Asian Economic and Financial Review

Transcription

1 Asian Economic and Financial Review, 014, 4(5): Asian Economic and Financial Review journal homepage: hp://aessweb.com/journal-deail.php?id=500 THE IMPACTS OF INFLATION DYNAMICS AND GLOBAL FINANCIAL CRISES ON STOCK MARKET RETURNS AND VOLATILITY: EVIDENCE FROM NIGERIA C.G Amaefula Deparmen of Saisics, Federal Universiy of Technology Owerri, Imo Sae, Nigeria B.K Asare Deparmen of Mahemaics, UsmanuDanfodiyo Universiy Sokoo, Sokoo Sae, Nigeria ABSTRACT The paper invesigaes he impacs of inflaion dynamics and global financial crises on Sock marke reurns and volailiy in Nigeria. The daa ses on monhly All Shares Index Prices of NSE, and consumers price index (CPI) cover he period of January, 1985 o December, 010. The GARCH (1, 1) model wih mulivariae regressors were adoped and he resul shows ha in he condiional mean equaion; inflaion exers insignifican posiive impac on sock marke reurns, inflaion volailiy exers significan posiive impac on sock marke reurns and during he global financial crises, inflaion exers significan negaive effec on sock marke reurns. In he condiional variance equaion, boh inflaion and is volailiy have negaive effecs on he volailiy of sock marke reurns, hough significan for inflaion and insignifican for inflaion volailiy. And during he global financial crises, inflaion has significan posiive impac on he condiional variance of sock marke reurns. The resul implies ha sock marke reurns can serve as a good hedge agains volaile inflaion, bu his hedging propensiy ends o be illusive if exernal shocks like he global financial crises affec he sock marke condiion negaively. However, i becomes imperaive for invesors o pay aenion o variaion of inflaionary changes when predicing sock marke reurns and is vulnerabiliy. Keywords: Inflaion, Global financial crises, Sock Reurns, Volailiy, GARCH, Mulivariae regressors. 1. INTRODUCTION The relaionship beween macroeconomic indicaors and sock marke reurns has been invesigaed for over wo decades now. Bu he recen global financial crises ha affeced many financial insiuions across he globe has surged he need for more inensive research for beer macroeconomic policy ha engenders economic and financial sabiliy via empirical findings across he globe. According o Sanusi Lamido Sanusi (010), he world economy was hi by an 641

2 Asian Economic and Financial Review, 014, 4(5): unprecedened financial and economic crisis in , ipped ino recession by he sub-prime crisis in he US in Augus 007. This crisis led o he collapse of many world renowned financial insiuions and even caused an enire naion o be rendered bankrup. In Nigeria, he economy falered and he banking sysem experienced a crisis in 009, riggered by global evens. The sock marke collapsed by 70% in and many Nigerian banks had o be rescued. And here is no doub ha he global financial crises have necessiaed sudden changes in macroeconomic policies of he governmen in he recen imes in order o preven economic recession. This sudy differs significanly from pas sudies on he relaed subjec maer in many ways; firsly, i examines he impacs of inflaion rae and is volailiy on sock reurns and sock reurns volailiy. Here, volailiy of inflaion is measured using he condiional variance from GARCH (1, 1) model insead of using mere changes in inflaion rae as assumed by some pas researchers. There are wo jusificaions for his; (i) volailiy of any variable per say, canno be negaive unlike firs log differencing in he variable, since miss-specificaion of volailiy could resul o spurious, unreliable esimaes and misleading inerpreaions and (ii) financial daa exhibis nonsaionary characerisics and o achieve saionariy, firs log differencing in mos cases is used, his should no in any way be inerpreed as volailiy of such variable. Secondly, he sudy will also look a he impac of inflaion rae on sock reurn during he period of global financial crises. Thirdly, he sudy will employ a mulivariae regressor GARCH specificaion which will help in checking wheher here exiss direc volailiy spillover effec. The res of he paper is organized as follows; secion presens he lieraure review, secion 3 deals wih mehodology and variable definiion, secion 4 describes daa analysis and resuls and secion 5 presens he conclusion and policy implicaions.. LITERATURE REVIEW Many lieraure reviews have evolved on he relaionship beween inflaion and sock reurns. Some of hese papers are heoreically based on classical economic hypoheses while ohers are based on empirical findings.engle (004), Engle and Rangel (005), Rizwan (007) esablished a srong predicive power of inflaion on sock marke volailiy and reurns. Habibullah e al. (009) examined he impac of inflaion and oupu growh on sock marke reurns and volailiy in seleced Asian counries, namely India, Japan, Korea, Malaysia and Philippines. Using monhly daa from 1991 o 004 and by employing GARCH (1, 1) model, hey found ha macroeconomic volailiy, which is measured by movemen in inflaion and oupu growh; have a weak predicive power for sock marke reurns and volailiy in hese counries. Fama (1981) hypohesized ha he negaive relaion beween real socks reurns and inflaion observed during he pos-1953 period were he consequence of proxy effecs. Sock reurns are deermined by forecass of more relevan real variables, and negaive sock reurn-inflaion relaions are induced by negaive relaions beween inflaion and real aciviy. However, Firh (1979) for UK and Adam and Tweneboah (008) for Ghana repored a significan posiive relaionship beween inflaion (CPI) and sock reurns. These resuls provide a 64

