سیاست پولی و شاخص شرایط پولی

Several central banks have recently adopted a Monetary Conditions Index (MCI) to guide monetary policy under floating exchange rates. This paper discusses some analytical and practical questions raised by MCIs. Furthermore, using data for Australia, Canada and New Zealand, which all operate monetary policy under floating rates and with an inflation target, it estimates the responses of the central banks to exchange rate changes. The results reveal clear differences between central banks: while the Reserve Bank of Australia does not appear to respond, the Bank of Canada and the Reserve Bank of New Zealand, who use the MCI as an operating target, do respond quite strongly to movements in the exchange rate.

The day-to-day conduct of monetary policy in a small open economy is arguably more difficult under a floating than under a (credibly) fixed exchange rate regime. While the daily management of policy under fixed rates mainly involves moving short-term rates in order to maintain the exchange rate in the appropriate part of the band, under floating rates policy makers need to determine to what extent interest rates should react to exchange rate changes.1 This requires them to form opinions about the sources of the shocks affecting the exchange rate, their effects on macroeconomic conditions, and whether to attempt to offset them. A further question that arises is how they should judge the stance of policy when the exchange rate changes. Needless to say, these are difficult tasks. In this paper, we discuss a number of issues concerning the role of the exchange rate in the conduct of monetary policy in small open economies under flexible exchange rates. Recently, several central banks have adopted a monetary conditions index, MCI, in order to incorporate the exchange rate in the design and implementation of monetary policy. While the MCI is defined in the same way in all countries – as a weighted average of a short-term interest rate and an exchange rate, real or nominal – its exact use differs between central banks. For instance, the Bank of Canada and the Reserve Bank of New Zealand use it as an operating target: that is, in light of the inflation outlook they form a view of a desirable range for the MCI and use policy-controlled interest rates to achieve it. In contrast, in Sweden, Norway, Finland and Iceland it is merely used as one indicator among many of the stance of policy. Despite its increasing popularity, research on the economic underpinnings of MCIs has only recently started to appear.2 In this paper we attempt to make some progress in this direction. The paper consists of three sections. In the second section we use a simple model to derive the optimal feedback rule of a central bank which cares about output and inflation (Section 2.1). We demonstrate that in this specific model, the feedback rule can be written in terms of an MCI, that is, the central bank can minimise its objective function by setting a weighted average of interest and exchange rates in response to changing macroeconomic conditions. Moreover, we show that such an operating target for the MCI has the advantage that it automatically achieves the desired monetary policy stance in the presence of credibility shocks to the exchange rate. In Section 2.2 we discuss some problems concerning the use and construction of MCIs. We argue that the concept of an MCI and its use as an operating target is based on a particular view of the transmission mechanism, which may be a poor approximation of the actual working of the economy. In accordance with central bank practice, our model suggests that the optimal weight on the exchange rate depends on the elasticities of aggregate demand with respect to the real interest rate and real exchange rate. Thus, the larger the impact of exchange rate changes on aggregate demand, the more weight should the central bank attach to the exchange rate. However, this model may be deficient in a number of ways. First, monetary policy may affect inflation through other transmission channels than the output gap, for instance through the direct effect of exchange rate changes on import prices. If so, the attractive result that the weight on the exchange rate depends solely on the elasticities of aggregate demand may no longer hold true. Secondly, the lags with which the exchange rate and the interest rate affect aggregate demand may differ. While our model assumes that monetary policy affects the economy without a lag, in practice such lags are important. The usual practice of central banks to focus on the one- to two-year horizon for estimating both the interest rate and exchange rate elasticities implicitly imposes the constraint that changes in exchange and interest rates affect aggregate demand equally quickly. Of course, this assumption may be wrong in actual economies. A final concern arises from the fact that there is no reason to assume that only a short-term interest rate and the exchange rate should enter the MCI. The use solely of short-term interest rates and exchange rates in the MCI was first adopted by the Bank of Canada because research indicated that the output gap played an important role in determining inflation, and that monetary policy affected the gap primarily through the real short-term interest rate and the exchange rate. Of course, in other countries the relative importance of these two variables may be very different, and potentially other variables such as long-term interest rates or equity prices should (also) enter. In sum, the design of the MCI should depend on the exact nature of the monetary transmission mechanism in the economy in question. Since this is likely to differ between countries and over time, so should the definition of the MCI. Section 2.3 turns to a second issue which arises from the fact that the exchange rate is both a relative price which affects the demand for domestic and foreign goods, and an information variable that central banks can potentially use to distinguish between exchange rate movements that warrant a change in short-term interest rates (such as a shift in the credibility of monetary policy) and those that do not (such as an increase in the demand for domestic goods). Since central banks in general do not know the sources of the shocks hitting the economy, we solve the model assuming that the policy makers have imperfect information about the causes of exchange rate changes and demonstrate that in this case the optimal MCI-weight depends on the relative importance of different shocks. Thus, in countries in which exchange rate changes are typically due to aggregate demand disturbances, monetary policy makers will tend not to change interest rates in response to exchange rate changes. By contrast, in countries where exchange rate changes in the past have largely been due to shifts in the credibility of monetary and fiscal policy, central banks will offset exchange rate changes by altering short-term interest rates. This provides one explanation why several central banks, including the Reserve Bank of Australia and the Bank of England, have rejected the use of an MCI. In Section 3 we turn to our empirical work. The use of an MCI raises two sets of issues for empirical analysis. The first of these is whether we have sufficiently good estimates of the interest and exchange rate elasticities of aggregate demand to determine the appropriate weight for the exchange rate in the MCI. This question was recently addressed by Ericsson et al. (1997), who studied equations used by the central banks in Canada, New Zealand, Norway and Sweden to motivate the weights used in their MCIs. The authors demonstrated that these equations provide very poor estimates of the relative size of the two elasticities, suggesting that the data provide little guidance as to the appropriate weight for the exchange rate. Rather than revisit this question, we focus on the second set of issues which concerns how in practice central banks respond to exchange rate changes. To do so, we estimate simple reaction functions for Australia, Canada and New Zealand, where monetary policy is conducted using explicit inflation targets but where the role of the exchange rate in the conduct of policy appears to differ. After a brief review of the main features of the monetary policy framework in these countries, we discuss the econometric results. These suggest that there are differences between the central banks in the extent to which they respond to exchange rate movements. In Section 4 we provide some conclusions.

