本研究以基礎樣條函數(basic spline model)為基礎,來估計台灣公債市場的利率期限結構。實證結果發現,基礎樣條數對於台灣公債市場債券利率期限結構的估計,具有相當不錯之配適結果,其判定係數高達94.03%。然而台灣公債市場存有流動性不足的問題,因此本研究同時考慮此項限制,以Subramanian(2001)之流動性加權因子來修正基礎樣條函數,實證結果發現考慮流動性限制的修正基礎樣條函數,能大幅提高判定係數達97.6%,顯見加入流動性限制之修正基礎樣條函數更具有高度配適能力。此外,為減少樣條函數在較長到期日出現震盪過大的缺失,本文同時採用Waggoner(1997)所提出之variable roughness penalty (VRP)來修正基礎樣條函數,惟實證結果發現,此一模型對於利率期限結構配適結果之改善相當有限。
This paper uses the Basis spline model to fit the term structure of interest rates in the Taiwan Government Bond market. The empirical results show that the fitting performance of B-spline model is quite well, which has a mean R-square higher that 94.03%. However, the Taiwan Government bond market is an illiquid market, it must consider the illiquidity constraint when estimating the term structure of interest rates. This paper uses the liquidity-weighted objective functions is better than the original one, and the mean R-square increases to 97.6%. Based on the results of this study, we conclude the fitting performance of B-spline model could be effectively improved considering the illiquidity constraint. Meanwhile, in order to reduce oscillations at longer maturities for spline estimating models, we also adopt the variable roughness penalty (VRP) to modify the orginal B-spline model. But, unfortunately, we conclude that the term structure fitting improvement is quite limited.