3 Asian Economic and Financial Review, 014, 4(5): sharp conras o empirical works ha have found a significan negaive relaionship beween sock reurns and expeced inflaion. Kononikas and Ioannidis (005) show ha an inflaion argeing regime wih srong ineres rae reacion o inflaion should lead o lower sock marke volailiy. Gjerde and Sæem (1999) uilized mulivariae VAR approach on he Norwegian monhly daa from heir resul showed ha inflaion has a negaive effec on sock reurns, a senimen which was shared by Spyrou (001) who examined he emerging economy of Greece, during he 1990s. On he conrary, Choudhry (001) in his sudy on four high inflaion counries, Argenina, Chile, Mexico and Venezuela for he sample period from 1981:1 o 1998:6 provided evidence of a posiive relaionship beween curren sock marke reurns and curren inflaion. This resul confirms ha sock reurns ac as hedge agains inflaion. Saunders and Tress (1981) indicae ha Ausralian nominal sock reurns and inflaion are relaed in a significanly negaive fashion, implying ha socks are exremely poor inflaionary hedges for he invesor. In addiion, he sudy indicaes a mainly unidirecional relaionship beween inflaion and sock reurns, wih price level charges leading he equiy index in ime. Aliyu (011) used GARCH model in assessing he impac of inflaion on sock marke reurns and volailiy using monhly ime series daa from he wo Wes African counries, ha is, Nigeria and Ghana. He also invesigaed he impac of asymmeric shocks using he QGARCH model developed by Senana (1995), in boh counries. His daa covers he period of 1998M1 o 010M5 for Nigeria and 1999M1 o 010M5 for Ghana. Resuls for Nigeria show weak suppor for he hypohesis ha bad news exer more adverse effec on sock marke volailiy han good news of he same magniude, while a srong opposie case holds for Ghana. Furhermore, inflaion rae and is hree monh average were found o have significan effec on sock marke volailiy in he wo counries. Douglason (01) sudied relaionship beween inflaion and sock marke reurns in Nigeria using monhly and quarerly daa from Applying a simple regression model and esing for Fisher s effec, his findings seem o sugges ha sock marke reurns may provide an effecive hedge agains inflaion in Nigeria. Taofik and Omosola (013) examined he long run relaionship and dynamic ineracion beween sock reurns and inflaion in Nigeria using monhly All Shares Price Index from NSE and Consumer Price Index from Applying he Auoregressive Disribued Lag (ARDL) model heir resul showed evidence of he exisence of long-run relaionship beween sock reurns and inflaion. The shor-run dynamic model also revealed ha he speed of convergence o equilibrium is moderae, implying ha here is a shor-run relaionship beween sock reurns and inflaion. 3. METHODOLOGY AND VARIABLE DEFINITION The daa ses consis of monhly All Shares Index prices of he NSE marke and consumers price index (CPI) were all obained hrough ranscripion from he published (Cenral Bank Nigerian, 010) and he daa ses cover he period of 1985 M1-010 M1. Sock reurns and 643

