آیا بانک مرکزی باید به حرکات نرخ مبادله واکنش نشان دهند ؟ تجزیه و تحلیل نیرومندی قوانین سیاست ساده تحت عدم قطعیت نرخ ارز

This paper evaluates the performance of simple policy rules in an open economy. By introducing a high degree of exchange rate uncertainty we find that policy rules with an important feedback from movements in the real exchange rate are very robust to uncertainty about the true exchange rate model. A closed economy rule performs badly in most exchange rate specifications. This is in contrast to the findings of many other studies. In our view, this result is due to the fact that these studies assume a known and reliable relationship between the exchange rate and the interest rate.

In recent years there has been considerable progress in the field of monetary policy analysis. A growing academic literature explores simple policy rules expressed in terms of interest rate instruments as guides for monetary policy under a strategy of flexible inflation targeting. Since most inflation targeting countries today are small open economies, the role of the exchange rate for the conduct of monetary policy is a central issue. In particular, the question whether the exchange rate (in whatsoever form) should enter the policy rule or not is still a matter of debate in the literature. Thus, the evaluation of so-called open economy policy rules has become an important extension to the closed economy analysis of interest rate rules. Empirically, the tendency of central banks to indirectly influence the exchange rate by interest rate adjustments is largely confirmed (even for developed countries) by work on monetary policy rules. One strand of evidence results from the estimation of structural VARs in which, among other dynamic relationships such as aggregate demand, an equation for the monetary policy instrument has to be specified. For example, Clarida and Gertler (1997) reported estimates according to which the Bundesbank responded to a depreciation of the real exchange rate with a rise in short-term interest rates. Based on a small-scale model of the Australian economy, Brischetto and Voss, 1999 and Dungey and Pagan, 2000 found that the Reserve Bank of Australia reacts with the short-term interest rate to movements in the exchange rate. Another strand of empirical evidence results from the direct estimation of monetary policy rules. Clarida et al. (1998) found a small but significant reaction of the nominal interest rate of the Bundesbank (1979–1993), the Bank of Japan (1979–1994) and the Bank of England (1979–1990) to the real exchange rate. Gerlach and Smets (2000) estimated interest rate policy rules according to which the Reserve Bank of New Zealand and the Bank of Canada respond significantly with the short-term interest rate to changes in the nominal exchange rate, whereas the Reserve Bank of Australia does not. Investigating the inflation targeters Australia, Canada, New Zealand, Sweden and the United Kingdom, Hüfner (2004) found that the exchange rate term in the policy rule is only significant for the United Kingdom and New Zealand. He explains the differences to the study of Gerlach and Smets (2000) mainly by a somewhat larger sample period. For the emerging market economies Chile, Israel, South Africa, the Czech Republic and Mexico, Ades et al. (2002) also found significant (and, in comparison with the developed economies of the aforementioned studies, larger) exchange rate coefficients in the interest rate policy rule. These findings are confirmed by a recent study of Mohanty and Klau (2004). In contrast to the rather clear-cut results from empirical studies, the results from numerical simulations of calibrated open economy macro models are mixed. By adding an exchange rate term to a simple, Taylor-type policy rule, Ball, 1999, Svensson, 2000 and Batini et al., 2001 find a small improvement of the macroeconomic performance of a central bank’s interest rate policy. In contrast to this, Côté et al. (2002) come to the result that using an open economy monetary policy rule often increases the value of the loss function. Taylor (1999c) gets somewhat mixed results in his multi-country study, favouring open economy rules for some countries and rejecting their usefulness for other countries. In a recent overview, he ultimately comes to the conclusion that “research to date indicates that monetary policy rules that react directly to the exchange rate, as well as to inflation and output, do not work much better in stabilizing inflation and real output and sometimes work worse than policy rules that do not react directly to the exchange rate” (Taylor, 2001, p. 267). In our view the problem of most of these numerical simulation studies is that they disregard the fact that economists know little about the determinants of exchange rate movements and the interaction between exchange rates and other fundamental variables in systems of independently floating exchange rates. A notable exception is the paper of Leitemo and Söderström (2005), which also studies the impact of exchange rate uncertainty on the conduct of monetary policy. They conclude that a simple Taylor-type policy rule is sufficient to stabilize a small open economy, also under exchange rate model uncertainty. The present paper takes up the idea of Leitemo and Söderström (2005). In contrast to their paper we substantially increase the degree of exchange rate uncertainty. By doing so we seek to provide a rationale for the observed central banks’ interest rate response to exchange rate movements. We investigate whether this uncertainty has any influence on the structure of the policy rules that central banks should commit to in a small open economy. Following McCallum (1988) we try to identify policy rules that possess a high degree of robustness against these uncertainties in the sense that they perform well across a range of alternative models. Our results indicate rather clearly that, due to the introduction of a high degree of exchange rate uncertainty, open economy rules become superior to simple policy rules that only react to inflation and output. By following an open economy policy rule a central bank adopts a strategy that insulates the economy from the uncertainties stemming from the mostly unknown and unreliable relationship between the nominal exchange rate and the nominal interest rate or other macroeconomic variables. The remainder of this paper proceeds as follows. We begin in Section 2 by presenting a standard Neo-Keynesian open economy macro model typically used by academics and central banks for the evaluation of monetary policy rules. In this model, to which we refer as the baseline model, the path of the nominal exchange rate is determined according to uncovered interest parity (UIP). We will reproduce the result usually obtained from numerical simulations by evaluating the performance of six simple policy rules. As the exchange rate mainly determined by the interest rate itself, a separate interest rate reaction to exchange rate movements is redundant. Section 3 analyzes the impact of exchange rate uncertainty on the performance of monetary policy rules. We first define the types of uncertainty stemming from the poor knowledge about the determination of exchange rates. In particular, we will focus on deviations from UIP. As alternatives to UIP we propose exchange rate specifications that either show a much better fit in empirical studies that allow for deviations from the rational expectations hypothesis by introducing backward-looking expectations, or that simply display purely random exchange rate behavior. Apart from the exchange rate specification, all the models are identical with respect to the IS equation and the Phillips curve. We then evaluate the extent to which the performance of the six policy rules, which are derived from the baseline model is affected by the risk that instead of uncovered interest parity another exchange rate model, is a better description of actual exchange rate behavior. We will show that this exchange rate uncertainty impacts on the conduct of monetary policy on two levels. First, the exchange rate appears as an own source of shocks, which conveys independent information to the policy maker. And second, the transmission of interest rate impulses on the central bank’s final targets via the exchange rate channel is subject to a high degree of uncertainty. We will then identify the characteristics of those policy rules that perform reasonably well over all exchange rate specifications. In Section 4 we check whether the results obtained in Section 3 depend on the choice of the central bank’s preferences towards inflation and output stabilization and on the choice of the baseline model. As in Leitemo and Söderström (2005) we introduce partially forward-looking behavior on the part of consumers and price setters by using a hybrid IS equation and a hybrid Phillips curve. We then do the same robustness analysis as with the purely backward-looking Neo-Keynesian open economy macro model. Section 5 summarizes the main results

