Monetary Policy and Output-inflation Volatility Interaction in Nigeria: Evidence from Bivariate GARCH-M Model

Transcription

1 Moneary Policy and Oupu-inflaion Volailiy Ineracion in Nigeria: Evidence from Bivariae GARCH-M Model Onyukwu E. Onyukwu (corresponding auhor) Deparmen of Economics, Universiy of Nigeria, Nsukka, Enugu Sae, Nigeria Tel: Emmanuel O. Nwosu Deparmen of Economics, Universiy of Nigeria, Nsukka Enugu Sae, Nigeria Diejomaoh Io African Insiue for Applied Economics 54 Nza Sree, Independence Layou, Enugu, Nigeria Received: Ocober 11 s, 2011 Acceped: Ocober 19 h, 2011 Published: Ocober 30 h, 2011 Absrac This aricle repors on a recen sudy ha applies bivariae GARCH mehodology o invesigae he exisence of a radeoff beween oupu growh and inflaion variabiliy in Nigeria and o ascerain he impac of moneary policy regime changes (from direc conrol regime o indirec or marke based regime) on he naure of he volailiy radeoffs. Invesigaions reveal he exisence of a shor run radeoff relaionship beween oupu growh and inflaion wihin and across boh regimes. However, no srong evidence of long run volailiy relaionship could be esablished. Our resuls furher reveal ha regime changes affeced he magniude of policy effecs on oupu and inflaion. Moneary policy had a sronger effec on oupu growh han on price sabiliy during he period of direc conrol while i has a much larger impac on inflaion during he curren period of marke-based regime. Also volailiy of oupu and inflaion became more persisen during he period of indirec conrol. Keywords: Moneary Policy, Oupu-Inflaion Volailiy, Bivariae GARCH-M Model 46

2 1. Inroducion and Background o he Sudy Over he pas four decades, economiss and policy makers have shown considerable ineres in undersanding causes of macroeconomic volailiy and how o reduce i. Concepually, macroeconomic insabiliy refers o condiions in which he domesic macroeconomic environmen is less predicable. I is of concern because unpredicabiliy hampers resource allocaion decisions, invesmen, and growh. This aricle focuses on he volailiy ineracion of he growh rae of oupu and he levels of inflaion raes. Changes in he behavior of hese endogenous variables usually reflec changes in he macroeconomic policy environmen as well as exernal shocks. Boh Okonjo-Iweala & Phillip (2006) and Balini (2004) have described he Nigerian macroeconomic environmen as one of he mos volaile among emerging markes. Among emerging marke economies, Nigeria exhibis he highes inflaion and exchange rae variabiliy, he lowes oupu volailiy, and an ineres rae volailiy ha is slighly smaller han ha of Souh Africa and much smaller han ha of Brazil, bu slighly larger han ha of Chile (Balini 2004). Nigeria offers unique opporuniy o he sudy of oupu-inflaion volailiy ineracion and is relaionship wih moneary policy. This is due o he fac ha moneary policy conduc in Nigeria has winessed wo alernaive regimes since he mid-1970s, he direc conrol regime and indirec or marke-based regime wih differen relaive weighs aached o oupu growh and price sabiliy objecives. 1.1 Saemen of he Problem This sudy evaluaes he effec of moneary policy regime changes, by esimaing a bivariae GARCH-M model of oupu and, hus examines he naure of oupu-inflaion variabiliy rade-off as well as volailiy persisence which are of major ineres in macroeconomic policy debaes. Our sudy furher ascerains he efficacy of moneary policy regime change from direc o indirec approaches (Noe 1) in reducing macroeconomic volailiy, paricularly in he ligh of he Taylor curve radeoff hypohesis. Thus, our basic research objecives are: To ascerain if here is evidence of oupu-inflaion volailiy rade-off in Nigeria; To invesigae if a change in moneary policy regime affecs he naure of oupu growh-inflaion volailiy radeoff. To ascerain how moneary policy shocks affec inflaion and oupu growh variabiliy dynamics. Wha is remaining of his aricle has been organized ino hree secions. Secion wo is devoed o an overview of relevan heoreical framework and mehodology of analysis. Secion hree presens and discusses he resuls of our analyses. Secion four summarizes he sudy findings and makes some policy recommendaions. 47

