Volatility of Exchange Rate and Export Growth in Pakistan: The Structure and Interdependence in Regional Markets

Transcription

1 Volailiy of Exchange Rae and Expor Growh in Pakisan: The Srucure and Inerdependence in Regional Markes Khalid Musafa Assisan Professor Deparmen of Economics, Universiy of Karachi and Mohammed Nisha, PhD Professor and Chairman, Finance and Economics Insiue of Business Adminisraion, Karachi Absrac The sudy empirically invesigaes he effec of exchange rae volailiy on expors growh beween Pakisan and leading rade parners. The counries are seleced o deermine he bilaeral relaionship beween Pakisan and he oher counries under various regional economic blocks such as SAARC, ASEAN, European, and Asia-Pacific regions. Coinegraion and Error Correcion echniques are used o esablish he empirical relaionship beween exchange rae volailiy and expors growh, using quarerly daa from 1991:3 o 2004:2. The resul indicaes ha he volailiy of exchange rae has negaive and significan effecs boh in he long run and shor run wih major rade parners namely UK and US. Similar paern was observed in case of Ausralia, Bangladesh, and Singapore, where he volume of rade wih Pakisan is comparaively consisen and less volaile. The relaionship beween expors growh and exchange rae volailiy for India and Pakisan is observed only in long run perspecive. However, of counries like New Zealand and Malaysia no empirical relaionship is observed beween expor growh and exchange rae volailiy. Key words: Exchange Rae, Volailiy, Expor growh, Regional inegraion 1. Inroducion. The impac of exchange rae volailiy on he volume of inernaional rade has been sudied inensively since he lae 1970 s when he exchange rae moved from fixed o flexible exchange rae, means facing a volaile real exchange rae. The heory says ha higher exchange rae volailiy will reduce rade by creaing uncerainy abou fuure profi from expor rade. By using he forward markes and by managing he iming of paymens and receips he firm can reduce he uncerainies in he shor run. In he long run, exchange rae volailiy may also affec rade indirecly by influencing firm s

2 invesmen decision. However, he commercial invesors have limied possibiliies of rading claims o fuure operaional cash flows. Thus hey are forced o shif away from risky markes. According o hese argumens, raders are risk averse, and hedging is expensive or impossible; herefore, exchange rae volailiy will reduce risk adjused profi from foreign rade. 1 The high degree of volailiy and uncerainy of exchange rae movemens since he beginning of he generalized floaing in 1973 have led policy makers and researchers o invesigae he naure and exen of he impac of such movemens on he volume of rade. However, hese sudies deals wih he effecs of exchange rae volailiy on rade flows have yielded mixed resuls. On one hand, a number of sudies have argued ha exchange rae volailiy will impose coss on risk averse marke paricipans who will generally respond by favouring domesic o foreign rade a he margin. The argumens views raders as bearing undiversified exchange risk; if hedging is impossible or cosly and raders are risk averse, risk adjused expeced profis from rade would fall when exchange risk increases. Pakisan follows he flexible exchange rae sysem since A he iniial sage he flucuaion of exchange rae is very nominal. However, expors evolved broadly in line wih oal world impors. Pakisan s share in world impors was sable during he las 24 years, ranging beween a minimum of 0.12 percen in 1980 and a maximum of 0.18 percen in In he share was 0.17 percen. This suggess ha Pakisan s expors performance was based on he volailiy of exchange rae. Only one empirical sudy is available regarding o Pakisan s conex. Kumar and Dhawan (1991) esimaed he exchange rae volailiy on Pakisan expors o he developed world from 1974 o They found ha volailiy of exchange rae adversely effec on expor demand. They also invesigaed he hird counry effec and suggesed ha Japan and Wes Germany ac as he alernae marke for Pakisan s expor o he Unied Saes and Unied Kingdom. The objecive of his paper is o invesigae he effec of exchange rae volailiy on expors growh beween Pakisan and oher leading rade parners during The counries are seleced from various regions o capure he varying impac of level and degrees of bilaeral relaionship beween Pakisan and oher counries. Therefore regional counries included are SAARC (India and Bangladesh), ASEAN (Singapore and Malaysia), European (UK), and Asia-Pacific (Ausralia and New Zealand) and Norh America (US). The res of he paper is organised such ha second secion describes he. The daa descripion is provided in secion hree followed by discussion of resuls in secion four. The summary and concluding remarks are given in secion five. 1 A very risk averse exporer who worries abou he decline in revenue may expor more when risks are higher. On he oher hand a less risk averse individual may no be considered wih wors possible oucome and, considering he reurns on expors less aracive, may decide o expor less when risks are higher. 1

