An Empirical Relationship between Exchange Rates, Interest Rates and Stock Returns

Transcription

1 An Empirical Relaionship beween Exchange Raes, Ineres Raes and Sock Reurns Auhor Paramai, Sudharshan Reddy, Gupa, Rakesh Published 2013 Journal Tile European Journal of Economics, Finance and Adminisraive Sciences Downloaded from hp://hdl.handle.ne/10072/58658 Link o published version hp:// EJEFAS_56.hml Griffih Research Online hps://research-reposiory.griffih.edu.au

2 European Journal of Economics, Finance and Adminisraive Sciences ISSN Issue 56 January, 2013 EuroJournals, Inc hp:// An Empirical Relaionship beween Exchange Raes, Ineres Raes and Sock Reurns Sudharshan Reddy Paramai PhD Candidae, Deparmen of Accouning, Finance and Economics Griffih Business School, Griffih Universiy, Ausralia Rakesh Gupa Senior Lecurer, Deparmen of Accouning, Finance and Economics Griffih Business School, Griffih Universiy, Ausralia Absrac In his paper sudy aims o invesigae he relaionship beween call money raes, exchange raes and sock reurns from he perspecive of India. We use monhly daa for he ime span of April 1992 o March This provides sufficien daa se for he empirical analysis. Resul from Granger causaliy es evidences bidirecional relaionship beween call money raes and exchange raes. I is also idenified ha call money raes and exchange raes Granger cause sock reurns and did no find reverse causaliy from sock reurns o call money and exchange raes. To explore, lead-lag ineracion among he variables sudied we employed VAR models. Resuls sugges ha here is subsanial lead-lag relaionship from call money raes o exchange raes and sock reurns. Similar relaionship also found from exchange raes o call money raes and sock reurns. However, here is no evidence of lead-lag causaion from sock reurns o call money and exchange raes. Findings of his sudy are useful for he invesors and policy makers. In invesors sandpoin, hey can uilize his hisorical informaion of call money raes and exchange raes for predicing he movemens of sock reurns. Similarly, policy makers can sabilize he sock marke flucuaions by adoping appropriae policies owards ineres raes and exchange raes for ime o ime. Keywords: Direcion of causaliy and Lead-lag ineracion 1. Inroducion In he early 1990s, he financial secor reforms have been iniiaed in India, which ensured he relaxaion of foreign invesors resricions and opened up domesic financial markes for he inernaional pracices. Subsequenly, he Indian sock markes (Bombay Sock Exchange and Naional Sock Exchange) have winessed as he mos acive and emerging sock markes of he world. Paricularly, hese sock markes have araced he invesors across he globe by expanding heir horizons, which resuled in erms of lised companies, shareholders, volume of rade and marke capializaion. This New Economic Policy (1991) has helped he Indian sock markes o grow coninuously and expand. Gupa and Basu (2007) provide an ineresing saisics on he developmen of Bombay Sock Exchange (BSE) and Naional Sock Exchange (NSE) over he period. The economic heory suggess ha ineres raes, inflaion, money supply, price level and oher macro elemens are imporan variables in undersanding he deeds of sock prices and also for Elecronic copy available a: hp://ssrn.com/absrac=

3 169 European Journal of Economics Finance and Adminisraive Sciences - Issue 56 (2013) predicing he movemens and rends in exchange raes. Basically, here are wo heoreical argumens in his regard. The firs approach argues ha currency depreciaion will resul in higher expors and his evenually raises he sock prices and hen leads o higher corporae profis in he shor-run. Oher argumen explains he relaionship beween exchange raes and sock prices in a porfolio adjusmen pah. This heory argues ha whenever here is a change in sock prices will resul in porfolio adjusmen. If he sock prices are in he upward direcion hen here is high possibiliy ha more foreign capial can inflow. In oher words, if he sock prices are in downward direcion will resul in reducion of corporae as well as counry wealh. This will lead o reducion in demand for money in he economy; in hese circumsances cenral bank auhoriies will ake decisions o reduce he ineres raes o relieve his siuaion. This lower ineres rae can encourage capial ouflow o ake over he advanages of higher ineres raes in oher counries. Finally, his heory suggess ha a lower sock price may lead o currency depreciaion (Kuy, 2010). There has been considerable aenion on wheher he exchange raes and sock prices have any empirical relaionship. Paricularly, his issue has become more aracive among he economiss, invesors and policy makers afer he Asian Financial Crises (1997). Empirically, i is argued ha if he exchange raes and sock prices are iner-relaed hen i is quie possible o preven such crises by looking a he direcion of causaliy e.g. if he causaliy runs from exchange raes o sock prices hen he crises can be avered by implemening he appropriae policies owards conrolling he exchange rae flucuaions. In oher words, if he causaliy runs from sock prices o exchange raes hen he policy makers can keep an eye on sabilizing he sock markes by enforcing he desirable economic policies. This is also an imporan in he invesors poin of view o know he direcion of causaliy beween exchange raes and sock prices. If he causaliy is idenified beween hese variables hen invesors can use his hisorical informaion of one marke for predicing he behavior of oher marke. Similarly, he relaionship beween ineres raes and sock prices has been exensively sudied by invesors, policy makers and researchers. I evidences from he lieraure ha ineres rae is one of he mos effecive and significan facor in deermining he sock prices. Theoreically, i is conended ha here is an inverse relaionship beween ineres rae and sock prices and boh end o move in an opposie direcions. Since, las hree decades here are several sudies have been examined he relaionship beween exchange raes, ineres raes and sock prices. Largely, hese sudies have repored mixed resuls in naure and some oher sudies have drawn inconclusive resuls. The relaionship beween exchange raes and sock prices were examined by Aggarwal (1981), Giovannini and Jorion (1987), Solnik (1987) and Smih (1992). The evidences of hese sudies indicae ha here is significan posiive relaionship beween he variables sudied. Similarly, he sudies of Soenen and Hennigar (1988), Muhammad and Rasheed (2002), Bhaacharya and Mukherjee (2003), Rahman and Uddin (2009) also invesigaed he connecion beween exchange raes and sock prices and found negaive relaionship among he variables. Some oher sudies have also aemped o invesigae he causal relaionship beween he above variables and esablished bidirecional relaionship (Bahmani-Oskooee and Sohrabian, 1992; and Kumar, 2010) and unidirecional relaionship (Abdalla and Murinde, 1997; Mishra, 2004; and Alagidede e al., 2011). Finally, Ong and Izan (1999); Nieh and Lee (2001) sudies revealed ha here is no empirical associaion beween exchange raes and sock prices. Paramai and Gupa (2011) also invesigaed he relaionship beween sock prices and economic growh. Their empirical resuls sugges ha economic growh has significan influence on sock marke developmen. In he same way, here are oher sudies have invesigaed he relaionship beween ineres rae and sock prices. The empirical resuls of Harasy and Roule (2000), Wong e al. (2005) sudies provided evidences of long-run relaionship beween ineres rae and sock prices. Similarly, Campbell (1987), Shanken (1990), Arango e al. (2002), Hsing (2004), Rigobon and Sack (2004), Uddin and Alam (2007), Alam and Uddin (2009) sudies have esablished significan negaive relaionship beween ineres rae and sock prices. While some oher sudies have found weak causaliy (Farsio and Fazel, 2010) and inconclusive resuls (Lee, 1997) among he same variables. Overall, i is undersood from he exising sudies ha he empirical resuls are mosly assored and indeerminae. Given hese Elecronic copy available a: hp://ssrn.com/absrac=

