Asian Journal of Empirical Research

Transcription

1 . Asian Journal of Empirical Research journal homepage: hp://aessweb.com/journal-deail.php?id=5004 THE DYNAMIC RELATIONSHIP BETWEEN THE FOREIGN EXCHANGE EXPOSURE AND STOCK MARKETS: EVIDENCE DURING THE GLOBAL ECONOMIC CRISIS Nik Muhamad Naziman Ab RAHMAN Yusrina hayai Nik Muhd NAZIMAN 2 Amri Ab. RAHMAN 3 ABSTRACT This paper examines he dynamic linkages beween he foreign exchange and sock markes for five Eas Asian counries, including Hong Kong, Japan, Malaysia, Singapore, and Thailand. While he lieraure suggess he exisence of significan Ineracions beween he wo markes, our empirical resuls show ha, in general, exchange raes Granger-cause sock prices wih less significan causal relaions from sock prices o exchange raes. Furhermore, his one-way Granger causaliy effec from exchange raes o sock prices becomes less significan during he US financial crisis of Our resuls also sugges ha, here is insignifican long-run oulook (no coinegraion) excep for Hong Kong, implies ha hese financial asses share on common rends in heir economy sysem and hence hey will move apar in he long-run for counries ha have higher rade size exchange rae flucuaions end o exhibi significan influence on he equiy marke, regardless of he exchange rae arrangemen sysem and he degree of capial conrols during he US financial crisis of Keywords: Foreign exchange Exposure, Sock markes, Global crisis INTRODUCTION Argumen of wheher sock prices and exchange raes are relaed or no, has received considerable aenion afer he Eas Asian crises. The financial crisis ha affecs he counries affeced saw Universii Teknologi MARA (UiTM) Kelanan Campus, Buki Ilmu, 8500 Machang, Kelanan, Malaysia naziman925@kelanan.uim.edu.my 2 Universii Teknologi MARA (UiTM) Kelanan Campus, Buki Ilmu, 8500 Machang, Kelanan, Malaysia yusrina@kelanan.uim.edu.my 3 Universii Teknologi MARA (UiTM) Kelanan Campus, Buki Ilmu, 8500 Machang, Kelanan, Malaysia amri@kelanan.uim.edu.my 763

2 Asian Journal of Empirical Research, 3(6)203: urmoil in boh currency and sock markes. If exchange raes and sock prices are relaed and he causaion runs from exchange raes o sock prices hen crises in he sock markes can be prevened by conrolling he exchange raes. Likewise hese developing counries can exploi such a link o arac/simulae foreign porfolio invesmen in heir own counries. Similarly, if here is a causal relaionship from sock prices o exchange raes hen auhoriies can focus on domesic economic policies o sabilize he sock marke. Also hese will benefi he invesors if he wo markes/prices are relaed hen hey can use his informaion o predic he behavior of one marke using he informaion on oher marke. Review from he relaed lieraure shows ha he sock prices-exchange rae relaionship has focused on examining his relaionship for he developed counries wih very lile aenion on he developing counries. The findings are ambiguous. Smih (992), Solnik (987), and Aggarwal (98)) found ha is a significan posiive relaionship beween sock prices and exchange raes while Soenen and Hennigar (998) have repored a significan negaive relaionship beween he sock prices-exchange rae relaionship. On he oher hand, Franck and Young (972), Eli Barov and Bodnor (994) in heir sudies found ha here is very weak or no associaion beween sock prices and exchange raes. In relaion o causaion, he evidence is also mixed. Abdalla and Murinde (997) have found causaion runs from exchange raes o sock prices while Ajayi and Mougoue (996) repored ha a reverse causaion. Bahmani-Oskooee and Sohrabian, (992) exposed ha he relaionship beween sock prices and exchange raes is bidirecional. Based on he heoreical consensus, here is no proper relaionship beween sock prices and exchange raes eiher. In evaluaing he porfolio balance models for examples, he exchange rae deerminaion posulae a negaive relaionship beween sock prices and exchange raes and ha he causaion runs from sock prices o exchange raes. The exchange raes play he main role in balancing he demand for and supply of asses. Due o an increase in domesic sock prices lead individuals o demand more domesic asses. Buying more domesic asses local invesors would sell foreign asses (hey are relaively less aracive now), causing local currency appreciaion. Due o an increase in wealh can rise in domesic asse prices will also lead invesors o increase heir demand for money, which in urn raises domesic ineres raes. Then his will again leads o appreciaion of domesic currency by aracing foreign capial. In conras for he same negaive relaionship is increase in foreign demand for domesic asses due o sock price increase would also cause a domesic currency appreciaion. While he argumen, if a posiive relaionship beween sock prices and exchange raes wih direcion of causaion running from exchange raes o sock prices can be explained as follows: a domesic currency depreciaion makes local firms more compeiive, leading o an increase in heir expors and heir sock prices. The facors ha cause changes in exchange raes may be differen from he facors ha cause changes in sock prices. Under such scenario, i is clear ha here is no empirical or heoreical consensus on he issue of 764