3 The period covered by he sudy is 1981 o 2007, essenially due o daa availabiliy. The sudy period cus across wo moneary policy regimes, hus affording us he opporuniy o make sub-sample comparisons. The sudy uses quarerly daa in he analysis in order o capure subsanial variabiliy in he variables of ineres. 2. Mehodology of Sudy On he relaionships beween moneary policy, oupu volailiy, and inflaion volailiy, one heory holds ha he volailiy of oupu and inflaion will be smaller he sronger moneary policy reacs o inflaion han oupu gap (Gaspar & Smes 2002) (Noe 2). In his case i is cusomary o assume ha he economy s social loss funcion is he sum of he variance of inflaion and oupu. If he economy is hi by demand shocks, he cenral bank will never face a rade-off beween oupu sabiliy and inflaion sabiliy, ha is, a moneary policy acion ha reduces oupu volailiy is consisen wih sable inflaion. However, in a more general case when he economy is hi by demand shocks as well as supply shocks, he cenral bank faces an inescapable radeoff beween oupu sabiliy and inflaion sabiliy if i chooses, (as is commonly he case), o minimize he social loss funcion. Thus, moneary policy acion which reduces he variance of inflaion will increase he variance of oupu, and vice versa. These hypoheses have been joinly sudied using bivariae GARCH-M class of models (Founas e al 2002; Grier e al 2004; Lee 2002). According o Lee (2004), he GARCH approach has wo major advanages over he convenional measure of volailiy, such as moving sandard deviaions and squared residual erms in vecor auoregression (VAR) models. The firs advanage is ha condiional volailiy, as compared o uncondiional volailiy, beer represens perceived uncerainy which is of paricular ineres o policy makers. The second advanage is ha he GARCH model offers insighs ino he hypohesized volailiy relaionship in boh he shor run and he long run. Whereas ime-varying condiional variances reveal volailiy dynamics in he shor run, he model also generaes a long run measure of he oupu-inflaion covariance ha will be helpful in evaluaing moneary policy radeoffs. These inform he use of bivariae GARCH model in his sudy. 2.1 Our Empirical Models Following Founas e al. (2002), Grier e al. (2004) and Lee (2002) we use a bivariae GARCH model o simulaneously esimae he condiional variances and covariance of inflaion and oupu growh in order o address our firs and second research quesions. We employ he following bivariae VAR (p) model for esimaing he condiional means of oupu and inflaion: 48

4 Χ where ε such ha H Φ Χ Υ Φ ε 0 1i = Φ + Φ Χ + Φ Y, Φ [ Φ,Φ ] [ y, π ] is a = = = 2j /Ω ~ N(0, H = (h, h is a ) andωis informaion se available up o ime -1 ) + ε 2x1 vecor of consans is a 2 x 1 vecor of real oupu growh y and inflaion rae π vecor of addiional explanaory variables such as inflaion uncerainy, ec are vecors of 2x2 marices of parameers o be esimaed. [ ε, ε ] is a 2x1 vecor of oupu and inflaion innovaions. y 0 yo π p i= 1 π0 1i y i π p 2j j= 1 j The vecor of condiional variances of oupu and inflaion are specified as follows: H ' = C 0 C 0 + A H 1 A+ B ε 1 ε 1 B + D F 1 D whereh = (h y, h ) π is a 2x1vecor of hecondiiona C, A,B,andD are2x2upper riangular maricesof parameers; andfis a vecor of explanaory variables, ha is,f = ( MPR, Oilprices,ec). Moreexplicilyhemariceof parameerscan bespecifiedin upper riangular formas: C 0 c 11 c 12 a 11 a 12 b 11 b ; B 12 δ ; A ; andd 11 δ 12 = = = =. 0 c 22 a 21 a 22 b 21 b 22 δ 21 δ 22 l variancesof oupu andinflaion,. 2.2 Economic Meaning of Coefficiens and Apriori Expecaions The diagonal elemens in marix C o represen he means of condiional variances of oupu growh and inflaion, while he off diagonal elemen represens heir covariance. The Parameers in marix A depic he exens o which he curren levels of condiional variances are correlaed wih heir pas levels. In specific erms, he diagonal elemens (a 11 and a 22 ) reflec he levels of persisence in he condiional variances; a 12 capures he exen o which he condiional variance of oupu is correlaed wih he lagged condiional variance of inflaion. For he exisence of oupu-inflaion volailiy rade-off, he variable is expeced o have negaive sign and be saisically significan. The parameers in marix B reveal he exens o which he condiional variances of inflaion and oupu are correlaed wih pas squared innovaions; b 12 depics how he condiional variance of oupu is correlaed wih he pas innovaion of inflaion. This measures he exisence of cross-effec from an oupu shock o inflaion volailiy. In order o esimae he impac of moneary policy on he condiional variances, we include one period lagged change in he Cenral Bank s moneary policy rae (MPR) (or VAR-based generaed moneary surprises) in he vecor F. The resuling coefficiens in marix D measure he effecs of hese variables on inflaion and oupu volailiy. For moneary policy o have a rade-off on he condiional variances, he diagonal elemens have o alernae in signs. In order o address he hird research quesion we augmen he VAR model specified in (1) by adding 49

