Causality between Money and Prices:Evidence from Pakistan

Transcription

1 MPRA Munich Personal RePEc Archive Causaliy beween Money and Prices:Evidence from Pakisan Fazal Husain and Tariq Mahmood Pakisan Insiue of Develomen Economics 1998 Online a h://mra.ub.uni-muenchen.de/2720/ MPRA Paer No. 2720, osed 7. June 2007

2 The Pakisan Develomen Review 37 : 3 Par II (Winer 1998). 37:4, Causaliy beween Money and Prices: Evidence from Pakisan FAZAL HUSAIN and TARIQ MAHMOOD I. INTRODUCTION The relaionshi beween money and rices has been a debaed issue among economic schools of hough aricularly beween he Moneariss and he Keynesians. The moneariss claim ha changes in money socks cause changes in rice levels. In oher words, he direcion of causaion runs from money o rices imlying ha rices can be conrolled hrough money suly. The keynesians, on he oher hand, argue ha money is imoran bu is no resonsible for changes in rice levels. Insead, srucural facors lay imoran role suggesing ha money suly is no an effecive insrumen o conrol rice changes. The causal relaionshi beween money and rices has been exensively esed in various counries. For examle, Brillembourg and Khan (1979) examined his relaionshi in USA. Using Sims rocedure for he eriod , hey found unidirecional causaliy running from money o rices. Similar direcions of causaion are reored by Lee and Li (1983) and Ramachandran and Kamaiah (1992) who invesigaed he causal relaionshi in Singaore and India resecively. On he oher hand, Aghevli and Khan (1978), while invesigaing he causal relaionshi in Brazil, Columbia, he Dominican Reublic, and Thailand, found bidirecional causaliy beween money and rices in hese counries. The money-rice relaionshi has also been invesigaed in Pakisan. Ali (1986) found he validiy of he Quaniy Theory of Money in redicing rice levels. The causal relaionshi was formally invesigaed by Jones and Khilji (1988). Using monhly daa on alernaive measures of money, M1 and M2, and rices, Wholesale Price Index (WPI) and Consumer Price Index (CPI), for he eriod and Fazal Husain and Tariq Mahmood are boh Research Economiss a he Pakisan Insiue of Develomen Economics, Islamabad.

3 37:4, 1156 Husain and Mahmood alying Granger es, hey found significan imacs of money lags on WPI and CPI lags on money measures wihou any feedback. On he oher hand, Khan and Siddiqui (1990), using Sims rocedure and quarerly daa for he eriod , found unidirecional causaliy from M1 o CPI bu bidirecional beween M2 and CPI. Finally, Bengali, Khan, and Sadaqa (1997), aking care of ime series roeries and using quarerly daa from , found bidirecional causaliy beween money measures (M1 and M2) and CPI. This aer re-examines he causal relaionshi beween money and rices in Pakisan using recen daa on money and rices and aking care of ime series roeries. Earlier sudies, wih he exceion of Bengali e al. (1997), ignore he ime series asec and may lead o wrong conclusion regarding causaliy. None of he sudies uses daa beyond 1990 whereas, he 1990s has been he eriod of high growh in rices and money as well as banking and financial secor reforms. The aer is organised as follows. The nex secion describes he daa sources. Secion III oulines he mehodology o es for he ime series roeries as well as he causaliy. The emirical resuls are resened in Secion IV. The las secion conains he summary and conclusion. II. DATA SOURCES Monhly daa on money socks and rices were used from July 1981 o June Three measures of money socks were aken. These are, M0 = Currency in Circulaion. M1 = M0 + Demand Deosis + Oher deosis wih SBP. M2 = M1 + Time Deosis + Residens Foreign Currency Deosis. The daa on hese measures, in million ruees, were aken from various issues of he Monhly Saisical Bullein of he Sae Bank of Pakisan (SBP). The daa on wo measures of rices, Wholesale Price Index (WPI) and Consumer Price Index (CPI), wih base eriod , were aken from he various issues of he Monhly Saisical Bullein of he Federal Bureau of Saisics, Saisics Division. The WPI is consruced by he wholesale rices of food, raw maerials, fuel and lubricans, manufacures, and building maerials. The CPI is based on reail rices covering 61 markes in 25 ciies and 460 commodiies. III. METHODOLOGY In he firs se, he saionariy of variables included in he analysis was esed by a Uni Roo es. In his conex, he Augmened Dickey Fuller (ADF) es was used and hree models were esimaed. Model I (wihou any consan and rend)