3 2. Lieraure Review and Theoreical Framework The conradicory resuls abou he impac of exchange rae volailiy on inernaional rade are observed. Sudies ha suppor he hypohesis ha he volailiy of exchange rae reduces he volume of rade are included Cushman (1983, 1986, 1988); Akhar and Hilon (1984); Kenen and Rodrick (1986); Thursby and Thursby (1987); De Grauwe (1988); Pere and Seinherr (1986); Koray and Lasrapes (1989); and Arize (1995). On he oher hand, Hooper and Kohlhagen (1978), Gour (1985), Bailey, Tavlas and Ulan (1987), and Asseery and Peel (1991) found no evidence abou he impac of exchange rae volailiy on rade. Hooper and Kohlhagen (1978) was he firs sudy o analyze sysemaically he effecs of exchange rae uncerainy on he rade. They invesigaed bilaeral and mulilaeral rade among developed counries during They measured exchange rae risk by sandard error of nominal exchange rae flucuaions. They could no esablish any significan impac of exchange rae volailiy on he volume of rade. They measured he exchange rae risk volailiy as he sandard error of nominal exchange rae funcion. Laer Cushman (1983) inroduced he real exchange rae raher han nominal exchange rae and found negaive relaion among he exchange rae volailiy and volume of rade. In anoher sudy Cushman (1986) inroduced also he hird counry effec and argued ha he recogniion of hird counries in he analyical framework implies ha he effec of exchange rae variabiliy on bilaeral rade flows no only depend upon he exchange rae risk experienced by he counry under consideraion bu also depend upon he correlaion of he exchange rae flucuaions by oher counries. Akher and Hilon (1984) examined he bilaeral rade beween Wes Germany and US. They deermined ha he exchange rae volailiy has a significan negaive impac on he expors and impors of wo counries. However, he volailiy of exchange rae has been measured by he sandard deviaion of effecive exchange raes. Gour (1985) rejeced he resul of Akher and Hilon (1984). He added he counries in Akher and Hilon (1984) models i.e. France Japan, UK and increasing he sample period and he measures of exchange rae risks. He did no observe any significan relaion beween exchange rae volailiy and volume of rade on he bilaeral rade flows. His resul is idenical o IMF (1984) sudy on his issue. Chowdhury (1993) invesigaed he impac of exchange rae volailiy on he rade flows of he G-7 counries in conex of a mulivariae error-correcion model. They found ha he exchange rae volailiy has a significan negaive impac on he volume of expors in each of he G-7 counries. Baak, Mahmood, and Vixahep (2002) invesigaed he impac of exchange rae volailiy on expors in four Eas Asians counries (Hong Kong, Souh Korea, Singapore, and Thailand). Their resuls indicaed ha exchange rae volailiy has negaive impacs on expors in boh he shor run and long run periods. The empirical evidences regard he impac of exchange rae volailiy on expor growh o developing counries are inconclusive as hey have explained variaion in exchange rae policies and level of growh Bahmani-Oskooee (1984, 1986); Coes (1981); and Rana 2