4 170 European Journal of Economics Finance and Adminisraive Sciences - Issue 56 (2013) empirical and heoreical argumens, he presen sudy will examine he empirical relaionship beween call money raes, exchange raes and sock reurns in Indian conex by using monhly daa from April 1992 o March The remainder of his sudy is organized as follows: Secion 2 provides empirical lieraure review on he relaionship beween exchange raes, ineres raes and sock prices. Secion 3 presens empirical mehodology for analysis. Secion 4 displays naure and sources of daa, variables and empirical resuls of he sudy. Finally, secion 5 provides summary and conclusion. 2. Lieraure Review This sudy has carried ou a broad lieraure survey on boh developed and emerging economies; o review, a broader heoreical and empirical lieraure on he connecion beween exchange raes, ineres raes and sock prices. This undersanding is very imporan for conducing an empirical analysis. The lieraure survey has been divided ino wo pars; par one explains he relaionship beween exchange raes and sock prices and, par wo presens he connecion beween ineres rae and sock prices The Exchange Raes and Sock Prices A sudy by Aggarwal (1981) made an aemp o explore he relaionship beween changes in he dollar exchange raes and change in indices of sock prices. His sudy used monhly daa from 1974 o 1978 on sock prices and effecive exchange raes for he USA. Resuls show a posiive correlaion and his relaionship is sronger in he shor run han in he long run. Giovannini and Jorion (1987) sudy also arrived wih similar resuls of Aggarwal (1981) in case of he USA. In he same way, he relaionship beween sock prices and exchange raes were invesigaed by Bahmani and Sohrabian (1992) in he conex of he USA for he period of 1973 o The Granger causaliy es resuls confirmed ha here is a shor run relaionship beween he exchange raes and sock prices. While, coinegraion es resuls indicae ha here is no evidence of long-run relaionship among he same variables. Likewise, Rahman and Uddin (2008) explored he relaionship beween sock prices and exchange raes in he case of Bangladesh for he period of June 2003 o March Their sudy resuls concluded ha here is no long-run relaionship beween sock prices and exchange raes by employing Johansen coinegraion es. The Granger causaliy es suppored shor-run relaionship beween he variables and ha runs from sock prices o exchange raes of US dollar and Japanese yen bu no he euro and pound serling. The long-run and shor-run relaionship beween exchange raes and sock prices were examined by Kumar (2010). This sudy uses linear and non-linear Granger causaliy ess for invesigaing causal relaionship and also employed coinegraion es for idenifying long-run relaionship among he sudied variables. Resuls of Granger causaliy (linear and non-linear) es sugges ha here is a bi-direcional relaionship beween sock prices and exchange raes. Similarly, he resuls of coinegraion es evidence no long-run relaionship beween he variables. Kuy (2010) also examined he relaionship beween sock prices and exchange raes for he period of January 1989 o December 2006 in he case of Mexico. The empirical resuls of his sudy conclude ha here is a shor run relaionship beween he sudied variables and ha runs from sock prices o exchange raes and found no long run relaionship beween same variables. Anoher sudy by Alagidede e al. (2011) aemped o find ou he underlying connecion beween foreign exchange markes and sock markes in Ausralia, Canada, Japan, Swizerland and he Unied Kingdom for he period of January 1992 o December Resuls of coinegraion ess reveal ha here is no long-run relaionship beween he variables. The Granger causaliy es resuls indicae ha here is unidirecional relaionship from exchange raes o sock prices in he case of Canada, Swizerland and Unied Kingdom. The causal relaionship from sock prices o exchange raes was found only in his case of Swizerland. Furher, his sudy employed Hiemsra-Jones es o find

