Asian Journal of Empirical Research

Transcription

1 . Asian Journal of Epirical Research journal hoepage: hp://aessweb.co/journal-deail.php?id=5004 REAL MACROECONOMIC VARIABLES AND STOCK PRICES: EVIDENCE FROM TURKEY H. Aydın OKUYAN 1 ABSTRACT This paper invesigaes he relaionship beween reel acroeconoic variables and sock prices in Turkey. Consupion expendiures, indusrial producion index, eployen level and fixed invesens are used as indicaors of real econoic aciviy and consupion price index as an indicaor of inflaion. The ARDL bounds esing is applied o he long-run relaionship beween he variables. Our epirical evidence reveals he validaion of Proxy hypohesis in case of Turkey. Keywords: Proxy hypohesis, Sock prices, Real acroeconoic variables, Turkey. INTRODUCTION The relaionship beween he real acroeconoic variables and he sock prices has been invesigaed by any sudies in exiing finance lieraure. Unil he Proxy hypohesis- whose basis was buil by Faa and Schwer, (1977) and developed by Faa (1981), i has been claied ha an increase in he real econoic aciviy causes a rise in inflaion and also declines sock prices due o he negaive relaionship beween inflaion and sock prices. However, Proxy * hypohesis developed by Faa, (1981) reveals ha here is a posiive relaionship beween sock prices and real econoic aciviy. In conras o previous approaches, Proxy hypohesis says ha an increase in he real econoic aciviies causes no an incline bu decline in inflaion [Faa and Schwer, (1977); Schwer, (1981)]. The decline in inflaion resuls in an increase in sock prices due o he negaive relaionship beween he.price of a sock ay be defined as discouned value of is fuure cash flows. A decline in expeced cash flows or an increase in he real ineres rae (discoun facor) causes a decrease in sock prices. Cash flows of he firs ove according o he real 1 Deparen of Business Adinisraion, Balikesir Universiy Bandira FEAS Turkey E-ail: aokuyan@balikesir.edu.r * Proxy is defined as agen, represenaive in he dicionary. Faa uses he real sock reurns as an agen, represenaive of real econoic aciviies. In oher words, real reurns of socks are he represenaive indicaors of real econoic aciviies. 654

2 econoic aciviy. In he periods of rise in real econoic aciviy, expeced cash flows and herefore price of socks ay rise. The rise in he econoic aciviy also increases inflaion. In case of inflaion, he auhoriies usually ipleen ani-inflaionis, conracionary policies o slow down he deand. These policies increase he real ineres rae; decrease he expeced cash flows of he firs which also ean a decrease in sock prices. As a resul, he iniial increase in he sock prices will have been balanced afer he ipleenaion of acroeconoic policies. In conclusion, he effec of a acroeconoic variable on sock prices depends on he relaionship beween fuure cash flows and inflaion. A variable which has a negaive effec on cash flows and/or a posiive effec on inflaion will be in a negaive relaionship wih he sock prices. Friedan, (1977) argued ha inflaion decreases he producion efficiency. Reducion in efficiency slows down he real econoic aciviy and he cash flows of he firs. Malkiel, (1979) saed ha flucuaions in he inflaion increase he risk of owning a sock and herefore he found a negaive relaionship beween inflaion and sock prices. Cukieran, (1983) iplied ha here is a posiive relaionship beween anicipaed inflaion, flucuaions in inflaion, unanicipaed inflaion and acual inflaion. Proxy hypohesis has been esed in any epirical sudies since Faa, (1981). Geske and Roll, (1983) Kaul, (1987) and Barnes e al. (1999) found evidence supporing his hypohesis in heir sudies. Mc Queen and Roley, (1993) claied ha he negaive relaionship beween real econoic aciviies and sock prices appears only under soe specific circusances. They saed ha news regarding an incline in econoic aciviies reduces he sock prices in booing econoies, bu increases sock prices in weak econoies. Park, (1997) approached o he subjec in a differen aspec in his sudy regarding USA. He enioned abou a negaive relaionship beween he econoic variables ha igh be he indicaor of fuure inflaion and sock prices. Park, (1997) exposed ha eployen level is he os iporan variable which igh be he indicaor of fuure inflaion. He found a negaive relaionship wih sock prices. Merikas and Merika, (006) exained he relaionships beween he socks and real acro econoic variables in Gerany in period. They eployed he VAR odel and used eployen growh, oupu growh, invesen, indusrial producion and reail sales as real acroeconoic variables. In heir sudy, hey claied ha he real econoic variables which are in a posiive relaionship wih inflaion would be in a negaive relaionship wih sock prices. They found a posiive relaionship beween sock prices and eployen level on he one hand, and a negaive relaionship beween he sock prices and econoic growh which is in a posiive relaionship wih inflaion on he oher hand. According o his resul, hey concluded ha here is a posiive relaionship beween real econoic variables and sock prices, where as here is a negaive relaionship beween real econoic variables which igh be he indicaor of fuure inflaion and which are also in a posiive relaionship wih inflaion- and sock prices. Ki and In (005) invesigaed he relaionship beween inflaion and noinal and real reurns of socks in USA. As a resul, hey deeced a posiive relaionship beween inflaion and reurns of socks in shor (1 onh) and long (18 655

