auhor. 3. Resuls of frs dfference form of he varable wll be avalable upon reques o he auhor. 4. See Table 1 n Pedron (1999, p. 660). 5. Ths approach s adoped n response o an anonymous referee of hs journal who argued ha even f he panel conegraon ess unanmously rejec he null hypohess of no conegraon, hs does no mean ha expors and mpors are conegraed n each counry. Reference Ahmed, S. and Roger, J.H. (1995), Governmen budge defcs and rade defcs: Are presen value consrans sasfed n long-erm daa?, Inernaonal Fnance Secon Dscusson Paper No. 494 (January), Washngon, D.C.: Federal Reserve Board. Arze, A.C. (2002), Impors and expors n 50 counres: ess of conegraon and srucural breaks, Inernaonal Revew of Economcs and Fnance, Vol. 11, pp Bahman-Oskooee, M. and Rhee, H.J. (1997), Are expors and mpors of korea conegraed?, Inernaonal Economc Journal, Vol. 11, pp Bohn, H. (2007), Are saonary and conegraon resrcons really necessary for he emporal budge consran?, Journal of Moneary Economcs, Vol. 54, pp Cho, I. (2001), Un roo ess for panel daa, Journal of Inernaonal Money and Fnance, Vol. 20, pp Cho, I. (2006), Combnaon un roo ess for cross-seconally correlaed panels, n P.C.B. (Ed.), Economerc Theory and Pracce: Froners of Analyss and Appled Research.

18 Olaolu Rchard Olayen, and Avral Kumar Twar The susanably of rade accouns of he ASEAN-5 counres. Journal of Chnese Economc and Foregn Trade Sudes (forhcomng) D Ioro F., and Fachn S. (2011), A panel conegraon sudy of he long-run relaonshp beween savngs and nvesmens n he OECD economes, , Mnsry of Economcs and Fnance (MEF) Workng Papers No 3, February. Erbaykal, E. and Karaca, O. (2008), Is urkey s foregn defc susanable? conegraon relaonshp beween expors and mpors, Inernaonal Research Journal of Fnance and Economcs, 14, pp Founas, S. and Wu, J.L. (1999), Are he u.s. curren accoun defcs really susanable?, Inernaonal Economc Journal, Vol. 13, pp Greendge, K., Holder, C. and Moore, A. (2011), Curren accoun defc susanably: he case of Barbados, Appled Economcs, Vol. 43, pp Hakko, C. and Rush, M. (1991), Is he budge defc oo large?, Economc Inqury, Vol. 29, pp Hamor, S. (2009), The susanably of rade accouns of he g-7 counres, Appled Economcs Leers, Vol. 16, pp Holmes, M.J. (2006), How susanable are oecd curren accoun balances n he long-run?, Mancheser School, Vol. 74, pp Holmes, Mark J. and Panagods, T. (2009), Conegraon and asymmerc adjusmen: some new evdence concernng he behavor of he u.s. curren accoun, The B.E. Journal of Macroeconomcs, Vol. 9 No. 1, Arcle 23. Hurln, C. (2004), Nelson and plosser revsed: a re-examnaon usng oecd panel daa, Documen de Recherche no. 2004/23, Laboraore d Econome d Orleans. Hurln, C. (2010), Wha would nelson and plosser fnd had hey used panel un roo ess?, Appled Economcs, Vol. 42 No. 12, pp Hused, S. (1992), The emergng u.s. curren accoun defc n he 1980s: a conegraon analyss, Revew of Economcs and Sascs, Vol. 74, pp Im, K.S., Pesaran, M.H. and Shn, Y. (2003), Tesng for un roos n heerogeneous panels, Journal of Economercs, Vol. 115, pp Irandous, M. and Ercsson, J. (2004), Are mpors and expors conegraed? an nernaonal comparson, Meroeconomca, Vol. 55, pp Kao, C. (1999), Spurous regresson and resdual-based ess for conegraon n panel daa, Journal of Economercs, Vol. 90, pp

