Analysis of the Relationship between Oil Prices and Exchange Rates in Tehran Stock Exchange

Transcription

1 Internatonal Journal of Research n Busness Studes and Management Volume 1, Issue, December 014, PP 818 ISSN (Prnt) & ISSN (Onlne) Analyss of the Relatonshp between Ol Prces and Exchange Rates n Tehran Stock Exchange Samane Shadab Department of Management, Zanjan Branch, Islamc Azad Unversty, Zanjan, Iran Amr Gholam Department of Management, Zanjan Branch, Islamc Azad Unversty, Zanjan, Iran gholam.amr.ac@gmal.com Abstract: The money and captal markets, as pllars of the fnancal sector, have the duty of rasng captal for real economy sector. The effcency of fnancal sector leads to optmal allocaton of scarce resources to economc actvtes. The money market s a place for shortterm fnancng and captal market s a place for longterm fnancng of enterprses. Stock Exchange s a captal market n whch fnancal assets are traded as securtes. The recesson and boom of stock market n some countres affect not only domestc economy, but also the global economy. Alternatvely, the major polcy makngs affect the stock market n each country. For makng decson to nvest n the stock market, prce s an mportant crteron whch reflects nvestors' purchasng power and fnancal strength of the companes whch are members of the stock market. Macroeconomc varables are mportant factors that affect the stock prce. Therefore, ths study amed to examne the relatonshp among ol prces, exchange rates, nflaton rate, and stock prce ndex n Tehran Stock Exchange. For ths purpose, the stock market has been studed n detal. Then, the theoretcal foundatons of research were studed whch ncluded portfolo theory, the theory of Fsher, arbtrage prcng theory, and the theory of commodty market approach. Also, the relatonshp between the varables for the 1month perod: 13911: 1376 was analyzed by Model (VAR) and the method of varance analyss. The causalty relatonshp between varables was studed usng Granger causalty test (1987) and Wald test. The results showed that n the long term and short term, all varables except the ol shocks had a sgnfcant relatonshp wth the Stock Exchange ndex. Based on the results of these tests, none of the varables used n ths study had not causal relatonshp wth ol shock. Therefore, the results were consstent wth the expected theoretcal studes. Usng Johansson contegraton tests and Granger s error correcton model, t was found that there was longterm causal relatonshp among the logarthm of the exchange rate, logarthm of nflaton, and the logarthm of stock prce ndex. The greatest mpact of nflaton was on ndex logarthm. Keywords: ol prce, exchange rate, stock market, causalty method, Granger test 1. INTRODUCTION The recent surge n ol prces over the last 8 years has great nterests n the relatonshp among ol prces, fnancal markets, and the economy. The cost of crude ol usng Texas crude ol was calculated to be near the prce of $ 4.9 a barrel n 00. Durng ths perod, unlke other modes of exchange and ncrease of stock market prces, the Amerca dollar was dropped. Although there are artcles about the relatonshp between ol prces and stock prces and the relatonshp between ol prces and the exchange rate, the relatonshp between the two has not been well studed, especally n the feld of stock prces. Scentfcally, the ol prces could affect stock prces n several ways. The prce of a share n a country and n every perod of tme equals to the present value of current spendng and future costs (Huang Masoulys, and Stahl, 010). The ol prce may affect the stock prces drectly by nfluencng on cash flow or ndrectly through ts effect on nterest rates used to reduce the present value of payments and future costs. In the absence of a complete replacement, the affects of producton factors wll rase ol prces and the costs of carryng out an economc actvty and n the nonol companes wll reduce the profts. The ncreased ol prces appear n ncreased prce of the fnal products and servces. Ths reduces the demand for goods and servces and wll result n reduced proftablty. The longterm and stable economc development requres the moblzaton and allocaton of fnancal resources n the natonal economy. As a key element of the captal market, Stock IJRBSM 8