4 (1983). Bahmani-Oskooee (1984, 1986) found ha exchange rae has a significan impac on rade flows of seleced developing counries even in periods when mos of hem had pegged exchange raes. Coes (1981) and Rana 1983) analysed his issue on he basis of Hooper-Kohlhagen (1978) sudy using annual daa. Coes (1981) examines Brazilian expors (as a proporion of he oal value added) in 9 primary and 13 manufacuring secors for His resul indicaed ha he significan reducion in exchange rae uncerainy in he Brazilian economy during he crawling peg period migh have conribued as much as he changes in prices oward explain he greaer openness of he economy afer Rana (1983) sudy is he mos horough sudy in conex of developing counries. He reached he same resuls regarding he impor volumes of a number of Souheas Asian counries some of which are also included in he Bahamani- Oskooee (1984) sample. Rana (1983) esimaed he impor demand funcion for each counry in he sample. He concluded ha he increase in exchange rae risk has a significan negaive impac on impor volumes. He did no analyze expor volumes in he same manner alhough hey are likely o be of greaer ineres. Kabir (1988) used he sandard regression model o invesigae he Bangladesh expor demand funcion. He found evidence for income inelasic demand for expors. Ahmed, Haque and Talukder (1993) esimaed an expor demand funcion using co inegraion and error correcion model. Their resuls are similar o Kabir (1988) resul regarding o expor demand funcion for Bangladesh Expor. However, hey concluded ha he cos efficiency by lowering price migh no boos he expor demand significanly. Bayes, Hossein and Rahman (1995) has hypohesized ha Bangladesh expor supply is a funcion of relaive prices of is expors and he capaciy oupu of he radable secor. They have esimaed he demand and supply models of expors wih annual daa and found ha Bangladeshis expor is highly sensiive o he income growh of is rading parners and esimaed ha a 10% rise in a foreign income would raise he demand for Bangladeshi expors by 23%. 3. Economeric Models and Specificaion I is concluded ha differen sudies have differen resuls. The reasons are differen mehodology, he differen sample period, and differen esimaion echniques. The characerisics of he above sudies are: hey did no recognize ha rade flows and he variables explaining he as relaive price measures and oupus are likely o be nonsaionery; he economeric mehodology used in hese sudies used only he problems of e shor run perspecive, ha is why if resul found any evidence regarding o he relaionship beween volailiy and rade flows i is mos likely medium or shor run relaionship. Based on he above discussion he following equaion is esimaed: X = ξ + ε 0+ 1 ξ. i + ξ2 p + ξσ 3 (1) 3

5 where X denoes real expors from Pakisan o oher counries seleced in differen regions, p is he real bilaeral exchange rae reflecing he price compeiiveness, i is he manufacuring producion index of imporing counry which is he proxy for GDP, because he quarerly daa on GDP is no available. σ is he exchange rae volailiy. The sign of ξ1 is expeced o be posiive and he sign of ξ2 is also o be posiive because higher exchange rae implies a lower relaive price ha increases expor. In order o ensure consisency in daa, he expors of Pakisan measured in local currency and o conver o real expor we used expor uni index, which is based on Pakisan currency. Real expors of Pakisan define as X EX i = Ln X EXUVi 100 (2) Where X is he real expor of Pakisan in domesic currency uni naural logarihm EX is he monhly nominal expors of Pakisan in domesic currency and EXUV is he index of expor uni of Pakisan and is he ime period. Indusrial producion index ( i ) is used as a proxy for GDP of imporing counry because non-availabiliy of quarerly daa on GDP. Many sudy has been used he indusrial producion index as proxy variable e.g. Baum, Calagy and Ozkan (2002). The variable i is he naural logarihm of he indusrial producion index of an imporing counry. Bilaeral rade beween wo counries depends upon he exchange rae and he relaive price level of wo rading parner counries. Hence real exchange rae is calculaed on he basis of hese variables. The real exchange rae is p i = Ln E i CPI x CPI Where Pakisan and oher rading parners. j i pi is he real quarerly exchange rae beween in naural logarihm beween Ei is he nominal quarerly exchange rae: CPIi and CPI is he consumer price index number of Pakisan and an imporing counry j j respecively. Various sudies provide he formula for he measuremen of exchange rae risk. However, in his sudy he sandard deviaion of exchange rae risk is used which is also used by Akhar and Hilon (1984) and Baum, Calagyan and Ozkan (2002). The exchange rae volailiy define in naural logarihm (3) 4

