Volume 12 Issue 4 Article 3 12-31-2010 The impact of monetary policy and exchange rate regime on real GDP and prices in the Republic of Macedonia Besnik Fetai Zeqiri Izet Follow this and additional works at: https://www.ebrjournal.net/home Recommended Citation Fetai, B., & Izet, Z. (2010). The impact of monetary policy and exchange rate regime on real GDP and prices in the Republic of Macedonia. Economic and Business Review, 12(4). https://doi.org/10.15458/ 2335-4216.1256 This Original Article is brought to you for free and open access by Economic and Business Review. It has been accepted for inclusion in Economic and Business Review by an authorized editor of Economic and Business Review. THE IMPACT OF MONETARY POLICY AND EXCHANGE RATE REGIME ON REAL GDP AND PRICES IN THE REPUBLIC OF MACEDONIA BESNIK FETAI* IZET ZEQIRI** ABSTRACT: Th is paper investigates the relative costs and benefi ts associated with intro- ducing a more active monetary and a diff erent exchange rate regime in the Republic of Macedonia. In this fi nding, the econometrics result show that introducing a more active monetary policy and a diff erent strategy of the exchange rate targeting in order to promote rapid economic growth could easy disturb macroeconomic stability (aft er having achieved it at a substantial cost) without any signifi cant economic benefi ts. Th erefore, introducing a more active monetary policy and a diff erent strategy of the exchange rate regime is likely to incur more costs than benefi ts, since changes of the monetary policy and exchange rate regime type do not show a persistent eff ect on real GDP, while changes of money stock and exchange rate regime do show a strong and persistent eff ect on prices level. Key words: money stock, exchange rate, Structural Vector Autoregressive JEL classifi cation: M31 1. INTRODUCTION Monetary policy and its eff ect on real economic activity have traditionally attracted great attention from many researchers. It is well established that, in the long term, changes in monetary policy will aff ect price levels, i.e. the rate of infl ation. Th erefore, economists agree that the main long-term goal of monetary policy should be to main- tain low and stable price levels (Abel, Bernanke and Smith, 2003). However, in the short term, monetary policy remains powerful tools for aff ecting real economic activ- ity via several channels. * Southeast European University, Faculty of Business and Economics, Ilindenska bb, 1200, Tetovo, Macedo- nia, e-mail: b.fetai@seeu.edu.mk ** Southeast European University, Faculty of Business and Economics, Ilindenska bb, 1200 Tetovo, Macedo- nia, e-mail: i.zeqiri@seeu.edu.mk ECONOMIC AND BUSINESS REVIEW | VOL. 12 | No. 4 | 2010 | 263–284 ECONOMIC AND BUSINESS REVIEW | VOL. 12 | No. 4 | 2010264 Th e theory and the empirical evidence (based on the Structural Vector Autoregressive Model approach) have established that there are several channels through which the ef- fects of monetary policy are transmitted to real economic activity (though much of this evidence comes from developed countries). In contrast to the conventional theory and empirical evidence from developed economies, both the theory and the empirical evi- dence pertaining to countries in transition suggest a potential weakness and a potential instability of the conventional channels (the short-term interest rate and the base money, and, through it, the money stock) of monetary transmission during transition because of structural and institutional defi ciencies – in particular underdeveloped fi nancial sys- tems and higher dollarization. Additionally, the empirical literature suggests that the monetary transmission mechanism is diff erent among countries in transition depend- ing on characteristics of individual national economies, such as size, openness, develop- ment of the fi nancial sector, the level of dollarization, and other factors. In addition, as transition is a dynamic trend marked by permanent qualitative changes, the literature suggests that the monetary transmission mechanism is an endogenous or dependent variable with regular features that is changeable over time. Th is point needs to ongo- ing research to further analyze the eff ect of monetary policy on real GDP and prices, in order to assess the relative costs and benefi ts associated with introducing a more active monetary policy. Regarding the exchange rate regime, the question of the optimal monetary regime for small open economies is still unanswered (Ribnikar, 2004). According to Ribnikar, there is no optimal monetary regime; it depends on the circumstances of the country. For in- stance, in Slovenia there was a managed fl oating exchange rate. Economists have not been able to determine whether these countries should use fl oating or fi xed exchange rates. With respect to such countries in transition, the exchange rate has oft en played a funda- mental role in macroeconomic stabilization. However, in recent years, globalization and changes in policy orientation have resulted in closer international trade and fi nancial linkages, which in turn have led to mobility of capital, i.e. capital infl ow and outfl ows, generating potential external shocks and increasing the pressure for additional fl exibil- ity. Th e Republic of Macedonia, in particular, has faced many systemic changes, such as the liberalization of the capital account in 2003, becoming a member of the World Trade Organization in 2002, and gaining candidacy status for joining the European Union in 2004. Under such circumstances, any investigation of monetary policy and exchange rate regime must address a seemingly incompatible trinity: the liberalization of capi- tal movement, fi xed exchange rates, and independent monetary policy (Obstfeld 1998 and Mishkin 2003). Since the liberalization of capital accounts took place, the exchange rate could easily become a target of speculative attack (sudden large capital infl ows), which in turn could lead to negative impacts on real economic activity due to increases in and fl uctuation of the interest rate and fl uctuation of foreign exchange reserve – for- eign exchange reserve being important for international liquidity. On the other hand, the solution of simply shift ing the exchange rate from a fi xed exchange rate to a more fl exible one, depreciating the domestic currency in order to settle the problem of defi cit in the current account, and thereby promoting fast economic growth, could easily dis- turb macroeconomic stability without any real short-term economic benefi ts (Ribnikar B. FETAI, I. ZEQIRI | THE IMPACT OF MONETARY POLICY AND EXCHANGE RATE REGIME ON ... 