2 Samane Shadab & Amr Gholam Exchange market plays an mportant role n collectng the funds and transferrng them to ndvduals and unts who request funds. The success of Stock Exchange and ts attractveness to potental nvestors s possble by ncreased yelds and stock prces of companes lsted on the stock exchange. The effcency and stock prces n stock market of all countres of the world s nfluenced by dfferent factors. Several factors ncludng companes condtons (such as assets and labltes, management, development projects, and uncertanty of future profts), stock market characterstcs (such as the presence of speculators, stock tradng mechansms, lack of nformaton, change of opportunty cost, and determnng the value of shares), and macroeconomc factors (such as exchange rates, nflaton, money supply, ol prces, and bank s nterest rates) has a great nfluence on developments n the stock market. Understandng the factors affectng ths market mechansm such as macroeconomc varables and changes n prces of rval has mportant role n the predcton of stock market behavor. Therefore, ths study ams to evaluate the mpact of macroeconomc varables and compettor assets of stock market on stock prce ndex n Tehran Stock Exchange market. Usng the nformaton of stock prce ndex varables n stock exchange market, Tehran, exchange rate, ndex of vehcles prces, and housng prce ndex for the perod , Barazande (1997) examned the mpact of mentoned macroeconomc varables on stock prces ndex. The results showed that the mentoned varables have small partcpaton n the stock prce ndex changes. Ths result ndcates that dsrupton and volatlty of the currency and vehcles markets s not strongly applcable to the stock market. Delav and colleagues (008) examned the longterm relatonshp between ol prces and economc growth wth quarterly data n the perod 1989 to 007 n Iran. The results show that n Ira as an exporter of ol, ol shocks effect asymmetrcally the economc growth. Ths means that GDP does not ncrease sgnfcantly by Increasng of ol prce. Abrsham and Rahm (004) studed the shortterm and longterm factors that determne the actual rate of exchange at three commodtes framework. They concluded that n longterm, there was a relatonshp between mports real exchange rate, the terms of trade, the share of nvestment, central bank reserves, the degree of openness of the economy, and consumpton spendng of government. In ther study, Abbasan and Moradpour (008) nvestgated the effect of exchange rate, trade balance, nflaton rate, currency, and nterest rates on the Tehran Stock Exchange general ndex usng quarterly data for the years 1998 to 005. In ths study, the contegraton method, error correcton model, mpulse response functons, and varance analyss were used. The results showed a postve mpact of exchange rate and trade balance n the long term on stock exchange and negatve mpact of nflaton rate, lqudty, and nterest rate on the stock exchange. Based on the error correcton model estmaton, 1 percent of mbalance s adjusted n each perod. Chen, Roll and Ross (1986) nvestgated the relatonshp between some macroeconomc varables and stock market behavor n New York Stock Exchange for the perod In ths study, from the perspectve of ratonal expectatons and effcent markets theory, the mpact of systematc rsk on stock market ndex has been nvestgated. The results showed that real per capta consumpton and ol prces ndex have not sgnfcant mpact on stock prce ndces. Also, nflaton n the short term has a negatve relatonshp wth stock prce and n the long term, t has postve relatonshp wth stock prce. Mukherjee and Naka (1995) examned the dynamc relatonshp between macroeconomc varables (offcal exchange rate, money supply, nflaton rate, ndustral producton, longterm government bond rate, and the rate of deposts) and the Japanese stock market prce ndex. Ths study was estmated usng a vector error correcton model (VECM) n a system of seven equatons for the perod Based on the results, ths assumpton was confrmed that there s a longterm equlbrum relatonshp between stock prces and macroeconomc varables. Muradoglu and Matn (1996) assessed the longterm relatonshp between stock prce ndex of the Istanbul Stock Exchange and varables ncludng nterest rate, exchange rate, nflaton rate, and money supply for the perod usng the Engle Granger and Johansen Juselus approaches. Based on research results, the longterm relatonshp between these varables over the studed perod s verfed. The drecton of longrun relatonshp between stock prce ndex and money volume ndex s postve and among the stock prces ndex, exchange rates, nterest rates, and nflaton rates s negatve. Internatonal Journal of Research n Busness Studes and Management 9

3 Analyss of the Relatonshp between Ol Prces and Exchange Rates n Tehran Stock Exchange Studyng the effect of exchange rate changes on stock market n Ghana, Salfu and et al (007) showed that 55 percent of studed companes were affected by the prce change of dollar Amerca. Ths relatonshp was statstcally sgnfcant. Also, ths study showed that there s drect relatonshp between more stock returns of companes and changes n the dollar exchange rate. By analyzng the weekly data of stock returns and exchange rates for four Swedsh companes n forest ndustry, Rolseth (1996) showed that n December 199 to December 1995, the exchange rates had a negatve mpact on effcency of companes shares.. METHODOLOGY Ths s a causal correlaton study and uses the exstng data. The sample data ncludes stock prce ndex, exchange rate of Tehran s free market, nflaton rate, and Iran s ol prce. They have been obtaned from ran s natonal accounts data publshed by the Central Bank and the Tehran Stock Exchange for monthly tme seres of 