|
Monetary Policy Strategy in Rwanda( Télécharger le fichier original )par Serge Musana Mukunzi University of Kwazulu Natal - Maitrise 2004 |
4.2 Estimation resultsThe results from OLS estimation of our model are displayed in Table (4.3). Table 4.3: ORDINARY LEAST SQUARES ESTIMATION *************************************************************************** Dependent variable is HD 20 observations used for estimation from 1997Q1 to 2001Q4 *************************************************************************** Regressor Coefficient Standard Error T-Ratio[Prob] C .041926 .018295 2.2917[.041] HD(-1) -.55205 .24191 -2.2820[.042] IG .87528 .34904 2.5077[.028] IG(-1) -.76760 .29924 -2.5651[.025] YD .57167 .57095 1.0013[.336] YD(-1) -.14445 .52748 -.27385[.789] DEX .032610 .65972 .049429[.961] DEX(-1) -1.9057 .76955 -2.4763[.029] *************************************************************************** R-Squared .72147 R-Bar-Squared .55900 S.E. of Regression .040975 F-stat. F( 7, 12) 4.4405[.012] Mean of Dependent Variable .0099828 S.D. of Dependent Variable .061702 Residual Sum of Squares .020147 Equation Log-likelihood 40.6254 Akaike Info. Criterion 32.6254 Schwarz Bayesian Criterion 28.6425 DW-statistic 2.1915 Durbin's h-statistic *NONE* *************************************************************************** Diagnostic Tests *************************************************************************** * Test Statistics * LM Version * F Version * ******************************************************************************* * * * * * A:Serial Correlation*CHSQ( 4)= 8.6706[.070]*F( 4, 8)= 1.5307[.282]* * * * * * B:Functional Form *CHSQ( 1)= .74496[.388]*F( 1, 11)= .42558[.528]* * * * * * C:Normality *CHSQ( 2)= 1.2503[.535]* Not applicable * * * * * * D:Heteroscedasticity*CHSQ( 1)= 5.7353[.017]*F( 1, 18)= 7.2371[.015]* *************************************************************************** A:Lagrange multiplier test of residual serial correlation B:Ramsey's RESET test using the square of the fitted values C:Based on a test of skewness and kurtosis of residuals D:Based on the regression of squared residuals on squared fitted values Source: Microfit outputs Considering the full sample period 1997-2001, the coefficient of YGt, YGt-1 and DEXt are insignificant and carry unexpected signs and the model fails the serial correlation and heteroscedasticity test. When reducing the model by taking out all the insignificant variables the final preferred model is specified as following: Mt= 0 + 1Mt-1 + 2IGt + 3IGt-1+ 4DEXt-1 + t (5) This complies with most of the diagnostic statistics regarding no serial correlation, good functional form, normality and the absence of heteroscedasticity. On the other hand, the R2 and the adjusted R2 are fairly good and the signs of the estimated coefficient in the relation to their prior expectation are satisfactory.
The results are shown in Table (4.4) as the following: Table 4.4: ORDINARY LEAST SQUARES ESTIMATION *************************************************************************** Dependent variable is HD 20 observations used for estimation from 1997Q1 to 2001Q4 *************************************************************************** Regressor Coefficient Standard Error T-Ratio[Prob] C .058537 .014898 3.9293[.001] HD(-1) -.46291 .22639 -2.0448[.059] IG .96117 .30511 3.1502[.007] IG(-1) -.64710 .26722 -2.4216[.029] DEX(-1) -2.4506 .55757 -4.3951[.001] *************************************************************************** R-Squared .64415 R-Bar-Squared .54926 S.E. of Regression .041425 F-stat. F( 4, 15) 6.7881[.003] Mean of Dependent Variable .0099828 S.D. of Dependent Variable .061702 Residual Sum of Squares .025740 Equation Log-likelihood 38.1755 Akaike Info. Criterion 33.1755 Schwarz Bayesian Criterion 30.6861 DW-statistic 2.3383 Durbin's h-statistic *NONE* ************************************************************************ Diagnostic Tests *************************************************************************** * Test Statistics * LM Version * F Version * *************************************************************************** * * * * * A:Serial Correlation*CHSQ( 4)= 3.9091[.418]*F( 4, 11)= .66808[.627]* * * * * * B:Functional Form *CHSQ( 1)= 1.9189[.166]*F( 1, 14)= 1.4858[.243]* * * * * * C:Normality *CHSQ( 2)= 2.0604[.357]* Not applicable * * * * * * D:Heteroscedasticity*CHSQ( 1)= 1.4580[.227]*F( 1, 18)= 1.4154[.250]* ******************************************************************************* A: Lagrange multiplier test of residual serial correlation B: Ramsey's RESET test using the square of the fitted values C:Based on a test of skewness and kurtosis of residuals D:Based on the regression of squared residuals on squared fitted values (Source: Microfit output) The regression is significant as demonstrated by the F-statistic, which provides a test that the true value of the slope coefficients are simultaneously equal to zero (the p-value 0.003 is almost zero). The goodness of fit, as measured by the coefficient of determination R2, indicates how well the sample regression line fits the data. The fit indicates that about 64.4% of changes in monetary stock aggregate are explained by monetary stock aggregate in the one period lagged, the current inflation gap, the one period lagged inflation gap and the one period lagged change in the nominal exchange rate. When examining each explanatory variable it was found that: - The response coefficients on inflation gap are about 0.96 and -0.65 respectively the contemporaneous and the one period lagged value. - Both coefficients are statistically significant at the 5% level of significance. - The response on the change in the one period lagged value of nominal exchange rate is(-2.45). - Judging by the t-ratio, the coefficient is significantly different from zero at the all-conventional level of significance. - The response coefficient on the one period lagged value of the monetary stock aggregate