265 and Bole, 2006). Th is suggests a need for ongoing analyses of the eff ects of exchange rate policy on real GDP and prices in order to assess the relative costs and benefi ts associated with introducing diff erent exchange rate regimes. Although the role of monetary policy and exchange rate regime type is limited in in- fl uencing economic growth, the role of such policies and regimes in aff ecting economic growth – especially the costs and benefi ts of introducing a more active monetary and a diff erent exchange rate regime (one using infl ation targeting) – has received growing at- tention in the Republic of Macedonia. Th erefore, in this paper, I focus on identifying the eff ects that the monetary policy and exchange rate regime have on real GDP and prices in the Republic of Macedonia. Based on the data from 1997 to 2008, my empirical research is supported by empirical testing using the most-used methodology, such as SVAR. Th e theoretical and empirical litera- ture concerning SVAR, both in the developed countries and the countries in transition, provides the foundation of my empirical research on the Republic of Macedonia. Th e reminder paper is organized as a follows: Section II explains the reviews the litera- ture on the eff ect of the monetary policy and exchange rate regime; Section III econo- metrics model for testing the eff ect of money stock and exchange rate on real GDP and prices; Section IV conclusions. 2. REVIEW OF LITERATURE Th e eff ective setting and manage of monetary policy depend basically upon proper as- sessment and understanding of the eff ect of monetary policy on relevant macroeconomic variables (such as real GDP and infl ation) through various monetary channels. Th eory and empirical evidence have established the existence of several transmission mecha- nisms or channels through which monetary policy can aff ect real economic activity. For the Republic of Macedonia as a country in transition, we examine the well-known con- ventional monetary channel by which the eff ect of monetary policy is transmitted to real economic activity: the money supply. We do not take into the consideration, the interest rate channels, since, the money market is not eff ectively functioning in the Republic of Macedonia, and they would not accurately refl ect, market-type behaviour, so we design a model utilizing the money stock and exchange rate. Regarding the exchange rate channel, we include the exchange rate into the model since the majority of the empirical research confi rms that exchange rate channels seem to play an important role in the monetary transmission mechanism in countries in transition. Th us, with regard to the Republic of Macedonia and other similar countries in transition, it is impossible to evaluate the eff ect of monetary policy without taking into account the exchange regime. Th e exchange rate channel is important in the monetary transmission mechanism for several reasons, e.g. the credibility of the exchange rates for the economic agents and the actual behaviour of the exchange rate for the level of currency substitu- ECONOMIC AND BUSINESS REVIEW | VOL. 12 | No. 4 | 2010266 tion. Eduardo and Berg (2000) show that the higher the level of dollarization or currency substitution in a country, the less eff ective will be the traditional set of monetary policies of the central bank. Actions on the part of the monetary authority relating to money market rates, reserve requirements, and refi nancing may turn out to have a insignifi cant eff ect on real GDP and infl ation. Th e empirical evidence in developed countries is mixed relating to the eff ect of money supply on real GDP and prices. Recently, Giovanni and Gordani (2006) reconsider the role of money in output and prices in the U.S. and their results suggest that shocks to monetary aggregates have substantial and persistent eff ects on output and prices. Exam- ining the economy of New Zealand, Cîtu (2003) fi nds a similar pattern to those seen in the developed countries, whereby money has an eff ect on output and prices in the short term. Applying VAR methodology, Hafer and Kutan (2001) examine 20 countries - in- cluding developed economies and countries in transition - and their results also suggest that money plays a signifi cant role in explaining the behaviour of real output. Th e empirical evidence in transitional countries shows that the money supply channel is relatively weak in transmitting the dynamic eff ect of monetary policy on real economic activity. However, I believe that the money supply constitutes a relatively useful indicator in measuring the dynamic eff ect of monetary policy on real GDP and prices in Republic of Macedonia. Starr (2004) examines four countries in transition using VAR methodol- ogy, and reaches the conclusion that money shocks do tend to increase output although the eff ect is not statistically signifi cant. Th at is, she fi nds the eff ects of money supply on output in the Ukraine and Belarus to be insignifi cant, while noting a transitory real eff ect on output in Russia and Kazakhstan. Belullo’s research (1999) into the impact of money on real economic activity fi nds that expansionary monetary policy does not have an eff ect on real economic activity in Croatia. Hristov (2004) examines two countries in transition (Th e Czech Republic and Poland) and fi nds that contractionary money shocks lead to persistent changes in the prices and onset of a decline in output in both countries. Regarding exchange rate channel: the early literature on the choice of exchange rate re- gimes proposes that the smaller and more “open” an economy is (i.e. the more depend- ent it is upon exports and imports) the better it will be served by the adoption of a fi xed exchange rate regime. A later approach to the choice of exchange rate regimes looks at the eff ects of various random disturbances on the domestic economy. In general, a fi xed exchange rate regime is preferable if the disturbance in the economy is predominantly monetary, for example in the form of changes in the demand for money which aff ect the general level of prices. Th e main reason why the Republic of Macedonia and other similar small countries in transition have pegged their currencies to their leading trading partner’s currency is the unstable demand for money. Th e central banks in those countries were experienced dif- fi culties in achieving their fi nal goals of maintaining price stability through monetary strategy targeting the growth rate of money; thus, the Republic of Macedonia abandoned this strategy in 1995 and adopted a monetary strategy targeting the exchange rate. B. FETAI, I. ZEQIRI | THE IMPACT OF MONETARY POLICY AND EXCHANGE RATE REGIME ON ... 