1391: The data analyss was conducted usng econometrcs and tme seres wth model (VAR) to nvestgate the Granger causalty accordng to Valet statstc. To nvestgate the long term relatonshp between the varables, the contegraton and related tests such as Granger contegraton method (1978), Joselus Johansson contegraton method (1991), and self notng wth wde lags (1999) were used based on the dynamc relatonshp and longterm equlbrum relatonshp. Theoretcal framework (the effects of ol prce and exchange rate) The exchange rate fluctuatons nfluence the total demand of the economy through the mport, export, and supply, and the supply of economy through the channel of mportng ntermedate goods cost. Therefore, the result of these two effects n producton and prces depends on the ntal condtons of the countres economy. The effect of exchange rate fluctuatons can be nvestgated by total demand usng exports and mports elastcty. In ths method, accordng to Marshall and Lerner condton, f the total exports and mports elastcty wll be greater than one, the devaluaton of the money (exchange rate ncrease) wll lead to mprovement n trade balance and consequently n GDP. If the total of ths elastcty wll be less than one, the value of money wll ncrease (exchange rate decrease) and trade balance wll mprove. Thus, the effect of exchange rate changes on the demand sde depends on the elastcty of exports and mports. The declne n nvestment secton s one of the other factors of demand sde whch s affected by changes n exchange rates. In most developng countres, the domestc nvestment depends greatly on mportng captal goods whch wll be exploted after gettng combned wth domestc captal and resources. In such crcumstances, the mport costs wll ncrease wth ncrease of exchange rate and decrease of domestc money value. Wth any decrease n mports of captal goods, the domestc nvestment and subsequently the total demand wll be reduced (Qetmr and Sherafatan Jahrom, 007: 6). In addton, many developng countres have large external debts due to recevng foregn loans. The devaluaton of money n these countres has ncreased ther debt n terms of domestc currency. The extenson of these debts pressure has lead to the loss of resources n producton and declne n GDP (Bahman Oskooee, et al (006), p50). In the commodtes market, the postve shocks cause an ncrease n the prce of mported goods and decrease n the prce of mported goods. As a result, demand for domestc goods wll ncrease (Kazeroon and Rostam, 007: 180). On the supply sde, t can be argued that n developng countres, postve shocks to the exchange rate (natonal currency devaluaton) wll lead to an ncrease n the cost of mported ntermedate goods and, therefore, the mport of ntermedate goods wll be more expensve. Ths may have a negatve effect on producton. Also, the ol and ts prce affects on ol exporter and mporter countres through varous mechansms. In the olmportng countres, the ol prce nfluence through two channels of supply and demand on real actvtes. In terms of ol supply sde, t can be sad that crude ol s a major factor n producton. By ncreasng of ol prce, the producton costs wll ncrease and producton wll decrease. On the demand sde, t can be sad that by ncrease of ol prce, the consumpton wll reduce because the avalable ncome (n olmportng countres) s decreased. Also, the ncrease n prce leads to a decrease n nvestment because the ncrease n ol prce ncreases cost of frms (Jn, 01: 9998). Gven ther economc structure, the petroleum exportng countres are not much mpacted by above ol prces trends. The ol prce mpacts through dfferent mechansms on the economes of these countres. The ncrease n ol prces stmulates both the supply and demand sdes n the oldependent Internatonal Journal of Research n Busness Studes and Management 10

4 Samane Shadab & Amr Gholam economes. But because of government subsdes and support systems, the costs of actvtes whch use energy (petroleum and petroleum products) as nput to producton wll not ncrease. As a result, t does not shft the macro supply curve and only stmulate the demand (Jongwook, 013: 73). The ncrease of ol prce causes more revenues to be transferred from ol mportng countres to olexportng countres. Also, the ol sector s one economc part of the petroleum exportng countres and has a large share of value added n these countres (Delavar et al, 008: 69). Rsng of ol prces wll boost ths sector and attract domestc and foregn nvestments whch lead to an ncrease n total producton. Also, the olexportng countres have mostly state economy and run by ol revenues. Therefore, major nvestments n nfrastructure and other nvestments are fnanced from the state's ol revenues. The declne of ol prces n olexportng countres wll reduce the government's ol revenues. Snce current expendtures have relatvely low vscosty and they cannot easly be reduced when there s a declne n ol revenues, the declne n ol revenues wll reduce nfrastructure nvestments. Ths reduces the producton n socety. However, studes