is about (-0.46) and judging by the t-ratio, the coefficient is significantly different from zero at the 10% level of significance. According to those estimation results, the interpretation is that: (1) Considering the one period lagged value of the monetary stock aggregate (Mt-1), one can observe that a one percentage point change in the previous period monetary stock aggregate results in about 0.46% point change in the current monetary stock aggregate in the opposite sense, holding other things the same. This means that, there is one to 0.46 inverse relationship between the previous and the current monetary stock aggregate. (2) Considering the current and the one period lagged inflation gap: It is noticed that for a given change in the one period lagged deviation of the inflation from its target, the monetary stock aggregate reacts by a change of about 0.65 units. This impact seems statistically significant since the t-ratio is significant and the coefficient carries the right expected sign. This means that, if inflation were 1% point above its target, the Central Bank would decrease the monetary stock aggregate by 0.65% in terms of reaction, holding other things the same. Thus, the Central Bank was reacting to one period lag on the deviation of inflation from its target by a weight of about 0.65. Regarding the current period of inflation gap, the coefficient 0.96 can be interpreted as the change in the value of monetary stock aggregate following a unit change in inflation gap in the same period. That is, the monetary stock aggregate was changing in the same sense as the price level by 0.96% following a 1% change in prices. (3) Considering the exchange rate: The result shows that the Central Bank of Rwanda reacts to 1% change in the exchange rate by a change of 2.45% in the monetary stock aggregate inversely, holding other things the same. In the light of the above findings, the Rwanda Central Bank's behavior in respect to the weight carried by the inflation gap in the current period and the one period lagged should be interpreted cautiously. Indeed, from the estimated of equation (4), the coefficient affecting the variable IGt is positive and significant whereas the one affecting the variable IGt-1 is negative and significant as well. This seems to show that the Central Bank reacts mainly to the inflation focusing on the one period lagged of the state of the economy while the positive contemporaneous relationship between the monetary stock and the inflation gap could be seen as the period of economic state adjustments. That is, dynamically speaking, in the current period the change in price leads the monetary policy to adjust the monetary base toward its trend. Overall, the change in the one period lagged in the exchange rate seems to be more influential than the inflation rate in explaining the change in monetary policy since the coefficient of the former carries the expected sign and its t-ratio is relatively very significant. When looking at the result more closely with the objective of highlighting the variable that has influenced monetary policy decisions over the period of study, it is apparent that the monetary authorities were mainly concerned with the exchange rate. This is relevant given the importance of the exchange rate in a small, open, and developing country especially Rwanda, in the present case. Indeed, as noted previously, the exchange rate policy in Rwanda aims at approaching a balanced level of the exchange rate, to stabilise prices, to ensure a support for the growth and to connect Rwanda's foreign exchange market to the international market. These objectives are pursued under a controlled flexible policy regime, that is, the exchange rate can fluctuate from day to day but the Central Bank attempts to influence the exchange rate by buying and selling currencies in the foreign exchange market. The impact of such interventions is to affect the monetary base. According to this fact, the exchange rate considerations play a great role in the conduct of monetary policy and this has been shown through the estimation results. The exchange rate coefficient from the results leads one to consider that a depreciation of Rwanda Francs may lead the Central Bank to pursue a kind of contractionary monetary policy and similarly, an appreciation in Rwanda Francs may lead to pursue a high rate of money growth and such reaction consists basically in acquiring or selling of international reserve. Given the fact that Rwanda has a strong dependence on assistance from multilateral financial institutions which in its turn has a real impact on the balance of payment, apparently, the importance of reacting to exchange rates seems to be relevant in Rwanda since it could limit the pressure exerted on the Rwanda currency in order to meet international prices and the debt service management. In addition, these findings about the exchange rate influence on monetary policy are consistent with the state of the Rwanda's economy from 1995 when it started to benefit from financial assistance from international institutions in the context of the Enhanced Structural Adjustment Facility (ESAF) and the Poverty Reduction and Growth Facility (PRGF). This may lead one to think that changes in the flow of international assistance could contribute to the significant changes in official reserves for the country. As a consequence, an objective of Rwanda currency stability should play an important role given the link between the foreign market operation and the change in monetary base and thereby the behavior of prices level. In this specific context, one could consider that monetary authorities were not as concerned about reacting to changes in the inflation as to the exchange rate simply because when focusing on the exchange rate, the Central Bank has also attempted to stabilize prices given the fact that the depreciation of Rwandan Francs and the increase in international prices affects the inflation rate.