267 A fl exible exchange rate is preferable if disturbances are predominantly real factors or factors that originate abroad and aff ect the relative prices of domestic goods. In addition, the literature in general indicates that a small open economy is better served by a fi xed ex- change rate than by a fl exible exchange rate regime. Most economists claim that the best exchange rate regime is one that stabilizes macroeconomic performance, i.e. one that minimizes fl uctuations in output, domestic prices and other macroeconomic variables. Th e question of the optimal monetary regime for small open economies has yet to be defi nitively answered, however. Following the path of most literature regarding U.S. and Western Europe countries, the few VAR studies which have been carried out on Central Eastern European and South Eastern European countries use the same identifi cation in their models regarding the eff ect of exchange rate regimes on real GDP and infl ation. McCarthy (2000) analyzes the eff ect of exchange rate changes and import price fl uc- tuation on producer and consumer prices in six industrialized OECD countries from 1976:1 to 1998:4. Th e impulse response function and variance decomposition show that the exchange rate has had a modest eff ect on domestic prices throughout the post- Bretton Woods era. He also fi nds that pass-through is somewhat stronger in countries with larger import shares. It can be seen that in large developed countries, the exchange rate channel does not play a signifi cant role in the transmission of the dynamic eff ect of monetary policy. Cushman and Zha (1997) analyze monetary policy in Canada and fi nd that the dynamic responses to identifi ed monetary policy shocks are consistent with standard theory and highlight the exchange rate as a transmission channel of monetary policy. Kim and Rubini (2000) and Maćkowiak (2003) analyze small developed countries and fi nd that the exchange rate channel did play an important role in transmitting the dynamic eff ect of monetary policy. Cîtu (2003) examines New Zealand and fi nds that exchange rate channels play an important role in the transmission of the dynamic ef- fect of monetary policy. By this empirical evidence, it can be seen that in the big and developed countries the exchange rate does not play a signifi cant role in transmitting the dynamic eff ect of monetary shock on real economic activity, while it does play such a role in smaller developed countries. Recently, exchange rate channels have been examined in two ways in the small econo- mies in transition: fi rstly, in terms of the pass-through eff ect of nominal exchange rate changes, via import prices, on prices in small and open economies, whereby a deprecia- tion of domestic currency causes price level to rise; secondly, in terms of the possible implications of diff erent exchange rate regimes on monetary strategy. Billmeier and Bonato (2002) examine Croatia using VAR and VECM with the model in- cluding manufacturing and retail indices, the exchange rate nominal anchor, monetary aggregate, and output gap. Th ey fi nd a signifi cant role for the exchange rate in the level of prices. Kuijs (2002) fi nds almost similar results for Slovakia. Hristov (2004) discovers an interesting result by analyzing monetary and exchange rate regime in two post-transition countries, the Czech Republic and Poland. By 1997, those countries had changed their ECONOMIC AND BUSINESS REVIEW | VOL. 12 | No. 4 | 2010268 monetary strategy from one of exchange rate targeting to infl ation targeting. Hristov fi nds that the changes in the exchange rate regimes did not produce marked diff erences in macroeconomic variables. Th us, he claims that the behaviour of macroeconomic vari- ables remains constant irrespective of which regime is adopted. Mayes (2003) analyzes the Baltic States and fi nds that the exchange rate does not show any eff ect on real GDP in the Latvia case, whereas it has transitory eff ects on real GDP in the cases of Estonia and Lithuania. Horváth and Maino (2006,) examine the dynamic eff ect of monetary policy on real GDP and prices in Belarus. Th e model they employ incorporates four variables: price, money, exchange rate peg, and real GDP. Th ey fi nd that exchange rates have a strong eff ect on prices – i.e. a strong pass-through eff ect – but do not aff ect real GDP. In addition, money causes infl ation but does not have an eff ect on real GDP. My model is very similar to the model employed by Horváth and Maino (2006), Billmeier and Bonato (2002), and Kuijs (2002). Th e Republic of Macedonia, Croatia, and Belarus all have the same monetary strategy of targeting the exchange rate. In the Republic of Macedonia, like in the other small countries in transition, I expect the exchange rate channel to be more powerful than money stock channel in the monetary transmission mechanism. 3. ECONOMETRIC MODEL FOR TESTING SHORT-TERM DYNAMIC EFFECTS OF MONEY STOCK AND EXCHANGE RATES ON REAL GDP AND PRICES IN THE REPUBLIC OF MACEDONIA Both theory and empirical evidence suggest that exchange rate channels seem to play an important role in the monetary transmission mechanism in almost all countries in transition. In section II, we explain that both theoretical and empirical research ef- forts imply a potential weakness and a potential instability of the conventional channel (money stock and short-term interest rate) of monetary transmission during transi- tion. We introduce exchange rate channels into the model in order to get more infor- mation about the eff ect of money stock or money supply on real economic activity in the Republic of Macedonia. At the same time in this empirical research, two channels of the dynamic eff ect of exchange rate can be identifi ed. First, the direct channel of the exchange rate: that it aff ects infl ation via the import prices pass-through eff ect. Th at is, changes in the nominal exchange rates directly aff ect import prices, which in turn cause domestic prices to rise. Second, the indirect channel of the exchange rate: that it aff ects real GDP through the balance of payments. It is important to recognize these possible implications of diff erent exchange rate regimes, i.e. the costs and benefi ts of introducing a diff erent exchange rate regime. Th e issue at hand is whether the ex- change rate still plays a signifi cant role in maintaining macroeconomic stability in the Republic of Macedonia. For this purposes, we perform several tests: diagnostic tests (JB-test, LB-test and ARCH- test); a test for VAR order; a test for reaction of the real GDP and prices to money stock and exchange rate disturbances, i.e. the dynamic eff ect of money stock and exchange rate B. FETAI, I. ZEQIRI | THE IMPACT OF MONETARY POLICY AND EXCHANGE RATE REGIME ON ... 269 disturbances on real GDP and prices; and forecast error variance decomposition test of money stock and exchange rate disturbance on real GDP and prices. 3.1 Data in empirical research We are limited to using data from the period of 1997:01-2008:12, in our work, because if we were to use time series before 1997, we think that our research would be of lesser qual- ity due to high rate of infl ation experience in the period 1992-1995. Because of the narrow time series in countries in transition, we use monthly data rather than quarterly in order to perform a more observation points. Almost all empirical research involving countries in transition follows this approach, owing to the short time periods inherent to such evalu- ation (excepting Mayes, 2003, who used annual data). Bernanke and Mihov (1998); and Christiano, Eichenbaum and Evans (1996 and 1999) furthermore show that inferences drawn from quarterly data are congruent with inferences gathered from monthly data. Th e variables used in our model are: manufacturing prices index (MPI), retail prices in- dex (RPI), money stock (M1), exchange rate (EXCH.R), and real gross domestic product (real GDP). All data is expressed in logarithmic form (denoted ln). Hence, coeffi cients on the logged levels measure constants of elasticity. Th e short-term interest rate has no in- cluded in the model since it does not refl ect the market type behaviour in the Republic of Macedonia. Th e short term interest rate has no means to be included in the model (Bill- meier and Bonato 2002) when countries use currency boards or employ the exchange rate as a nominal anchor. Th e sources of data are mainly from the NBRM, the Ministry of Finance, and the Offi cial State Statistics Bureau of the Republic of Macedonia. 