5 171 European Journal of Economics Finance and Adminisraive Sciences - Issue 56 (2013) ou he non-linear causaliy. Resuls show ha causaliy is found from sock prices o exchange raes in he case of Japan and exchange raes o sock prices in Swizerland. Some oher sudies such as; Ong and Izan (1999) and Nieh and Lee (2001) also examined he relaionship beween exchange raes and sock prices in case of G-7 counries and resuls sugges ha here is no significan relaionship beween he observed variables in he long run. The relaionship beween sock prices and effecive exchange raes for he ime span of 1980 o 1986 was examined by Soenen and Hennigar (1988) for he perspecive of he USA. Resuls evidence a significan negaive relaionship among he variables. Granger e al. (2000); Caporale e al. (2002); Savarek (2005) and Pan e al. (2007) provided elaboraive informaion on he linkages beween sock prices and exchange raes. The changes in sock prices will have influence on he movemens of exchange raes. In oher way, i is explained ha he sock prices can lead he exchange raes wih a negaive correlaion. Precisely, a decrease in sock prices may reduces domesic wealh, which leads o lower he domesic money demand and ineres raes and also his furher guide o lower he demands of foreign invesors for domesic asses and currency. The changes in demand and supply of currencies may lead o depreciaion of domesic currency and encourages capial ou flow. In oher words, when he sock prices move upward, hen his will fascinae foreign invesors o diversify heir invesmen inernaionally and gain he profis. Thus, i will lead o appreciae domesic currency and raises capial inflow. A sudy of Abdalla and Murinde (1997) explored he relaionship beween exchange raes and sock prices in he emerging markes of India, Korea, Pakisan and Philippines for he ime span of January 1985 o July This sudy resuls provide an evidence of unidirecional relaionship and ha runs from exchange raes o sock prices in case of India, Korea and Pakisan. However, his sudy could no find any relaionship beween exchange raes and sock prices in case of he Philippines. Mishra (2004) analyzed wheher foreign exchange markes and sock marke are relaed each oher in case of India by using daa from April 1992 o March The Granger causaliy es resuls indicae ha here is unidirecional relaionship and ha runs from exchange raes o demand for money and ineres rae. Though, his sudy couldn find any causal relaionship beween exchange raes and sock reurns. The resuls of VAR model indicae ha he observed variables are relaed each oher bu here is no consisency in his regard. Furher, his sudy employed forecas error variance decomposiion mehod o confirm he relaionship among he variables and resuls reveal ha each variable is influenced by oher variable (s). In a similar way, Aydemir and Demirhan (2009) invesigaed he relaionship beween sock prices and exchange raes for he period of February 23 rd, 2001 o January 11 h, 2008 by using daily daa in case of Turkey. Their sudy resuls reveal ha here is a bidirecional relaionship and also evidenced ha here is a posiive and negaive causal relaionship beween exchange raes and sock marke indices. On he oher way, negaive causaliy occurs from exchange raes o all sock marke indices. Economic heory suggess ha he changes in foreign exchange can have a significan impac on sock prices by alering cash flow, invesmen and profiabiliy of he firms. However, here is no empirical harmony in his regard and mosly resuls are indecisive (Joseph, 2002; and Vygodina, 2006). Similarly, Rahman and Uddin (2009) examined he dynamic relaionship beween sock prices and exchange raes for he period of January 2003 o June 2008 in hree Souh Asian emerging counries viz. Bangladesh, India and Pakisan. The resuls of Johansen Co-inegraion and Granger Causaliy ess revealed ha here are no long run and shor run relaionship beween sock prices and exchange raes, respecively The Ineres Raes and Sock Prices The relaionship beween ineres rae and sock prices has been empirically invesigaed by Zhou (1996). The regression resuls of his sudy saes ha he ineres rae has an imporan implicaion on sock reurns i.e., paricularly in he longer horizons. Wong e al. (2005) invesigaed he long-run relaionship beween macroeconomic variables and leading sock indices of Singapore and he Unied Saes for he period of January 1982 o December The Granger causaliy es resuls indicae

6 172 European Journal of Economics Finance and Adminisraive Sciences - Issue 56 (2013) ha he performance of sock marke may have an impac on he adjusmen of cenral bank s moneary policy. Resuls of coinegraion es imply ha here is a long-run relaionship i.e., Singapore sock prices wih he ineres rae and money supply. However, his sudy could no find similar resuls in case of he Unied Saes for he same period. Harasy and Roule (2000) provided evidence of longerm relaionship beween sock prices, dividends (earnings) and ineres rae (long-erm) in 16 counries bu could no provide he same resuls for Ialian marke. Bren e al., (1989) argue ha he ineres raes are useful in forecasing he sign and he variance of he excess reurns of socks. Similarly, Arango e al. (2002) aemped o invesigae he dynamic relaionship beween he Bogoa sock marke reurns and shor-erm ineres rae (inerbank loan ineres rae) by using daa from January 1994 o February Sudy finds evidence of nonlinear and inverse relaionship among he share prices and ineres rae. Alam and Uddin (2009) aemped o explore he empirical relaionship beween ineres rae and sock prices of fifeen developed and developing counries and also focused on finding he weak form efficiency of he share markes. This sudy uses daa from January 1988 o March The evidences of his sudy reveal ha ineres rae has negaive relaionship wih he share prices for all he counries. Furher, i is idenified ha he changes of ineres rae have a significan negaive relaionship wih he changes of share prices for six counries. Finally, he uni roo es resuls confirm ha none of hese counries sock markes follow random walk (weak form efficiency). Hsing (2004) sudy found ha here is an inverse relaionship beween ineres rae and sock prices. Rigobon and Sack (2004) empirically invesigaed he impac of moneary policy on asse prices. Their analysis reveals ha he increase in shor-erm ineres raes will negaively affec he sock prices. The linear relaionship among ineres rae and sock prices was invesigaed by Uddin and Alam (2007). Their sudy reveals ha ineres rae has negaive relaionship wih he share prices and furher i evidences ha changes of ineres rae has subsanial negaive associaion wih he changes of share prices. Leon (2008) examined he effecs of ineres raes volailiy on sock marke reurns and volailiy for he period of January 31 s o Ocober 16 h Resuls indicae ha here is a negaive and significan associaion beween condiional marke reurns and ineres raes. Furher, i is evidenced ha here is a posiive connecion among condiional variance of reurns and ineres raes, bu however his relaionship is no significan. Campbell (1987) and Shanken (1990) sudies found ha he nominal Treasury bill yield (one-monh) is negaively associaed wih he fuure sock reurns. Anoher sudy by Lee (1997) invesigaed he connecion beween ineres rae (shor-erm) and sock prices. His sudy aim was o predic he excess reurns on he sock index wih he shor-erm ineres rae, bu analysis found he associaion beween hese wo variables is inconsisen over he ime. This indicaes ha he direcion is no he same always and changes gradually from one direcion o oher such as; negaive o posiive and even someimes here is no relaionship beween hem. Farsio and Fazel (2010) empirically invesigaed he effec of ineres rae on sock prices in he UAE by using daa from June 2006 o January This sudy evidences from Granger causaliy es ha here is a lack of causal associaion beween hese wo variables and regression model shows ha ineres raes do no have a robus explanaory power in forecasing he sock prices in he UAE. Overall, his lieraure survey indicaes ha here are numerous sudies have been aemped o invesigae he link beween exchange raes, ineres raes and sock prices in boh developed and emerging financial markes. The empirical resuls of hese sudies are mosly inconclusive in naure. In he ligh of he above lieraure survey, he purpose of presen sudy is o conribue o he exising lieraure on he grounds of connecion beween call money raes, exchange raes and sock prices. To he bes of our knowledge, in Indian conex here are some sudies have aemped o find ou he relaionship beween macro variables and sock prices in general, and here is no specific sudy on he relaionship beween call money raes, exchange raes and sock prices, in paricular. Hence, his moivaes us o invesigae he empirical relaionship beween call money raes, exchange raes and sock reurns in Indian perspecive.