3 Asian Journal of Empirical Research, 3(6)203: wheher sock prices and exchange raes are relaed and he direcion of causaion if hey are relaed. Objecives of he sudy The primary objecive of his sudy is o invesigae he dynamics ineracions beween exchange raes and sock prices using recen developmens in ime series modelling. This sudy invesigaes wheher hese naional sock markes in he periods are moving in andem and in equilibrium or hey depar permanenly from each oher in shor and long run. The Asian emerging markes seleced for analysis are Malaysia, Thailand Singapore and world major equiy markes are Hong Kong and Japan in his region. The research employed he Johansen and Juselius (990) mulivariae approach o es for long-run marke inegraion. The es is powerful o esimae he inerdependence of muli-counries sock markes. Tes on serial co-movemen of sock prices will esablish wheher a sable long-run coinegraing relaionship exiss among he seleced markes. The dynamic approach suggesed by Johansen and Juselius (990) allows marke srucure o change hroughou he sample period. We exend he analysis o assess long and shor-run marke inegraion or o gain insigh ino he shor-run and long run lead-lag or causal relaionships beween hese major ASIAN emerging markes and world markes by using hree separae model of five-dimensional vecor auoregressive (VAR) model including vecor error-correcion modelling (VECM). The dynamic VECM represenaion provides us wih a framework o es for he emporal causal dynamics in he Granger sense among he price indexes hrough boh shor-run and error-correcion channels of causaion (Granger, 988). Shor-run VAR marke inegraion es will deermine wheher prices in differen markes respond immediaely o changes in oher equiy markes. This sudy differs from mos of he exising lieraure in he following aspecs: () The sudy employs more recen daily observaion covering he period of he US financial crisis from January 2008 o February (2) The sudy considers he dynamic relaionships of marke liberalisaionon he long run relaionships among he equiy markes. (3) The mulivariae Johansen and Jusellus (990) maximum likelihood procedure is supplemened by vecor error correcion modelling mehods o analyse dynamic aspec of markes inegraion. METHODOLOGY The prerequisie condiion for he series o be coinegraed is ha he series mus have he same order of inegraion. The order of inegraion of a series is deermined by he number of imes ha he series mus be difference before achieving saionary. A series, Y is said o be inegraed of order d if he series achieves saionary afer differencing d imes and denoed as Y ~ I(d). For insance, if price series (Y ) is no saionary a is level bu becomes so afer firs differencing, (i.e. Y Y - is saionary) we describe his as Y ~I(). If Y is saionary a is level before firs 765