3 onhs) ers and negaive relaionship in he ediu er. They reached o conclusion ha he relaionship beween hese wo variables depends on he lengh of he er and change in ie. Adrangi e al. (1999) invesigaed he validiy of Proxy hypohesis in Peru and Chile. They saed ha inflaionis pressure decreases he fuure cash flows of he firs and decreases he curren values of fuure cash flows by increasing he noinal discoun raes. They found a negaive relaionship beween inflaion and real sock reurns and concluded ha Proxy hypohesis is valid in hese counries. Besides he views supporing Proxy hypohesis, here are also couner views in he lieraure. McCarhy e al. (1990) invesigaed he relaionship beween he expeced real reurns of socks and expeced real econoic aciviy in England, USA and Gerany and could no find a significan relaionship. Cochran (1993) for USA ha Proxy hypohesis is no valid. Caporale and Jung (1997) esed he Proxy hypohesis for USA and deeced -in conras o Proxy hypohesis- a posiive relaionship beween real reurns of socks and inflaion. I is seen ha conradicory resuls are obained in he lieraure and a consensus is no reached abou he validiy of he proxy hypohesis. The ai of his sudy is o es he relaionship beween he real acroeconoic variables and sock prices in Turkey under he frae of Proxy hypohesis. In order o es his relaionship, bounds esing approach developed by Paseran e al. (001) is applied. This sudy consiss of four secions. Firs secion is he inroducion. Model, he daa se and ehodology are inroduced in he second secion. The long-run relaionship beween he sock prices and acroeconoic variables is invesigaed by he help of co inegraion ehod. Third secion explains he epirical resuls and akes an evaluaion. And secion four concludes. MODEL, DATA SET AND METHODOLOGY The odel which was developed by Merikas and Merika (006) will be used in presen sudy using Turkish daa. The epirical equaion of odel is coposed as follows: LSP LCE LIPI LEL LFI LCPI (1) In he equaion, SP, CE, IPI, EL, FI and CPI refer o sock prices, consupion expendiures, indusrial producion index, eployen level, fixed invesen and consuer price index respecively. ISE 100 index is used as he indicaor of sock prices. Quarerly daa is used covering he period of 1987Q1-01Q9 and i is obained fro he Cenral Bank of Republic of Turkey selecronic daa disribuion syse. Naural logarihs of he variables have been aken and all he variables are used in heir logarihic for. L in he odel iplies ha he variables are logarihic. 656

4 Since facing a spurious regression proble aong hese series which include a uni roo, soe ehods are suggesed o solve his proble. One of he is aking he differences of he series and hen puing he ino regressions. However, in his case we are confroned wih a new proble. This ehod leads o he loss of inforaion ha is iporan for he long-run equilibriu. As long as he firs differences of he variables are used, deerining a poenial long run relaionship beween hese variables becoes ipossible. These all issues are covered by bounds esing approach developed by Pesaran e al. (001). The bounds esing apporach is superior o all oher radiional coinegaion aparoches on wo grounds. Firsly, i can be applied irrespecive of wheher he regressors are I(0), I(1), or even inegraed of he sae order.secondly, Bounds esing procedure is robus for coinegraion analysis wih sall saple sudy. EMPIRICAL RESULTS AND EVALUATION Co-inegraion ess In order o apply he Bounds esing approach, firsly an unresriced error correcion odel (UECM) is fored. The for of his odel adaped ino our sudy is as follows: LSP 0 1i i i i 3i i 4i i 5i i i1 i0 i0 i0 i0 () i0 LCPI 6i LSP i LSP 7 LCE 1 LCE 8 1 LIPI LIPI 9 1 LEL 10 LEL 1 LFI 11 LFI 1 LCPI 1 1 F-es is applied on firs period lags of dependen and independen variables o es he exisence of he co-inegraion relaionship. Basic hypohesis for his es is esablished as (H 0 :α 7 =α 8 =α 9 =α 10 =α 11 =α 1 =0). The calculaed F-saisic is copared wih boo and upper criical value generaed by Pesaran e al. (001). If he calculaed F-saisic is lower han lower boo criical value hen here is no co-inegraion relaionship beween he series. If he calculaed F-saisic is beween he lower and upper criical values, no exac opinion can be ade and here is a need o apply oher co-inegraion es approaches. Lasly, if he calculaed F-saisic is higher han he upper criical value, here is a co-inegraion relaionship beween he series. Afer a co-inegraion relaionship is observed beween he series, Auoregressive Disribuion Lag (ARDL) odels are esablished o deerine he long er and shor er relaionships. In UECM odels, represens nuber of lags. Criical values like Akaike, Schwarz and Hannan-Quinn are used o deerine he nuber of lags. Also he duraion of he lag which provides he salles criical value is idenified as he odel s duraion of lag. However, if he odel esablished wih he duraion of lag in which he seleced criical value is he salles involves an auocorrelaion, duraion of lag, which gives he second 657