4 Asian Economic and Financial Review, 014, 4(5): inflaion rae changes were defined as follows; sock reurns is given as sr 100 (log SPI log SPI 1) and inflaion rae is given as ifr 100 (log CPI log CPI 1). The simple Generalized Auoregressive Condiional Heeroscedasiciy GARCH (1, 1) model inroduced by Bollerslev (1986) was used o measure he volailiy of inflaion rae which of he form: ifr (1) () 1 1 Where ~ GED(v,0, ), The ail parameer v > 0 and he Generalized Error Disribuion (GED) is a normal disribuion if v and fa-ailed if v <, 0, 1 and ( ) 1 shows ha he model is covariance saionary. And 1 and are he lagged squared error 1 erm and he lagged condiional variance respecively Model Specificaion This expresses he impac of inflaion rae, is volailiy and inflaion rae dummy (which represens he impac of inflaion rae during he period of global financial crisis) on sock reurns and volailiy and i is given as sr c0 c1ifr c ifr, c3difr, (3) ifr 1 ifr, 3D ifr, (4) where c 0, c1, c, c3,,, 1, and 3 are consan parameers. ifr, is he inflaion volailiy. The consans 1, and 3 are he impacs of inflaion rae, inflaion rae volailiy and inflaion dummy respecively on he sock reurns volailiy. Their values can be posiive or negaive. If he value of is significan hen, here exis volailiy ransmission from inflaion o he sock reurns. Noe ha ~ GED(0, ). 4. DATA ANALYSIS AND DISCUSSION OF RESULTS In Figure 1, sock reurns changes showed concenraion around zero. I appears ha more of his concenraion occurs much more from 000 o 010 han from 1985 o Also more 644

5 Asian Economic and Financial Review, 014, 4(5): oulying observaions occur from 000 o 010 han any oher ime wihin he period under review. Specifically in 009, sock reurns changes occur mos accouning for he larges oulying observaion in he process. Figure shows he plo of inflaion rae changes. The plo exhibi an evidence of large changes clusered in beween 1987 and This period shows larges oulying observaions in he series. While moderae changes wih concenraion around zero is observable beween he periods of 1996 o 010. An examinaion of he summary saisics in Table 1 clearly shows ha he average monhly reurns are posiive. And from he monhly sandard deviaion, i is apparen ha sr is more volaile han ifr. The saisics show excess kurosis, since he kurosis of all he variables exceeds he normal value of 3, and he skewness is posiive for inflaion rae and negaive for sock reurns. These show fa ails and sharper peaks han he normal disribuion. This lepokuric 645ehavior is no capured by an ARCH process wih a normal disribuion. The Jarque-Bera normaliy es confirms he non normaliy of he disribuions of all he variables. Therefore he ARCH model wih generalized error disribuion can be employed o address he excess kurosis. Ellio Rohenberg, and Sock Poin Opimal Ellio e al. (1996) Tes was adoped o check he order of inegraion of he variables. The resul of uni roo es in Table, shows ha he null hypoheses of he exisence of a uni roo in sr and ifr is rejeced a 1%, 5% and 10% respecively. Hence, he variables are saionary. The lag orders used in he ERS Tes were suggesed by he model selecion crieria. Table 3 above gives he parameer esimaes of inflaion volailiy. The volailiy of inflaion rae can be explained by approximaely 10% and 85% of he news abou inflaion volailiy in he pas period and he previous period volailiy of inflaion respecively. Boh values of lagged squared error erm and lagged condiional variance are significan a 1% level. The volailiy of inflaion is porrayed in Figure 3. The plo showed evidence of volailiy clusering beween 1988 and Wihin his period he volailiy of inflaion rae exhibied a highes peak in And from 1997 o 010, he volailiy of inflaion rae has mainained a consisen low level wih concenraion around zero. The resul in Table 4 shows ha in he condiional mean equaion, inflaion and is volailiy all exer posiive impacs on sock marke volailiy, significan for inflaion volailiy and insignifican for inflaion rae. And during he global financial crises, inflaion has significan negaive impac on sock marke reurns. This resul indicaes ha inflaion has weak predicive impac on sock marke reurns bu inflaion volailiy exers srong predicive impac of sock marke reurns. This also implies ha sock marke reurns can serve as a good hedge agains volaile inflaion. Bu his hedge ends o be illusive if exernal shock like he global financial crises affecs he condiion of he sock marke. In he condiional variance equaion, he resul shows ha boh he lagged 645

