Pre and post crisis analysis of stock price and exchange rate: Evidence from Malaysia

Transcription

1 MPRA Munich Personal RePEc Archive Pre and pos crisis analysis of sock price and exchange rae: Evidence from Malaysia A.H. Baharom and M.S. Habibullah and Royfaizal R.C. Universii Pura Malaysia 1. June 2008 Online a hp://mpra.ub.uni-muenchen.de/12445/ MPRA Paper No , posed 1. January :24 UTC

2 PRE AND POST CRISIS ANALYSIS OF STOCK PRICE AND EXCHANGE RATE: EVIDENCE FROM MALAYSIA by A.H. Baharom, R.C. Royfaizal and M.S. Habibullah 1 ABSTRACT The furore and chaos creaed by he Asian financial crisis have ignied many sudies on numerous subjecs, and i is believed ha he crisis has changed he way naions being adminisered and policies formed and implemened especially hose regarding moneary and fiscal policies. Johansen (1991) coinegraion mehod was used and he period was divided ino wo sub periods, albei pre crisis and pos crisis. The resuls obained are similar wih a number of pas lieraures poining o no long run relaionship beween sock price and exchange rae for boh periods. Keywords: Sock price, exchange rae, Asian financial crisis, Coinegraion. JEL code: G14, F INTRODUCTION The Asian Financial Crisis in sruck he economy of many Asian counries and hese Asian counries suffered coninuous currency depreciaion and heir sock markes were badly hi. Thai Bah was he firs o be hi, and soon i spread o is neighbors. In Malaysia, he radiionally sable ringgi collapsed and i was followed by he collapse of he KLSE whereby KL composie index which decreased by 44.8% in he second half of 1997 (Azman-Saini e al., 2006). The issue of wheher he sock prices and exchange raes are dynamically relaed has excied he ineres of many researchers. Bahmani-Oskooee and Sohrabian (1992), and Nieh and Lee (2001) found ha here are no long-run relaionship beween sock prices and exchange raes. While he former sudied abou he scenario in he Unied Saes, he laer s sudy was on he dynamic relaionship for G-7 counries, namely Canada, France, Germany, Ialy, Japan, Unied Saes, and he Unied Kingdom. A sudy on Malaysia, Azman-Saini e al. (2006) found 1 All auhors are affiliaed wih he Deparmen of Economics, Faculy of Economics and Managemen, Universii Pura Malaysia, Serdang, UPM Selangor. The corresponding auhor is Baharom Abdul Hamid, Lecurer. Tel: (603) ; baharom@econ.upm.edu.my

3 ha during he crisis period, exchange raes led sock price. In heir sudy hey used Granger non-causaliy es proposed by Toda and Yamamoo (1995). However, his paper explore he long run relaionship beween sock price index and exchange rae (pre and pos crisis) using he mulivariae Johansen-Juselius mehod. The oher disinc difference is, while Azman Saini e al. use bilaeral exchange rae, our sudy employ he real effecive exchange rae (REER) due o some inrinsic reasons. Malaysia is one of he mos open and fas growing economies in Souh Eas Asia, hence sock marke and foreign exchange marke play an inegral role in conribuing o he financial developmen and economic growh in Malaysia. The volailiy of exchange raes and sock prices would be a hindrance o a healhy and warraned growh. Thus, he undersanding of relaionship beween sock prices and exchange raes is of umos imporance and very vial o capure he co-movemen or rends of hese wo imporan macroeconomic variables. There are wo schools of houghs on his subjec, whereby he radiional approach explained ha he exchange rae leads he sock price, and he changes of exchange raes would affec he value of a firm via compeiiveness, asses and liabiliies. Similarly, he changes of exchange raes would evenually affec he firm s profis or loss and hus, he value of is equiy. In conras, porfolio approach claims ha acually sock prices lead exchange raes. When sock prices decreases, i reflecs a reducion in invesors wealh, and his ulimaely leads lower of demand for money, which in urn causes he depreciaion of he currency. Under porfolio approach, he main assumpion is ha sock price is inversely relaed wih he exchange raes. This paper is organized as follow, whereby in he nex secion, we discuss he mehod and sources of daa used in analysis. The hird secion will be on he resuls of he sudy while he concluding las secion conains our conclusion. 2.0 DATA AND METHODOLOGY The sample period is segregaed ino wo periods, which is January 1988 o June 1997 for he pre crisis period and July 1998 o December 2006 for he pos crisis period. The REER was chosen in his sudy because he Malaysian Ringgi was pegged o he US dollar righ afer he crisis and his migh conribue o a misconsrued finding. The efficiency of he sock marke is also in quesion, if here exis any long run relaionship, and hen he sock marke can be assumed inefficien due o inerdependency. In empirical economics macroeconomic variables comprises of non saionary series. Treaing non saionary variables in empirical analysis is imporan so ha he resuls of spurious regression can be avoided. According o he concep of coinegraion, wo or more non-saionary ime series share a common rend, hen hey are said o be coinegraed. The heoreical framework highlighed are expressed as follows: he componen of he vecor Y = (y 1,y 2,,y n ) are considered o be coinegraed of order d,b, denoed Y ~ CI (d,b) if (i) all he componen Y are saionary afer n difference, or