5 salles criical value is aken. If he auocorrelaion proble sill coninues, his process is susained unil he proble is solved. In his sudy, axiu duraion of lag has been aken as 8. The nuber of lags which iniize he Akaike inforaion crierion is found as 8, bu afer applying he LM es, i has been deeced ha 8 lags cause an auocorrelaion proble. In ha case, he nuber of lags is deerined as 7, which saisfies he second iniu value of Akaike inforaion crierion. Afer perforing an LM es, again an auocorrelaion is deeced.then he hird value iniizing he Akaike crierion which is 3 is deerined as he nuber of lags. No auocorrelaion proble has been observed afer LM es. Crieria and es values are given in Table- 1. Table 1: Lag lengh selecing M AIC X BG * * BG x : Breusch-Godfrey auocorrelaion es. * shows 1% significance level iplying auocorrelaion beween error ers. Table : Resuls of bound es Criical values (significance % 5) K F saisics Lower bound Upper bound K: nuber of independen variables in equaion (). Criical values were obained fro able CI (iii) Pesaran a al. (001) Afer he nuber of lags is deerined, co-inegraion beween series is invesigaed by he bounds esing approach. F-saisics calculaed wih UECM odel are copared wih he lower and upper criical levels by Pesaran e al. (001). The bound es resuls are given in Table-.As i is seen fro Table, a co-inegraion relaionship has been deeced beween he series, because F-saisic exceeds he upper bounds value of Pesaran e al. (001). Tha a co-inegraion relaionship has been deeced beween he series, Auoregressive Disribuion Lag (ARDL) odel can be esablished o deerine long-er and shor-er relaionships. As can be seen fro Table-3, none of he econoeric probles such as auocorrelaion, heeroscedasiciy, conflic o noral disribuion and odel specificaion error has been observed.observing he long er coefficiens in Table-4, a posiive long er relaionship has 658

6 been deeced beween sock prices and saisically significan independen variables which are indusrial price index, eployen level and fixed invesens where as a negaive relaionship is seen beween he sock prices and inflaion. In addiion, a significan relaionship beween he consupion expendiures and sock prices is no e.the error correcion variable ECT(-1) has been found negaive and also saisically significan (see Table-5). Table 3: ARDL (4, 0, 1, 4, 1, 4) esiaion resuls Variables Coefficien T saisics LSP(-1) * LSP(-) LSP(-3) LSP(-4) ** LCE LIPI LIPI(-1) LEL LEL(-1) LEL(-) LEL(-3) LEL(-4) ** LFI LFI(-1) ** LCPI ** LCPI(-1) * LCPI(-) LCPI(-3) ** LCPI(-4) C Resuls of Diagnosic Tess R R 0.77 X BG 5.074[0.375] NORM ().304[0.33] WHITE (1) 0.054[0.14] X RAMSEY (1) 0.738[0.336] BG, NORM,, RAMSEY areauocorrelaion, noraliy, heeroscedasiciy, odel specificaion WHITE error es saisics, respecively. 659

7 Table 4: ARDL (4, 0, 1, 4, 1, 4) long-er coefficiens Variables Coefficiens T saisics LCE LIPI ** LEL ** LFI ** LCPI ** C ** Noe: * and ** significan a1% and 5% respecively Table 5: Error correcion odel Variables Coefficiens T saisics DLSP(-1) * DLSP(-) * DLSP(-3) * DLCE DLIPI DLEL DLEL(-1) ** DLEL(-) DLEL(-3) ** DLFI DLCPI DLCPI(-1) * DLCPI(-) DLCPI(-3) C ECT(-1) * Noe: * and ** significan a1% and 5% respecively CONCLUSION In his sudy, he Proxy hypohesis developed by Faa, (1981) has been esed for Turkey. According o Faa (1981), here is a posiive relaionship beween he real acroeconoic variables and sock prices. This relaionship eans ha an increase in he real econoic aciviy affecs he inflaion negaively. And because of he negaive relaionship beween he inflaion and sock prices, sock prices will be affeced posiively. Our resuls show a posiive relaionship beween he sock prices and he real acroeconoic variables in long run. Also a significan negaive relaionship has been deeced beween he sock prices and inflaion. Under he ligh of his evidence, i can be said ha Proxy hypohesis developed by Faa, (1981) is valid for Turkey. This paper resuls ha he variables which are he indicaors of real econoic aciviy such as 660