6 Asian Economic and Financial Review, 014, 4(5): squared error erm and lagged condiional variance are posiive and well specified hough no significan. Inflaion and is volailiy all have negaive impac on he condiional variance of sock reurns, significan for inflaion and insignifican for inflaion volailiy. Also, during he period of global financial crises, inflaion exers significan posiive impac on he condiional variance of sock reurns. This indicaes ha inflaion has srong predicive impac on sock marke reurns. This resul opines wih ha of Aliyu (011) for Nigeria. The diagnosic es of ARCH LM resul shows ha here is no ARCH in he squared residuals up o 15h lag; Ljung-Box Q-saisics show ha here is no serial correlaion in he squared residuals up o 15h lag. Comparing he values of LJB, skewness and kurosis ( , and ) in Table1 wih heir values ( , and ) in Table3 respecively, shows ha he laer is smaller. This indicaes ha he model is fairly specified and adequae. The Wald es in Table 3 suggess ha he model is covariance saionary. 5. CONCLUSION AND POLICY IMPLICATIONS The paper examines he impacs of inflaion dynamics and global financial crises on sock marke reurns and volailiy in Nigeria. The simple GARCH (1, 1) model specificaion wih mulivariae regressors was adoped. And he resul shows ha in he condiional mean equaion, inflaion and is volailiy all exer posiive impacs on sock marke volailiy, and i is significan for inflaion volailiy and insignifican for inflaion rae. And during he global financial crises, inflaion has significan negaive impac on sock marke reurns. This resul indicaes ha inflaion has weak predicive impac on sock marke reurns bu inflaion volailiy exers srong predicive impac of sock marke reurns. In he condiional variance equaion, he resul shows ha inflaion and is volailiy all have negaive impacs on he condiional variance of sock reurns; i is significan for inflaion and insignifican for inflaion volailiy. Also, during he period of global financial crises, inflaion exers significan posiive impac on he condiional variance of sock reurns. This indicaes ha inflaion has srong predicive impac on sock marke reurns. The resul implies ha sock marke reurns can serve as a good hedge agains volaile inflaion, bu his hedging propensiy ends o be illusive if exernal shocks like he global financial crises affec he condiion of he sock marke negaively. Also, he resul implies ha variaion in inflaionary pressure over a period of ime increases he risk of invesmen and his forces sock reurn o decrease. And if exernal shocks increases, he average sock marke reurns ends o fall via inflaionary changes. However, i becomes imperaive for invesors o pay aenion o variaion of inflaionary changes when predicing sock marke reurns and is vulnerabiliy. REFERENCES Adam, A.M. and G. Tweneboah, 008. Macroeconomic facors and sock marke movemen: Evidence from Ghana. MPRA Paper No (008), (008), posed 17. March 01 / 08:34. Available from hp://mpra.ub.uni-muenchen.de/1156/. 646