have shown that ol prces have asymmetrc effects on olexportng countres. Ths means that the rate of reducton n producton when ol prces decrease s not the same as the rate of ncrease n producton when ol prces ncrease. Also, the effect of ol prce on the economy of ol exportng countres s also nvestgated n the resource curse lterature. The resource curse phenomenon refers to multple deleterous effects whch are appled from rsng prces of ol and other natural resources on the economc, socal, and poltcal sectons of exporter communtes. In ths regard, the economsts rased the ssue of Dutch dsease. Accordng to the Dutch dsease phenomenon, f the economy faces wth an ncrease n export prces of basc commodtes such as crude ol, ths wll lead to an ncrease n revenue and ncrease n domestc demand. The man response of the economy n ths tme s ncrease of labor demand and subsequently ncreases n wages. As the prce of nontrade products ncreases, thus the ncreased wages wll decrease the export sector benefts. Fnally, the effects of sudden ol prce shocks wll lead to real exchange rate devaluaton. Ths would reduce the country's compettveness n the nternatonal arena, decrease productvty n the tradable sector of the economy, and reduce the value added n these sectors (Abbasan et al., 007: 110). In general, theoretcally, the volatlty of ol prces has dfferent effects n opposte drectons on economc growth n developng countres and the general effect s dependent on the outcome of these effects. 3. ANALYSIS OF RESEARCH The requred data n ths study are extracted for the years 1997 to 01 from monthly tme seres of the Central Bank and the Statstcal Center of Iran. In the studes whch are based on these data, t s assumed that the tme seres data are statonary. However, these data are statonary f ther mean and varance wll be constant over tme and the covarance between ther tme ntervals depends on only tme nterval and the gap between the two perods. Otherwse, the studed varables are non statonary and due to facng wth spurous regresson phenomenon, the conventonal tests such as t and F do not have enough credts. The man reason for ths s the desre of tme seres varables to ascendng and descendng tme oscllatons whch causes the result of statstcal correlaton between these varables and the R value wll not be relable. When usng tme seres data, therefore, t s needed to test the statc varables (Gujarat, 1998). In ths study, the Augmented Dckey Fuller statstc and unt root test are used to nvestgate the statonary of used varables. Also the vablty of varables s studed by Phllps Perron unt root test to ensure about the valdty of generalzed DckeyFuller test results due to the possblty of a structural break n the pattern varables. Frst, the software EVews 5 was used to nvestgate the vablty of varables assumng the ntercept at 5% probablty level. Ths software allows automatcally the choce of optmal number of lags to elmnate seral correlaton n the resduals based on the Schwarz Info Crtera, HannanQunn nformaton crteron, generalzed Dckey Fuller, Fnal Predcaton Error crtera, and a measure of LR. In ths study, the Schwarz Info Crtera s consdered to be determnng crtera of optmal lag. The studes suggest that the logarthm varables of nflaton rate are vable and the logarthm of ol prce s dfferent at data level. The results also show that the logarthm varables of stock general ndex and logarthm of exchange rate at data level are nonvable and get vable by a dfference I (1) (Table 1). Internatonal Journal of Research n Busness Studes and Management 11

5 Analyss of the Relatonshp between Ol Prces and Exchange Rates n Tehran Stock Exchange Table1. The generalzed DckeyFuller test n the frst order level and dfference of varables Varable Level Frst dfference order Source: the research results Model wth lattude from orgn statstc Crtcal value Result LnTepx I(1) LnExchange I(1) LmInflaton I(0) dlnol I(0) LnTepx I(0) LnExchange I(0) As mentoned above, due to structural falure n varables and not consderng ths falure n generalzed Dckey Fuller test, the Phllps Perron test n the software EVews 5 s used to ensure the results of Dyky Fuller test. The PhllpsPerron test results show that the results of generalzed Dckey Fuller test are not affected by structural falures, the unt root n varables does not result from structural falure, and the logarthm varables of general stock ndex and the logarthm of stock exchange actually has a unt root. Table. PhllpsPerron test at the frst order level and dfference of varables Model wth lattude from the orgn Result Crtcal value I(1) I(1) I(0) I(0) I(0) I(0) Source: research results Determnng the optmal lag length Statstc Varable LnTepx LnExchange LmInflaton dlnol LnTepx LnExchange Level Frst dfference As you know, the optmal lag VAR models can be determned by Schwarz nformaton crtera, Hnnan Queen Informaton crtera, generalzed Dckey Fuller, maxmum lkelhood, and fnal predcton error crtera. The EVews 5 software enables