中国利率期限结构平滑样条拟合改进研究

Research on improving smoothing spline method to fit China's term structure of interest rates

在FNZ模型和W aggoner模型的基础上,结合中国国债市场的发展现状,从拟合对象、粗糙度惩罚项的函数形式、参数估计方法、最优化目标函数形式以及估计样本等五方面进行修正和改进,提出利用可变粗糙度惩罚项三次平滑样条改进模型来拟合中国的利率期限结构,并利用上交所国债市场2002年1月1日至2003年12月31日的国债收盘价格数据对该模型进行实证研究.结果发现,笔者提出的改进模型能够较合理、有效地估计较为完整的中国静态利率期限结构.

This paper considering the current situation of the treasury security market in China, improves the FNZ model and Waggoner model from five aspects： the fitted object, the function in describing the roughness penalty, the method of estimating parameters, the optimal function and the sample. Then, we propose to approximate China＇ s term structure of interest rates with the improved variable roughness penalty cubic smoothing spline approach. Also, this paper carries out an empirical work with the daily treasury security close prices data which are extracted from the SSE treasury security market. Finally, both the in-sample and out-of- sample results indicate the improved model, can reasonably fit the relatively integrated static term structure of interest rates in China.