6 σ ij n 1 = Ln n 1 k = 1 2 ( RER RERi ) (4) ik Where σ i is he volailiy of real exchange rae and RER i is he quarerly exchange rae of Pakisan and RERi is he quarerly average of real exchange rae. We es real expor ( X ) of Pakisan wih real exchange rae volailiy ( σ ) wih he combinaion of he real bilaeral exchange rae ( p ) and indusrial producion index ( i ). If X and equilibrium relaionship, hen σ are considered o be sochasic rends and if hey follow a common long run X and σ should be coinegraed. Coinegraion is a es for equilibrium beween non-saionary variables inegraed of same order. According o Engle and Granger (1987), coinegraed variables mus have an ECM represenaion. The main reason for he populariy of coinegraion analysis is ha i provides a formal background for esing and esimaing shor run and long run relaionships among economic variables. Furhermore, he ECM sraegy provides an answer o he problem of spurious correlaion. If X and σ are coinegrae, an ECM represenaion could have he following form. n n n n 0 + α1b 1 + α 2i X i + α3i σ 1 + α 4 i 1 + α5 i = 0 i = 0 i = 0 i = 0 X = α p + e (5) i Where B 1 is an error correcion erm. In equaion (1) x, σ and e are saionary, a firs difference implying ha here righ hand side mus also be saionary. I is obvious ha equaion 1 composes a bi-variae vecor auoregression (VAR) in firs difference augmened by he error correcion erms B 1 indicaing ha ECM and coinegraion are equivalen represenaions. According o Granger (1988) in a coinegraed sysem of wo series expressed by an ECM represenaion, causaliy mus run in a leas one way. Wihin he ECM formulaion of equaion (1) X does no granger cause σ if α 1 =α 3 = Daa The daa used in his sudy is quarerly covered from 1991:3 o 2004:2. The daa for nominal expors ( EX ) is aken from various issues of Saisical Buellon presened by Federal Bureau of Saisics governmen of Pakisan. The daa for expor uni value of Pakisan ( EXUV ), he indusrial producion index of imporing counry ( i ), consumer price index of Pakisan ( CPI ) and consumer price index of imporing counry ( CPI ) i are aken from various issues of Inernaional Financial Saisics (IFS) of Inernaional j 5

7 Moneary Fund (IMF). The daa for nominal exchange rae aken from several issues of Monhly Saisical Bullein published by Sae Bank of Pakisan. 5. Esimaion and inerpreaion of Resuls Table 1 shows he expor of Pakisan o Ausralia, Bangladesh, India, Malaysia, New Zealand, Singapore, UK and US during he sudy period. The daa shows ha a large porion of rade goes o he US and UK. However, consisency is found in Newzealand, Ausralia, Bangladesh, and India. Coinegraion es requires he series of all variables o be saionary. Augmened Dickey Fuller (1979) uni roo es, which also checks for serial correlaion is performed. The resuls presened in able 2 indicaes ha series of all four variables are each I(1) wih consan and ime rend in he daa a he level. Subsequenly Johanson (1988, 1991) coinegraion es is employed. This es is more appropriae when more han wo variables are used in he equaion and i also can make use of I(0) variables. The null hypohesis is ha here can be r coinegraing vecors among four variables sysem ( X, σ, p, i ) for all counries, which are aken in he sudy periods. The es saisics implies he presence of one coinegraing relaionship for all four variables in all counries. The ADF saisics of a he level of all series are lower han he criical value which implies he presence of uni roos of all four variables i.e.each I(1). However, he resuls derived form firs difference of he variables rejec he null hypohesis of a uni roo a leas five percen level of significance. The coinegraing vecors are given in able 4, which shows ha for each counry he impac of indusrial producion is posiively relaed o he volume of expors excep New Zealand, Singapore, and UK. The expeced sign of i is posiive. I indicaes ha he higher he economic aciviy in imporing counry, he higher he demand for expors. However he negaive sign shows ha he higher economic aciviy in imporing counry leads o decrease in he volume of expors. This implies ha ha Pakisani commodiies are considered as inferior goods in New Zealand, Singapore, and UK. The value of coefficien i is ranging from o The relaion of real exchange rae o he volume of expor is expeced o be posiive. I indicaes ha a higher real exchange rae implies a lower relaive price, and as resul he volume of expors increases. Empirical evidence shows ha he posiive signs for is relaionship in case of Ausralia, Bangladesh, Malaysia, Singapore, and UK. Whereas negaive signs are for India, New Zealand, and US. I implies ha our demand for expors in hese counries is inelasic. The volailiy of exchange rae has expeced negaive relaionship wih real expor in all counries. I suppors o he sudy of Cushman (1983, 1986, 1988); Akhar and Hilon (1984); Kenen and Rodrick (1986); Thursby and Thursby (1987); De Grauwe (1988); Pere and Seiner (1986); Koray and Lasrapes (1989); and Arize (1995). The causal relaionship beween X and σ are presened in ables 5 wihin he ECMs form. Amos hree lags are used for each independen variable o conserve degree of freedom and AIC is used for model selecion. While error correcion erms B -1 appearing 6