automatc calculaton and determnaton of the optmal lag. Thus, the values of LR, FPE, AIC, SC, HQ crtera are provded n Table 3. The results of above crtera show that the second lag s the optmal lag for vector autoreturn model. Table 3. The values of dfferent crtera for determnng the optmal lag n VAR model Result K= HQ Source: research results SC AIC FPE Autoregresve vector model and the results of Granger causalty test results and Valet statstc: Snce the optmal lag of autoregresve vector model, based on Schwartz crteron, s equal to, the autoregresve model s estmated wth two lags (equatons (51) to (54)). LnTepxt c 1 t LnTepx t LnEx t LnInflaton tdlnol t 1 1 t 1 1 t 1 1 t LR NA order Lag (15) Internatonal Journal of Research n Busness Studes and Management 1

6 Samane Shadab & Amr Gholam LnExt c 1 t LnTepx t LnEx t LnInflaton tdlnol 1 t 1 t 1 t 1 t LnInflaton t c 1 t LnTepx t LnEx t LnInflaton tdlnol t 1 3 t 1 3 t 1 3 t LndOlt c 1 t LnTepx t LnEx t LnInflaton tdlnol t 1 4 t 1 4 t 1 4 t Wth the help of Valet test, the above equatons are used to study the Granger causalty test from the varables on the rght to the varables on the left sde. For example, the smultaneous sgnfcance of lag coeffcents n logarthm varable of exchange rate (B 1 ) n equaton (51) through the Wald test (wth the dstrbuton of X ) ndcates the presence of Granger causalty relatonshp from exchange rates to stock prce ndex. The results of Wald test for the coeffcents of varables lag n equatons (5 1) to (54) are summarzed n Table 4. Table 4. Wald test results Concluson LnEx LnTepx LnInflaton LnTepx dlnol LnTepx LnTepx LnEx LnInflaton LnEx dlnol LnEx LnTepx LnInflaton LnEx LnInflaton dlnol LnInflaton LnTepx dlnol LnEx dlnol LnInflaton dlnol Source: research calculatons Wald statstc (x ) (Prob) Affectng varable LnEx LnInflato n dlnol LnTepx LnInflato n dlnol LnTepx LnEx dlnol LnTepx LnEx LnInflato n Dependent varable LnTepx LnEx LnInflato n dlnol Based on the expermental results presented n Table 4, there s a causal relatonshp from the exchange rate and bankng nflaton rate to the stock prce ndex. There s a undrectonal causal relatonshp from stock prce ndex and exchange rate to nflaton rate. Also, there s a undrectonal causal relatonshp from nflaton to the exchange rate. Accordng to Fsher's theory theoretcal prncples, the commodty markets approach and dentfyng factors affectng the stock prce ndex, there s causal relatonshp from exchange rate and nflaton ndex to the general stock market ndex. Also, the results of Zadefar and Samm (008) confrm the undrectonal causalty relatonshp from nflaton rate to the stock market ndex. Mohammad, Taghav, and Barazandeh (008) also confrm undrectonal causalty relatonshp from exchange rate to the total stock market ndex. Also t s expected that there wll be causal relatonshp from exchange rate to nflaton rate. The reduced local currency value wll ncrease the mport prce and the cost of mported nputs and wll mpact on the producton and domestc prces. Therefore, the ncrease n mport prces due to devaluatons of domestc currency s one reason for the ncrease n domestc nflaton. The studes of Abasnejad and Teshn (004) also confrm the mpact of exchange rate on nflaton. One of the man tasks of the stock exchange n the economy s helpng to absorb lqudty and reducng nflaton n the communty. Therefore, t s not unexpected that a causal relatonshp wll exst from the stock prce ndex to nflaton rate. Internatonal Journal of Research n Busness Studes and Management 13

7 Analyss of the Relatonshp between Ol Prces and Exchange Rates n Tehran Stock Exchange However, ol prce n the economy of Iran s an exogenous varable and t s a functon of global ol supply and demand and ts nternal varables such as changes n stock prces, exchange rates, and nflaton rates wll not change world ol prces. The central bank n Iran controls the exchange rates and prevents from great changes n assets and drects the substantal fnancal nvestments of nvestors to the market. For example, after one perod of ncrease n ol prce, the government sells one part of the foregn currency obtaned from ol sales to the central bank and ncreases the net assets of the Central Bank and thus ncreases the monetary base. Before the exchange rate mutates because of declne n ol prces and the demand of fnancal nvestors ncreases, the central bank ntervene n the currency market because of havng suffcent fnancal resources and controls the exchange rate wthn the target area. The mportance of ol revenues n the economy of Iran and ts mpact on GDP and other macroeconomc varables s undenable. Ol prce volatlty s one of the man sources of fluctuatons n ol producng countres. The ncrease of ol prces after 1973 had a sgnfcant mpact on the economes of olproducng countres. Durng ths perod, the foregn exchange earnngs from ol sales rose sharply and lead to the growth of prces, wages rates, and the mport rates of Petroleum Exportng Countres. Most studes on the effects of ol