4 * Y = ρ Y Causaliy beween Money and Prices 37:4, δi Y i e (1) + Model II (wih consan bu no rend) * Y = α + ρ Y 1 + δi Y i e. (2) + Model III (wih consan and rend) Y = α + β + ρ Y * 1 + δi Y i e (3) + The ADF saisic is he -value associaed wih he esimaed coefficien of ρ *. The acceance of null hyohesis (H 0 : ρ * = 0) imlies he resence of uni roo which in urn imlies nonsaionariy. The hyohesis is rejeced if ρ * is negaive and significan. The es was erformed for all he series of money and rices where boh he original series and he differences of he series were esed for saionariy. The co-inegraion beween he wo series was esed by running he OLS regression, called he co-inegraing regression: Y = α + β X + e. (4) Then he series of residuals, e, from his regression was esed for saionary. Saionariy in e imlies co-inegraion beween Y and X, and can be esed in many ways, as discussed in Engle and Granger (1987). One way is o aly he ADF es (defined above) on e. The second and a quick way is o look a he Durbin-Wason saisic of he co-inegraing regression. If i is close o zero hen his imlies ha e is no saionary and herefore Y and X are no co-inegraed. Engle and Granger (1987), hough, roosed several mehods o es for co-inegraion, however, recommended he ADF rocedure. The acceance of co-inegraion beween wo series imlies ha here exiss a long run relaionshi beween hem. However, his relaionshi may be disurbed by shor run deviaions from equilibrium and hus an Error Correcion Model (ECM) may be an aroriae framework which is an exension of he Granger causaliy es where an error correcion erm is inroduced ino he es, ha is, q Y = α1+ ρ e 1+ β Y i 1+ δ j X j 1 (5) 1 i j= 1 q X = α2+ ρ e 1+ β Y i 1+ δ j X j 1 (6) 2 i j= 1 where e 1 is an error correcion erm reresening he long run relaionshi. ρ 1 and ρ 2 are considered he seed of adjusmen coefficiens. Furhermore, a leas one of hese coefficiens mus be significan in order he ECM o hold. If boh coefficiens

5 37:4, 1158 Husain and Mahmood are significan, his will sugges ha boh series exer long run relaionshi, ha is, here exiss a feedback mechanism beween hem. If, however, only ρ 1 is found o be significan, his will sugges ha X off drives Y oward long run equilibrium bu no he oher way around. Moreover, hese coefficiens mus be negaive in order he las eriod osiive deviaion from long run rend o have negaive effec in his eriod and hus ushing i back oward he rend. The lagged erms of Y and X, aeared as exlanaory variables, indicae shor run dynamics or cause and effec relaionshi beween he wo series. Thus, if he lagged coefficiens of X aear o be significan in he regression of Y, his means ha X affecs Y. Similarly, he oosie holds if he lagged coefficiens Y are significan in X. If none of he lagged coefficien is significan anywhere his imlies ha here is no cause and effec relaionshi beween he wo series. IV. EMPIRICAL RESULTS In he firs se all he series were esed for he uni roos. In his conex, he ADF was alied o boh he original series (in log) and he firs differences. The resuls are reored in Table 1. Table 1 Uni Roo Tess (Augmened Dickey Fuller) for he Period July 81 June 98 Original Series Fis Differences Model I Model II Model III Model I Model II Model III ln CP * 13.72* 13.61* ln WP * 11.08* 11.06* ln M * 9.65* 10.26* ln M * 10.73* 10.65* ln M * 4.49* 4.37* Criical Values (0.05) CP = Consumer Price Index, WP = Wholesale Price Ind ex, M0 = Currency in Circulaion. M1= M0 + Demand and Oher Deosis, M2 = M1 + Time Deosis + Foreign Currency Accouns. The resuls sugges he acceance of he resence of uni roos in he original series indicaing ha none of he original series is saionary. However, he resence of uni roos is conclusively rejeced in he firs differences of he series for all he variables in all he models. This suggess ha all he money and rice variables are inegraed of order one. Since all he series are inegraed of same order, here is a ossible chance of co-inegraion among he series. Nex, co-inegraing regressions were esimaed and are reored in Table 2. Furher, he series of residuals were obained from each regression and ADF was alied o es for he resence of uni roos in hese residuals. These resuls are also reored in Table 2.