Standard open economy macro models are usually based on the assumption that the exchange rate is determined on an efficient foreign exchange market with forward-looking and rational behavior on the part of the international financial markets’ participants. Uncovered interest parity defines a known and reliable relationship between changes in the central bank’s operating target and the exchange rate so that in addition to the interest rate channel a second important transmission channel of monetary impulses – the exchange rate channel – can be exploited by the central bank. The majority of the empirical literature, however, comes to the result that in the short and medium run (which is the most relevant for the conduct of monetary policy) the behavior of exchange rates cannot be explained and predicted by any of the existing models. In particular, uncovered interest parity does not find much empirical support. This finding raises the question of how the conduct of monetary policy in such an environment of uncertainty about the true determination of exchange rates is affected. The intention of the present analysis is to provide a rationale for the observed widespread use of the so-called open economy policy rules, according to which the interest rate directly responds to movements in the exchange rate. Our approach is based on a standard, open economy macro model typically employed for the analysis of monetary policy strategies. The consequences of freely floating exchange rates are evaluated in terms of a social welfare function, or, to be more precise, in terms of an intertemporal loss function containing a central bank’s final targets output and inflation. In order to take account of the poor empirical evidence of uncovered interest parity, we question the basic assumption underlying most open economy macro models that the foreign exchange market is an efficient asset market with rational agents and we model the central bank’s decision making process as being confronted by a high degree of exchange rate uncertainty. Exchange rate uncertainty is defined as the risk that instead of uncovered interest parity another exchange rate model is a better description of the exchange rate behavior at a certain moment in time. The main lesson that can be drawn from the analysis of monetary policy under freely floating exchange rates and exchange rate uncertainty is that exchange rate uncertainty provides a rationale for adopting an open economy policy rule. This finding is in sharp contrast to the traditional literature on policy rules in open economies. Most of the studies conducted in this field only attach a minor importance to the possibility of interest rate feedback to the exchange rate movements. The study of Leitemo and Söderström (2005), for example, which also analyzes monetary policy under exchange rate uncertainty, comes to the conclusion that as long as there are no extreme parameterizations of the uncertainty scenario, closed economy policy rules are an efficient and robust guide for monetary policy in an open economy. The main difference to their analysis is that we substantially increased the degree of exchange rate uncertainty to account for the little knowledge about the true exchange rate behavior by extending the set of possible exchange rate specifications. We showed that even with quite realistic parameters underlying the uncertainty scenarios, the closed economy policy rule and some of the open economy policy rules perform poorly in terms of the loss they produce. By contrast, the use of an open economy policy rule with an important exchange rate feedback from contemporaneous and lagged movements in the real exchange rate performs reasonably well over all exchange rate specifications. And this result also holds for alternative preferences of the central bank towards inflation and output stabilization and for a range of alternative baseline models that incorporate hybrid expectations on the part of consumers and price setters and a different lag structure in the Phillips curve. Even though the ranking of the policy rules in terms of the loss they produce under exchange rate uncertainty is somewhat less clear-cut than with a purely backward-looking IS equation and Phillips curve, there is still a clear advantage of following a policy rule that responds to contemporaneous and lagged movements in the real exchange rate. Thus, in a world in which we allow for deviations from the assumption of perfectly functioning foreign exchange markets and in which we assume a central bank taking these deviations into account and behaving so as to reach its final targets, the rationale for using open economy policy rules is the monetary policy maker’s quest for a robust interest rate policy rule that performs comparatively well across a range of alternative exchange rate models.