3.2 Econometrics model and result 3.2.1 Testing the short-term dynamic eff ect of money stock and exchange rate on real GDP and prices: SVAR Sims-approach Th e empirical research has some features in common with previous empirical research regarding both developed countries and countries in transition. Th e specifi cation of the model is: We start with identifying our model as a fi ve-dimensional vector. In the equation above, the vector xt includes fi ve variables: the exchange rate (EURO), the manufacturing prices index (MPI), the retail prices index (RPI), the money stock (M1), and the real gross do- mestic product (real GDP). Th e is the vector of the constant. All variables are expressed in logarithmic form to satisfy the theoretical assumptions of constant elasticity models. ECONOMIC AND BUSINESS REVIEW | VOL. 12 | No. 4 | 2010270 During the period of investigation, there are notable structural monetary shocks and episodes of internal and external political turbulence. In order to eliminate the negative impact of internal and external shocks, we include several vector dummy variables, such as: the devaluation of the denar by 16.1 percent, the war between Serbia and Kosovo, the value-added tax, and ethnic confl ict in the Republic of Macedonia. Th e structural model is composed of fi ve equations. Moreover, the series are estimated consistently in levels with OLS (ordinary least square). Th e variables in the model are divided into two blocks: the non-policy vectors, including the log of MPI, the log of RPI, and the log of real GDP; and the policy vectors, including the log of M1 and the log of EXCH.E. Vector et is the vector of structural disturbance. MPI and RPI are included in the model for two reasons: First, we expect a strong link to emerge between the exchange rate and RPI and MPI. Second, the fi nal goal of the NBRM is price stability; therefore, introducing these vari- ables contributes to eliminating the so-called “price puzzle”. In our empirical research we use Choleski which is just identifi ed, and the number of coeffi cients of the matrix B0 is 10 (lower triangular), which can be estimated in the mone- tary VAR (M-VAR henceforth) with unity on the main diagonal. Th ereby, the covariance of the matrix will be a diagonal matrix. Before examining impulse response function and variance decomposition of monetary disturbance, we must select VAR order and perform diagnostic tests. To begin, we make a visual inspection of the time series. All of the time series show trends, with the exception of exchange rate, which is fi xed with the exception of devalua- tion in 1997. We are interested in performing tests for seasonality, i.e. whether or not the time series exhibit seasonality. Th e model is explained in detail in Appendix I. In Appendix I, we show the model proposed by Gardner (1985) in order to reveal which time series shows seasonality. As we can see from the visual inspection of the time series, only real GDP displays seasonality. Th is conclusion leads also to the result of Gardner’s methodology. Hence, we make a seasonal adjustment of real GDP, and in my further research, we use seasonally adjusted real GDP. On the other hand, MPI, RPI, M1, and EXCH.E do not display seasonality, and therefore we use these time series without sea- sonal adjustments. To continue, we include these time series only in VAR level (Sims-approach) – estimated by OLS. In Appendix II, we show the result of the routine tests of VAR: tests for selection of VAR order and for checking the “quality” of the VAR based on its residuals, such as: JB- test for normality distribution, LB-test and LM-test for autocorrelation, and ARCH test for the presence of heteroscedasticity in the VAR’s residuals. We test VAR-order according to criteria such as FPE, AIC, HQ and SC. Using these criteria, we select appropriate models that best fi t the data. Th e tests show that by all B. FETAI, I. ZEQIRI | THE IMPACT OF MONETARY POLICY AND EXCHANGE RATE REGIME ON ... 271 criteria the optimal order is VAR (2). Th erefore, we use VAR of order 2 in my further research. Concerning the matrix of residual correlation of diagonal elements, they are rather close to zero, such that no contemporaneous correlation is being ignored by the VAR. Th erefore, we can conclude that: there are contemporaneous correlations residual between MPI ,̀ RPI, M1, IR and real GDP or contemporaneous and intertemporal cor- relation between the residuals of the variables. Upon visual inspection, the residuals of almost all series exhibit a number of statisti- cal outliers, such that we do expect signifi cant non-normality. On the other hand, the result of the JB-test for normality distribution of residuals shows that the H0 hypothesis of normality distribution cannot be rejected for MPI, RPI and real GDP at a signifi cance level of 5%, while it is rejected for EXCH.E and M1. Th ere are problems with the EXCH. E. and M1 equations; however, the time series in the short term is quite sure to have this kind of problem since it includes monthly data with a great deal of noise. Furthermore, in models with many points of observation, there are oft en instances of non-normality of distribution of the residuals. Concerning the LB-test for non-signifi cant residual au- tocorrelation, there are no statistically signifi cant autocorrelated residuals and no visible patterns (see appendix II). Th e ARCH-test strongly rejects the assumption of hetero- scedacity of VARs residuals. Finally, we can conclude that, despite an unstable VAR due to the inclusion of non-sta- tionary time series in the model, the diagnostic test is satisfactory and consistent with the assumption of white noise process with constant variance over time. 3.2.1.1 Th e dynamic eff ect of money stock and exchange rate disturbance on real GDP and prices: Choleski decomposition We employ a recursive VAR approach in order to analyze the dynamic eff ect of money stock and exchange rate on real GDP and prices. In addition, we use Choleski decompo- sition which is consistent with the methods used by McCarthy (2000), and Cîtu (2003) in their research on developed countries, and also with the methods used by Billmeier and Bonato (2002), Mayes (2003), and Horváth and Maino (2006) in examining small countries in transition via the recursive VAR approach or Choleski decomposition. In order to identify shocks or their respective impulse-response functions via Choleski decomposition, the variables need to be given a plausible ordering. Following McCarthy (2000), we assume a recursive ordering with some small modifi cations, mostly due to the diff erent characteristics of the national economy. Th e aforementioned authors assume that international supply