Comparing these findings with the suggestion of the Taylor rule, it would appear that the weight of the inflation gap does not correspond to the Taylor rule's description which sets the weight greater than one. In addition, the estimates indicated a neglect of output gap as a goal variable under the period of study (see Table 4.3). These results may indicate that the Taylor rule adapted to the context of monetary policy in Rwanda expresses that the economy behaved by giving much importance to the exchange rate than to inflation and neglected the output change. CHAPTER 5 CONCLUSION AND SUGGESTION This dissertation intended to study how monetary policy was conducted in Rwanda. The task has been accomplished by designing and estimating a Taylor rule, monetary policy reaction function for the National Bank of Rwanda over the period 1997-2001. Applying Ordinary Least Squared (OLS) on data taken from the National Bank of Rwanda, the Ministry of Economic and Finance of Rwanda and the IMF together, the study shows that the National Bank of Rwanda has had a monetary policy over the years with the monetary stock aggregates (M1) as the principal instrument. The Rwanda Central Bank's reaction function can be characterised by: - A contemporaneous inflation gap weight of (0.96), which is positively related to the monetary stock aggregate, - A previous quarter inflation gap weight of (-0.65), which is negatively related to the monetary stock aggregate, - A previous quarter change in exchange rate weight of (-2.45), which is negatively related to the monetary stock and - A previous quarter monetary stock aggregate weight of (-0.46), which is also negatively related to the current monetary stock aggregate. Judging these results according to the importance of each variable weight, one may be inclined to contend that while the National Bank of Rwanda reacted to the inflation from the previous quarter and the contemporaneous period, there was a much stronger response to the change in the previous quarter's exchange rate. In general, such strong response of the National Bank of Rwanda to the exchange rate may reflect the economic environment in which the monetary policy was operating, because international aid has flow into Rwanda since 1995 in the context of various economic programs undertaken with the help of the International Monetary Found (IMF) or World Bank (WB). In addition, the estimate results indicated a neglect of output gap as a goal variable. Given the fact that the Rwanda Central Bank claimed to be following the objective of preserving the internal and the external value of the currency in order to maintain harmony between the pace of the money creation and that of economic growth it may be suggested that the Central Bank of Rwanda should give more consideration to the way the output changes, that is, the Central Bank should respond to the variation of output gap following the influence of its change in the economic state. In addition, the results of this study of course, are backward looking, in the sense that they represent the relationships that existed so far in the data. It is worth noting that a forward-looking model may enable the implantation of a more successful monetary policy rule for Rwanda and there may be areas for future research.