8 as regressors reflec long run dynamics or in oher words he sysem converges o he long run equilibrium implied by coinegraing regression. The coefficien of B -1 represens he response of he dependen variables in each period o deparure from equilibrium. The coefficiens on he lagged values of X, σ, ι, and p are shor run parameers measuring he shor run immediae impac of independen variable on X. The resuls indicaes ha he error correcion erms B 1 are negaive sign and saisically significan in case of Ausralia, Bangladesh, India, and Singapore. I indicaes ha a measure of he average speed a which expor volume adjuss o a change in equilibrium condiions. The absolue values of he error correcion erms indicae ha he movemen of real expor owards eliminaing disequilibrium wih in a quarer varies from one counry o anoher. e.g. in case of Ausralia only 246% of he adjusmen occur in one quarer while 80.7% in Bangladesh, 3.5% in India, 37.8% in Singapore and 69.1% for US. and UK. The negaive sign are observed for error correcion model in case of New Zealand and Malaysia, which provide he evidence of adjusmen. However, he values are saisically insignifican. The coefficien on he indusrial manufacuring producion (i ) and real exchange rae on real expor show how he average speed of expor adjuss or i may differ. I depends he adjusmen in response o indusrial producion or real exchange rae. The resul is ambiguous regarding o he relaion ship beween real exchange rae and expors demand and indusrial producion.. Our main focus is o see he impac of exchange rae volailiy on expor of all counries, which are aken in his sudy. I indicaes he ambiguous resuls, e.g. in case of Ausralia and Singapore he resul shows negaive and significan impac on real expor. However he esimaion of he oher counries show he saisically insignifican resul. The reason is ha he Pakisan economy is Dollar economy and is expors and impors depend on Dollar. Tha is why bilaeral exchange rae is less effec on real expor. However he resul regarding o US is negaive and insignifican even Pakisan economy is Dollar economy. I is an imporan empirical finding. 6. Summary and Concluding Remarks Impac of exchange rae volailiy on expors growh beween Pakisan and leading rade parners has been invesigaed. The counries are seleced under various regional economic blocks such as SAARC, ASEAN, European, and Asia-Pacific regions.. Coinegraion and Error Correcion echniques are used o esablish he empirical relaionship beween exchange rae volailiy and expors growh, using quarerly daa from 1991:3 o 2004:2. The resul indicaes ha he volailiy of exchange rae has negaive and significan effecs boh in he long run and shor run wih UK, US, Ausralia, Bangladesh, and Singapore, where he volume of rade wih Pakisan is comparaively consisen and less volaile. The relaionship beween expors growh and exchange rae volailiy for India and Pakisan is observed only in long run perspecive. However, of counries like New Zealand and Malaysia no empirical relaionship is observed beween expor growh and exchange rae volailiy. 7

9 REFERENCES Ahmed, M., Haque, N., and Talukder, S. I. (1993), Esimaing Expor Demand Funcion for Bangladesh: An Applicaion of Co-inegraion and Error Correcion Modeling, The Bangladesh Developmen Sudies, 21, Akhar, M., and Hilon, R. S. (1984), Effecs of Exchange Rae Uncerainy on German and U.S. Trade, Federal Reserve Bank of New York Quarerly Review, 9, Arize, A. (1995), The Effecs of Exchange Rae Volailiy on U.S. Expors: An Empirical Invesigaion, Souhern Economic Journal, 62, Asseery, A., and Peel, D. A. (1991), The Effecs of Exchange Rae Volailiy on Expors, Economic Leers, 37, Baak, S., Al-Mahmood, A., and Vixahep, S., (2002), Exchange Rae Volailiy and Expors from Eas Asian Counires o Japan and U.S., (Manuscrip). Universiy of Japan. Bahmani-Oskooee, M. (1984), On he Effecs of Effecive Exchange Raes on Trade Flows, Indian Journal of Economics, 256, Bahmani-Oskooee, M., (1986), The Deerminans of Trade Flows: The Case of Developmen Counries, Journal of Developmen Economics, 20, Bailey, M. J., Tavlas, G. S., and Ulan, M. (1987), The Impac of Exchange Rae Volailiy on Expor Growh: Some Theoreical Consideraion and Empirical Resuls, Journal of Policy Modeling, 9, Baldwin, R. and Krugman, P. (1989), Persisen Trade Effecs of Large Exchange Rae Shocks, Quarerly Journal of Economics, 104, Baum, C. F., Musafa, C., and Ozkan, N., (2001), Exchange Rae Effecs on he Volume of Trade Flows, An Empirical Analysis Employing High-Frequency Daa, (Manuscrip). Boson College. Bayes. A. M., Hossain. I., and Rahman, M. (1995), Independen Review of Bangladesh s Developmen Exernal Secor, Cenre for Policy Dialogue, Dhaka, Bangladesh. Chowdhury, A. R. (1993), Does Exchange Rae Volailiy Depress Trade Flows? Evidence from Error-Correcion Models, The Review of Economics and Saisics. 75,