6 Causaliy beween Money and Prices 37:4, 1159 Table 2 Resuls from Coinegraion Tess Y on X Cons. Sloe CRDW ADF CP on M M0 on CP CP on M M1 on CP CP on M * M2 on CP * WP on M M0 on WP WP on M M1 on WP WP on M * M2 on WP * CRDW = Coinegraing Regression Durbin Wason saisic. ADF = Augmened Dickey Fuller es. Criical Value for ADF is The able shows ha he null hyohesis of no co-inegraion beween he wo series is acceed in all cases, exce beween rice variables and M2 measure of money. This suggess ha here exis a long run relaionshi beween rices and moneary asses, M2. Finally, o exlore he long run relaionshi beween M2 and rices furher, ha is, o know he shor run dynamics as well as he cause and effec relaionshi, he Error Correcion Model was emloyed and he resuls are reored in Table 3. Table 3 Resuls from Error Correcion Models Y on X -values Signif F-Values Signif -Values Signif F-Values Signif Lags = 6 Lags = 12 CP on M M2 on CP WP on M M2 on WP Lags = 18 Lags = 24 CP on M M2 on CP WP on M M2 on WP Lags = 30 Lags = 36 CP on M M2 on CP WP on M M2 on WP The able shows he -values for he error correcion erm along wih he significance level. The able also shows he F-values for he lags of indeenden

7 37:4, 1160 Husain and Mahmood variable along wih he significance level. The model was esed for six differen lags seleced arbirarily. These lags range from 6 monhs o 36 monhs. The able shows ha he error erm has he correc negaive sign in he regressions of rice variables on money. Furher i is significan uo 10 ercen in all cases where wholesale rices are regressed on money whereas i is significan for lags 6, 30 and 36 in he case of CPI. Regarding lags value of money, hese are found significan u o 10 ercen a lags 6, 12, and 18 when wholesale rices are regressed on money and a lags 12, 18, 30, and 36 in he case of CPI. None of he lags of rices are significan in any money equaion. The error erm in he money equaion, alhough significan in some cases, does no have he negaive sign anywhere. Overall, he resuls sugges a unidirecional causaliy running from money o rices. Furher, i seems ha he moneary exansion has a greaer imac on wholesale sale rices comared o CPI. This is somewha similar o he finding of Jones and Khilji (1988) where hey found he imac of money growh on WPI, bu no on CPI. However, heir finding of feedback from CPI o money growh is no suored here. V. SUMMARY AND CONCLUSIONS The objecive of he aer is o re-examine he causal relaionshi beween money and rices in Pakisan using recen daa and emloying coinegraion and error correcion models. Two measures of rices (CPI and WPI) and hree measures of money socks (M0, M1, and M2) were aken for he eriod July 1981 o June The analysis indicaes a long run relaionshi beween rices and M2 definiion of money. The oher definiions of money do no seem o be relaed wih rices. The analysis furher suggess a unidirecional causaliy running from money o rices and hus suoring he moneariss s claim regarding he role of money. The resuls, however, are differen from he findings of earlier sudies in Pakisan which generally reor bidirecionl causaliy beween rices and money. REFERENCES Aghevli, B., and M. S. Khan (1978) Governmen Deficis and he Inflaionary Process in Develoing Counries. Inernaional Moneary Fund Saff Paers Ali, Shauka (1986) Causaliy beween Money Suly and Price Level: Evidence in Theory and Pracice. Kashmir Economic Review Bengali, K., A. Khan, and M. Sadaqa (1997) Money, Income, Prices, and Causaliy: The Pakisani Exerience. (Unublished Paer.) Brillembourg, A., and M. S. Khan (1979) The Relaionshi beween Money,

8 Causaliy beween Money and Prices 37:4, 1161 Income, and Prices: Has Money Maered Hisorically? Journal of Money, Credi and Banking Engle, R., and C. Granger (1987) Co-inegraion and Error Correcion: Reresenaion, Esimaion and Tesing. Economerica 55: Jones, J., and N. Khilji (1988) Money Growh, Inflaion, and Causaliy (Emirical Evidence for Pakisan The Pakisan Develomen Review 28: Khan, A. H., and A. N. Siddiqui (1990) Money, Prices and Economic Aciviy in Pakisan: A Tes of Causal Relaion. Pakisan Economic and Social Review Lee, S. Y., and W. K. Li (1983) Money, Income, and Prices and heir Lead-Lag Relaionshi in Singaore. Singaore Economic Review Pakisan, Governmen of (Various Issues) Monhly Saisical Bullein. Federal Bureau of Saisics. Pakisan, Governmen of (Various Issues) Monhly Saisical Bullein. Sae Bank of Pakisan. Ramachandran, M. and B. Kamaiah (1992) Causaliy beween Money and Prices in India: Some Evidence from Coinegraion and Error Correcion Models. Singaore Economic Review