4 inegraed of order d and noed as Y ~ I(d). (ii) presence of a vecor β = (β1,β 2,,β n ) in such ha linear combinaion βy = β 1 y 1 + β 2 y β n y n whereby he vecor β is named he coinegraing vecor. A few major characerisics of his model are ha he coinegraion relaionship obained indicaes a linear combinaion of non-saionary variables, in which all variables mus be inegraed of he same order and lasly if here are n series of variables, here may be as many as n-1 linearly independen coinegraing vecors. Johansen s (1991) coinegraion es is adoped o deermine wheher he linear combinaion of he series possesses a long-run equilibrium relaionship. The numbers of significan coinegraing vecors in non-saionary ime series are esed by using he maximum likelihood based λ race and λ max saisics inroduced by Johansen and Juselius (1990). The advanage of his es is ha i uilises es saisic ha can be used o evaluae coinegraion relaionship among a group of wo or more variables. Therefore, i is a superior es as i can deal wih wo or more variables ha may be more han one coinegraing vecor in he sysem. Prior o esing for he number of significan coingraing vecors, he likelihood raio (LR) ess are performed o deermine he lag lengh of he vecor auoregressive sysem. In he Johansen procedure, following a vecor auoregressive (VAR) model, i involves he idenificaion of rank of he n x n marix in he specificaion given by: Y k 1 i 1 i Y i Y k (1) where Y is a column vecor of he n variables, is he difference operaor, Γ and are he coefficien marices, k denoes he lag lengh and δ is a consan. In he absence of coinegraing vecor, is a singular marix, which means ha he coinegraing vecor rank is equal o zero. On he oher hand, in a coinegraed scenario, he rank of could be anywhere beween zero. In oher words, he Johansen coinegraion es can deermine he number of coinegraing equaion and his number is named he coinegraing rank. The Johansen Maximum likelihood es provides a es for he rank of, namely he race es (λ race ) and he maximum eigenvalue es (λ max ). Firsly, he λ race saisic es wheher he number of coinegraing vecor is zero or one. Then, he λ max saisic es wheher a single coinegraion equaion is sufficien. Boh es saisics are given as follows: race race ( r) T p i r 1 ln(1 ˆ) ( r, r 1) T ln(1 ˆ r 1) (2) (3) where p is he number of separae series o be analysed, T is he number of usable observaions and λ is he esimaed eigenvalues.

