آیا جنبش شرکتی بیش از بازدهی اوراق قرضه بین کشورها وجود دارد؟

This paper examines the issues of excess volatility and excess comovement of interest rates among global bond markets. The base model of interest rate behavior is the expectations theory of the term structure. The empirical evidence presented in the paper indicates that 10-year government bond yields in five major markets—the United States, Japan, Germany, the United Kingdom and Canada—have in the past displayed both excess volatility and excess comovement relative to the base model. This suggests that term premia at the long end of the term structure are both time-varying and positively correlated across markets.

In early 1994 considerable differences existed between the cyclical position of the United States, where monetary policy was tightened in February, and those of Japan and Germany. However, long-term interest rates in the G-3 countries displayed a striking tendency to move together before, during and after 1994. The relatively high degree of comovement of long rates between these countries, coupled with the global nature of the 1994 bond market reversal, suggests that there may be an international component to bond yield fluctuations beyond that attributable to common movements in short-term interest rates and inflation. The existence of an international component to risk premia might explain the apparent tendency of bond yields to covary excessively between markets. This paper investigates this possibility by examining the historical behavior of bond yields in five major markets—the United States, Japan, Germany, the United Kingdom and Canada. The base model of interest rate behavior is the expectations theory of the term structure. Most empirical investigations of interest rates which have taken the expectations theory of the term structure as the base model have focused on the US market and have come to the conclusion that long-term bond yields deviate from the predictions of the model.2 A much smaller, but rapidly growing, literature tests the expectations theory with interest rate data for other countries. A recent example is the study by Hardouvelis (1994), which examines the behavior of 10-year government bond yields for G-7 countries. Hardouvelis concludes that bond yields in the majority of G-7 countries deviate from the predictions of the expectations theory.3 The main contribution of the present paper is an investigation of the comovements of bond yields between countries. With this goal in mind, the paper tests for excess volatility of interest rates and excess comovement of interest rates between countries by examining the behavior of long-term interest rates divided by their own 5-year moving average. The study of these stochastically detrended series allows for an investigation of the relatively high frequency movements in interest rates. Also, the use of detrended interest rate data implies that the present study is not as vulnerable to the criticism that the sample period is too short compared with studies that do not detrend. The empirical results can be briefly summarized as follows. First, the restrictions that the expectations theory of the term structure imposes on the behavior of bond yields in groups of countries allow the rejection of the model at high levels of statistical significance in the case of every country examined. Thus, this paper provides evidence against the expectations theory as a model of the behavior of long-term government bond yields in global bond markets. Secondly, the empirical evidence suggests that term premia at the long end of the term structure are positively correlated across markets. This is consistent with the existence of an international component to global bond yield fluctuations beyond that attributable to common movements in short-term interest rates and inflation. The rest of the paper is organized as follows. Section 2 develops a simple empirical framework which nests the expectations theory of the term structure, as it applies to interest rate behavior within a group of countries, in a more general model. The particular alternative hypothesis entertained regarding the joint behavior of bond yields is Shiller's (1989) notion of excess comovement of asset prices and is related to the concept of excess volatility of an asset price. Loosely speaking, an asset price displays excess volatility, relative to a specific asset pricing model, if there are deviations of the asset price from the theoretical values predicted by the model. Excess comovement of two asset prices is the case where two asset prices display excess volatility and the deviations of prices from predicted values are positively correlated across assets. Section 3 examines the historical behavior of 10-year government bond yields in the United States, Japan, Germany, the United Kingdom and Canada within the context of the empirical framework presented in Section 2. Using this framework, measures of excess volatility and excess comovement of bond yields are estimated and their statistical significance judged on the basis of the small sample distributions of the relevant test statistics computed from Monte Carlo simulations. The results of the Monte Carlo simulations indicate that over the period examined the comovements of bond yields between all countries studied were excessive, given the inter-country correlations of changes in short-term interest rates. Some conclusions are offered in Section 4.

This paper examines the issues of excess volatility and excess comovement of interest rates among global bond markets. The base model of interest rate behavior is the expectations theory of the term structure. The restrictions that the base model imposes on the behavior of yields within groups of countries allow the rejection of the model at high levels of statistical significance in the case of every country examined. In particular the empirical results indicate that, over the period studied, bond yields in the major markets displayed excess volatility relative to the base model. The empirical results also indicate that bond yields in the major markets displayed excess comovement relative to the base model. Taken together, these findings suggest that term premia at the long end of the term structure are both time-varying and positively correlated across markets. The finding that long bond term premia are positively correlated across markets suggests that the factors responsible for time variation in term premia have an international component. One potential explanation for this finding is uncertainty with regard to the conduct of monetary policy. The sample period studied, from the 1960s to the early 1990s, is one that shows a build up of inflation and then a reduction of inflation to very low levels in all the countries studied. Apparently, in all these countries the monetary authorities were more tolerant of high inflation in the earlier years than in the later period. With uncertainty about the commitment of monetary authorities to fighting inflation changing so much at the same time around the world, it seems highly plausible that there is naturally a factor in risk premiums that is common to all of the countries studied.