در پاسخ مطلوب سیاست های پولی به شاخص های جنجالی
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|24953||2003||23 صفحه PDF||سفارش دهید||10653 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Journal of Monetary Economics, Volume 50, Issue 3, April 2003, Pages 501–523
We describe a behavior of a central bank when its measures of current inflation and output are subject to measurement errors, in a framework of optimizing models with nominal price stickiness. In our model, a central bank sets the interest rate equal to its current estimate of the so-called Wicksellian natural rate of interest. This is shown to imply that the interest rate responds to the central bank's estimates of both current inflation and output gap, as advocated by Taylor (1993). It is also shown that the noise contained in the indicators justifies a degree of policy cautiousness. A reduced-form representation of optimal policy should exhibit interest-rate smoothing, which is often found in the empirical literature on monetary policy reaction functions.
In the recent studies of monetary policy, simple policy rules have received attention as a means to a more transparent and effective monetary policy. The policies of this family, which is originally proposed by Taylor (1993), are feedback rules which vary interest rate responding to current and lagged output gap, inflation, and interest rates. There has been considerable amount of research suggesting that simple feedback rules achieve good results in simulated small macroeconomic models.1 At the same time, many empirical studies report that the policy specifications of this kind fit the actual behavior of the central banks in several countries.2 An issue in the implementation of rules is the availability of relevant data. A central bank has only preliminary measures of current conditions with considerable measurement errors when it makes its policy decision. Accurate measures of these variables, which are required for the implementation of an optimal rule, are not known until much later. For example, Orphanides (1998a) reports that more than 30 percent of the fluctuations in the preliminary measures of the output gap and the GDP deflator may be caused by measurement errors. Thus the measurement problem of this kind is not negligible. This information constraint would be particularly important for a central bank which follows a Taylor-type policy rule. When the central bank has only preliminary data, the interest rate chosen by a feedback rule is affected by measurement errors. A good candidate for practical policy feedback rule might have undesirable properties once we recognize the fact that the central bank has only noisy data. In this paper, we describe the behavior of a central bank under data uncertainty, using a simple dynamic sticky price model. Specifically, we consider a central bank under Svensson's (1999) flexible inflation targeting regime, formulate how the central bank optimally extracts information about economic shocks from noisy indicators, and derive some properties of the optimal monetary policy. As is shown in Svensson and Woodford (2003b), even if the central bank and the private agents have asymmetric information, certainty equivalence holds in the sense that optimal policy response to estimates of the state of the economy is independent of degree of uncertainty. In our model, this implies that the central bank sets the nominal interest rate equal to the conditional estimate of the so-called “Wicksellian natural rate of interest” conditional on its information set. Next we consider optimal information extraction from noisy indicators, and derive an optimal policy in terms of observable variables. Here it is shown that separation principle does not hold under our information structure, since estimation is not independent of policy chosen (Svensson and Woodford, 2003b). We show that the interest rate should respond to the central bank's estimates of both current inflation and the current output gap, as advocated by Taylor (1993), even if the bank's objective is purely to stabilize inflation. It is also shown that, when an indicator is noisy, the central bank should respond to it cautiously. A classic paper by Brainard (1967) shows that uncertainty about parameters (i.e., multiplicative uncertainty), and hence uncertainty about policy multipliers, may make policy-makers more cautious. He also argues that additive uncertainty may not justify policy cautiousness. On the contrary, we show that noise contained in indicators is also a reason for policy cautiousness, even if it is additive uncertainty. However, this does not contradict to the certainty equivalence property. We also consider the optimal response of the interest rates to lagged endogenous variables. Optimal policy has a unique representation in terms of a feedback rule that depends only on observed endogenous variables. It is shown that the optimal policy rule of this form exhibits a degree of interest rate smoothing. This result is derived from the central bank's optimal response to the persistence of economic shocks and not from some other structural reason to smooth interest rate. The structure of the paper is as follows. The next section explains the model used throughout the paper. In Section 3, we discuss how a central bank extracts relevant information from noisy indicators, and derive the optimal response of the central bank to noisy indicators. The optimal response to lagged variables is derived in Section 5. Finally, we discuss the determinacy of equilibrium when the central bank commits itself to the optimal feedback policy rule. Section 7 concludes.
نتیجه گیری انگلیسی
This paper explicitly analyzes the effect of noisy data on the properties of the optimal policy in a framework of a simple optimizing model with nominal price stickiness. In our model, the central bank estimates the demand and supply shocks by observing noisy inflation and output measures. Based on these estimates, the central bank chooses the interest rate equal to its estimate of the natural interest rate. We show that, when an indicator is noisy, the central bank should respond to it more cautiously. Hence the noise contained in data offers a reason for policy cautiousness. Furthermore, our optimal policy has a unique representation in terms of a feedback rule that depends only on observed variables. The optimal feedback policy rule of this form involves a lagged interest rate. This stems from the fact that the central bank needs to use the past interest rate as well as past inflation and output in order to identify the past exogenous shocks. We also have shown that the feedback rule of this form is a good candidate for the implementation of the optimal policy, since a commitment to this feedback rule results in a unique equilibrium. We have also discussed the implications of our results for empirical studies of monetary policy. The behavior of an optimizing central bank should be dependent on its information set and the structure of the economy. The exogenous variation in our optimal feedback policy rule consists solely of measurement errors. Therefore, it would be difficult to distinguish whether an observed deviation from the estimated reaction function represents a response to measurement errors or a discretionary policy change.35 Empirical studies of monetary policy need to consider this issue. Finally, we considered in this paper a discretionary regime where a central bank expects itself to re-optimize at each successive date. We believe discretionary optimization is a reasonable description of the actual behavior of an optimizing central bank. However, optimization does not necessarily imply a fully optimal policy, because a fully optimal policy is often time inconsistent. Indeed, Aoki (2002) shows that the optimal policy is no longer time-consistent when the central bank's information is imperfect, even if the optimal policy is time consistent under perfect information. The analysis of an optimal policy which requires commitment under noisy information is taken up in Aoki (2002).