shocks are exogenous shocks to the exchange rate by way of import prices. Shocks in the exchange rate will instantly aff ect the manufacturing prices index and retail prices index, whereas the central bank reaction function with money stock is ranked at the end of the ordering of the variables (Cîtu, 2003). Some research regarding countries in transition has employed shocks in the oil price index or exchange rate as the fi rst variable (Billmeier and Bonato 2002). Th erefore, with respect to the afore- ECONOMIC AND BUSINESS REVIEW | VOL. 12 | No. 4 | 2010272 mentioned paper, my own methodology diff ers in that we do not include a measure of import prices due to a lack of relevant data in the Republic of Macedonia. Billmeier and Bonato (2002) have the same problem, and he also omits import prices (in the case of Croatia). Given that Macedonia is a small open economy, and an insignifi cant power in the world market, we expect the transmission of import prices to be complete over a rather short time horizon. In contrast to the work of McCarthy (2000), which analyzes only the eff ect of the ex- change rate on prices, we follow Horváth and Maino (2006) and other authors who fo- cus on countries in transition, and we include real GDP and prices. Th at is, the direct channel of the eff ect of exchange rate on prices and the indirect channel of the eff ect of exchange rate on real GDP through the balance of payments. We are interested in analyz- ing both channels of the dynamic eff ect of exchange rate on real economic activity in the Republic of Macedonia. It is assumed that an unexpected change in the exchange rate instantly aff ects the manu- facturing prices index, while MPI instantly aff ects RPI. Since the central bank in Re- public of Macedonia addresses its fi nal goal of price stability through the exchange rate (EXCH.E-VAR), it reacts instantly to changes in the exchange rate and price indices with its operative targets: the base money and through it on the money stock M1. Moreover, according to McCarthy (2000), and Cîtu (2003), central banks react to changes in the exchange rate and prices indices. Concerning the ordering of the variables to follow the exchange rate, we address this in the fi rst section (monetary policy M-VAR). Th e ordering of the variables is as follows: Th e recursive approach (Choleski decomposition) is constructed like this: Th e fi rst period of the matrix shows that any unexpected change in the exchange rate will instantly aff ect the two prices indices, and then the central bank responds by the operational target M1 in order to maintain exchange rate stability as an intermediate target. Hence, by stabilizing the exchange rate, the central bank achieves the main goal of monetary policy: price stability. Th us, we can assume that the other variables do not react instantly to M1, but that M1 can instantly aff ect real GDP. In addition, real GDP can be aff ected by both channels – the money stock and the exchange rate. B. FETAI, I. ZEQIRI | THE IMPACT OF MONETARY POLICY AND EXCHANGE RATE REGIME ON ... 273 In addition, it can be seen from the matrix that the Choleski decomposition model is just identifi ed and that the number of coeffi cients of matrix B0 are 10 (lower triangle), which can be estimated in the VAR with unity on the main diagonal. Th erefore, the covariance of the matrix is a diagonal matrix. All variables are logarithmic and are estimated by the OLS, which produces residuals that are uncorrelated across the equations. Th e er- ror bands (interval of confi dence) corresponding to 95 percent probability intervals are computed by a Monte-Carlo simulation, following the methodology suggested by Sims and Zha (1999). Money supply channel as an indicator of the dynamic eff ect of monetary policy on real GDP and prices: Choleski-decomposition-M1-VAR-level Th e dynamic eff ects of money stock shock (disturbance) on real GDP and prices are reported in Figure 1 below. Th e vertical axis denotes the response of log MPI, log RPI, log M1, and log real GDP to a one percent shock in the money stock. Th e horizontal axis denotes time in months. As seen in Figure 1, real GDP response is insignifi cant to any money stock shock. Th at is, shock in the money stock does not generate a sig- nifi cant eff ect on real GDP. However, the response of the retail price index is signifi - cant at around 0.35 percent for 18 months. Th at is, money stock shock can cause an increase in the price level. As a result, we can conclude that dynamic eff ect of money stock does not have a signifi cant impact on real GDP, but that it can aff ect the price level. Furthermore, this result confi rms that money stock has an important infl uence in deter- mining infl ation in the Republic of Macedonia. My conclusion is in line with most fi nd- ings concerning countries in transition, e.g. Belullo (1999); Horváth and Maino (2006); Starr (2004); who all funded that money stock does not have an eff ect on real GDP, but that it does aff ect price level. Th is fi nding is also consistent with most other fi ndings regarding countries in transition, e.g.; Ceccheti and Krause (2001); and Elbourne, Kiviet and Bas (2003) who all conclude that the action of monetary policy in transitional economies may render traditional pol- icy tools less eff ective than a neoclassical view would suggest. Th is result also is in line with the monetarist view that an increase of money growth by one percent causes the price level to rise by 0.35 percent. Moreover, this result also supports the view that the pri- mary role of monetary policy should be to control infl ation in the Republic of Macedonia, since the money stock does not show a signifi cant eff ect on real GDP, while it highlights a strong eff ect on price level. ECONOMIC AND BUSINESS REVIEW | VOL. 12 | No. 4 | 2010274 FIGURE 1: Dynamic eff ect of money stock on real GDP and prices: Choleski decomposition M1-VAR level Source: Author’s calculations Th e result shows that the base money and through it money stock is endogenous to the level of infl ation in the Republic of Macedonia. Figure 1 shows that the price level can – in the absence of changes in the other variables – are returned to its baseline trend within 18 months via the endogenous money stock adjustment. Th is result is consistent with my fi nding that the base money and through it the money stock in the period of investiga- tion is an endogenous variable that adjusts to the demand for money, and therefore in this model it is driven by infl ation. We expect this empirical research to demonstrate that the money stock channel is weak as an independent channel of monetary policy in the Republic of Macedonia. Th is is a consequence of the fact that banking and fi nancial sectors are still characterized by shallow levels of fi nancial intermediation, the fi nancial sector is underdeveloped, the banking sector suff ers from low levels of competition, and the economy has a high de- gree of dollarization. Th us, the impact of money supply on economic outcomes does not B. FETAI, I. ZEQIRI | THE IMPACT OF MONETARY POLICY AND EXCHANGE RATE REGIME ON ... 