تغییرات تورم، میزان بازدهی و اثرات آستانه
|کد مقاله||سال انتشار||تعداد صفحات مقاله انگلیسی||ترجمه فارسی|
|24026||2004||13 صفحه PDF||سفارش دهید|
نسخه انگلیسی مقاله همین الان قابل دانلود است.
هزینه ترجمه مقاله بر اساس تعداد کلمات مقاله انگلیسی محاسبه می شود.
این مقاله تقریباً شامل 19210 کلمه می باشد.
هزینه ترجمه مقاله توسط مترجمان با تجربه، طبق جدول زیر محاسبه می شود:
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : International Review of Economics & Finance, Volume 13, Issue 2, 2004, Pages 187–199
Using interest rate yield spreads to explain changes in inflation, this paper investigates whether such relationships can be modeled using two-regime threshold models. Implementing a robust test to detect evidence of a threshold, we find that the hypothesis of linearity is generally rejected. We find that the inflation–yield spread relationship at most horizons is more pronounced when the yield curve is inverted, which is usually associated with periods of tight monetary policy. This implies that monetary policy may have an asymmetric effect on inflation
This paper investigates whether threshold effects exist in the relationship between inflation changes and interest rate yield spreads in the United States using the technique developed by Hansen (1996) to test for a threshold whose location is unknown a priori. Testing for threshold effects allows us to gauge whether this relationship is more pronounced for certain levels of the interest rate yield spread. Through the introduction of a threshold variable, one is capable of capturing potential nonlinearities in the relationship within a tractable and intuitively appealing parametric model. Our interest in nonlinearities in general, and threshold models in particular, is motivated by two factors. First, this work extends and complements the leading indicator literature of the interest rate yield spread, such as the works of Frankel and Lown (1994) and Mishkin, 1990a, Mishkin, 1990b and Mishkin, 1991. These authors generally find, within a linear framework, that the term structure of interest rates at longer horizons contains useful information about inflation changes. However, detection of significant nonlinearities can potentially improve the forecasts of future directions of inflation extracted from the term structure of interest rates, since linear models can simply be viewed as constrained nonlinear models. This paper therefore provides some preliminary evidence on whether nonlinearities are likely to be of value in extracting inflation information from interest rates. Second, it has been noted by Bernanke and Blinder (1992), Blinder (1998), Johnson and Keleher (1996), Laurent, 1988 and Laurent, 1989, and others that the difference between long-term and short-term interest rates can provide a measure of the stance of monetary policy. This argument follows since monetary policy can readily influence short-term rates while long-term rates are generally market driven and do not react hastily to everyday policy actions. Furthermore, the long-term rate can act as a proxy for the equilibrium short-term rate, a Wicksellian natural rate of sorts, so that the difference between a long-term and a short-term rate can be viewed as a measure of the relative tightness of policy. As such, some of the yield spreads examined in this paper can be thought to capture the effects of monetary policy, so our results can be compared to other empirical studies on the asymmetric effects of monetary policy, such as the studies of Karras (1996), Karras and Stokes (1999), Morgan (1993), and Rhee and Rich (1995). These authors generally find that expansionary monetary policy, measured using either money or interest rates as the policy variable, has a weaker impact on output growth than a contractionary policy. The idea of asymmetries in monetary policy was most clearly explained by Friedman (1968), who stated that policy was akin to pushing on a string when expansionary and pulling on it when restrictive. Firms and consumers have an incentive to reduce their investments and consumption when interest rates rise, since projects may no longer be profitable or goods may no longer be affordable. When interest rates fall, however, there is no guarantee that firms will immediately increase their investments, since projects that were initially profitable remain profitable and new projects that can take advantage of the lower interest rate environment may require substantial time to be developed and implemented. If there is a link between real economic activity and the rate of inflation through a Phillips curve, and if asymmetries in the effects of policy on the economy exist, then by Friedman's argument we would expect inflation to be less responsive to expansionary policy than to contractionary policy. A threshold model lends itself nicely to the detection of asymmetries, such as those proposed in the above discussion on the effects of monetary policy. The magnitude of the asymmetry can be readily interpreted within the parametric threshold model, since the parameters vary discretely according to the level of the yield spread. More elaborate nonlinear models, such as the nonparametric kernel regressions or neural networks