5 moneary policy variable in he vecor X and hen compue he generalized impulse responses and generalized variance decomposiions o analyse he shor run dynamic response of oupu growh and inflaion moneary policy shocks/innovaions. The generalized variance decomposiion and impulse response funcions are unique soluion and invarian o he ordering of he variables in he VAR (Pesaran & Shin 1998). Also, i has been argued, however, ha in he shor run unresriced VARs perform beer han a coinegraing VAR. For example, Naka & Tufe (1997) sudied he performance of VECMs and unresriced VARs for impulse response analysis over he shor-run and found ha he performance of he wo mehods is nearly idenical. We adop unresriced VARs in aemping o answer our hird research quesion because of he shor-erm naure of he variance decomposiion and impulse response analysis. AIC and SBC will be used for lag order selecion. 2.3 Mehod of Esimaion and Daa Sources The models are esimaed by he mehod of Broyden, Flecher, Goldferb and Shanno (BFGS) simplex algorihm. We employed he RATS sofware in he esimaion of he sysem. This is due o he fac ha RATS has an inbuil bivariae GARCH sysem ha suppors simulaneous esimaion and hus is able o implemen differen resricions ha migh be assumed on he variance-covariance srucure of he sysem. The daa used for he esimaions are quarerly daa from he Cenral Bank of Nigeria Saisical Bullein of various years, he Annual Repor and Saemen of Accoun of various years. Moneary Policy Measure: According o Nnanna (2001) he Minimum Rediscoun Rae (MRR) is he nominal anchor, which influences he level and direcion of oher ineres raes in he domesic money marke. Is movemens are generally inended o signal o marke operaors he moneary policy sance of he CBN. Similarly, Agu (2007) observes ha he major policy insrumen for moneary policy in Nigeria is he minimum rediscoun rae (MRR) of he Cenral Bank and noes ha while boh ineres and inflaion raes are high, a worrisome problem in he observed response o hese macroeconomic imbalances is he lack of policy consisency and coherence. This could be on accoun of inadequae informaion on he naure and size of impac of he MRR on key macroeconomic aggregaes. This ype of inconsisency in he conduc of moneary policy is likely o increase raher han sabilize macroeconomic volailiy. Hence, we shall use MRR (MPR) as a measure of moneary policy sance. 3. Presenaion of Resuls and Discussion Table 1 shows he summary saisics of he variables used in he sudy. Inflaion and GDP growh raes are calculaed as annualized quarerly growh raes of consumer price index (CPI) and real GDP respecively. As he able indicaes, average nominal GDP was almos wo imes greaer in period compared o period. Bu, he consumer price index for he period, , was almos weny fold of ha in he period, For his reason average real GDP for he period 1995 o 2007 was lower han ha of However, he average growh rae of real GDP was higher in he second period compared o he firs; while average inflaion rae for he second period was lower compared o he firs period. This shows 50