7 173 European Journal of Economics Finance and Adminisraive Sciences - Issue 56 (2013) 3. Empirical Models 3.1. Uni Roo Tess This sudy uses he convenional uni roo ess such as; Augmened Dickey-Fuller (Dickey and Fuller, 1979) es, Phillips-Perron (Phillips and Perron, 1988) es and he KPSS (Kwiakowski e al. 1992) ess. All of hese uni roo ess are used o es wheher he daa conains uni roo (non-saionary) or is a saionary process. A series is said o be saionary if he mean and auo co-variances of he series do no depend on he ime facor. Any series ha is no saionary hen i is said o be non-saionary. A series is said o be inegraed of order d which can be denoed by I (d), means ha i has o be differenced d imes before i becomes saionary. Oherwise, if a series by iself, le say saionary a levels, wihou having o be differenced, hen ha is said o be I (0). In he case of boh ADF and PP ess, he null hypohesis of non-saionary (uni roo) is esed agains he alernaive hypohesis of saionary. For he Augmened Dickey-Fuller (ADF) ess; consider a simple AR (1) process: ' y y 1 x (3.1) y Where is he observed variable and x are opional exogenous regressors which may consis of consan or a consan and rend, and are parameers o be esimaed, and he are assumed o 1 be whie noise (i.e., zero mean and consan variance). If, y is a non-saionary series and he variance of y 1 increases wih ime and approaches infiniy, on he oher hand if, hen y is a y (rend) saionary series. Now, subracing equaion (1) boh sides wih 1, hen we ge: ' y y 1 x (3.2) Where 1. The null and alernaive hypoheses can be wrien as; H 0 : = 0 ( is uni roo) H1 : 0 ( is saionary) and can be evaluaed using he convenional -raio for : /( se ( )) (3.3) Where: is he esimae of, and se ( ) is he coefficien sandard error. The equaion (2) is valid only if he series is an AR (1) process, oherwise le say, if he series is correlaed a higher order lags, hen he assumpion of whie noise ( ) disurbances and is violaed. Thus, he ADF es consrucs a parameric correcion for higher-order correlaion by assuming ha he y series follows an AR ( p ) process and adding p lagged difference erms of he dependen variable y o he righ-hand side of he es regression, such as; ' y y 1 x 1 y 1... p y p (3.4) This augmened specificaion is hen used o es he above null hypohesis by using he -raio (1.3). Therefore, sudy uses MacKinnon (MacKinnon, 1996) criical values for ADF es and hen i has been evidenced ha ADF ess are sensiive o he selecion of lag lenghs. Thus, sudy deermines appropriae lag lengh by uilizing Schwarz informaion crieria (SIC). The Phillips-Perron (1988) es incorporaes an alernaive (non-parameric) mehod for conrolling serial correlaion when esing for a uni roo by esimaing he non- augmened Dickey- Fuller es equaion (3.2) and i is modifying he -raio of he coefficien so ha serial correlaion does no affec he asympoic disribuion of he es saisic. The modified -raio is he same as ha of ADF es for he asympoic disribuion of he PP es. Sudy uses Mackinnon (1996) lower-ail criical and p -values for his es.

8 174 European Journal of Economics Finance and Adminisraive Sciences - Issue 56 (2013) y The KPSS (Kwiakowski e al. 1992) es differs from above uni roo ess in ha he series is assumed o be (rend) saionary under he null hypohesis. The KPSS es is based on he y residuals from he OLS regression of on he exogenous variables x : y x ' u (3.5) The LM saisic can be defined as; 2 2 LM S ( ) /( T f 0 ) (3.6) f Where 0 is an esimaor of he residual specrum a zero frequency and where s () is a cumulaive residual funcion: S ( ) r 1 u r u y I is based on he residual of on he Kwiakowski e al. (1992, able-1) Granger Causaliy Tes x ' (3.7) (0). The criical values for he LM es are based The Granger (1969) causaliy procedure is explained as follows; he quesion of wheher y causes x is o see how much of he curren x can be explained by pas values of x and hen o see wheher adding lagged values of y can improve he explanaion. I is said ha x is Granger caused by y, if x can predic beer from pas values of x and y han from pas values of x alone. For a simple bivariae model, one can es he following equaion: x = y = n m 0 i y i j i 1 j 1 n x m x j y u (3.8) 0 i i j j (3.9) i 1 j 1 Where; he null hypohesis is ha y does no Granger causes x in he firs regression equaion and x does no Granger causes y in he second regression equaion Vecor Auoregression (VAR) Models The Vecor Auoregression (VAR) is commonly used for forecasing sysems of inerrelaed ime series and for analyzing he dynamic impac of random disurbances on he sysem of variables. The VAR models consiss a se of regression equaions in which all he variables are assumed o be endogenous. I is believed ha each endogenous variable is explained by is own lagged values and also he lagged values of oher endogenous variables which are included in he model. I is also assumed ha here are no exogenous variables in he model. Hence, only he lagged values of he endogenous variables appear on he righ side of he equaions. A sandard VAR model can be described as follows: C 10 11C 1 12C2 11S 1 12S2 11E 1 12E2 u1 (3.10) E 20 21E1 22E2 21S1 22S2 21C 1 22C2 u2 (3.11) S 30 31S1 32S2 31E 1 32E2 31C 1 32C2 u3 (3.12) C Where, is he call money rae a he ime period, E is he exchange rae and S is he ij, ij, ij sock price; and u 1 u, 2 u and 3 are whie noise disurbance erms wih E( u i are he parameers o be esimaed; ) = 0, (i E ( u u ) = 1,2), 1 2 = 0.