used by Tkacz (2000), can capture more subtle nonlinearities but lack the intuitive parametric interpretation of the threshold model. Our work differs from previous studies in three manners. First, instead of money, we use interest rate yield spreads to measure the effects of policy, since Blinder (1998), among others, emphasizes that policy is conducted through the control of short-term interest rates, not the money stock. Second, we do not restrict our threshold at a zero value of the yield spread, which would be considered the neutral policy setting. In earlier studies, the usual practice has been to test for asymmetry in tight or loose regimes, ignoring potential sources of asymmetry at different values of the policy variable. Instead, we consider a more general test that searches over a wide range of possible threshold values, maximizing the likelihood of uncovering any asymmetries should they exist. Our definition of asymmetry is therefore one in which the relationship between the yield spread and dependent variable is allowed to vary over any value of the spread. Finally, we test for threshold effects between interest rate yield spreads and inflation changes, not the rate of output growth. In a related study, Galbraith and Tkacz (2000) detect asymmetries between the yield spread and output growth in the United States and Canada; therefore, the present paper extends those results to a variable that is most closely monitored by policymakers, namely inflation. Thus, should a significant threshold be located, monetary policy would be said to have an asymmetric effect on inflation changes, differing when the yield spread is above and below its threshold value. Using the yields on securities with maturities ranging from 3 months to 10 years, we are able to construct 15 different long minus short yield spreads. The indicator properties of each of these spreads for changes in future CPI inflation is considered, and consistent with the studies of Mishkin, 1990a and Mishkin, 1990b, we find that spreads which jointly incorporate information from both short and long ends of the interest rate yield curve contain the most explanatory power for future inflation. For yield spreads that can be influenced to some degree by monetary policy, we find that significant thresholds emerge when the yield curve is relatively flat or inverted, namely periods of some monetary tightening. Below these thresholds, the effect of the yield spread on inflation changes is more pronounced than when it is above. This indicates that substantial monetary tightening is likely required in order to induce the type of asymmetry described by Friedman (1968). In Section 2, we present the framework used to analyze the inflation–yield spread relationship. The data used in this study, and the estimation results of some simple linear models, are also presented. The tests for threshold effects are implemented in Section 3, and the findings are discussed. Section 4 concludes.
نتیجه گیری انگلیسی
This paper first presents a standard model, relying on the foundations of the Fisher equation, to link changes in inflation to interest rate yield spreads. Positing that yield spreads may have asymmetric effects on inflation changes, we estimate threshold models for various forecasting horizons. This exercise is useful to exploit the information content of interest rate yield spreads and also to test for possible asymmetric effects of monetary policy on future inflation. We find that for policy-relevant horizons the relationship between long–short yield spreads and inflation changes is more pronounced when the spread is below some threshold, usually below 0.00. The consequence of asymmetry at such horizons is that the marginal benefits of a tightening of monetary policy, in terms of reduction in inflation, would be larger when monetary policy is already tight. This finding is consistent with earlier studies that have found the effects of monetary policy on output to be asymmetric. It should be noted, however, that if policymakers targeted the yield spread in a manner consistent with these findings, the Lucas Critique would apply. For this reason, these results should be of most value to forecasters who specialize in extracting information from financial variables. For nonpolicy horizons, such as those relating to either the short or the long end of the yield curve, we find that there is little information content at the short end and that both the magnitudes and the directions of long-term inflation forecasts are affected by thresholds. As a result, at long horizons, inflation forecasters should be vigilant when extracting information from long-term yield spreads, as both the magnitude and the direction of long-term inflation rates are dependent upon the level of yield spreads. Since we have detected some evidence of nonlinearities between yield spreads and inflation changes through a simple two-regime threshold model, in future work, it would be useful to compare linear and nonlinear models in an out-of-sample forecasting exercise. The set of nonlinear models can include the threshold model presented here in addition to general nonlinear models, such as nonparametric models and neural networks.