Term structure of interest rates for Malaysian fixed income security

This study investigates the term structure of interest rates for the Malaysian fixed income security. This study is relevant because Malaysia has one of the fastest growing bond markets in the region. The general objective of this research is to study the behavior of the term structure of interest rates of Malaysian fixed income security by focusing on the issues related to the predictive power of forward to the future spot rates, long rates to short rates, and forward rates to inflation. Another important objective is to explore the efficiency of Malaysian fixed income security market in terms of prices and yields and to find out to what extent they are driven by the market forces. The final objective here is to interpret the behavior of the yield curve of the term structure of interest rates of Malaysian fixed income security market and to find out which theories can explain the behavior of this market, in addition to testing the liquidity premium theory. This research differs from earlier studies in several ways. First, due to the existence of limited number of studies that focus on Malaysian bond market, this investigation addresses several issues in corporate bond market. In addition, this study examines the behavior of the term structure of interest rates in the government bond market. Second, this work compares and scrutinizes the behavior of the yields in both government and corporate bond markets. Third, an analysis on Malaysian corporate bond market seems to be less common in this context. A number of important research questions can be raised in this regard. Are the forward rates useful in the prediction of future spot rates and inflation? Does the market reflect the true value of the security by allocating fair yields and prices to investors? Which theory of the term structure can explain the behavior of Malaysia fixed income security market? The research has responded aforementioned questions by employing CIR as one-factor model as well as the extended BDT two factor model with the use of Financial Mathematical Computational tools in MATLAB. Here, it is worth mentioning that special code has been developed to calculate the zero curves and forward rates. The market participants in MGS and MCB based their investment decision on their own preferences according to their assets and liabilities. The interpretation of the TSOIR and yield curve shapes is influenced by the dominance of the institutional investors in the market as giant players. Interestingly, without a doubt the extended BDT two-factor model confirms the results obtained by CIR one-factor model especially in predicting the shapes of the compartmentalized market yields in both MGS and MCB which are humped curves with few exceptions. The yield-analysis by BDT indicates that the market mis-allocated the yield as the model fair yields for the entire issues of MGS are higher than the market current yields resulting in investors not getting what they deserve to get from their investments. In addition, the yield-analysis results by BDT for MCB are found to be similar to the result obtained from MGS except for B1 as the current market yield is higher than the model fair yield. Here, it also deserves emphasis that priceanalysis of MGS reveals that model fair prices are higher than the current market prices. However, MCB exhibited different behavior for its different classes as the model fair prices are found to be higher as well as lower than the market prices for a set of bond classes. Interestingly, the yield-analysis reveals that both CIR and BDT models produce similar results. Here, it should be emphasized that although CIR model is conventionally a one-factor model, it can successfully capture the dynamics of interest rates. The mean YTM analysis by CIR for MGS and the eight different classes of Malaysian corporate bond when comparing the different maturities within the same class of bond fail to support LPT proposition. While the mean-analysis result by BDT shows clear evidence of the liquidity premium when investors holding the MGS to the maturity date. The mean YTM reveals normal upward curve with the increase of the maturity spectrum. This probably compensates the risk-averse investors to offset the risk of uncertainty in long term investment. However, the mean-analysis results by BDT for the MCB shows mixed evidence. The spread-analysis which compares the same maturities within MGS and different classes of the corporate bond by CIR model shows mixed results. The analysis confirms some evidence of the LPT as the investors in some classes of the corporate bond in certain maturities will be getting positive access return in addition to the risk-free rates. In a similar vein the spread-analysis by BDT reveals that each maturity spectrum has different segmented behavior with inbuilt liquidity premium. The higher classes of MCB exhibited positive spread while medium and lower classes shows mixed spread results. The findings of the spread-analysis deserve a special emphasis in the sense that the MCB investors may formulate their investment strategy for their portfolio by using MGS as a benchmark. Turning to the predictive power of the variables of interest following findings deserve emphasis. The long rates obtained via CIR model could be incorporated in the model to predict the short rates. While Granger non-causality test between short and long rates for MGS suggest a unidirectional causality that runs from long rates to short rates. In addition, for the predictive power of the forward rates to future spot rates by BDT model, the pooled OLS does not seem to capture the dynamics of the variables of interest. However, there is strong bi-directional causality between the variables of interest. For the predictive power of the forward rates to the future inflation, the regression leads to misleading results for the standard errors. OLS does not capture the linear relation between the forward rates and inflation. Meanwhile the Granger non-causality tests between forward and inflation rates for MGS do not provide any causal relation between the forward and inflation rates. The bounds test results support the cointegration among the variables of short and long rates under CIR model at 10% significance level. The result for the variables of the forward and spot rates by BDT model reveals that the computed statistics are supportive for an existence of cointegration at 5% significance level. Finally the bounds test result for the variables of inflation and forward rates suggest that there is a strong cointegration. The cointegration analysis reveals significant empirical support for the validity of the expectation hypothesis in Malaysian fixed income security. Finally, the overall research outcome would be of great importance to investors and other market participants in formulating their own strategies based on the information which can reflect the true value of the securities.

Preview Text FEP 2012 16 IR.pdf Download (999kB) | Preview

Actions (login required)

View Item