5 3.0 EMPIRICAL RESULTS There have been numerous sudies examining he relaionship beween sock prices and exchange raes. And he resuls are mixed, such as sock prices affecing exchange raes as found by Abdalla and Murinde (1997), vice versa feedback effec as found by Ajayi and Mougoue (1996) or no relaionship found by Ibrahim (2000). However, he empirical resuls of long run relaionship beween sock prices and exchange raes are a bes mixed. This is because of differen daa se and mehodology used. In his sudy, he daa se consiss of monhly real effecive exchange raes (REER) and sock prices index (SP) for Malaysia covering he period from January 1988 o December In he analysis, he period of sudy is divided ino wo sample periods: firs, he pre-crisis period spanning from January 1988 o June 1997; and second, he pos-crisis period from July 1998 o December Before proceeding o he coinegraion es, he order of inegraion of he series was deermined using he Augmened Dickey-Fuller (ADF) uni roo es. Table 1 repors he resul of he uni roo es. The resuls clearly shows ha he null hypohesis of a uni roo canno be rejeced a he 5% level for REER and SP in heir levels. However, he null hypohesis is rejeced a 5% level when REER and SP one in heir firs-differences. Thus, hese indicaed ha REER and SP are inegraed of order one, I(1), in oher words hey are difference saionary process. Since he series are of he same order, we proceed o es he exisence of coinegraing relaions among he REER and SP using he Johansen coinegraion es. The resuls are repored in Table 2. The resuls indicae ha he null hypohesis of no coinegraion canno be rejeced in he pre- and pos-crisis periods. Therefore, we concluded ha here is no evidence of long run relaionship beween he real effecive exchange rae (REER) and he sock price index (SP) for boh periods, more specifically, pre and pos crisis periods. In addiion he resuls show ha he relaionship beween he sock price and he real effecive exchange rae did no change before and afer he Asian Financial crisis. 4.0 CONCLUSION AND IMPLICATION The resuls clearly show ha boh he variables of ineres, he REER and SP are no coinegraed for boh period, pre and pos crisis. Our findings suppor he resuls obained by Ibrahim (2000) bu in conras wih Azman Saini e al. (2006). This migh be due o differen ses of variables chosen and also differen mehod used in he analysis. The resul also clearly show ha Malaysia s sock marke can be considered efficien since i is found o independen of exchange rae. The findings of no long-run equilibrium beween sock price index and exchange rae imply ha Malaysian sock marke is informaional efficien. Under he efficien marke hypohesis, marke paricipans canno use he growh of he exchange raes as a rading rule in order o exploi abnormal profi. On he oher hand, if he variables are coinegraed or have a long-run equilibrium relaionship, hen he movemens in exchange rae can be exploiable as a rading rule o reap abnormal profi.

6 REFERENCES Ajayi, A.R., and Mougoue, M. (1996). On he dynamic relaion beween sock prices and exchange raes. The Journal of Financial Research, 19, Abdalla, I.S.A., and Murinde, V. (1997). Exchange rae and sock price ineracions in emerging financial markes: evidence on India, Korea, Pakisan and Philippines. Applied Financial Economics, 7, Azman-Saini, W.N.W., Habibullah, M.S., Law, S.H. and Dayang-Afizzah, A.M. (2006) Sock Prices, exchange raes and causaliy in Malaysia: a noe. The ICFAI Journal of Financial Economics, 5, Bahmani-Oskooee, M., and Sohrabian, A. (1992). Sock prices and he effecive exchange rae of he dollar. Applied Economics, 24, Ibrahim, M.H. (2000). Coinegraion and Granger causaliy ess of sock price and exchange rae ineracions in Malaysia. ASEAN Economic Bullein, 17, Johansen, S. (1991). Esimaion and hypohesis esing of coinegraion vecors in Gaussian vecor auoregressive models. Economerica 59, Johansen, S. and Juselius, K. (1990). The full informaion maximum likelihood procedure for inference on coinegraion wih applicaions o he demand for money. Oxford Bullein of Economics and Saisics, 52, Nieh, C.C. and Lee, C.F. (2001). Dynamic relaionship beween sock prices and exchange raes for G-7 counries. The Quarerly Review of Economics and Finance, 41, Oserwald-Lenum, M. (1992). A noe wih quaniles of he asympoic disribuion of he maximum likelihood coinegraion rank es saisics. Oxford Bullein of Economics and Saisics, 54, Schwer, G.W. (1987). Effecs of model specificaion on ess for uni roos in macroeconomic daa. Journal of Moneary Economics, 20, Toda, H.Y., and Yamamoo, T. (1995). Saisical inference in vecor auoregressions wih possibly inegraed processes. Journal of Economerics, 66,

7 Table 1: Resuls of he Uni Roo Tess ADF Tes Level Firs Difference Variables Consan, Consan, Consan, Consan, No Trend Trend No Trend Trend Before Crisis REER ** ** (2) (2) (0) (0) SP ** ** (1) (2) (0) (0) Afer Crisis REER ** ** (1) -1 (0) (0) SP ** ** (1) (1) (0) (0) Noe: ** denoes significan a 5% significance levels, respecively. Figures in parenhesis ( ) refer o he seleced lag lengh. The lag lengh was arbirarily seleced using SIC. Table 2: Resuls of Coinegraion Tes Vecors r = 0 r 1 Pre-crisis resuls (Jan, June, 1997), Lags = 4 Trace es % level Max-Eigen es % level Pos-crisis resuls (July, Dec, 2006), Lags = 4 Trace es % level Max-Eigen Tes % level Noe: r indicaes he number of coinegraing vecors. Trace and Max-Eigen denoe he race saisic and maximum eigenvalue saisic. The criical values are obained from Oserwald-Lenum (1992). Lag selecion (k) is based on Schwer (1987) formula.