اندازه گیری وابستگی متقابل سیاست های پولی

This paper measures the degree of monetary policy interdependence between major industrialized countries from a new perspective. The analysis uses a special data set on central bank issued policy rate targets for 14 OECD countries. Methodologically, our approach is novel in that we separately examine monetary interdependence due to (1) the coincidence in time of when policy actions are executed from (2) the nature and magnitude of the policy adjustments made. The first of these elements requires that the timing of events be modeled with a dynamic discrete duration design. The discrete nature of the policy rate adjustment process that characterizes the second element is captured with an ordered response model. The results indicate there is significant policy interdependence among these 14 countries during the 1980–1998 sample period. This is especially true for a number of European countries which appeared to respond to German policy during our sample period. A number of other countries appeared to respond to U.S. policy, though this number is smaller than that suggested in preceding studies. Moreover, the policy harmonization we find appears to work through channels other than formal coordination agreements.

This paper investigates the empirical regularities of monetary policy setting among a sample of 14 industrialized countries, with special focus on the international interdependence among these policies. A new empirical approach is proposed, which avoids the pitfalls implied by the VAR approach that traditionally has been used to address this issue. This approach allows us to look in a new way at the nature of policy coordination and the relative leadership roles played by the U.S. Federal Reserve, the Bank of Japan, and the German Bundesbank. International monetary policy coordination has become a subject of renewed interest of late. In particular, Obstfeld and Rogoff (2002) initiated a re-exploration of international policy coordination using models with microeconomic foundations, representing a significant methodological departure from past literature. They conclude that in this context the welfare gains of coordination are likely to be quantitatively small in comparison to the gains from domestic stabilization policy. In rebuttal, subsequent theoretical work has suggested a variety of economic features that could potentially generate greater motivation for nations to coordinate their monetary policies. See for example, Benigno, 2002, Canzoneri et al., 2001 and Clarida et al., 2002. Given this theoretical controversy, it is natural to ask the empirical question of how much coordination we observe in actuality. There is a history of empirical research on this question. Studies focusing on the major industrial countries generally find evidence that the U.S. acts as a leader for policy makers in certain countries, but the mechanisms through which this coordination takes place are often unclear. (See Dominguez, 1997, Furman and Leahy, 1996, Chung, 1993, Burdekin and Burkett, 1992, Burdekin, 1989 and Batten and Ott, 1985.) Another branch has focused on coordination among European countries, generally finding that Germany had a limited leadership role among European countries prior to monetary union. (See Garcia-Herrero and Thorton, 1996, Katsimbris and Miller, 1993, Biltoft and Boersch, 1992, Karfakis and Moschos, 1990, von Hagen and Fratianni, 1990a and von Hagen and Fratianni, 1990b.) Our paper represents a significant methodological departure from this pre ceding empirical literature. We build upon the recently developed methodology of Hamilton and Jordà (2002), which has been successfully employed to study monetary policy in the U.S. We extend this methodology to a set of countries, and we explicitly allow for interdependence among these national policies. Our approach is based on a novel data set on overnight, interbank interest rate targets that central banks in our sample use to communicate and operationalize monetary policy. Methodologically, we first argue against traditional time series techniques based on dynamic conditional correlations in semi-structural vector processes, which suffer from the common identification assumptions that mar the monetary vector autoregression (VAR) literature. Perhaps more importantly however, we show via Monte Carlo experimentation that the peculiar statistical properties that policy rate targeting imbue on market interest rates tend to severely distort these measures of association. In particular, these experiments demonstrate that, for example, Granger causality tests often will misrepresent the true nature of existing monetary policy linkages that are measured with market interest rate data. These methodological pitfalls prompt us to pursue a more modest approach but one which is immune to the deficiencies described above. In particular, our goal is to determine whether G-3 policy moves help predict the timing of domestic policy adjustments and the direction in which interest rates are modified, all conditional on domestic macroeconomic conditions. Exploring these issues poses special econometric challenges that arise from the irregular nature in which policy rate targets are adjusted over time as well as the discrete nature in which these adjustments are made. Hence, we will use the autoregressive conditional hazard (ACH) model proposed in Hamilton and Jordà (2002) to tackle the predictability in the timing whereas a discrete ordered response model specification will permit us to deal with the discreteness of policy rate adjustments. The empirical results are elucidating along several dimensions. First, our study can be seen as a broad investigation of individual central bank behavior, and in this sense, it provides important lessons to the literature that explores monetary policy rules. The special nature of the data and the particulars of our empirical design suggest that Taylor type rules do not universally describe monetary policy well, once the natural inertia of interest rates and macroeconomic data has been filtered. In particular, while U.S. policy responds to inflation in a statically significant manner, we were unable to find similar evidence among any of the other countries, including Germany, which traditionally enjoyed a reputation for an aggressive stance against inflation. Further, while it is a current debate among monetary theorists whether it is optimal for policy rules to include the exchange rate, we find evidence that this variable has in practice played a very prominent role in the policy rules of many countries. With regard to monetary policy interdependence, Germany emerges as a significant factor for a block of European countries, in agreement with most past work. This indicates that policy interdependence existed among this European core long before the creation of the explicit monetary union. Further this European interdependence entailed more than simply responding to exchange rate movements, as required by the European Exchange Rate Mechanism (ERM), but appears to have involved direct responses by policy makers to changes in German policy. We also find evidence that a limited number of countries respond to U.S. policy changes. This list includes Germany and the UK, but in contrast to some past work, Japanese policy does not appear to respond in a significant way to the U.S.

The contributions of this paper span in several dimensions. At a methodological level, we demonstrated that interest rate targeting modifies the time series properties of market interest rate data in a manner that renders traditional vector time series techniques inapplicable. Our Monte Carlo evidence complements the results reported in Rudebusch (1995) regarding the effects of interest rate targeting on term structure regressions. The solution we propose to this problem is based on an uncommon data set on policy rate target data and consists on decomposing central bank policy into the decision to adjust versus the actual type of adjustment. This strategy is based on previous work by Hamilton and Jordà (2002) for the U.S. and is modified and used here for the first time on other countries to detect monetary policy interdependence. This empirical design has several advantages that go beyond the econometric pitfalls discussed early on. In particular, recent research by Rudebusch, 2002 and Van Gaasbeck, 2001 suggest that Taylor rules may appear to accurately describe central bank behavior due to the inherent persistence in macroeconomic data. We completely dispense with this artifact by concentrating instead on the snap shots that target rate adjustments provide of actual central bank policy. Hence, our results are elucidating, not only with regard to the issue of monetary policy linkages, but also as an alternative analysis of domestic monetary policy in industrialized OECD countries. We find evidence of a clear European core, that responded to German policy long before the creation of a formal monetary union. Further this interdependence seems to have involved more than just attempts to stabilize exchange rates, but instead involved more direct responses to German policy shifts. We also find evidence of a distinct European periphery, including Italy, Switzerland and the UK, which did not respond to German policy. Finally, some countries respond to U.S. policy shifts, notably Germany and the UK. But we did not find any evidence that Japan responds to the U.S. Our results also shed light on the debate whether countries should include the exchange rate in their policy rules—quite a number of countries in practice do assign a very prominent weight on this variable.