10 Coes, D. (1981), The Crawling Peg and Exchange Rae Uncerainy, in J. Williamson (ed) New York: SDT. Marin s Press, Cushman, D. O. (1983), The Effecs of Real Exchange Rae Risk on Inernaional Trade, Journal of Inernaional Economics, 15, Cushman, D. O. (1986), Has Exchange Risk Depressed Inernaional Trade? The Impac of Third Counry Exchange Risk, Journal of Inernaional Money and Finance, 5, Cushman, D. O. (1988), US Bilaeral Trade Flows and Exchange Rae Risk During he Floaing Period, Journal of Inernaional Economics, 25, DeGrauwe, P. (1988), Exchange Rae Variabiliy and he Slowdown in Growh of Inernaional Trade, IMF Saff Papers, 35, Dixi, A. (1989), Hyseresis, Impor Peneraion, and Exchange Rae Pass-hrough, Quarely Journal of Economics, 104, Gour, P. (1985), Effecs of Exchange Rae Volailiy on Trade: Some furher Evidence, IMF Saff Papers, 32, Hooper, P., and Kohlhagen, S. W. (1978), The Effecs of Exchange Rae Uncerainy on he Price and Volume of Inernaional Trade, Journal of Inernaional Economics, 8, Houhakkar, H. S., and Magee, S. (1969), Income and Price Elasiciies in World Trade, Review of Economics and Saisics 51, Inernaional Moneary Fund, (1984), Exchange Rae Volailiy and World Trade, IMF Occasional Papers No. 28 Kabir, R. (1988). Esimaing Impor and Expor Demand Funcion: The case of Bangladesh, The Bangladesh Developmen Sudies, Kenen, P. T., and Rodrick, D. (1986), Measuring and Analyzing he Effecs of Shor Run Volailiy in Real Exchange Raes, The Review of Economics and Saisics, 68, Koray, F., and Lasrapes, W. D. (1989), Real Exchange Rae Volailiy and U.S. Bilaeral Trade: A VAR approach, The Review of Economics and Saisics, 71, Kumar, R. and Dhawan, R. (1991), Exchange Rae Volailiy and Pakisan s Expor o he Developed World, World Developmen, 19,

11 Pere, E., and Seinherr, A. (1989), Exchange Rae Uncerainy and Foreign Trade, European Economic Review, 33, Rana, P. (1983), The Impac of Generalized Floaing on Trade Flows and Reserve Needs, Seleced Asian Developing Counries, New York Garland Publishers. Thursby, M. C., and Thursby, J. G. (1987), Bilaeral Trade Flows, Linders Hypohesis, and Exchange Risk, The Review of Economics and Saisics. 69, Warner, D. F., and Kreinin, M. E. (1983), Deerminans of Inernaional Flows, The Review of Economics and Saisics, 65,

12 TABLE # 1 EXPORT OF PAKISTAN Years Ausralia Bangladesh India Malaysia Expor % Expor % Expor % Expor % TABLE # 1 (Cond.) EXPORT OF PAKISTAN Years Singapore New Zealand Unied Kingdom Unied Sae Expor % Expor % Expor % Expor %

13 TABLE # 2 AUGMENTED DICKEY FULLER TEST (ADF) Variables Wih Inercep Wih inercep and rend N Criical values Level Is diff. Level Is diff. 1% 5% 10% AUSTRALIA Real Expor Real Ex. Rae IPI Sigma BANGLADESH Real Expor Real Ex. Rae IPI Sigma INDIA Real Expor Real Ex. Rae IPI Sigma MALAYSIA Real Expor Real Ex. Rae IPI Sigma NEW ZEALAND Real Expor Real Ex. Rae IPI Sigma SINGAPORE Real Expor Real Ex. Rae IPI Sigma UNITED KINGDOM Real Expor Real Ex. Rae IPI Sigma UNITED STATE Real Expor Real Ex. Rae IPI Sigma