shocks on macroeconomc varables have been conducted n ol mporters ndustralzed countres. In these countres, hgher ol prces as one of the factors n producton have led to nflaton and recesson smultaneously. The reduced ol prces are also acted as a postve shock on the supply sde. In the studes of Petroleum Exportng Countres, the Dutch dsease lterature s consdered to be as the most mportant theoretcal foundaton. Accordng to studes conducted n Iran, the role of ol revenues of government n nflaton s not onedmensonal and t s completely multdmensonal. These earnngs have mportant effects on nflaton such as development of mports n the sectons whch are dependent on mported goods, ncrease n ol revenues due to ts ablty to gan foregn currency, and the country's economc boom through the mport of captal goods and raw materals. The ncrease n consumpton goods mports hurts domestc producton. The government controls temporarly the nflaton through commodtes lke fruts that have the ablty to be mported. However, n the case of goods such as housng that cannot be mported, the ol revenues ncrease wll ncrease the prces of these goods. Also, the ncreasng dependence on mports ncreases the mported nflaton role (n relaton to the ncrease n nternatonal prces) n domestc nflaton. The revenue from nflaton whch s avalable for government can be consumed n the form of current and development expendtures n communty. The ncreased ol revenue ncreases government expendtures. It ncreases strongly the aggregate demand n socety drectly and ndrectly. Due to structural constrants n the supply of goods and servces, the gap between demand and supply n the communty ncreases and nflaton rses. Wth rsng of ol revenues, the monetary base ncreases nflaton through foregn assets of the central bank. And by reducng of government's ol revenues, the fnancng of budget defct also causes nflaton by expandng the monetary base. Gven the above, generally, the prce of ol and ol revenues affect nflaton through channels of exchange rate, budget defct, mports, mported nflaton, monetary base, GDP growth, and nvestment n the publc sector. In some cases, they have opposte effects on nflaton. Therefore necessarly, the ncrease or decrease of prces and ol revenues do not cause Inflaton. But, t s the management of the revenues n these channels that can be the cause of nflaton (Hossenpoor). Response Functons: As you know, the estmated coeffcents may dffcultly be nterpreted n Model VAR, especally when the coeffcents symbols change accordng to varable lag. For ths purpose, the reacton exctaton functon s estmated and on ts bass, the behavor of varables s examned over tme. The reacton exctaton functon shows the effect of an endogenous varable reacton to changes n one of the dsrupton (stmulaton) terms durng the tme (Noferest, 1999). Snce the purpose of ths study s nvestgatng the nfluence of exchange rate, nflaton, and ol prces on the Stock Exchange, the response functon of total stock ndex s presented (Fgure 1). Internatonal Journal of Research n Busness Studes and Management 14

8 Samane Shadab & Amr Gholam Response to Cholesky One S.D. Innovatons.08 Response of LNTEPIX to LNTEPIX.08 Response of LNTEPIX to LNINFLATION Response of LNTEPIX to LNEX.08 Response of LNTEPIX to DLNOIL Fgure 1. Stmulaton response dagrams The frst dagram shows the effect of one standard devaton shock from the stock ndex on tself. As you can see, the effect of ths varable on tself has been postve durng the 10year perod. Its effect does not dsappear durng ths perod. The second dagram shows the effect of one standard devaton shock from the nflaton on stock ndex. Accordng to ths dagram, the effect of nflaton on stock ndex has been negatve and ts effect has not been dsappeared durng 10 perods. The thrd dagram shows the effect of shock from exchange rate on stock ndex. Accordng to ths dagram, the effect of ths varable s negatve and ts effect s low, such that ts effect almost dsappears durng 10 perods. The fourth dagram shows the effect of shock from ol prce on stock ndex. Accordng to ths dagram, the effect of ths varable s low n ntal perod and gets more durng next perods and ts effect remans constant and does not dsappear durng 10 perods. Table 5. The results of varance analyss Varance Decomposton of LNTEPIX: DLNOIL LNEX LNINFLATION LNTEPIX S.E. Perod Internatonal Journal of Research n Busness Studes and Management 15