4 Asian Journal of Empirical Research, 3(6)203: difference, hen we describe i as Y ~I(0). Thus he very beginning sep in he coinegraion analysis is o deermine he order of inegraion of he series. We ADF and PP uni roo es o es he saionary properies of he variables. Schwer (987) and laer on, Campbell and Perron (99) noed ha ADF is beer for small-sample daa se. In esing he order of inegraion using ADF approach, he following wo ADF regression equaions could be esimaed: Y = 0 Y = 0 L + Y + i Y i + () + Y i + T + i Y i + (2) 2 L i Where Y is he firs difference of he series, 0 is inercep, and 2 are consan, and are disurbance erms, T is ime or rend variable and L is he number of lagged erms. To ensure disurbance erm and are approximaely whie noise, a sufficien number of lagged differences L should be esimaed. The opimum lag lengh L may be deermined by using he Akaike Informaion Crieria (AIC) suggesed by Akaike (977). The null hypohesis is ha he level of he series, Y, conains a uni roo H 0 : Y is I() and he alernaive hypohesis is ha H : Y is no I(). We rejec he null hypohesis when is found o be negaive and saisically significan. The rejecion (or accepance) of he null hypohesis is made by calculaing a -raio of o is sandard error. The criical value for he es is compared o criical values provided by Fuller (976). The uni roo es in level is only necessary bu no sufficien condiion for he series o be inegraed of order one, (I ()). To conform ha he series is I(), hen he sufficien condiion has o be esed using uni roo es on he firs difference for equaions and 2. We follow given regression for empirical analysis: L Y = 0 + Y +, i Y i + (3) i Y = 0 + Y + T + i Y i + (4) 2 L i Where Y is he firs difference of he series. The null hypohesis is H o : Y ~I(), which is rejeced in favour of I(2) if is found o be negaive and saisically significan from zero. This es is known as uni roo es in firs difference. Phillips Perron (PP) uni roo es proposed by Phillips and Perron (988) is more robus in he sense ha PP allows for wide variey of serial correlaion and ime dependen heeroskedasiciy. I is also has been considered o be powerful es o moderae and small sample size. The PP es esimaes he following equaions for a series Y : 766

5 Asian Journal of Empirical Research, 3(6)203: Y Y (5) Y Y 2 (6) Where Y is he firs difference of Y -, is rend variable. In equaion 5, for Y o be saionary, he adjused -saisic Z( * ) should be negaive and significanly difference from zero. For Y o be saionary around linear rend in equaion 5, he adjused -saisic Z(^ ) should be negaive and significanly differen from zero. The criical value for PP ess is given in MacKinnon (99). Like he ADF es, he PP es is also sensiive o he choice of runcaed lag parameers. The crieria discussed in Schwer (989) may be used o deermine he appropriae lag lengh in he PP ess. Johansen coinegraion approach We apply he Johansen maximum-likelihood approach o examine he coinegraion beween he variables. This approach permis in examining linear resricion for coinegraing esimaes (Perman, 99). To illusrae his approach, le Y be a vecor of N ime series variables, each of which is inegraed of order. Assume ha Y can be modelled by he vecor auoregression, Y Y... Y k k Where = T (7) Here Y is Nx vecor of sochasic variables; all Y -k are assumed predeermined; is a Nx vecor of consan; is a vecor of normal disribued error wih zero mean and consan variance; and k is he maximum number of lag lengh processing he whie noise. The lag lengh of k is chosen by using he Akaike Final Predicion Errors (FPE) crierion. In brief, he echnique chooses he lengh which minimise he forecas error of he series. The following formulaion is used: FPE = [( T 2 k) /( T k)] (8) where T is he number of observaions, k is he number of lags and 2 is variance. The sysem of equaion 7 can be rewrien in he firs difference and in he reduce form as follows: Y Y e (9) Y k k k Where = -[ I i ], ( i =,.., k- ) And = - [ I k ] Equaion 9 is in he form of radiional VAR model of Sims (980) in firs differences excep for he Y -k erm. The marix is called he long-run impac marix. This erm deermines wheher or no, and o wha exen, he sysem of equaion is coinegraed. The rank of he marix shows he 767

