中文摘要：

以多因素CIR为基础，运用卡尔曼滤波模拟和估计沪市国债利率的期限结构。卡尔曼滤波方法可有效处理不可观察的向量（状态向量），能综合多因素CIR模型的时间序列特性和横截面信息，可以充分利用可获得的期限结构的其他各种信息。采用的国债利率为用息票剥离方法得到的零息票利率，原始数据为1997年6月27日-2003年2月25日期间交易所的国债现货和国债回购的周交易价格数据，源自中国国债网。多因素模型能很好地刻画收益率曲线的动态变化，统计结果支持四因素CIR模型。实证发现，中国国债期限结构先升后降、而后再次升降，呈明显的驼峰状；长期利率与短期利率差距不大，收益率期限结构扭曲.这也表明合理的国债利率期限结构的确尚未形成。

英文摘要：

The objective of this paper is to simulate and estimate the nominal term structure of interest rates in Shanghai Stock Exchange, basing on the multi-factor versions of the Cox-Ingersoll-Ross（CIR） and utilizing Kalman Filter approach. Zero-coupon interest rate is gained by bootstrapping, employing weekly trade data of governments debt from 27 June 1997 -25 March 2002 in Shanghai Stock Exchange （SSE）. Multi-factor models are necessary to characterize the changing shape of the yield curve over time, and the statistical tests support the case for four factor models. The empirical results indicate that the yield curve of the government bond in SSE looks like a hump. There are not significant differences between short-term interest rates and long term rates, and shape of the yield curve of government＇s debt in SSE is distorted. It also indicates that the reasonable term structure of interest rates is not formed in SSE.