275 yet operate in the Republic of Macedonia the same way as it does in developed countries, namely by the asset price eff ect, the wealth eff ect, the bank-lending eff ect, and the fi rms’ balance sheet eff ect. Particularly, the asset price eff ect does not work in Republic of Mac- edonia, since assets such as bonds, shares, real estate, and other domestic assets are not closely tied to economic outcomes in the Republic of Macedonia, so little impact is felt from changes in the values of these assets, which may arise from changes in the money stock. One reason domestic assets are not closely tied to economic outcomes in the Republic of Macedonia is that the government sells treasury bills only to the banking system, not to households and fi rms. Such examples of structural and institutional defi cien- cies, in particular underdeveloped fi nancial system, support the reasoning of most authors in explaining why monetary transmission mechanism in countries in transi- tion is weak (Elbourne Kiviet and Bas, 2003; Krause, 1999; Ceccheti and Krause, 2001; and Juks, 2004, ). Th is result is also consistent with other fi ndings regarding economies with signifi cant degrees of currency substitution. In this context, higher dollarization in the Republic of Macedonia can also weaken the eff ect of expansionary monetary policy on bank lending channels. An increase in the money stock in the Republic of Macedonia causes a decrease of foreign exchange reserves in the foreign exchange market, as a result of currency sub- stitution between domestic currency and foreign currency, and this process may lead to deterioration of foreign exchange reserves. As a result, an increase in the money stock will not boost domestic credit, but it can leak in the form of capital outfl ows, which results in very little or no increase in the amount of credit the banking system extends to the private sector. In addition, the dollarization in the Republic of Macedonia is mo- tivated by asset substitution, both real and fi nancial assets. Many prices of real estate and consumer durable goods are to some extent indexed to foreign currency, and resi- dents use the foreign currency (as domestic currency) for buying and selling real estate (houses, lending, cars). As to fi nancial assets, residents deposit or hold large proportions of their savings in foreign currency deposits either in the banking system or outside of the banks1, and banks provide loans that are either denominated in foreign currency or indexed to foreign currency. Under such circumstances, the scope of monetary policy to increase exogenously the money supply over the demand for money in order to maximiz- ing real GDP is limited. Th erefore, an increase of money stock does not mean an increase of the purchasing power, but rather only a substitution of currencies. In summary, an increase in the money supply does not have a signifi cant eff ect on real GDP by either the asset price eff ect, the wealth eff ect, the bank-lending eff ect, or the fi rms’ balance sheet eff ect, but it will cause a decrease in the level of foreign exchange reserves in the foreign exchange market, as a result of currency substitution between domestic currency and foreign currency in the Republic of Macedonia. 1 Perhaps, as the people have had problem with exchange rate from 1992 to the middle of 1997, they do not trust domestic currency, and so many of them keep their savings in foreign currency. ECONOMIC AND BUSINESS REVIEW | VOL. 12 | No. 4 | 2010276 Exchange rate channel as an indicator of the dynamic eff ect of exchange rate regime on real GDP and prices: Choleski-decomposition-EXCH.E-VAR-level Th e eff ects of exchange rate shock or disturbances are displayed in Figure 2. Th e vertical axis denotes the response of log MPI, log RPI, log M1, and log real GDP to a one percent shock in the exchange rates by the initial period. Th e horizontal axis denotes time in months. Th e dynamic eff ect of a one percent shock in the exchange rate generates a permanent increase in manufacturing and retail prices and money stock, whereas the response of real GDP is insignifi cant. Th e shock in the exchange rate triggers: (i) a depreciation of the exchange rate; (ii) a sharp and rapid increase of the manufacturing prices index; (iii) an increase of the retail prices index; (iv) an insignifi cant eff ect on real GDP; and (v) an increase of money stock. Th e manufacturing prices index responds in the fi rst month, and thereaft er it shows a permanent increase aft er twenty-four months of around 0.52 percent. Th e retail prices index shows a signifi cant response, and it continues to show a permanent increase aft er twenty-four months of around 0.59 percent. Money stock shows a signifi cant eff ect in the fi rst month and a permanent increase aft er twenty-four months of around 0.32 percent. Th e results of the empirical research suggest that the direct channel of the exchange rate has a strong pass-through eff ect on prices, but that the indirect channel of the exchange rate does not have an eff ect on real GDP. Figure 2 highlights this potentially strong pass- through exchange rate eff ect on prices in the Republic of Macedonia. Th ere is strong transmission of the eff ect of changes in the nominal exchange rate via import prices to prices in the economy; therefore, a deprecation of domestic currency causes price levels to rise approximately 0.59 percent. Even within the fi rst month, manufacturing prices react to changes in the nominal exchange rate, which refl ects a strong pass-through ef- fect of exchange rate changes into domestic prices via import prices. On the other hand, a depreciation of the domestic currency does not show a signifi cant eff ect on real GDP in Republic of Macedonia. In addition, the fi ndings are consistent with those from other empirical studies in that the monetary transmission mechanisms are diff erent between developed and small countries in transition. Due to relatively high dollarization of the domestic economy (as- set substitution), we see a large pass-through eff ect from the exchange rate to prices. In contrast to the results of McCarthy (2000), which show that changes in the exchange rate have a modest eff ect on domestic prices in developed countries, the empirical evidence regarding countries in transition does not seem to suggest the same is true for them, pos- sibly due to a lack of credibility of the monetary authorities, higher dollarization levels and/or the structural element of the price-taking nature of the fi rms in the international market. My result is therefore consistent with most other fi ndings regarding small countries in transition, such as those from: Billmeier and Bonato (2002), Kuijs (2002), Coricelli and B. FETAI, I. ZEQIRI | THE IMPACT OF MONETARY POLICY AND EXCHANGE RATE REGIME ON ... 277 Jazbec et al.