6 ha he economy performed beer on he average under marke-based moneary regime compared o he conrolled regime. The sandard deviaion of inflaion is lower in he second period implying lower uncondiional volailiy and here is no significan change in he uncondiional volailiy of real GDP growh which appears o sugges ha real quarerly GDP flucuaions are modes over he wo periods bu slighly higher in he second period. The minimum rediscoun rae, a measure of moneary policy sance of he Cenral Bank of Nigeria (CBN) was on he average lower during he conrolled regime period compared o he indirec regime period. This suggess ha moneary policy became igher during he period of indirec or marke-based regime aimed specifically o conrol inflaion by reducing he rae of money growh. Oil prices being used o conrol for exogenous shocks in he model had a low average during he period of conrolled regime. A ha ime lower oil prices affeced GDP growh adversely as he economy enered ino a recession ha led o he inroducion he Srucural Adjusmen Programme (SAP), which furher depressed he economy. The adverse supply effec resuling from SAP and lower oil revenue ogeher wih igher conrols on ineres rae helped o pu a srain on he economy. During he period of indirec moneary approach, oil prices began o increase rapidly in he inernaional marke and his resuled in posiive oupu growh for mos of he period and quick recovery of he economy from he adverse effecs of SAP. However, he Cenral Bank has been very cauious wih rising oil prices and as a resul has been seing he minimum rediscoun rae in order o accommodae he adverse effecs of he rise in oil prices on inflaion. Table 2 shows he ess for serial correlaion, auoregressive condiional heeroskedasiciy effec, and normaliy. A series of Ljung-Box (1978) ess for serial correlaion suggess ha here is a significan amoun of serial dependence in he daa. Oupu growh is negaively skewed, and inflaion is posiively skewed and oupu growh failed o saisfy he Jarque-Bera ess for normaliy (Jarque & Bera 1980). The ARCH ess also reveal he presence of firs order serial dependence in he condiional variances of oupu growh and inflaion suggesing ha our applicaion of GARCH (1, 1) model is appropriae o he daa. Valid inference from GARCH model requires ha he variables be saionary, a leas in heir condiional means (Lee 2002). As a resul, uni roo ess were conduced on he variables using he Dickey & Fuller (1981) mehodology. The analysis, however, was evenually based on he augmened Dickey-Fuller uni roo ess. The resuls are presened in able 3 below. The resuls show ha only inflaion rae is level saionary while oher variables are saionary in heir firs differences. Table 4 shows he resuls from GARCH esimaions of he wo sub-samples and he overall sample. The resuls would help o address our concerns in he firs and second research quesions, which are: wheher or no he change in he approach o moneary policy in Nigeria from direc o indirec led o a change in volailiy ineracions, ransmissions and radeoff beween oupu growh and inflaion. The resuls are in wo pars. The firs par shows he esimaions of he condional mean equaion while he second par shows he ime-varying condiional variance equaion which is of paricular ineres o us. The firs par of he condiional mean and condional variance eqauions represens oupu growh while he second vecor denoes inflaion rae. The condiional mean for GDP growh shows ha 51

7 inflaion volailiy does affec oupu growh negaively and his is saisically significan across he wo regimes bu was highly significan during he period of indirec regime. Bu he effec of inflaion volailiy on inflaion was no cerain as he cofficien was neiher sable nor significan across he wo periods. Pas levels of inflaion are found o have posiive effecs on currren inflaion levels and his is significan across he wo samples while he coefficien was almos he same across he wo periods. Oil price shocks do no have any meaningful effec on inflaion bu do have posiive and significan effecs on oupu growh especially in he second period. This may be due o he fac ha oil price increases which characerized mos of he sample period from 1995 o 2007 may have conribued posiively o he oupu growh in Nigeria as a largely oil dependen economy. This resul should no be surprising as mos sudies have found posiive oupu effec of oil price shocks in major oil-exporing counries. We are especially ineresed in he condiional variance equaions. The esimaes show ha he long run or uncondiional volailiies of inflaion and oupu growh were higher in he firs period han in he second period. The esimaes for he mean covariance erms are negaive and bu significan only in he firs sample period. The esimaes also indicae ha over he sample periods, he mean uncondiional variance of oupu is significan across he wo sample periods. The parameers in marix A show he exens o which he curren levels of condiional variance are correlaed wih heir pas levels. The higher esimaes in he second period seem o sugges ha a curren shock will have relaively long lasing effecs on he fuure levels of he condiional variances of oupu growh and inflaion han i had in he firs sample period. The esimaes also reveal ha following change in moneary policy regime, inflaion volailiy and oupu volailiy have relaively become more persisen in Nigeria. The off-diagonal elemens in A ha is A (1, 2), on he oher hand, reveal he exen o which he condiional variance of one variable is correlaed wih he lagged condiional variance of anoher variable. The esimae for A (1, 2) appears wih he expeced negaive sign and is saisically differen from zero for all sub-samples and he overall sample. This confirms he exisence of Taylor-curve volailiy radeoff in Nigeria. Ineresingly he esimae for he sample period 1981 o 1994 is relaively larger han he esimaes for he sample period 1995 o 2007 suggesing ha low inflaion variabiliy is now associaed wih higher oupu gap variabiliy. This seems o sugges ha CBN s effors o sabilize prices or specifically o arge low inflaion mus come a a heavy cos of oupu flucuaions. This resul is herefore, consisen wih he finding by Caselnuovo (2006) ha he igher he moneary policy, he higher is he inflaion-oupu gap volailiy. The parameers in B marix reveal he exens o which he condiional variances of inflaion and oupu are correlaed wih pas squared innovaions (deviaions from heir condiional means). Of paricular ineres is he off-diagonal elemens B (1, 2) and B (2, 1) which depic how he condiional variance of inflaion is correlaed wih he pas squared innovaions of oupu. In he firs sample period here is a posiive and significan volailiy cross-effec from inflaion o oupu. While in he second period here is posiive and significan volailiy cross effec from oupu growh o inflaion variabiliy. In order o address he hird research quesion we compued he impulse responses and variance 52