BIBLIOGRAPHY Books Backhouse, R. (1999). Applied UK macroeconomics. Massachusetts: Basil Blackwell Blinder, A. S. (1999). Central Banking in theory and practice. Massachusetts: The MIT Press Daniels, J. , Vanhoose, D. (1999). International monetary and financial economics. Cincinnati: South-Western College Publishing Darnell, A. C. (1994). A dictionary of econometrics. Brookfield: Edward Edgar Publishing limited. Dornbusch, R. , Fisher, S. , Startz, R. (2001). Macroeconomics. 8th Edition. New-York: McGraw-Hill Edwards, G. T. (1984). How economic growth and inflation happen. London: Macmillan Press Enders, W. (1995). Applied econometrics time series. New-York: John Wiley and Sons. Frowen, S. F. (1993). Monetary theory and monetary policy. London: The Macmillan press. Gujarati, D. (1992). Essentials of econometrics. New-York: McGraw-Hill. Mankiw, N. G. (2000). Macroeconomics. 4th Edition. New-York: Worth Publisher Meyer, L. H. (1980). Macroeconometrics, a model bulding approach. Cincinnati: South Weatern Publishing Co. Mishkin, F. S. (1997). The economics of money, banking and financial. 5th Edition. Massachusetts: Addison-Wesley Pesaran, M. H., Pesaran, B. (1997). Microfit 4.0. Oxford: Oxford University Press. Wallich, H. C. (1982). Monetary policy and practice. Lexington: D. C. Heath Walsh, C. E. (1998). Monetary theory and policy. Massachusetts: The MIT Press. Journals, Working Papers Alexander, W. E., Thomas, J. T. , Balino, Enoch, C. (1995). An Adopting Indirect Instrument of Monetary Policy. IMF Occasional Paper No 126. Washington. Clark, T., E. (1994). Nominal GDP Targeting Rules: can they stabilize economy? Federal Reserve Bank of Kansas City. Economic Review, Third Quarter 1994 Federal Reserve System and Monetary policy (1979). Congressional digest, Vol 58 issue 4: p101 2p (AN 10584253) Gidlow, R. M. (1998). Instrument of monetary policy in South Africa. south Africa Reserve Bank, January 1998. Government of Rwanda (2001). Third United Nations Conference on the Least Developed Countries. Brussels, 14-20 May 2001. Action Prgramme for the Development of Rwanda 2001-2010. Hsing, Y. (2004). Estimating the monetary policy reaction function for Canada. A VAR model. Southeastern Louisiane University. Vol.3 No.1 Appril (2004) Judd, J. P., Motley, B. (1993). Economic Review Federal Reserve Bank of SanFrancisco. 1993. Pages: 3-11. Judd, J. P. , Rudebusch, D. (1998). Taylor Rule and the Fed: 1970-1997. FRBSF Economic Review 1998, No3 Martinez, L. , Sanchez, A. , Werner, A. (2000). Monetary policy and the transmission mechanism in Mexico. BIS paper No8. Paper presented at the seminar on stabilization and monetary policy organized by the Bank of Mexico, November 2000. Mishkin, F. S., Posen, A. S., McDonough, W. (1997) Special Issue on Inflation Targeting. Economic Policy Review. Vol3, No3. Federal Reserve Bank of New-York, 1997. Mishkin, F. S. (1995). Symposia on the monetary transmission mechanism. The journal of Economic Perspective. Vol.9 No 4 (1995): 3-10 Musinguzi, P. , Opondo, H. (1999). Bank of Uganda Staff Paper. Vol 1. No 1. 1999. Musinguzi, P. , Obwona, M. B., Stryker, D. J. (2000). Monetary and exchange rate policy in Uganda. African Economic Policy Discussion Paper No23, April 2000. Republic of Rwanda (2000). Rwanda Development Indicator. Ministry of Finance, Kigali. Republic of Rwanda (2002). Rwanda Development Indicator. Ministry of Finace, Kigali Sanchez-Fung, J. R. (2002). Estimating a monetary policy reaction function for the Dominican Republic. Discussion Paper No 02/01 Department of Economics. University of Kent, UK. Small, M. M. ,de Jager S. The monetary transmission mechanism in South Africa. Occasional Paper No.16, September 2001, South African Reserve Bank. Stryker, J. D. , Rajaratnam, B. , Cassim, F. (2000) The effect of fiscal, monetary, and exchange rate policy on the South African growth and employment. TIPS 2000 Annual Forum September 2000. Svensson, L. E. O. Inflation Targeting as a Monetary Policy Rule. Journal of Monetary Economics 43 (june 1999), pp. 607-54. Taylor, J. B. The monetary transmission mechanism: an empirical framework. The Journal of Economic Perspectives, Vol 9 No 4 (1995): 11-26. Worrel, D. (2000). IMF Working