(2004), Mayes (2003), and Horváth and Maino (2006). Results from the em- pirical research of these authors is covered in the review of the literature on exchange rate regime type, where we mention that they fi nd that exchange rate channels play a more signifi cant role than do other channels in transmitting the dynamic eff ect of mon- etary policy to real GDP and prices. FIGURE 2: Dynamic eff ect of exchange rate on real GDP and prices: Choleski decomposi- tion EXCH.E-VAR-level Source: Author’s calculations ECONOMIC AND BUSINESS REVIEW | VOL. 12 | No. 4 | 2010278 Th e result refl ects the NBRM’s monetary strategy of targeting the exchange rate. Figure 2 shows that the prices level cannot be returned back to its baseline trend by endogenous ex- change rate adjustments. Th us, this suggests that the exchange rate has been directly deter- mined more by the NBRM than by the prices or real GDP – for example devaluation of the denar against the deutschmark in 1997. Also, this fi nding is consistent with the defi ciency of short-term economic determinants of the exchange rate. Hence, in the absence of changes in the other variables, the prices level can be returned to its baseline trend gradually through changes in the rate of infl ation and more rapidly by money supply adjustment. As an alter- native, the prices level has to be brought into line with the exchange rate target. Th e result is consistent with the evidence that the base money and through it the money stock during the period of investigation is a predetermined endogenous variable, so employing rapid endog- enous adjustments of money supply in order to return the prices level to its equilibrium is also consistent with endogenety of money supply to exchange rate targeting. Th erefore, the money supply is a predetermined endogenous variable to infl ation and exchange rate move- ment through the NBRM’s intervention in the foreign exchange market. In addition, I expected a potentially strong pass-through eff ect of exchange rate on prices in the Republic of Macedonia. Th is is made clear when assessing the character- istics of the Republic of Macedonia’s economy, such as: small open economy, high de- gree of dollarization at around 51.50 percent (see the topic of dollarization in the second chapter), a large imports share (particularly of raw materials), and the lack of infl uence of the Republic of Macedonia in the world economy. Moreover, many prices, mainly of property and consumer durable goods, are to some extent indexed to the exchange rate. Wages are even indexed to the exchange rate in some economic sectors. Th erefore, my result is consistent with the features of the national economy in the Republic of Macedo- nia. Hence, these results suggest that monetary policymakers in the Republic of Macedonia must take into account these features concerning the eff ects of the monetary policy trans- mission mechanism on the economy. Finally, the result shows that any change in the current monetary strategy of exchange rate targeting carries a likely risk of fi nancial instability, due to higher dollarization in Republic of Macedonia, and such changes would adversely aff ect the NBRM’s ability to control infl a- tion due to the higher pass-through eff ect of the exchange rate regime change on prices. 4. CONCLUSIONS Th e main objective of this paper is to examine the eff ect of monetary policy and ex- change rate regime type on real GDP and prices in the Republic of Macedonia over the period from 1997 to 2008. Th e research includes the main conventional transmission channel of monetary policy (money supply), as well as the exchange rate channel, that are assumed to operate in the Republic of Macedonia. Based on the available theoretical and empirical evidence, we employ SVAR methodology in order to examine the relative costs and benefi ts associated with introducing a more active monetary policy and a diff erent exchange rate regime in the Republic of Macedonia. B. FETAI, I. ZEQIRI | THE IMPACT OF MONETARY POLICY AND EXCHANGE RATE REGIME ON ... 279 For countries in transition with short spans of data (which are sometimes of questionable quality), empirical results are to be indicative rather than defi nitive. With that caveat in mind, my main fi ndings and their implications are as follows: Assessing the relative costs and benefi ts associated with introducing a more active mon- etary policy and a diff erent exchange rate regime in the Republic of Macedonia, using SVAR methodology, show that introducing such policies in order to promote rapid eco- nomic growth could easy disturb macroeconomic stability (aft er having achieved it at a substantial cost) without any signifi cant economic benefi ts. Th erefore, introducing more active monetary policy and a diff erent exchange rate regime is likely to incur more costs than benefi ts, since changes of money stock and exchange rate regime type do not show a persistent eff ect on real GDP, while changes of money stock and exchange rate regime do show a strong and persistent eff ect on prices level. (i) Th e empirical analyses reveal that money supply prove to be weak as independent chan- nels of monetary transmission in the Republic of Macedonia, and therefore the result does not suggest that money supply are useful to the NBRM as independent instruments of monetary policy. Th is is a consequence of the fact that the bank and fi nancial sectors are still characterized by shallow levels of fi nancial intermediation – the fi nancial sector is underdeveloped, the banking sector suff ers from a lack of competition, and the economy has a high degree of dollarization. Th erefore, an increase of money supply does not show any signifi cant eff ect on real GDP via either the asset prices eff ect or the wealth eff ect, the bank-lending eff ect and the fi rms’ balance sheet eff ect, while it does have a strong eff ect on prices level. Th erefore, the result suggests that the primary role of monetary policy should be to control the rate of infl ation in the Republic of Macedonia since changes in the money stock did not show a signifi cant eff ect on real GDP, while they exhibit a strong and persist- ent eff ect on prices level. (ii) As for the exchange rate regime, all results show that changes in the exchange rate exhibit a potentially strong pass-through eff ect on domestic prices via import prices. A depreciation of the domestic currency against the Euro causes a sharp and rapid increase in manufactur- ing prices, an increase in the retail prices index and an insignifi cant eff ect on real GDP. Since the Republic of Macedonia achieved macroeconomic stability at a substantial cost, the em- pirical result suggests that the stability of the exchange rate is very important for macroeco- nomic stability because it highlights a potentially strong pass-through eff ect on the domestic prices level. Without a doubt, changing the type of the exchange rate regime carries a likely risk of fi nancial instability due to higher dollarization. Such changes also adversely aff ect the NBRM’s ability to control infl ation, due to the strong pass-through eff ect of the exchange rate changes on domestic prices. It is probably not worthwhile to do anything that may return Republic of Macedonia to infl ation, which the fl exible exchange rate regime may do, since the high cost of stabilization will once more be born by the people. Since the exchange rate reveals a strong potential eff ect on prices level, the results suggest that abandoning