9 175 European Journal of Economics Finance and Adminisraive Sciences - Issue 56 (2013) 4. Daa and Empirical Resuls The presen sudy uses monhly daa for he ime span of April, 1992 o March, This comprehensive daa se enhances he accuracy of he empirical resuls. The required daa on call money raes (monhly weighed average) and exchange raes (monhly US Dollar average agains Indian rupee) are colleced from RBI (Reserve Bank of India) s daabase on Real-Time Handbook of Saisics on he Indian Economy (HBS). Similarly, he daa on monhly closing price indices of BSE- SENSEX and NSE-S&P CNX Nify are colleced from BSE (Bombay Sock Exchange) and NSE (Naional Sock Exchange) official websies, respecively. As an iniiaive, he colleced daa on all he variables are ransformed ino naural logarihms (ln) before he analysis commence. Table 1: Summary Saisics Variables Mean Median Max. Min. Sd. Dev. Skewness Kurosis Jarque- Bera Tes Call M. Raes Exchange Raes Nify Sensex Noe: The esimaion of summary saisics is carried ou on he naural logarihm daa series of all he observed variables The summary saisics of his sudy are presened in above able 1. This indicaes ha all he series are having posiive mean. The series of call money raes, nify and sensex are having posiive skewness and exchange raes series shows negaive skewness. This implies ha he posiively skewed series are flaer o he righ as compared o he normal disribuion and while negaive skewed series is flaer o he lef. The kurosis values of call money raes are higher han he normal values of i and his sugges ha he kurosis curve is lepokuric. While he kurosis values of all oher series are less han he normal values of i, and his suggess ha he kurosis curve is playkuric. In general, value for skewness is zero and kurosis is hree when he observed series is perfecly normally disribued. Since, he resuls of his sudy indicae ha none of hese series are normally disribued. This view also suppored by Jarque-Bera (JB) es, he JB es is used o assess wheher he given series is normally disribued or no. Here, he null hypohesis is ha he series is normally disribued. Resuls of JB es find ha he null hypohesis is rejeced for all he variables and sugges ha all he observed series are no normally disribued. Prob. Table 2: Correlaion Marices Variables Call Money Raes Exchange Raes Nify Sensex Call M. Raes Exchange Raes Nify Sensex Noe: The analysis of correlaion marices is carried ou on he naural logarihm daa series of all he observed variables The above able provides informaion on correlaion beween he observed variables. The call money raes are negaively correlaed wih exchange raes and sock prices (Nify and Sensex). This is suggesing ha here is an inverse relaionship from call money raes o exchange raes and sock prices. In conrary, exchange raes are posiively correlaed wih sock prices. I signifies ha boh exchange raes and sock prices are moving in he same direcion. The resuls also reveal ha here is no high correlaion beween he idenified variables.

10 176 European Journal of Economics Finance and Adminisraive Sciences - Issue 56 (2013) Table 3a: ADF and PP Tes Resuls ADF PP Variables Level 1 s Difference Level 1 s Difference Wihou Wih Wihou Wih Wihou Wih Wihou Wih Trend Trend Trend Trend Trend Trend Trend Trend Call M. Raes -4.17* -6.82* * -6.72* Exchange Raes * * * * Nify * * * * Sensex * * * * Noe: Where (*) and (**) denoe significance level a 1 % and 5 %, respecively. ADF and PP ess examine he null hypohesis of a uni roo agains he alernaive of no uni roo. These wo ess are performed on he naural logarihm daa series. Table 3b: KPSS Tes Resuls Variables KPSS Level 1 s Difference Wihou Trend Wih Trend Wihou Trend Wih Trend Call M. Raes 1.005* Exchange Raes 1.420* 0.416* Nify 1.668* 0.347* Sensex 1.605* 0.349* Noe: Where (*) and (**) denoe significance level a 1 % and 5 % respecively. KPSS es ess he null hypohesis of saionary agains he alernaive hypohesis of non- saionary. This es is performed on he naural logarihm daa series. The able 3a and 3b display uni roo ess resuls, by encompassing he ADF (Augmened Dickey and Fuller, 1979), PP (Phillips and Perron, 1988) -saisics and KPSS (Kwiakowski e al., 1992) LM-saisic. The uni roo ess are performed on he naural logarihm daa series. The ADF and PP ess are carried ou on he assumpion ha he null hypohesis of a uni roo (non-saionary) is esed agains he alernaive hypohesis of no uni roo (saionary). These ess models are esimaed a he levels and firs-difference for boh wih and wihou rend variable in each case. A levels, he ADF and PP ess resuls of call money raes rejec he null hypohesis of uni roo a l % level of significance for boh wih and wihou rend variable. Hence, i suggess ha he call money raes are saionary a heir level. The ADF and PP ess resuls on he oher variables do no rejec he null hypohesis of uni roo a he 5 % level of significance for boh wih and wihou rend variable. This signifies ha hese series are non-saionary a heir levels. Therefore, we applied ADF and PP ess saisics on he firs differenced daa of exchange raes, nify and sensex. The firs differenced daa resuls rejec he null hypohesis of uni roo a 1 % significance level for boh wih and wihou rend for all he hree variables. Furher, we performed KPSS es on all he variables a heir levels and firs difference. This es presumes ha he null hypohesis of no uni roo (saionary) is esed agains he alernaive hypohesis of uni roo (non-saionary). The able 3b supplies KPSS es resuls, a levels he null hypohesis of no uni roo is rejeced for he model wihou rend a 1 % level of significance for all he variables, and for he model wih rend he null hypohesis is no rejeced a 5 % significance level for call money raes and rejeced 1 % level of significance for he remaining variables. This es resuls on firs differenced daa shows ha he null hypohesis is no rejeced for all he variables wih and wihou rend a 1 % significance level. Overall, i is undersood from he hree of uni roo ess ha he call money raes are saionary (no uni roo) a heir levels excep for he model wihou rend in he case of KPSS es. Similarly, he resuls of exchange raes, nify and sensex reveals ha he null hypohesis (non-saionary) has no been rejeced a heir levels and rejeced a heir firs differenced daa for wih and wihou rend variable.