8 decomposiions from he VAR specificaion augmened by including moneary policy variable as one of he endogenous variables in equaion (1.1) and hen using oil prices as exogenous variable. We compued separae impulse responses and forecas error variance decomposiions for each variable for he wo regime periods in order o undersand how oupu growh and inflaion respond o innovaions in moneary policy over he wo regimes (see Figure 1 and Table 5 in he appendix). The impulse response funcions are inerpreed in conjuncion wih he variance decomposiions. For example, in he period of direc conrol regime, inflaion responded negaively o innovaions in moneary policy bu he variance decomposiions show ha his response was no significan because moneary policy only accoun for a small par of he forecas error variance of inflaion and his seems o remain consan in he long run. During he period of indirec regime inflaion also responded negaively o innovaions o moneary policy bu moneary shocks accouned for larger par of is forecas error variance, which is almos wice ha of he direc conrol period. Real GDP growh rae responded posiively o innovaions in moneary policy during he direc approach. This may be due o he posiive effecs of low ineres raes pursued during mos of ha period. The variance decomposiion of oupu growh shows ha moneary policy accouned for a larger par of he forecas error variance of oupu growh during he period of direc conrol regime han i accouns for inflaion. This resul is expeced because he primary objecive of moneary policy hen was o achieve rapid oupu growh. However, inflaion innovaions had larger effec on oupu growh in boh periods of moneary regimes showing ha inflaion volailiy is very crucial in deermining movemens in oupu variance. During he wo regimes oupu growh responded negaively o shocks on inflaion. This finding is consisen wih he resuls in he GARCH esimaions. During he period of he indirec regime oupu growh responded negaively o innovaions o moneary policy and his shows here is exisence of policy radeoff beween inflaion and oupu growh. The pursuance of low inflaion objecive during he period of indirec regime does rigger off negaive oupu reacions. The variance decomposiion shows ha his reacion is no significan since moneary policy shocks accoun for an insignifican par of forecas error variance of oupu growh even in he long run. 4. Summary, Policy Recommendaions and Conclusion 4.1 Summary of Findings The sudy on which repor his aricle focuses invesigaed he exisence of radeoff relaionship beween oupu growh and inflaion in Nigeria and he impac of alernaive moneary policy regimes on inflaion and oupu growh. The sudy findings show evidence of shor-run radeoff relaionship beween he variabiliy of oupu growh and inflaion bu no evidence srong long run volailiy relaionship was found. The sudy also found ha moneary policy accouned for a larger par of he forecas error variance of oupu growh during he period of direc conrol moneary policy han in he period of indirec conrol moneary policy. This resul was expeced given ha he objecive of moneary policy during he direc conrol regime was o achieve rapid and sable oupu growh. On 53