6 Asian Journal of Empirical Research, 3(6)203: number of coinegraing vecors. If he value of he marix is r, hen here are r coinegraing relaionships among he elemens of Y. When r = 0, here is no long run relaionship among he price series. In he case of 0 < rank ( ) = r<p, where r is he rank of he marix and p is he number of variables in he sysem, here exis one or more coinegraing relaionship among he variables. Johansen s procedure is o deermine he rank of he marix by esing wheher he eigenvalues of, he esimaed of, are significanly differen from zero. If he marix is full rank, hen any linear combinaion of Y is saionary. If he rank ( ) = 0, he marix is null marix hen equaion 9 collapse o he radiional VAR model wih firs differences. To es he null hypohesis ha are a mos r coinegraing vecors in a se of p variables, firs regress Y on Y -, Y -2,, Y -k+ and oupu he residuals, D. For each and D has an n elemen. Second, regress Y -k on Y -, Y -2,, Y -k+ and oupu he residuals, L. For each and L has n elemens. Then compue squares of he canonical correlaion beween he wo residual, denoing hem as Q 2 i (Q 2 > Q 2 2>.> Q 2 i). The likelihood-raio es of he null hypohesis is obained by he race es defined as: p 2 Trace Tess = T ln( ) (0) i r Q i Where ime period is denoed by T. The null hypohesis of race es is o find eiher coinegraing vecor is equal o r or no. The null of r = 0 is es agains he alernae one of r, r p. This leads us o apply he maximal eigenvalue es. I is defined as: 2 Maximal Eigenvalue Tess = T ln( Q ) () r We compare criical values developed Oserwald-Lenum (992) wih he race and maximum eigenvalues. Empirical resuls Augmened Dickey-Fuller (ADF) and Phillips-Perron (PP) uni roo ess are employed o es for he saionariy of he macroeconomic series a level and hen firs difference of each series. The resuls of he ADF and PP ess a level are repored in Table-, by aking ino consideraion of rend variable and wihou rend variable in he regression. Based on Table-, he -saisics for all series from boh ADF and PP ess are saisically insignifican o rejec he null hypohesis of nonsaionary a 0.05 significance level. This indicaes ha hese series are non-saionary a heir level form. Therefore, hese variables are conaining a uni roo process or hey share a common sochasic movemen. This is consisen wih some previous sudies ha have been demonsraed he mos of he macroeconomics and financial series expeced o conain uni roo and hus are inegraed of order one, I(). When he ADF es is conduced a firs difference of each variable, he null hypohesis of non-saionary is easily rejeced a 0.05 significance level as shown in Table- 768

7 Asian Journal of Empirical Research, 3(6)203: A similar conclusion also comes from PP es. Therefore, we can conclude ha he series are inegraed of order one, and a higher order of differencing is no required o execue. Table : Uni roo analysis Counry Variable Aug.Dickey-Fuller Phillips-Perron Lags E Hong Kong S E * * S * * Japan Malaysia Singapore Thailand E S E S E S E S E S E S E S E S * * * * * * * -5.00* * * * * * * * * Noe: The null hypohesis is ha he series is non-saionary, or conains a uni roo. The rejecion of null hypohesis for boh ADF and PP ess are based on he MacKinnon criical values. * indicaes he rejecion of he null hypohesis of non-saionary a 5% significance level. The number of lag is se equal o one in order o avoid he problem of auocorrelaion ha is o ensure he error erms are uncorrelaed and enhance he robusness of he resuls. Since he variables are inegraed of order one, and hen we can proceed o conduc he mulivariae coinegraion es. Table 2: Mulivariae coinegraion analysis Counry Null Criical values (5%) hypohesis Trace Max- Trace Max - Hong Kong p= ** 7.655** p = Japan p= p = Malaysia p= p = Singapore p= p = Thailand p =

8 Asian Journal of Empirical Research, 3(6)203: p = Noe: The criical values are obained from Johansen and Juselius (990). ** indicaes significan a % level.* indicaes significan a 5% level. Table 3: Granger causaliy analysis Counry Null Hypohesis F-value P-value Hong Kong S E E S 8.654*** Japan S E E S Malaysia S E E S Singapore S E E S 5.43** Thailand S E E S 3.503** 0.03 The symbol.> implies does no Granger-cause. *, **, and *** denoe rejecions of he null hypohesis a 0%,5%, and % significance levels, respecively. We employ Johansen s (99) maximum likelihood mehod o examine wheher or no he exchange rae and sock price series for each counry are coinegraed. Table-2 repors he Johansen coinegraion es saisics. As can be seen in he able, here is only one coinegraion vecor beween he exchange rae and sock price series for Hong Kong, suggesing he Hong Kong dollar is coinegraed wih he Hang Seng sock marke index. For he oher four counries, no coinegraing vecor is found. Therefore, he resuls sugges a long-run equilibrium beween exchange raes and sock prices in Hong Kong. Consequenly, an error correcion erm should be included in he Granger causaliy es equaions for Hong Kong. The es resuls of Granger-causaliy beween exchange raes and sock prices are given in Table 3. For he sample period, he repored F-values sugges ha exchange raes significanly lead sock prices for five counries, including Hong Kong, Malaysia, Singapore, and Thailand. I is noeworhy ha hese five counries are also he ones ha have a larger size of inernaional rade. Thus, our findings seem o suppor he predicion ha sock prices for counries wih a high rade size end o be affeced more by exchange rae flucuaions. Excep Hong Kong, hese five counries also show a feedback effec from sock prices o exchange raes, alhough some of he F- saisics are significan only a a marginal level. No evidence of causal relaion beween he foreign exchange and equiy markes for Japan is found, which is consisen wih previous sudies ha focus on indusrialized counries. 770