11 177 European Journal of Economics Finance and Adminisraive Sciences - Issue 56 (2013) Table 4: Granger Causaliy Tes Resuls Null Hypohesis Number of Lags F- saisics Causal Relaion Exchange Raes do no Granger Cause Call Money Raes * Bidirecional Call Money Raes do no Granger Cause Exchange Raes ** Relaion Nify does no Granger Cause Call Money Raes Unidirecional Call Money Raes do no Granger Cause Nify *** Relaion Sensex does no Granger Cause Call Money Raes Unidirecional Call Money Raes do no Granger Cause Sensex *** Relaion Nify does no Granger Cause Exchange Raes Unidirecional Exchange Raes do no Granger Cause Nify *** Relaion Sensex does no Granger Cause Exchange Raes Unidirecional Exchange Raes do no Granger Cause Sensex *** Relaion Noe: Where * ** and *** indicae ha he null hypohesis is rejeced a 1 %, 5 % and 10 % level of significance, respecively. The appropriae lag lengh is seleced based on he Akaike Informaion Crierion (AIC). The Granger causaliy es is performed on he saionary daa series. The able 4 exemplifies Granger causaliy es resuls. This esimaion has been carried ou on he saionary variables (firs differenced daa has been used for all he variables excep call money raes) and appropriae lag lengh is seleced based on he Akaike Informaion Crierion (AIC). The es resuls demonsrae ha he null hypohesis of exchange raes do no Granger cause call money raes is rejeced a 1 % level; and he null hypohesis of call money raes do no Granger cause exchange raes is also rejeced a 5 % significance level. This implies ha here exiss a bidirecional relaionship beween exchange raes and call money raes. Similarly, he null hypohesis of Sock reurns (nify and sensex) do no Granger cause call money raes is no rejeced 5% or lower level bu he null hypohesis of call money raes do no Granger cause sock reurns is rejeced a 10 % significance level. This confirms ha here is unidirecional relaionship among call money raes and sock reurns, which runs from call money raes o sock reurns. In he same way, he null hypohesis of sock reurns do no Granger cause exchange raes is no rejeced a 5 % level, and he null hypohesis of exchange raes do no Granger cause sock reurns is rejeced a 10 % significance level. This also demonsraes ha here is unidirecional relaionship beween exchange raes and sock reurns ha runs from exchange raes o sock reurns. Largely, he Granger causaliy es resuls provide evidence ha here is a bidirecional relaionship beween call money raes and exchange raes. The es resuls also show ha sock reurns do no Granger cause neiher call money raes nor exchange raes. Despie of his, here is a unidirecional relaionship from call money raes and exchange raes o sock reurns. Table 5a: Vecor Auoregression (VAR) Models Variables Call Money Raes Exchange Raes Nify Call Money Raes * (0.067) (0.003) (0.015) Call Money Raes * (0.064) 0.008** (0.003) *** (0.014) Exchange Raes * (1.378) 0.224* (0.069) ** (0.306) Exchange Raes *** (1.402) (0.070) (0.311) Nify (0.312) *** (0.016) 0.213* (0.069) Nify (0.304) (0.015) (0.067) Consan 0.481* (0.093) (0.005) 0.049** (0.021) R-squared F-saisic Noe: The VAR models are performed on he saionary daa series and sandard errors are presened in parenheses.