9 Analyss of the Relatonshp between Ol Prces and Exchange Rates n Tehran Stock Exchange The analyss of varance showed that n the ntal perods, most of the changes n stock ndex varable are explaned by the varable tself. But as tme passes and durng subsequent perods, t s reduced and the explanatory of other varables ncreases. As you can see, the explanatory of nflaton s hgher than any other varable and reaches to 33% at tenth perod. Then, the greatest mpact s on the exchange rate. Investgatng the longterm and shortterm relatonshp of varables n the stock market ndex In ths secton, the Vector error correcton model, whch nvolves determnng the dynamc relatonshp and longterm relatonshp between the varables, s used to evaluate the longterm relatonshp among exchange rate, nflaton rate and ol prce shocks varables on stock market ndex. So frst, we conduct the Johansson contegraton test to determne the long term causal relatonshp. The presence of at least one contegraton vector between the research varables provdes the statstcal bass for usng error correcton model. If there wll be a contegraton vector, the vector error correcton model wll be estmated and ts results wll be analyzed. Johansson Contegraton test The Johansson contegraton test s used to nvestgate the longterm equlbrum relatonshp. In the contegraton analyss, frst the Johansson test was used to determne the number of contegraton vectors and the elements of normal contegraton vectors. The results are shown n Table 6. As you can see, accordng to Trace Statstc and Maxmum Egen Statstc, the studed varables are contegraton and at least one contegraton vector s approved. Thus, we can conclude that there s a longterm equlbrum relatonshp between the studed varables and Tehran Stock Exchange (TSE). Table 6. The results of contegraton usng Johansson test Prob MaxEgen Statstc Crtcal value Prob Trace Statstc crtcal level H 0 r=0 Source: research calculatons The exstence of contegraton between a set of economc varables provdes the statstcal bass for usng error correcton models. Accordng to Johansson test (Table 6), there s at least one contegraton vector for studed varables. The vector error correcton model (VECM) was used to evaluate the shortterm relatonshps between the varables n the model. Usng Evews software, the Granger causalty test was used to determne the drecton of causalty. Accordng to EngelGranger (1978), the longterm equlbrum values of contegraton varables get related wth shortterm fluctuatons as followng vector error correcton model. dlntepxt m 1 dindol m m m t t dlntepx d ln Ex dinflaton t Ectt 1 t t The sgnfcance of each coeffcent (excludng the error correcton coeffcent) n vector error correcton model s confrmed by Wald test. The sgnfcance of ndependent varables shortterm coeffcents ndcates the shortterm mpact of the ndependent varables on the dependent varable. The smultaneous sgnfcance of each of the dependent varables and the error correcton term Internatonal Journal of Research n Busness Studes and Management 16

10 Samane Shadab & Amr Gholam represents the long term causal relatonshp from that varable to the dependent varable. The sgnfcance of VECM coeffcent based on ttest shows that n each perod, a percentage of shortterm mbalances n stock prce ndex are moderated to reach to long term balance. In other words, ths coeffcent ndcates the tme that stock prce ndex takes to return to longterm equlbrum poston. The results of estmatng error correcton model are provded n Table 7 and Table 8. Table 7. shortterm Granger causalty test usng VECM model Dependent varable dlntepx H 0 Short term causalty test dlnex dlninflat on Testng the sgnfcance of varables lag coeffcents by Wald test Source: Results of Survey Table 8. Longterm Granger causalty test usng VECM Dependent varable dlntepx H 0 Coeffcent Standard devaton statstc t Source: Results of Survey dlndol Internatonal Journal of Research n Busness Studes and Management 17 Long term causalty test ECTt dlnex ECTt ] [ & ECT t 1 ] [ dlninflat on & ECT t 1 ] [ dlndol & ECT t 1 ] [ Accordng to above table, the changes n the logarthm of these varables cause Granger changes n the logarthm of stock prce ndex. The nsgnfcance of varaton coeffcent n the logarthm of ol shocks by Wald test ndcates that there s no shortterm causalty relatonshp from changes n ol shocks logarthm to changes n stock prce ndex logarthm. The correcton coeffcent of model error s smaller than one and ts sgn s negatve. Also, ts value shows that the adjustment n total stock market ndex n each perod s adjusted.49% of the total ndex market mbalances to longrun equlbrum values. Because the studed perod s monthly, as a result, about 9.88% of the mbalances n the stock prce ndex are adjusted each year. The smultaneous sgnfcance of changes n changes logarthm of exchange rate, nflaton change logarthm, and lag error correcton term ndcates that n the long, these varables are the cause of changes n the logarthm of stock prce ndex. 4. CONCLUSIONS AND RECOMMENDATIONS The results show that all varables, except the ol shocks and nflaton, are based on generalzed DckeyFuller test I (1). The results of Phllps Perron test also confrmed the DckeyFuller test results. It showed that the presence of unt root n varables s not due to structural falure, but ndeed the varables have a unt root. The Granger causalty test and the Wald test were used to determne causalty relatonshp by VAR model n whch the optmal lag had determned accordng to the Schwartz nformaton crteron. The results ndcate that there s shortterm