the exchange rate regime or depreciating the domestic currency would not be a wise strategy for promoting economic growth, since it would not create any economic benefi t, while macroeconomic in- stability would follow with well-known negative consequences for economic growth. ECONOMIC AND BUSINESS REVIEW | VOL. 12 | No. 4 | 2010280 APPENDIX I A. Data sources All data come from the NBRM, the, and the Macedonian Bureau of Statistics. “GDP”: gross domestic product, provided by the Macedonian Bureau of Statistics. “MPI”: manufacturing prices index, provided by the Macedonian Bureau of Statistics. “RPI”: retail prices index, provided by the Macedonian Bureau of Statistics. “M1”: money stock consists of the base money and balances held in chequing accounts (personal and current accounts), provided by the National Bank of the Republic of Mac- edonia (NBRM); “EXCH.EURO”: monthly average exchange rate of Macedonian Denar (MKD) per EURO, provided by the NBRM. B. Th e method and parameters for seasonality adjusting Th e seasonality of the variables will be tested by the method of exponential smoothing. Th e following table shows the model proposed by E.S.Gardener (1985, pp. 1-28) for dif- ferent combinations of the season and trend. smoothing the level of time series; trend; seasonal index error in period t; Th e estimation of parameters has utilized the simplex method, which minimizes the squares error 2 . Th e value of the parameters for GDP Model with TREND=Linear, SEASONAL=Multiplicative Estimated coeffi cients: alpha = 1.458067, gamma = -0.012734, and delta = -0.295535 parameter of smoothing level; seasonal period; parameter of smoothing seasonal. B. FETAI, I. ZEQIRI | THE IMPACT OF MONETARY POLICY AND EXCHANGE RATE REGIME ON ... 281 C. Figures FIGURE C1: Logarithms of time series MPI, RPI, M1 and GDP, except IR. Source: Author’s calculations APPENDIX II Endogenous variable: EXCH.EURO, MPI, RPI, M1, GDP Deterministic component: constant and dummy Criteria for VAR Order Selection FPE AIC HQ SC 1 2.55219e-020 -45.19921 -44.96213 -44.61536 2 1.15089e-020 -45.99863 -45.52194 -44.82461 3 1.24219e-020 -45.92892 -45.21007 -44.15830 4 1.33478e-020 -45.86932 -44.90570 -43.49554 5 1.59554e-020 -45.71120 -44.50016 -42.72757 VAR Residual Correlation EURO MPI RPI M1 GDP 1.0000 0.5764 0.1772 0.0853 -0.1319 0.5764 1.0000 0.1636 0.0277 -0.0191 0.1772 0.1636 1.0000 -0.0524 -0.0210 0.0853 0.0277 -0.0524 1.0000 -0.0358 -0.1319 -0.0791 -0.0210 -0.0358 1.0000 ECONOMIC AND BUSINESS REVIEW | VOL. 12 | No. 4 | 2010282 VAR Residual Analysis Skewness Kurtosis JB (2) LB (16) LM (16) ARCH (16) MPI -0.0148 3.3293 0.4236 21.1956 32.8244 14.5937 0.8091 0.0691 0.0078 0.5546 RPI 0.4247 3.3826 3.3630 21.3262 34.9456 7.3163 0.1861 0.0667 0.0040 0.9667 EURO 0.6025 21.9501 1397.1654 30.4409 42.4985 30.9899 0.0000 0.0041 0.0003 0.0635 M1 1.4352 8.4074 145.2319 39.6502 45.2594 8.3360 0.0000 0.0002 0.0001 0.9382 GDP 1.0178 5.0301 32.0263 30.0941 48.1023 12.2267 0.0000 0.0046 0.0000 0.7282 Residual Analyses B. FETAI, I. ZEQIRI | THE IMPACT OF MONETARY POLICY AND EXCHANGE RATE REGIME ON ... 283 REFERENCES Abel, A. B., Bernanke, B.S. & Smith, G. (2003). Macroeconomics (3rd ed.). Toronto: Pearson. Bernanke, B. & Mihov, I. (1998). Measuring Monetary Policy. Quarterly Journal of Economics, 113 (3), 869-902. Billmeier, A. & Bonato, L. (2002). Exchange Rate Pass-Th rough and Monetary Policy in Croatia. Interna- tional Monetary Fund, Working Paper 02/109. Belullo, A. (1999): Th e impact of money on real economic activity in Croatia. Zagreb: Faculty of Economic. Cecchetti, S.G. & Krause, S. (2001). Financial Structure, Macroeconomic Stability and Monetary Policy. National Bureau of Economic Research, Working Paper, No.7151. Christiano, L.J., Eichenbaum, M. & Evans, Ch. L. (1999). Monetary Policy shocks: What Have We Learned and What end?. In Taylor, J. B. & Woodford, M. (Eds.), Handbook of Macroeconomics vol. 1A. Amsterdam: Elsevier. Christiano, L. J., Eichenbaum, M. & Evans, Ch. L. (1996a). Th e Eff ect of Monetary Policy Shocks: Evidence from the Flow Founds. Review of Economics and Statistics, 78 (1), 16-34. Cîto, F. (2003). A VAR Investigation of the Transmission Mechanism in New Zealand. Reserve Bank of New Zealand. Cushman, D. O. & Zha, T. (1997). Identifying Monetary in a Small Open Economy Under Flexible Ex- change Rate. Journal of Monetary Economics, 39, 433-438. Eduardo, B. & Berg, A. (2000). Full Dollarization. IMF Economic Issues, No. 24. Elbourne de Haan, A., Kiviet, J. & Bas, M. C. (2003). Financial Structure and Monetary Policy in countries in transition. Mimeo: University of Groningen. ECONOMIC AND BUSINESS REVIEW | VOL. 12 | No. 4 | 2010284 Gardner, E. S.: Exponential Smoothing: Th e State of the Art. Journal of Forecasting, Vol. 4, (1985), pp.1-28. Giovanni, F. & Giordani, P. (2006). Reconsidering the Role of Money for Output, Price and Interest Rates. Th e Macro Lunch Seminar at IIES Stockholm University. Hafer, R. W. & Kutan, A. M. (2001). Determining and Th e Money-Output link: International Evidence. Working Paper, No.19, Center for European Integration Studies. Horváth, B. & Maino, R. (2006). Monetary Transmission Mechanism in Belarus. IMF Working Paper, 06 (2006)246. Hristov, A. (2004). Investigating the Eff ect of Monetary Policy in Post-Transition Economies: Th e Czech Repub- lic and Poland. http://www.diw.de/programme/jsp/MA.jsp?language=en&uid=ahristov Juks, R. (2004). Monetary Policy Transmission Mechanism a Th eoretical and Empirical Overview. In Th e Monetary Transmission Mechanism in the Baltic States (pp. 39-59) Tallinn. Kim, S. & Roubini, N. (2000). Exchange Rate Anomalies in Industrial Countries: A Solution with a Struc- tural VAR Approach. Journal of Monetary Economy, (45), 561-860. Kujis, L. (2002). Monetary Policy Transmission Mechanism and Infl ation in the Slovak Republic. IMF, Working Paper, No 02/80. Maćkowiak, W. S. (2003). External Shocks, U.S. Monetary Policy and Macroeconomic Fluctuations in Emerg- ing Markets. Berlin: Humboldt University Working Paper. Mayes, D. G. (2003). Th e Monetary Transmission Mechanism in Baltic States. Esti Pank-Bank of Estonia, pp. 61-107. McCarthy, J (2000). Pass-Th rough of Exchange Rates and Import Prices to Domestic Infl ation in Some Indus- trialized Economics. New York: Federal Reserve Bank Staff Reports, No. 11. Mishkin, F. S. (2003). Infl ation Targeting in Emerging Countries. NBER, Working paper, No. 7618. Obstfeld, M. (1998). Th e Global Capital Market: Benefactor or Menace? Journal of Economic Perspectives, 12 (4), 9-30. Ribnikar, I. & Bole, I. (2006). Th e Impact of a Large Current Account Imbalance on Monetary and Fiscal Policy: Th e Case of the Republic of Macedonia. Conference, National Bank of Republic of Macedonia, Dec- cusion Paper, Skpoje. Ribnikar, I. (2004). Exchange Rate Regimes and Monetary Arrangements. Journal of Economics and Busi- ness, 50 (2), 9-23. Sims, Ch. A. & Zha, T. (1998). Does Monetary Policy Generate Recessions?. Working Paper, No. 12, Federal Bank of Atlanta, pp. 98-12. Starr, M. A. (2004). Does Money Matter in the CIS? Eff ect of Monetary Policy on Output and Prices. Working Paper Series, No. 200410-02.