Paper. Money and Fiscal Coordination in Small Open Economies. WP/00/56. International Monetary Fund. Internet Aron, J. , Muellbaner, J. (2001). Estimating monetary policy rules for South Africa. Oxford. ( www.nottingham.ac.u/economics/research/seminars/paper/monpolpdf.) Accessed March 2004 Banque Nationale du Rwanda. (2002). Politique Monetaire par la BNR: Rôle et perspectives. ( www.bnr.rw/semipol/presentation/musoni.ppt). Accessed May 2004 Bernank, B. Mishkin, F. S , Milhov. What does the Bundesbank target? Europen Economic Review, 4 (june 1997). ( www.nber.org/papers/w5764.pdf). Accessed july 2004 Christensen, A. M. , Nielsen, H. B. (2003). Has US monetary policy followed the Taylor Rule? A cointegration analysis. 1988-2002. JEL classification: C32, E52. ( www.edge-page.net/jamb2003/jambore-copenhagen-Nielsen.pdf). Accessed May 2004 Economic New School ( www.cepa.newschool.edu/het/essays/monetarism/monetransmission.htm). Accessed May 2004 Esanov, A. , Merkl, C. , de Souza, V. (2004). A preliminary evaluation of monetary policy rules for Russia. Duestembrooker Weg 120 24 105 Kiel. Kiel Working Paper No.1201. ( www.uni-kiel-de/ifw/pub/kap/2004/kap120.pdf). Accessed July 2004 Federal Reserve Bank of SanFrancisco (2004). US monetary policy: an introduction. ( www.frbsf.org/publications/federalreserve/monetarypolicy.pdf.). Accessed March 2004 Hsing Y. (2003). Estimating the Bank of Japan's monetary policy reaction function. Working Paper. ( www.selu.edu/orgs/ije/guidelines.htm-16k) Accessed June 2004 IMF (2000). Policy for faster growth and poverty reduction in Sub-Sahara African and the role of the IMF. An IMF Issues Brief. ( www.imf.org) Accessed March 2004 IMF (2001) IMF concludes article IV consultation with Rwanda. Public information notice (PIN) No.01/31 March 27, 2001. ( www.imf.org) Accessed March 2004 IMF and Rwanda (1998-2001). Rwanda-enhanced structure adjustment facility economic and financial policy framework paper for 1998-2001. ( www.imf.org) accessed March 2004 IMF and Rwanda (1997-2001): Press release index for 1997-2001. ( www.imf.org) Accessed March 2004 King, M. (2003). What has inflation targeting achieved? ( www.missouri.edu./`ecoprm/ec413f03/yhch_/5.pdf) Accessed March 2004 Osterholm, P. (2003). The Taylor rule: a spurious regression? JEL classification: E52. ( www.nek.uu.se/pdf/wp2003_20.pdf). Accessed May 2004 Poole, W. (1999). Monetary Policy Rules? The Federal Reserve Bank of St Louis Review, Issues 1999 March/April, 1999 vol.81, No2. ( www.research.stlouisfed.org/publications/ereview/99/03/9903wp.pdf) Accessed April 2004 Sanchez-Fung, J. R (2000). Estimating a Taylor-type monetary policy reaction function for a small developing economy. JEL classification: E5 ( www.stanford.edu/johntaylor/papers/how+do+monetary+authorities+react+in+the+DR.pdf). Accessed April 2004 Setlhare, L. A. (2003) close look at Botswana's management of monetary policy: the reaction function approach. ( www.essa.org.za./download/paper.2003.htm-42k) Accessed June 2004 Svensson, L. E. O. (1998). Inflation targeting as monetary policy rule. Working paper 6790 JEL No E42, E52. ( www.missouri.edu/`ecoprm/ec413f03/yhcha-Is.pdf) Accessed April 2004 Taylor, J. B. (1999). The monetary transmission mechanism and the evaluation of monetary policy rules. Third annual international conference of Central Bank of Chile on monetary policy rules and transmission mechanism. September 20-21, 1999 WP No.87 ( www.bcentral.ch/estudios/doctrab.htm) Accessed May 2004 Taylor, J. B. (1998). An historical analysis of monetary policy rules. Working Paper 6768. Cambridge: National Bureau of Economic Research. ( www.nber.org/papers/w6768) Accessed April 2004
Walsh, C. E. (2001). The science(and art) of monetary policy. Federal Reserve Bank of Sanfrancisco. Economic Letter 2001-13, May 4, 2001. ( www.frbsf.org/publications/economics/letter/2001/el2001-13.html) Accessed April 2004 Worrel, D. (2000). Monetary and fiscal coordination in small open economies. IMF working paper. WPI/00/56 JEL classification No.E63, E58 ( www.imf.org/external/pubs/ft/wp/2000/wp0056.pdf) Accessed March 2004 APPENDIX1: PRELIMINARY DATA (1995Q1-2001Q4)
Source National Bank of Rwanda, Ministry of Finance and IMF
|
|