9 Asian Journal of Empirical Research, 3(6)203: CONCLUSIONS AND DEVELOPMENT In his paper, we examine dynamic linkages beween foreign exchange and sock markes in Hong Kong, Japan, Malaysia, Singapore, and Thailand. These five economies are significanly differen in erms of he size of each economy, degree of developmen, rae of growh, and mauriy of financial markes. Excep for Japan, which is a developed counry, he oher four economies are usually referred as semi- or newly indusrialized counries and all adop expor (or rade) promoion sraegies for economic developmen (see Frankel, Romer, and Cyrus (996)). Regarding he mauriy of financial markes, Hong Kong, Japan, and Singapore are considered developed markes, while Malaysia and Thailand are considered o be emerging markes. There are also differences in erms of capial marke liberalizaion and capial conrol. Hong Kong, Japan, and Singapore have no or lile resricions on foreign invesmens in heir equiy markes, bu Malaysia and Thailand sill do no have a compleely open equiy marke o foreigners. Firs, exising sudies rely primarily on evidence from analyzing indusrialized counries, wih less aenion paid o non-indusrial economies. The four newly indusrialized economies included in our sudy all purse an expor-led or rade-led approach o simulae economic growh, excep for Japan--an indusrialized counry, all have a very high share of rade in heir GDP. According o he goods marke heory, firms in hese counries may face higher exchange rae exposures han hose of indusrialized counries because of he size of rade and hence heir firm values would be affeced more by exchange rae flucuaions. Therefore, he robusness of previous sudies findings ha mainly relies on indusrialized Counries can be examined agains evidence from newly indusrialized counries ha expeced o have higher exchange rae exposures. Second, unlike developed counries, mos developing counries end no o adop a freely Floaing exchange rae sysem and have more capial conrols. None of he economies included in our sudy, wih he excepion of Japan, follow a freely floaing exchange rae arrangemen. I seems reasonable o expec ha for a counry ha does no employ a freely floaing exchange sysem, exchange raes migh no fully response o sock price movemens. Similarly, capial conrol migh reduce dynamic linkages beween foreign exchange and equiy markes. Thus, examining hese Eas Asian economies enables us o check he impac of he degree of financial marke liberalizaion and exchange rae arrangemen on he linkages beween foreign exchange and equiy markes. Third, a crisis in an economy, such as he US financial crisis of 2009, may aler he Naure of sock price-exchange rae relaions. Beginning in he early July of 2007 and over he subsequen one or wo years, several Eas Asian counries have been hrough a major depreciaion in heir currencies as well as a sock marke avalanche. The umbling down sock price and he plunging currency value during he crisis reinforce he convenional impression ha sock prices and exchange raes end o move in a andem, hough i is no clear wheher a causal relaion exiss from exchange raes o sock prices or he oher way around. Excep for Japan, our empirical resuls show a srong 77