In conducting monetary policy under floating exchange rates, monetary policy makers have to take into account the fact that changes in interest rates and exchange rates both affect aggregate demand for goods and services. To help assess the policy stance, a number of central banks have recently adopted MCIs, defined as a linear combination of a (real or nominal) short-term interest rate and exchange rate. Despite their growing popularity, however, there is relatively little work on their underpinnings. In this paper we have reviewed a number of questions that arise. Our main conclusions are as follows. Since interest rates and exchange rates both affect aggregate demand, it is appropriate to judge their joint effect on the economy. However, and as noted by a number of central banks, whether this can be done using a simple linear combination of a short-term interest rate and an exchange rate depends critically on the structural characteristics of the economy. First, while there is considerable evidence that these variables play an important role in the transmission mechanism in Canada, there is little reason to assume that this is necessarily the case in other countries. For instance, longer-term interest rates play a more important role in many European economies. Moreover, the uncertainty about the estimates of the MCI weights suggests it may not be appropriate to put too much emphasis on a particular number. Second, the use of an MCI with weights corresponding to the respective elasticities of aggregate demand presupposes that one can clearly distinguish between those exchange rate changes that are driven by changes in underlying supply and demand developments and those that are not. In the former case the central bank would optimally adjust its MCI target; in the latter case it would keep its MCI target unchanged and offset the effects on aggregate demand by adjusting interest rates. However, if it is not straightforward to distinguish between the sources of exchange rate changes, the optimal weight is smaller than implied by the elasticities of aggregate demand. In particular, when past experiences suggest that most of the exchange rate changes are due to changes in the equilibrium real exchange rate, it is optimal not to respond to exchange rate changes and the corresponding MCI-weight on the exchange rate will be close to zero.