This paper presents a model of the relationship between the term structure of interest rates and the expected inflation rate, and also an empirical analysis using data on yields on Japanese government bonds in the secondary market.
Concerning the mechanism for determining the yields on Japanese government bonds in the secondary market, Kuroda and Okubo [2], [3] have shown that the pure expectations theory satisfactorily explains the determination of yields in the negotiable government bond market and that the expected inflation rate is significantly reflected in the nominal interest rate (the "Fisher Effect"). Empirical investigation of (i) the "Liquidity Premium Hypothesis" (Hicks [12] ), (ii) the "Preferred Habitat Hypothesis", and (iii) the "Coupon Oriented Hypothesis" (Kuroda and Okubo [2]), all of which emphasize the segmentation of markets, was conducted. Results indicated that (i) and (ii) could clearly be rejected while (iii) had some explanatory power.
This paper builds upon empirical results of Kuroda-Okubo [2], [3]. Here, I first present a model which explicitly incorporates the "Fisher Effect" in relation to the term structure of interest rates. Forecast values of the "time series model" are used as expected rates of inflation, and these values, in turn, are used to estimate the magnitude of the "Fisher Effect" by term to maturity of government bonds. Next, the three hypotheses listed previously which emphasize the existence of market segmentation, are investigated for significance in explaining the yield on Japanese government bonds in the secondary market on a real interest rate basis after adjusting for the expected inflation rate.
A summary of the empirical results is as follows.
1) The magnitude of the "Fisher Effect" for the market yield on Japanese government bonds in the secondary market averaged about 0.3-0.4 for the sample period extending from June 1977 to December 1980 (quarterly data was used). In other words, expected inflation rates are not totally reflected in the yields on Japanese government bonds. Rather, only 30-40% of expected inflation rates are reflected in these yields.
2) Looking at government bonds in the secondary market by term to maturity, the longer the term to maturity is, the larger the "Fisher Effect" is. In other words, the "Fisher Effect" is larger for long term interest rates than for short term interest rates.
3) The "Liquidity Premium Hypothesis", the "Preferred Habitat Hypothesis" and the "Coupon Oriented Hypothesis" are all rejected as explanations of the determination of yields on negotiable government bonds on a real interest rate basis.
To date, theory and empirical analyses of the term structure of interest rates have left vague the relationship between the nominal rate of interest and the expected inflation rate (for example, Fisher [10] and Yohe's and Karnosky's distributed lag model [22] ). However, this paper, through the explicit modeling of this relationship, clarifies the magnitude of the "Fisher Effect" and the mechanism for determining the term structure of interest rates on a real interest rate basis.
It should be pointed out that two problems in the empirical analysis remain. The first is the question of how accurately expected inflation rates introduced through the "time series model" reflect people's "actual expectations" toward future inflation rates. Also, there is a question about the appropriateness of the "assumption of a constant equilibrium real short term interest rate" which was adopted to make the empirical analysis feasible. These two questions cast some doubt on the validity of the empirical results, and hence these results should be interpreted with great care.
Views expressed in the paper are those of the authors and do not necessarily reflect those of the Bank of Japan or Institute for Monetary and Economic Studies.