19 Olaolu Rchard Olayen, and Avral Kumar Twar The susanably of rade accouns of he ASEAN-5 counres. Journal of Chnese Economc and Foregn Trade Sudes (forhcomng) Kao, C. and Chang, M.-H. (1998), On he esmaon and nference of a conegraed regresson n panel daa, Workng Paper, Cener for Polcy Research, Syracuse Unversy. Konya, L. (2009), The susanably of he curren accoun n he czech republc, hungary and slovena, Emprcal Economcs, Vol. 36, pp Konya, L. and Sngh, J.P. (2008), Are ndan expors and mpors conegraed?, Appled Economercs and Inernaonal Developmen, Vol. 8, pp Levn, A., Ln, C. F. and Chu, C. (2002), Un roo ess n panel daa: asympoc and fnesample properes, Journal of Economercs, Vol. 108, pp Maddala, G. S. and Wu, S. (1999), A comparave sudy of un roo ess wh panel daa and a new smple es, Oxford Bullen of Economcs and Sascs, Vol. 61, pp Moon, H.R. and Perron, B. (2004), Tesng for a un roo n panels wh dynamc facors, Journal of Economercs, Vol. 122, pp Narayan, P.K. and Narayan, S. (2005), Are expors and mpors conegraed? evdence from 22 leas developed counres, Appled Economcs Leers, Vol. 12, Obsfeld, M. and Rogoff, K. (1996), Foundaons of Inernaonal Macroeconomcs, Massachuses, The MIT Press. Palm, F.C., Smeekes S, and Urban J.P., (2011), Cross-seconal dependence robus block boosrap panel un roo ess, Journal of Economercs, Vol. 163, pp Paparods, E. and Pols, D.N. (2003), Resdual-based block boosrap for un roo esng, Economerca, Vol. 71, pp Paparods, E. and Pos, D.N. (2001), The connuous-pah block-boosrap, In Asympocs n Sascs and Probably (M. Pur, ed.) pp VSP Publcaons, Zes, The Neherlands. Publcaons: Zes (NL). Parker, C., Paparods, E. and Pols, D.N., (2006), Un roo esng va he saonary boosrap, Journal of Economercs, Vol. 133, pp Pedron, P. (1999), Crcal values for conegraon ess n heerogeneous panels wh mulple regressors, Oxford Bullen of Economcs and Sascs, Vol. 61, pp Pesaran, H.M. (2007), A smple panel un roo es n he presence of cross-secon dependency, Journal of Appled Economercs, Vol. 22, pp Pesaran, M.H., (2004), General dagnosc ess for cross-secon dependence n panels, CESfo Workng Paper, 1233,

20 Olaolu Rchard Olayen, and Avral Kumar Twar The susanably of rade accouns of he ASEAN-5 counres. Journal of Chnese Economc and Foregn Trade Sudes (forhcomng) Robles A.C. (2013), EU servces rade wh ASEAN: owards polcy coherence for developmen?, Journal of European Inegraon, DOI: / Twar, Avral K. (2011), Are expors and mpors conegraed n nda and chna? an emprcal analyss, Economcs Bullen, Vol. 31 No. 1, pp Twar, Avral K. and Pandey, B. (2011), Susanably of rade defcs n Inda, The Emprcal Economcs Leers, Vol. 10 No.9, pp Twar, Avral. K. (2012), Reassessmen of susanably of curren accoun defc n Inda, Souh-Easern Europe Journal of Economcs, Vol. 10 No. 1, pp Twar, Avral. K. (2013), Are rade defcs susanable? Evdence from he ASEAN-fve, Inernaonal Journal of Socal Economcs, Vol. 40 No. 1, pp Upender, M. (2007), Long run equlbrum beween Inda s expors and mpors durng , Appled Economercs and Inernaonal Developmen, Vol. 7, pp Weserlund J. Edgeron D. (2007), A panel boosrap conegraon es, Economcs Leers, Vol. 97, pp Yn, F., and Hamor, S. (2011), The susanably of rade balances n Chna, Economcs Bullen, Vol. 31 No. 3, pp Appendx: Resdual-based Saonary Boosrap Procedure The boosrap resdual re-samplng algorhm proceeds as follows: 1. Esmae Eq. (4) by OLS o oban he resdual seres, u ˆ } ; 2. Esmae Eq. (5) by OLS o oban he AR(1) coeffcen ρˆ and he assocaed resdual seres { ˆ ε, }, where ˆ ε ˆ ˆ ˆ, = u, ρu, 1 ; 3. Consruc he pseudo resduals { ε& &, } by applyng he Resdual-based Saonary Boosrap o ˆ ε }. In order o preserve cross-seconal dependence, we apply he re- {, samplng algorhm o he enre cross-secon; 4. Oban pseudo resdual { u& &, } by cumulang { ε& &, } o obey he null hypohess of noconegraon; 5. Compue E X &, = ˆ µ + ˆ βim +, u&, ; 6. Esmae he conegrang regresson on he daase { E X &,, IM, } : X & && = && ˆ µ + ˆ β IM + u& ; E ˆ.,, 7. Esmae ρ& & by applyng Eq. (5) o he resduals { u& &, } ; {,

21 Olaolu Rchard Olayen, and Avral Kumar Twar The susanably of rade accouns of he ASEAN-5 counres. Journal of Chnese Economc and Foregn Trade Sudes (forhcomng) 8. Repea 3-7 B (large enough) mes