10 Asian Journal of Empirical Research, 3(6)203: causal relaion from Exchange raes o sock prices, despie ha several of hese counries such as Hong Kong follow a pegged exchange rae sysem. On he oher hands, here also exiss a causal relaion from he equiy marke o he foreign exchange marke for Malaysia, Singapore, and Thailand. However, for economies ha exhibi a bi-direcional causal relaion, i appears ha he impac from he foreign exchange marke o he equiy marke is much sronger han ha from he equiy marke o he foreign exchange marke. We also find a one-way causal relaion from exchange raes o sock prices for hese economies afer he US financial crisis of Our resuls remain similar when differen economeric mehods are used. REFERENCES Akaike, H. (977). On Enropy Maximizaion Principle. In P. R. Krishniah. (Ed.). Applicaion of Saisics. Norh-Holland, Amserdam. Abdalla, I. S. A. and V. Murinde, (997). Exchange Rae and Sock Price Ineracions in Emerging Financial Markes: Evidence on India, Korea, Pakisan, and Philippines. Applied Financial Economics, Vol. 7, pp Aggarwal, R. (98). Exchange Raes and Sock Prices: A Sudy of U.S. Capial Marke under Floaing Exchange Raes. Akron Business and Economic Review pp Ajayi, R. A. and M. Mougoue, (996). On he Dynamic Relaion beween Sock Prices and Exchange Raes. Journal of Financial Research, Vol. 9, pp Allayannis, G. and J. P. Weson, (200). The Use of Foreign Currency Derivaives and Firm Marke Value. Review of Financial Sudies, Vol. 4, pp Baek, E. and W. Brock, (992). A General Tes for Nonlinear Granger Causaliy: Bivariae Model. Unpublished working paper, Iowa Sae Universiy and Universiy of Wisconsin, Madison. Bahmani-Oskooee, Mohsen, and Ahmad Sohrabian. (992). Sock Prices and he Effecive Exchange Rae of he Dollar. Applied Economics, Vol. 24, pp Bahmani-Oskooee, Mohsen, and Ilker Domac. (997). Turkish Sock Prices and he Value of Turkish Lira. Canadian Journal of Developmen Sudies, Vol. 8, pp Bhandari, Jagdeep S. and Hans Genberg. (989). Exchange Rae Movemens and Inernaional Inerdependence of Sock Markes. Inernaional Moneary Fund Working Paper WP/89/44 (May). Choi, Jongmoo Jay. (995). The Japanese and U.S. Sock Prices: A Comparaive Fundamenal Analysis. Japan and he World Economy, Vol. 7, pp Barov, E. and G. M. Bodnar, (994). Firm Valuaion, Earnings Expecaions, and he Exchange- Rae Exposure Effec. Journal of Finance, Vol. 49, pp Bekaer, G. and C. R. Harvey, (2000). Foreign Speculaors and he Emerging Equiy Markes. Journal of Finance, Vol. 55, pp

11 Asian Journal of Empirical Research, 3(6)203: Bodnar, G. M. and W. M. Genry, (993). Exchange Rae Exposure and Indusry Characerisics: Evidence from Canada, Japan, and he USA. Journal of Inernaional Money and Finance Vol. 2, pp Brock, W. (99). Causaliy, Chaos, Explanaion and Predicion in Economics and Finance. In J. Casi and a Karlquis (Eds), Beyond Belief: Randomness, Predicion and Explanaion in Science, CRC Press, Boca Raon. Brock, W., J. Lakonishok, and B. LeBaron, (992). Simple Technical Trading Rules and he Sochasic Properies of Sock Reurns. Journal of Finance, Vol. 47, pp Campbell, J. Y. and P. Perron, (99). Pifalls and Opporuniies: Wha Macroeconomiss Should Know Abou Uni Roos. In Oliver J. Blanchard and Sanley Fischer, ed., NBER Macroeconomics Annual 99, USA: MIT Press, pp Chamberlain, S., J. S. Howe, and H. Popper, (997). The Exchange Rae Exposure of U. S. and Japanese Banking Insiuions. Journal of Banking and Finance, Vol. 2, pp Dornbusch, R. and S. Fischer, (980). Exchange Raes and Curren Accoun. American Economic Review, Vol. 70, pp Engle, R. F. and C. W. J. Granger, (987). Co-inegraion and Error Correcion: Represenaion, Esimaion, and Tesing. Economerica Vol. 55, pp Eun, C. and S. Shim, (989). Inernaional Transmission of Sock Marke Movemens. Journal of Financial and Quaniaive Analysis, Vol. 24, pp Frankel, J. A., D. Romer, and R. Cyrus, (996). Trade and Growh in Eas Asian counries: Cause and Effec? Naional Bureau of Economic Research Working Paper no Franck, P. and A. Young (972). Sock Price Reacion of Mulinaional Firms o Exchange Realignmens. Financial Managemen, Vol., pp Fuller, W. A. (976). Inroducion o Saisical Time Series. New York: John Wiley & Sons Gavin, M. (989). The Sock Marke and Exchange Rae Dynamics. Journal of Inernaional Money and Finance, Vol. 8, pp Granger, C. W. J. (969). Invesigaing Causal Relaions by Economerics Models and Cross Specral Mehods. Economerica, Vol. 37, pp Granger, C. W. J. (988). Some Recen Developmens in A Concep of Causaliy. Journal of Economerics, Vol. 39, pp Gregory, A. W. and B. E. Hansen, (996). Residual-based Tess for Coinegraion in Models wih Regime Shifs. Journal of Economerics, Vol. 70, pp Griffin, J. M. and R. M. Sulz, (200). Inernaional Compeiion and Exchange Rae Shocks: A Cross-Counry Indusry analysis of Sock Reurns. Review of Financial Sudies, Vol. 4, pp Gujarai, D. N. (988). Basic Economerics, Singapore: McGraw-Hill Book Co MacKinnon, J. (99). Criical Values for Coinergaion Tes. In Engle,R F and Granger, C W J(eds.). Long Run Economic Relaionships: Reading in Coinegraion, New York: Oxford Universiy Press. 773