and longterm causal relatonshp between the logarthm of exchange rate and the logarthm of nflaton rate to the stock prce ndex. The error correcton coeffcent s.49%. Ths ndcates that n each monthly perod and n each year, respectvely, only.49% 9.88% of mbalances n the stock prce ndex are removed n longterm. The weak and negatve relatonshp between exchange rate and stock prce ndex can be expressed n terms of commodty markets theory. In companes that are more exportorented, the ncrease n exchange rate would have a negatve mpact on the company's shares and the Stock Exchange. The causal relatonshp from exchange rate to nflaton rate s n accordance wth expectaton. Wth the devaluaton of local currency, the mport prce and the cost of mported nputs wll ncrease and wll affect producton and domestc prces. Therefore, the ncrease n mport prces due to devaluaton of local currency s one reason for the ncrease of domestc nflaton. The studes of Abasnejad and Teshkn (008) also confrm the mpact of exchange rate on nflaton. One of the man tasks of the

11 Analyss of the Relatonshp between Ol Prces and Exchange Rates n Tehran Stock Exchange stock exchange s absorbng lqudty and reducng nflaton n the communty. Therefore, the causal relatonshp from the stock prce ndex to nflaton rate s not unexpected. The mportance of ol revenues n the economy of Iran and ts mpact on GDP and other macroeconomc varables s undenable. In studes are conducted for ol exportng countres, the Dutch dsease s consdered to be the most mportant theoretcal foundaton. Accordng to studes conducted n Iran, the role of ol revenues n nflaton s not one dmensonal and t has multdmensonal nature. Generally, the prce of ol and ol revenues affect nflaton through channels of exchange rate, budget defct, mports, mported nflaton, monetary base, GDP growth, and nvestment n the publc sector. In some cases, they have opposte effects on nflaton. Therefore necessarly, the ncrease or decrease of prces and ol revenues do not cause Inflaton. But, t s the management of the revenues n these channels that can be the cause of nflaton. The central bank n Iran controls the exchange rates and prevents from great changes n assets. Accordng to the conclusons, the followng recommendatons are offered: The relatonshp between stock market ndex and exchange rate ndex s helpful for asset owners n optmal allocatng of portfolos or nvestors who are tryng to Hedage n front of external rsks. Usng the nput and output of captal, f possble, the nvestment n countres have a lower rsk than Iran to be consdered as an alternatve to stock market n Iran. Due to varous results n dfferent tme perods, t s recommended that the drecton of causalty relatonshps and the drecton of long and short term relatonshps to be studed n shorter tme perods and usng daly data to conduct the latest poltcal recommendatons. In order to address the poor performance of stock exchange n ts man task that s absorbng lqudty and contrbutng n economc growth, t s recommended that n addton to economc factors affectng the stock market actvty, the noneconomc varables also be examned to develop the opportunty for growth and prosperty of Tehran Stock Exchange. For empowerng the stock market prce ndex to descrbe macroeconomc condtons, fnally, t s suggested that some actons to be conducted for greater transparency of ths market n management level and n mechansms of market nformaton. REFERENCES [1]. Hsao, C. (1 981), "Autoregressve modelng and money ncome causalty detecton," Journal of Monetary Economcs, bp []. Kao, W., Ma, C. (1990), "On Exchange rate changes and Stock Prce Reacton," Journal of Busness, Fnance and Accountng, Vol3, No 17, pp [3]. Khan, M.Y. (009), "Stock Market n Work,' to be pcsented at 6 th Annual Conference of lslamc Economcs & Fnance, Jakarta. [4]. Kng, Robert. And Levne, Ross (1993), "Fnance and Growth: Schumpeter Mght Be Rght," Quarterly Journal of Economcs, Vol. 108, No.3, pp [5]. Macknnon, J. G., Haug, A. and Mchels, L. (1999), "Numercal Dstrbuton Functon of Lkelhood Rato Testes For Co ntegrauon, "Journal of Appled Econmetrcs, Vol.14, pp [6]. Panetta, F. (00), "The Stablty of the Relaton Between the Stock Market and Macroecnomc Forces, Economc Notes by Banca Monta de Pasch," Vol.31, No.3, pp [7]. Perron, P. (1989), "The Great Crash, the Ol Prce Shock and the Unt Root Hypothess," Econometrcal, Vol.57. [8]. Phllps, P.C.B. and Perron, P. (1988), "Testng for a Unt Root n Tme Seres Regresson, Bometrcal," Vol.75, pp [9]. Ross, S.A. (1976), "The Arbtrage Theory of Captal Asset Prcng," Journal of Economc Theory, Vol.13, pp [10]. Sadoesky, P.(1999), "Ol prce shocks and market actvty," Energy Economcs, Vol.1, No.5. [11]. Seddgh, HR, Lawler, KA, Katos, A.V. (000), "Econometrcs: A Parctcal approach," London, New York. [1]. Sharp, Wllam, F., Gordon, Alexander and Jeffry, Baley (1999), "Investments," 6rd edton, PrentcsHall Unversty. Internatonal Journal of Research n Busness Studes and Management 18