8 indusrial producion index, eployen level and fixed invesens are effecive on sock prices via inflaion. Therefore, hese variables can be used as leading indicaors in esiaing he sock prices. As a resul, i is obvious ha an invesor us ake acroeconoic variables and inflaion ino accoun in esiaing he fuure values of sock prices. REFERENCES Adrangi, B., Charah A. and Shank T. M. (1999). Inflaion, Oupu and Sock Prices: Evidence fro Lain Aerica. Managerial and Decisions Econoics, Vol. 0, pp Barnes, M., Boyd, J. H. and Sih, B. D. (1999). Inflaion and Asse Reurns. European Econoic Review, Vol. 43, pp Caporale, T. and Jung C. (1997). Inflaion and Real Sock Prices. Applied Financial Econoics, Vol. 7, No. 3, pp Cochran, S. J. and Defına R. (1993). Inflaion's Negaive Effecs On Real Sock Prices: New Evidence And A Tes Of The Proxy Effec Hypohesis. Applied Econoics, Vol. 5, No., pp Cukieran, A. (1983). Relaive Price Volailiy and Inflaion: A Survey and Furher Resuls. Carnegie-Rocheser Conference Series on Public Policy, Vol. 19, pp Engel, R. F. and Granger, C. W. J. (1987). Co-inegraion and Error Correcion Represenaion. Esiaion and Tesing. Econoerica, Vol. 55, No., pp Faa E. F. (1981). Sock Reurns, Real Aciviy, Inflaion and Money. The Aerican Econoic Review, Vol. 71, No. 4, pp Faa, E. F. and Gibbons, M. R. (1984). A Coparison of Inflaion Forecass. Journal of Moneary Econoics, Vol. 45, No. 4, pp Faa, E. F. and Schwer G. W. (1977). Asse Reurns and Inflaion. Journal of Financial Econoics, Vol. 5, No., pp Friedan, M. (1977). Nobel Lecure: Inflaion and Uneployen. Journal of Poliical Econoy, Vol. 85, pp Geske, R. and Roll, R. (1983). The Fiscal and Moneary Linkage beween Sock Reurns and Inflaion. The Journal of Finance, Vol. 38, No. 1, pp Johansen, S. (1988). Saisical Analysis of Coinegraion Vecors. Journal of Econoic Dynaics and Conrol, Vol. 1, No. /3, pp Johansen, S. and Juselius, K. (1990). Maxiu Likelihood Esiaion and Inference on Coinegraion-Wih Applicaions o he Deand For Money. Oxford Bullein of Econoics and Saisics, Vol. 5, No., pp Kaul, G. (1987). Sock Reurns and Inflaion: he Role of he Moneary Secor. Journal of Financial Econoics, Vol. 18, No., pp

9 Ki, S. and In F. (005). The Relaionship Beween Sock Reurns ad Inflaion: New Evidence Fro Wavele Analysis. Journal of Epirical Finance, Vol. 1, pp Lee, U. (1998). A Tes of he Proxy Effec Hypoesis: Evidence Fro he Pacific Basin Counries. Quarerly Journal of Business and Econoics, Vol. 37, No. 3, pp Malkiel, B. (1979). The Capial Foraion Proble in he Unied Saes. Journal of Finance, Vol. 34, pp McCarhy, J., Najand, M. and Seifer, B. (1990). Epirical Tess of he Proxy Hypohesis. Financial Review, Vol. 5, No., pp McQueen, G. abd Roley, V. V. (1993). Sock Prices, News and Business Condiions. The Review of Financial Sudies, Vol. 6, No. 3, pp Merikas, A. G. and Merika, A. A. (006). Sock Prices Response o Real Econoic Variables. Managerial Finance, Vol. 3, No. 5, pp Park, S. (1997). Raionaliy of Negaive Sock Price Responses o Srong Econoic Aciviy. Financial Analyss Journal, Vol. 53, No. 5, pp Pesaran, M. H. Shin, and Y. Sih, R. J. (001). Bounds Tesing Approaches o he Analysis of Level Relaionships. Journal of Applied Econoerics, Vol. 16, pp Schwer, G. W. (1981). The Adjusen of Sock Prices o Inforaion Abou Inflaion. Journal of Finance, Vol. 36, No. 1, pp