12 Asian Journal of Empirical Research, 3(6)203: Johansen, S. and Juselius, K. (990). Maximum Likelihood Esimaion and Inference on Coinegraion-Wih Applicaion o he Demand for Money, Oxford Bullein of Economics and Saisics, Vol. 52, pp Johansen, S. (988). Saisical Analysis of Coinegraion Vecors. Journal of Economic Dynamics and Conrol, Vol. 2, pp MacKinnon, J. (99). Criical Values for Coinergaion Tes. In Engle,R F and Granger, C W J (eds.). Long Run Economic Relaionships: Reading in Coinegraion, New York: Oxford Universiy Press. Dickey, D. A and W. A Fuller (98). Likelihood Raio Saisic for AuoregressiveTime Series wih a Uni Roo. Economerica. Vol. 49, pp Dickey, D. A. & Fuller, W. A. (979). Disribuion of he Esimaors for Auoregressive Time Series wih a Uni Roo. Journal of he American Saisical Associaion: Vol. 74, pp Campbell, J. Y. and P. Perron (99). Pifalls and Opporuniies: Wha Macroeconomiss Should Know Abou Uni Roos? In Oliver J. Blanchard and Sanley Fischer, ed., NBER Macroeconomics Annual 99, USA: MIT Press, pp Bahmani-Oskooee, Mohsen, and Ahmad Sohrabian. (992). Sock Prices and he Effecive Exchange Rae of he Dollar. Applied Economics, Vol. 24, pp Barov, E. and G. M. Bodnar, (994). Firm Valuaion, Earnings Expecaions, and he Exchange- Rae Exposure Effec. Journal of Finance, Vol. 49, pp Granger, C. W. J. (988). Some Recen Developmens in A Concep of Causaliy. Journal of Economerics Vol. 39, pp Phillips, P. C. B. & Perron, P. (988). Tesing for a Uni Roo in Time Series Regression. Biomerica, Vol. 75, pp Oserwald-Lenum, M. (992). A Noe wih Quaniles of he Asympoic Disribuion of he Maximum Likelihood of Coinegraion Rank Saisic. Oxford Bullein of Economics and Saisics, Vol. 54, pp Said, S. E. & Dickey, D. A. (984). Tesing for Uni Roo in Auogressive-Moving Average of Unknown Order. Biomerika, Vol. 7, pp Schwer, C. W. (987). Effecs of Model Specificaion on Tess for Uni Roos in Macroeconomic Daa. Journal of Moneary Economics, Vol. 20. pp Sim, C. A. (980) Macroeconomics and Realiy. Economerica, Vol. 48, pp Solnik, B. (987) Using Financial Prices o Tes Exchange Rae Models: A Noe. Journal of Finance, Vol. 42, pp Smih, C. (992). Sock Marke and he Exchange Rae: A Muli-Counry Approach. Journal of Macroeconomics, Vol. 4, pp Soenen, L. A. and E. S. Hennigar (988). An Analysis of Exchange Raes and Sock Prices: he U. S. Experience beween 980 and 986. Akron Business and Economic Review, pp