9 he oher hand, he response of inflaion o moneary policy changes in he period of indirec or marke-based regime was larger compared o is response during he period of direc conrol. Again, his resul is no surprising because he major focus of moneary policy in Nigeria during he period of indirec conrol was o achieve low inflaion. The resuls of he sudy furher reveal ha he volailiy of oupu growh and inflaion during he period of marke-based policy regime are more persisen compared o he period of direc conrols. Evidence of volailiy cross-effec from oupu o inflaion and vice versa was presen bu no significan across he wo sample periods. The resuls fail o esablish clearly any evidence o sugges ha moneary policy radeoff is a long run phenomenon. From hese findings, he following policy recommendaions could be made. 4.2 Policy Recommendaions Marke-based or indirec conrol approach o moneary policy conduc in Nigeria should be carefully examined regularly in order o ascerain is desirabiliy and workabiliy. This would help o deermine when changes in moneary policy sance acually affec he variabiliy of oupu and inflaion and in wha direcion. Policy makers should be careful no o believe oo fervenly ha he marke works in Nigeria. Policy changes could rigger off more volailiy han demand or supply shocks. This reflecs he fac ha marke imperfecion is very ypical of developing counries wih underdeveloped financial markes. In Nigeria, i is hard o believe ha inflaion is caused by excessive money growh or he growh of credi. Insead, inflaion has been driven largely by high cos of doing business, rising cos of energy prices and depreciaing exchange rae ha has made he cos of impored raw maerials exceedingly high. This has inensified he effec of adverse supply shocks on inflaion. Again, he size of he informal and non-moneized secor of he economy is quie subsanial making i possible for moneary policy o have a big impac. In such a macroeconomic environmen, ighening of moneary policy in response o high inflaion would exacerbae an already heaed environmen by increasing he cos of credi o firms ha depend on borrowing as he major source of finance. Our sudy reveals ha volailiy radeoff is higher during he period of indirec moneary regime han during he period of direc conrol and ha oupu is responding negaively o moneary shocks. This implies ha he Cenral Bank should be very cauious of he objecive of argeing low inflaion as such a policy could rigger off no only low oupu growh bu also high oupu growh volailiy. Finally, we sugges ha moneary policy insrumens be suppored by oher fiscal and physical measures such as ensuring ha energy cos, he cos of impored raw maerials and possibly he cos of housing are reduced hrough oher improved supply-side processes. While, aenion is being focus on low inflaion i is also perinen o realize ha high and sable oupu growh objecive is equally imporan o Nigeria as a developing economy. There should be a balance beween oupu growh objecive and low inflaion. The sudy reveals ha moneary policy objecive ha arges low inflaion would be likely achieved a a heavy cos in erms of adverse oupu growh effec. 54

10 4.3 Conclusion This sudy has shown ha here is very lile empirical evidence o sugges ha moneary policy regime change necessarily alers exising inflaion-oupu growh variabiliy radeoff. I could no find srong evidence of long run radeoff beween oupu growh and inflaion, which is required in order o ascerain he effeciveness of moneary policy regime changes from ha perspecive. This is no alogeher surprising as half of he sudies underaken on his same issue in oher jurisdicions have hus far found no evidence of long run policy radeoff (see Lee 2004 for example). However, mos sudies did find ha volailiy radeoff changed when moneary policy regime changed, his sudy seems o corroborae he same findings. I is perhaps imporan o observe here ha he availabiliy of good qualiy macroeconomic daa a shor-ime inervals like monhly or quarerly series remains a major challenge o policy-relevan research in Nigeria. However, he siuaion is no very much differen in mos oher developing counries. Using exrapolaed quarerly GDP daa in empirical sudies of his naure may influence he research oucomes since such daa were economerically generaed under cerain assumpions. Furher research is herefore recommended in his issue in he fuure, paricularly as high frequency and good qualiy daa begin o be available. I would indeed be very informaive o policy makers in Nigeria who are currenly experimening wih he adopion of inflaion argeing moneary policy regime o read his research oupu. I will perhaps assis hem in appreciaing he cos of such regime in erms of oupu growh volailiy. References Agu, C. (2007). Wha Does he Cenral Bank of Nigeria Targe: An Analysis of Moneary Policy Reacion Funcion in Nigeria. African Economic Research Consorium (AERC), Nairobi, Kenya. Balini, N. (2004). Achieving and Mainaining Price Sabiliy in Nigeria. IMF Working Paper, WP/04/97. Caselnuovo, E. (2006). Moneary Policy Swich, he Taylor Curve and Grea Moderaion, [Online] Available: hp://ssrn.com/absrac=880061]. Dickey, D. A., & Fuller, W. A. (1981). Likelihood Raio Saisics for Auoregressive Time Series wih a Uni Roo. Economerica, 49, Founas, S., Menelaos, K. & J. Kim (2002). Inflaion and Oupu Growh Uncerainy and heir Relaionship wih Inflaion and Oupu Growh. Economics Leers, 75, Founas, S. & Menelaos, K. (2007). Inflaion, Oupu Growh, and Nominal and Real Uncerainy: Empirical Evidence for he G7. Journal of Inernaional Money and Finance, 26, Fuhrer, J. (1997). Inflaion/Oupu Variance Trade-offs and Opimal Moneary Policy. Journal of Money, Credi and Banking, 29(2), Gaspar, V. & Smes, F. (2002). Moneary Policy, Price Sabiliy and Oupu Gap Sabilizaion. Inernaional Finance, 5(2), Grier, K. B., Henry, O. T., Olekalns, N. & Shields, K. (2004). The Asymmeric Effecs of Uncerainy 55