7 Asian Economic and Financial Review, 014, 4(5): Aliyu, S.U.R., 011. Does inflaion has an impac on sock reurns and volailiy? Evidence from Nigeria and Ghana. Inernaional Conference on Economics and Finance Research IPEDR (011), Singapore. 4. Bollerslev, T., Generalized auoregressive condional heeroscedasiciy. Journal of Economerics, 31(3): Cenral Bank Nigerian, 010. Saisical bullein, Websie: Choudhry, T., 001. Inflaion and raes of reurns on socks: Evidence from high inflaion counries. Journal of Inernaional Financial Markes, Insiuions and Money, 11(1): Douglason, G.O., 01. Relaionship beween inflaion and sock marke reurns: Evidence from Nigeria. Journal of Applied Saisics, 1(1): Ellio, G., J.R. Thomas and H.S. James, Efficien es for an auoregressive uni roo. Economerica, 64(4): Engle, R.F., 004. Risk and volailiy: Economeric models and financial pracice. American Economic Review, 94(3): Engle, R.F. and J.G. Rangel, 005. The Spline Garch model for uncondiional volailiy and is global macroeconomic causes, Mimeo, Presened a he World Congress of he Economeric Sociey, London. Fama, E.F., Sock reurns, real aciviy, inflaion, and money. American Economic Review, 71(4): Firh, M., The relaionship beween sock marke reurns and raes of inflaion. The Journal of Finance, 34(3): Gjerde, O. and F. Sæem, Causal relaion among sock reurns and macroeconomic variables in a small open economy. Journal of Inernaional Financial Markes, Insiuions and Money, 9(1): Habibullah, M.S., A.H. Baharom and Fong, 009. Predicive conen of oupu and inflaion for sock reurns and volailiy: Evidence from seleced Asian counries. MPRA Paper No , posed 6 January 01 / 07:14. Available from hp://mpra.ub.uni-muenchen.de/14114/. Kononikas, A. and C. Ioannidis, 005. Should moneary policy respond o asse price misalignmens? Economic Modelling, (6): Rizwan, M.F. and S.U. Khan, 007. Sock reurn volailiy in emerging equiy marke (kse): The relaive effecs of counry and global facors. Inernaional Review of Business Research Papers, 3(): Sanusi, L.S., 010. The Nigerian banking indusry: Wha wen wrong and he way forward. A convocaion lecure delivered a he Bayero Universiy, Kano, Nigeria. Saunders, A. and R.B. Tress, Inflaion and sock marke reurns: Some Ausralian evidence. The Economic Record, 57(156): Senana, E., Quadraic ARCH models. Review of Economic Sudies, 6(4): Spyrou, S.I., 001. Sock reurns and inflaion: Evidence from emerging marke. Applied Economics Leers, 8(7): Taofik, M.I. and M.A. Omosola, 013. The relaionship beween sock reurns and inflaion in Nigeria. European Scienific Journal, 9(4):

8 Inflaion rae (DLOG(CPI)) Sock reurns (DLOG(SPI)) Asian Economic and Financial Review, 014, 4(5): BIBLIOGRAPHY Cenral Bank of Nigeria, 010.Annual repor, Figure-1. Plo of monhly Sock reurns (sr) in Nigeria from Years Figure-. Plo of firs difference of he naural log of inflaion rae (ifr) in Nigeria Years 648

9 Asian Economic and Financial Review, 014, 4(5): Table-1.Summary Saisics Variables Δ log SPI (sr ) Δ log CPI (ifr ) Mean Sd.Dev Skewness Kurosis Jarque-Bera Prob Sum Sq. Dev Observaion Variable Table-.Uni Roo Tes using Ellio Rohenberg, and Sock Poin Opimal (ERS) sr ifr Deerminisic Terms C C, C C, Lags Tes criical values value 1% 5% 10% Remarks Saionary Saionary Saionary Saionary Table-3.GARCH (1, 1) esimaes of inflaion volailiy and Diagnosic es Saisics Coefficien z-saisic Prob. Remarks C 1.94E No significan Variance Equaion C 4.44E % significan RESID(-1)^ No significan GARCH(-1) % significan Diagnosic es ARCH(LM 15lag) No ARCH in he squared residuals up o 15 h lag Ljung-Box Q-saisics No serial correlaion in he squared residuals up o 15 h lag Wald es (α+β)< Covariance saionary Mehod: ML - ARCH (Marquard) - Generalized error disribuion (GED). Sample (adjused): 1985M0 010M1, Included observaions: 311 afer adjusmens. Convergence achieved afer 40 1eraion. 649

10 Inflaion rae(dlog CPI) volailiy Asian Economic and Financial Review, 014, 4(5): Figure-3.Plo inflaion rae volailiy using he condiional variance from GARCH (1, 1) model in Table Years Table -4.Impacs analysis of inflaion rae and inflaion rae volailiy on sock reurns and sock reurns volailiy using GARCH (1, 1) specificaion and Diagnosic es. Saisics Coefficiens z-value Prob. Condiional mean equaion C * ifr ifr, * D ifr, * Condiional Variance Equaion ** ifr * ifr, D ifr, ** Diagnosic Tes ARCH(LM, 15 lag) Ljung-Box Q-saisics LJB Skewness Kurosis Wald es(α+β)< Mehod: ML ARCH (Marquard) - Generalized error disribuion (GED). Sample (adjused): 1985M0 010M1. Convergence achieved afer 6 ieraions and he symbols (*) and (**) indicae 1% and 5% significan. 650