14 TABLE # 3 JOHANSEN CO-INTEGRATION TESTS FOR EXPORTS Trace Saisics Maximum Eigen value H 0 H 1 r =0 r 1 r 1 r 2 r 2 r 3 r r = 4 r=0 r=1 r 1 r =2 r 2 r=3 r 3 r =4 Ausralia Bangladesh India Malaysia NewZealand Singapore UK US CRITICAL VALUES H 0 H 1 Ausralia 5% 1% Bangladesh 5% 1% India 5% 1% Malaysia 5% 1% NewZealand 5% 1% Singapore 5% 1% UK 5% 1% US 5% 1% r =0 r r 1 r r 2 r r 3 r =

15 TABLE # 4 ESTIMATES OF THE COINTEGRATING VECTORS Normalized Coinegraing Coefficiens: 1 Coinegraing Equaion Ausralia (SE) Bangladesh (SE) India (SE) Malaysia (SE) New Zealand (SE) Singapore (SE) UK (SE) US (SE) C IPI REALER SIGMA TREND (7.727) (325.47) (27.741) (16.909) (0.007) (0.758) (0.034) (0.021) (0.037) (0.712) (0.927) (0.063) (0.001) (0.145) (0.083) (0.007) (1.453) (112.22) ( ) (5.251) (0.002) (0.692) (0.039) (0.003) (0.843) (44.356) (22.920) (0.908) (0.005) (0.398) (0.035) (0.005) 14

16 TABLE # 5 REGRESSION RESULTS FOR ERROR CORRECTION MODELS Variables AUSTRA LIA Consan (559.02) (2.23) R.Exp(-1) ** (2.18) (-2.61) R.Exp(-2) ** (2.80) (-2.19) R.Exp(-3) (2.387) (-1.61) IPI(-1) (72.52) (-1.78) IPI(-2) (91.02) (-1.425) IPI(-3) (68.66) (-1.47) R.ER(-1) -4148** ( ) (-2.801) R.ER(-2) ( ) (-1.799) R.ER(-3) ( ) (-1.689) Sigma(-1) ** (70.70) (-2.39) Sigma(-2) (119.79) (-0.42) Sigma(-3) ** (59.19) (-2.70) B ** (1.59) (-2.54) BANGL ADESH (0.11) (-1.84) (0.26) (1.18) (0.18) (-1.55) (0.16) (-0.48) (0.02) (1.82) (0.01) (1.87) (0.01) (2.70) (1.75) (0.23) (1.67) (0.07) ** (0.03) (-2.85) (0.04) (0.28) (0.04) (1.94) ** (0.316) INDIA (0.078) (-0.087) (0.27) (-2.82) (0.26) (-1.19) (0.00) (-2.74) (0.00) (-1.44) (0.21) (1.76) (0.19) (1.21) (0.06) (1.53) (0.06) (1.14) ** (0.009) (-3.64) MALAY SIA (0.043) (0.28) (0.22) (-2.30) (0.24) (-1.58) (0.192) (-0.515) (0.000) -(1.02) (0.004) (0.092) (0.00) (0.82) (0.62) (0.014) (0.055) (0.115) (0.040) (0.918) (0.049) (0.698) (0.047) (0.946) (0.04) (1.52) (0.035) -0.64) NEWZE ALAND (0.05) (1.44) (0.28) (-0.94) (0.78) (-0.14) (0.28) (-0.17) (0.10) (-1.04) (0.10) (-0.77) (0.019) (0.26) (0.36) (-1.67) (0.37) (-1.02) (0.37) (0.95) (1.61) (-0.53) (1.5) (-0.83) (1.55) (-0.05) (0.114) (-0.23) SINGAP ORE (0.046) (-0.163) (0.19) (-0.94) (0.166) (-0.97) (0.00) (-0.71) (0.0027) (0.572) (1.80) (1.097) (1.65) (-1.75) (0.049) (1.044) ** (0.052) (-3.98) ** (0.218) (-2.73) UK (0.03) (0.84) (0.16) (-2.80) (0.164) (-1.009) (0.00) (1.15) (0.00) (-0.24) 1.11 (0.99) (1.12) (0.976) (-0.94) ** (0.040) (-2.487) (0.46) (-1.27) ** (0.035) (-2.39) US (0.05) (-0.19) (0.29) (0.29) (0.226) (-0.14) (0.17) (-1.51) (0.025) (-0.544) 0.02 (0.026) (0.75) (0.028) (1.91) (1.15) (0.36) (1.77) (-1.049) 0.76 (1.19) (0.63) ** (0.06) (-2.32) (0.059) (-0.027) (0.056) (-0.30) ** (0.38) (-2.77) (-2.55) R Adjused R AIC N