11 on Inflaion and Oupu Growh. Journal of Applied economerics, 19, Jarque, C. M. & Bera, A. K. (1980). Efficien ess for normaliy, homoscedasiciy and serial independence of regression residuals. Economics Leers, 6, Lee, J. (2002). The Inflaion-Oupu Variabiliy Tradeoff and Moneary Policy: Evidence from a GARCH Model. Souhern Economic Journal, 69(1), Lee, J. (2004). The Inflaion-Oupu Variabiliy Tradeoff: OECD Evidence. Journal of Conemporary Economic Policy, 22(3), Ljung, G. M. & Box, G. E. P. (1978). On a Measure of Lack of Fi in Time Series Models. Boimerika, 66: Naka, A. & Tufe, D. (1997), Examining he Impulse Response Funcions in Coinegraed Sysems; Applied Economics, 29, Nnanna, O. (2001). Moneary Policy Framework in Africa: he Nigerian Experience. Cenral Bank of Nigeria, Working Paper. Okonjo-Iweala, N. & Philip Osafo-Kwaako. (2006). Nigeria s Economic Reforms: Progress and Challenges. Brookings Global Economy and Developmen, Working Paper, 6, Pesaran, M. H. and Y. Shin (1998). Generalised impulse response analysis in linear mulivariae models. Economic Leers, 58, Robers, J. M. (2006). Moneary Policy and Inflaion Dynamics. Inernaional Journal of Cenral Banking, Sepember, Taylor, J. (1979). Esimaion and conrol of Macroeconomic Model wih Raional Expecaions. Economerica, 47(5), Taylor, J. (1994). The Inflaion/Oupu Variabiliy Trade-off Revisied, in Goals, Guidelines and Consrains Facing Moneary Policymakers. FRB of Boson, Conference Series 38, Noes Noe 1. Oher sudies failed o disinguish he wo periods in heir analyses and hus evaluae he effeciveness of he CBN s moneary policy even over he period when moneary policy in Nigeria had overbearing poliical inerference. Noe 2. Gaspar, Vicor and Frank Smes (2002) Moneary Policy, Price Sabiliy and Oupu Gap Sabilizaion. Inernaional Finance, 5:2, Figures Impulse Response Funcions 56

12 0.08 Plo of responses of inflaion rae Real GDP Growh Rae MRR inflaion rae Figure 1. Plo of responses of inflaion rae Plo of responses of MRR Real GDP Growh Rae MRR inflaion rae Figure 2. Plo of responses of MRR Plo of responses of Real GDP Growh Rae Real GDP Growh Rae MRR inflaion rae Figure 3. Plo of responses of Real GDP Growh Rae

13 0.050 Plo of responses of inflaion rae Real GDP Growh Rae MRR inflaion rae Figure 4. Plo of responses of inflaion rae Plo of responses of MRR Real GDP Growh Rae MRR inflaion rae Figure 5. Plo of responses of MRR