12 178 European Journal of Economics Finance and Adminisraive Sciences - Issue 56 (2013) Table 5b: Vecor Auoregression (VAR) Model Variables Call Money Raes Exchange Raes Sensex Call Money Raes * (0.067) (0.003) (0.015) Call Money Raes * (0.064) 0.008** (0.003) *** (0.014) Exchange Raes * (1.378) 0.226* (0.069) ** (0.310) Exchange Raes *** (1.405) (0.070) (0.316) Sensex (0.309) (0.015) 0.196* (0.069) Sensex (0.304) (0.015) (0.068) Consan 0.478* (0.094) (0.005) 0.054** (0.021) R-squared F-saisic Noe: The VAR models are performed on he saionary daa series and sandard errors are presened in parenheses. The Vecor Auoregression (VAR) resuls are presened in he able 5a and 5b. This esimaion is carried ou on he saionary daa series. This sudy has esimaed ri-variae VAR models, wih one for each dependen variable in he sysem. Table 5a reveals ha he lagged values of call money raes are significan a 5 % and 10 % level for exchange raes and nify, respecively. This suggess ha here is causaion from call money raes o exchange raes and nify a 5 % and 10 % levels, respecively. There is also causaliy from exchange raes o call money rae and nify. No causaliy is idenified from nify o call money raes and for exchange raes here is causaliy a 10 % significan level. Similarly, able 5b also presens VAR model resuls (sensex). The lagged values of call money raes cause exchange raes and sensex a 5 % and 10 % significance level, respecively. The causaliy is also found from exchange raes o call money raes and sensex. However, here is no reverse causaliy from sensex o call money raes and exchange raes a boh he lags. Largely, he VAR model resuls indicae ha he causaliy is observed from call money raes o exchange raes and sock reurns (nify and sensex) a 5 % and 10 % significance level, respecively. In he same way, he causaliy is also observed from exchange raes o call money raes (1 % and 10 % levels a boh he lags, respecively) and sock reurns a 5 % level of significance. On he oher hand, here is no reverse causaliy from sock reurns o call money raes and exchange raes a 5 % significance level or lower level. Hence, his suggess ha he informaion is incorporaed more quickly ino call money raes and exchange raes han in he sock reurns. 5. Conclusion The presen sudy empirically invesigaed he relaionship beween call money raes, exchange raes and sock reurns in Indian perspecive. This sudy has used monhly daa from April 1992 o March 2011 which provides sufficien daa se for he empirical analysis. The resuls of Granger causaliy es presen ha here is a bidirecional relaionship beween call money raes and exchange raes. Furher, i evidences ha he call money raes and exchange raes Granger cause sock reurns a 10 % significance level. This suggess ha here is unidirecional relaionship and ha runs from call money and exchange raes o sock reurns. However, his sudy did no observe reverse causaliy from sock reurns o call money and exchange raes even a 10 % level. The Vecor Auoregression (VAR) models are employed for idenifying lead-lag ineracion among he observed variables. The resul of hese models reveals ha he causaliy is observed from he lagged values of call money raes o exchange raes and sock reurns a 5 % and 10 % level of significance, respecively. Similarly, he causaliy is also found from he lagged values of exchange raes o call money raes (a 1 % and 10 % level) and sock reurns (a 5 % level). In conradicion o his, he reverse causaliy is no found from

13 179 European Journal of Economics Finance and Adminisraive Sciences - Issue 56 (2013) sock reurns o call money and exchange raes a 5 % significance level. Overall, our sudy findings are very much imporan in he perspecive of invesors and policy makers. In he invesors poin of view, his hisorical informaion of call money raes and exchange raes can be uilized for predicing he movemens of sock reurns. Similarly, his findings also useful for he policy makers in sabilizing he sock marke flucuaions by adoping suiable policy measures owards ineres raes and exchange raes. References 1] Abdalla, I.S.A., and Murinde, V., (1997) Exchange Rae and Sock Price Ineracions in emerging Financial Markes: Evidence of India, Korea, Pakisan and Philippines Applied Financial Economics, 7, pp ] Aggarwal, R., (1981) Exchange Raes and Sock Prices: A Sudy of he US Capial Markes under Floaing Exchange Raes Akron Business and Economic Review, 12, pp ] Alagidede, P., Panagioidis, T. and Zhang, X. (2011) Causal Relaionship beween Sock Prices and Exchange Raes The Journal of Inernaional Trade & Economic Developmen, Vol. 20, No. 1, pp ] Alam, M. and Uddin, G.S. (2009) Relaionship beween Ineres Rae and Sock Prices: Empirical evidence from Developed and Developing Counries Inernaional Journal of Business and Managemen, vol. 4, No.3, pp. 43 o 51. 5] Arango, L.E., Gonzalez, A. and posada, C.E. (2002) Reurns and Ineres Rae: A Nonlinear Relaionship in he Bogoa Sock Marke applied Financial Economics, 12 (11), pp ] Aydemir, O., and Demirhan, E., (2009) The Relaionship beween Sock Prices and Exchange Raes: Evidence from Turkey Inernaional Research Journal of Finance and Economics, Issue 23, pp ] Bahmani,O. M., and Sohrabiab, A., (1992) Sock Prices and he Effecive Exchange Rae of he Dollar Applied Economics, 24, pp ] Bhaacharya, B. and Mukherjee, J., (2003) Causal Relaionship beween Sock Marke and Exchange Rae, Foreign Exchange Reserves and Value of Trade Balance: A Case Sudy for India This paper was presened a he Fifh Annual Conference on Money and Finance in he Indian Economy. 9] Bren, W., Glosen, R. L., and Jagannahan, R. (1989) Economic Significance of Predicable Variaions in Sock Index Reurns The Journal of Finance, 44 (5), pp ] Campbell, J.Y. (1987) Sock Reurns and he Term Srucure Journal of Financial Economics, 18, pp ] Caporale, G.M., Piis, N., and Spagnolo, N., (2002) Tesing for Causaliy-in-Variance: An Applicaion o he Eas Asian Markes Inernaional Journal of Finance and Economics 7 (3), pp ] Chow, E.H., Lee, W.Y. and Sol, M.S., (1997) The Exchange Rae Risk Exposure of Asse Reurns Journal of Business, 70, pp ] Dickey, D.A. and Fuller, W.A., (1979) Disribuion of he Esimaors for Auoregressive Time Series wih a Uni Roo Journal of he American Saisical Associaion, 74, pp ] Farsio, F. and Fazel, S. (2010) The Impac of Ineres Raes on Sock Prices in he UAE European Journal of Managemen, Vol. 10, Issue, 3. 15] Frank, P. and Young, A., (1972) Sock Price Reacion of Mulinaional Firms o Exchange Realignmens Financial Managemen, 1, pp ] Giovannini, A. and Jorion, P., (1987) Ineres Raes and Risk Premia in he Sock Marke and in he Foreign Exchange Marke Journal of Inernaional Money and Finance, 6, pp ] Granger, C.W.J., Huang, B-N. and Chin-Wei, Y. (2000) A Bivariae Causaliy beween Sock Prices and Exchange Raes: Evidence from Recen Asian Flu The Quarerly Review of Economics and Finance 40, pp

14 180 European Journal of Economics Finance and Adminisraive Sciences - Issue 56 (2013) 18] Granger, K.R., (1969) Invesigaing Causal Relaions by Economeric Models and Cross- Specral Mehods Economerica, 37, pp ] Gupa, R., and Basu, P. (2007) Weak Form Efficiency In Indian Sock Markes Inernaional Business & Economic Research Journal, 6, pp ] Harasy, H. and Roule, J. (2000) Modeling Sock Marke Reurns Journal of Porfolio Managemen 26 (2). 21] Hsing, Y. (2004) Impacs of Fiscal Policy, Moneary Policy and Exchange Rae Policy on Real GDP in Brazil: A VAR Model Brazilian Elecronic Journal of Economics, 6 (1). 22] Joseph, N.L., (2002) Modeling he Impacs of Ineres Rae and Exchange Rae Changes on UK Sock Reurns Derivaives Use, Trading and Regulaion 7 (4), pp ] Kumar, M. (2010): Causal Relaionship beween Sock Price and Exchange Rae: Evidence from India Inernaional Journal of Economic Policy in emerging Economies, vol. 3, No. 1, pp ] Kuy, G. (2010) The Relaionship beween Exchange Raes and Sock Prices: The Case of Mexico Norh American Journal of Finance and Banking Research, Vol. 4, No. 4, pp ] Kwiakowski, D., Phillips, P.C.B., Schmid, P. and Shin, Y., (1992) Tesing he Null Hypohesis of Saionariy agains he Alernaive of a Uni Roo: How sure are we ha Economic Time Series have a Uni Roo, Journal of Economerics, Vol. 54, pp ] Lee, W. (1997) Marke Timing and Shor-erm Ineres Raes Journal of Porfolio Managemen, 23 (3), pp ] Leon, N.K. (2008) The Effecs of Ineres Raes Volailiy on Sock Reurns and Volailiy: Evidence from Korea Inernaional Research Journal of Finance and Economics, Issue, 14, pp ] MacKinnon, J.G., (1996) Numerical Disribuion Funcions for Uni Roo and Coinegraion Tess Journal of Applied Economerics, 11, pp ] Mishra, A.K. (2004) Sock Marke and Foreign Exchange Marke in India: Are hey Relaed? Souh Asia Economic Journal, 5 (2), pp ] Muhammad, N. and Rasheed, A. (2002) Sock Prices and Exchange Raes: Are They Relaed? Evidence from Souh Asian counries The Pakisan Developmen Review, 41:4, pp ] Nieh, C., and Lee, C., (2001) Dynamic Relaionship beween Sock Prices and Exchange Rae for G-7 Counries Quarerly Review of Economics and Finance, 41, pp ] Ong, L.L., and Izan, H.Y., (1999) Sock and Currencies: Are They Relaed? Applied Financial Economics 9, pp ] Pan, M-S., Fok, R.C-W., and Liu, Y.A., (2007) Dynamic Linkages beween Exchange Raes and Sock Prices: Evidence from Eas Asian Markes Inernaional Review of Economics and Finance 16, pp ] Paramai, S.R., and Gupa, R. (2011) An Empirical Analysis of Sock Marke Performance and Economic Growh: Evidence from India Inernaional Research Journal of Finance and Economics, 73, pp ] Phillips, P.C.B. and Perron, P., (1988) Tesing for a Uni Roo in Time Series Regression Biomerika, Vol. 75, pp ] Rahman, Md. L. and Uddin, J. (2008) Relaionship beween Sock Prices and Exchange Raes: Evidence from Bangladesh Inernaional Journal of Business and Managemen, Vol. 3, No. 9, pp ] Rahman, Md. L. and Uddin, J. (2009) Dynamic Relaionship beween Sock Prices and Exchange Raes: Evidence from Three Souh Asian Counies Inernaional Business Research, Vol.2, No. 2, pp ] Rigobon, R. and Sack, B. (2004) The Impac of Moneary Policy on Asse Price Journal of Moneary Economics, 51, pp ] Roll, R. (1992) Indusrial Srucure and he Comparaive Behavior of Inernaional Sock Markes Indices Journal of Finance, 47, pp

15 181 European Journal of Economics Finance and Adminisraive Sciences - Issue 56 (2013) 40] Shanken, J. (1990) Ineremporal Asse Pricing Journal of Economerics, 45, pp ] Smih, C., (1992) Sock Marke and he Exchange Rae: A Muli-Counry approach Journal of Macroeconomics 14, pp ] Soenen, L.A. and Hennigar, E.S., (1988) An Analysis of Exchange Raes and Sock Prices: The US Experience beween 1980 and 1986 Akron Business and Economic Review, winer, pp ] Solnik, B., (1987) Using Financial Prices o Tes Exchange Rae Models: A Noe Journal of Finance, 42, pp ] Savarek, D., (2005) Sock Prices and Exchange Raes in he EU and he USA: Evidence of heir Muual Ineracions Finance a úvûr Czech Journal of Economics and Finance, 55, pp ] Uddin, M.G.S. and Alam, M.M. (2007) The Impacs of Ineres Rae on Sock Marke: Empirical Evidence from Dhaka Sock Exchange Souh Asian Journal of Managemen and Sciences, 1 (2), pp ] Vygodina, A.V., (2006) Effecs of Size and Inernaional Exposure of he US Firms on he Relaionship beween Sock Prices and Exchange Raes Global Financial Journal 17, pp ] Wong, W-K., Khan, H. and Du, J. (2005) Money, Ineres Rae and Sock Prices: New Evidence from Singapore and he Unied Saes Graduae School for Global Leaders, Working Paper, No ] Zhou, C. (1996) Sock Marke Flucuaions and he Term Srucure Board of Governors of he Federal Reserve Sysem, Finance and Economics discussion Series, 03.