14 Real GDP Growh Rae inflaion rae MRR Figure 6. Plo of responses of Real GDP Growh Rae Tables Table 1. Summary Saisics of Sudy Variables Series Sample obs mean Sd Dev Minimum Maximum GDP 1981:1 2007: :1 1994: :1 2007: CPI 1981:1 2007: :1 1994: :1 2007: INFLAcbn 1981:1 2007: :1 1994: :1 2007: Oilprices 1981:1 2007: :1 1994: :1 2007: Mrr 1981:1 2007: :1 1994: :1 2007: LOGRGDP 1981:1 2007: :1 1994: :1 2007: GDPGRT 1981:1 2007: :1 1994:

15 1995:1 2007: Table 2. Tess for Serial Correlaion, ARCH and Normaliy Tess for Serial Correlaion, ARCH and Normaliy Series Q(8) Q(16) Q(24) ARCH(1) ARCH(2) JB-STAT GDPGRWT [0.0006] [0.000] 82.40[0.000] [0.000] 5.846[ ] 2.228[0.3283] INFLACBN [0.0000] [0.000] [0.00] [0.006] 40.51[0.0000] [0.07] Table 3. Uni Roo Tess of Variables of he Model Variable Level Firs Difference Lags Logrgdp * ** 1 Inflacbn ** ** 1 Oilprices ** 1 Logoilprices ** 1 MRR ** 1 GDPGRT ** 3 Table 4. Bivariae GARCH Esimaions of Oupu Growh and Inflaion BIVARIATE GARCH ESTIMATIONS OF OUTPUT GROWTH AND INFLATION CONDTIONAL MEAN EQUATIONS 1981:1 2007:4 1981:1 1994:4 1995:1 2007:4 Variables Mod1 Mod11 Mod21 CONSTANT ** * Trend ** GRGDP{1} INFLA_VOL ** * ** OILP_SHOCK ** ** CONSTANT ** ** Trend ** INFLACBN{1} ** ** ** INFLA_VOL ** OILP_SHOCK ** CONDITIONAL VARIANCE EQUATIONS C(1,1) ** ** ** C(1,2) ** C(2,2) ** A(1,1) ** ** A(1,2) ** ** ** 60

16 A(2,2) ** ** B(1,1) ** ** B(1,2) ** B(2,1) ** ** B(2,2) ** Convergence 72 Iers 54 Iers 55 Iers Table 5. Variance Decomposiions Decomposiion of Variance for Series MRR Sep Sd Error MRR INFLACBN LOGRGDP Decomposiion of Variance for Series INFLACBN Sep Sd Error MRR INFLACBN LOGRGDP

17 Decomposiion of Variance for Series LOGRGDP Sep Sd Error MRR INFLACBN LOGRGDP

18 Decomposiion of Variance for Series MRR Sep Sd Error MRR INFLACBN LOGRGDP Decomposiion of Variance for Series INFLACBN Sep Sd Error MRR INFLACBN LOGRGDP 63

19

20 This academic aricle was published by The Inernaional Insiue for Science, Technology and Educaion (IISTE). The IISTE is a pioneer in he Open Access Publishing service based in he U.S. and Europe. The aim of he insiue is Acceleraing Global Knowledge Sharing. More informaion abou he publisher can be found in he IISTE s homepage: hp:// The IISTE is currenly hosing more han 30 peer-reviewed academic journals and collaboraing wih academic insiuions around he world. Prospecive auhors of IISTE journals can find he submission insrucion on he following page: hp:///journals/ The IISTE ediorial eam promises o he review and publish all he qualified submissions in a fas manner. All he journals aricles are available online o he readers all over he world wihou financial, legal, or echnical barriers oher han hose inseparable from gaining access o he inerne iself. Prined version of he journals is also available upon reques of readers and auhors. IISTE Knowledge Sharing Parners EBSCO, Index Copernicus, Ulrich's Periodicals Direcory, JournalTOCS, PKP Open Archives Harveser, Bielefeld Academic Search Engine, Elekronische Zeischrifenbibliohek EZB, Open J-Gae, OCLC WorldCa, Universe Digial Library, NewJour, Google Scholar