基于动态Svensson模型的国债利率期限结构实证分析

Empirical Analysis of Dynamic Svensson Model in Treasure Bond Market

国债利率期限结构的动态演化特征对利率市场化顺利推进和产品定价有积极意义。使用非平滑FamaBliss方法对交易所市场的国债交易数据进行处理,可以获取实证研究的初始数据,然后扩展静态NSS模型为动态Svensson(LSCC)模型,该模型嵌套了动态Nelson-Siegel(LSC)模型。在利用状态空间模型与Kalman滤波估计出参数后,比较LSCC模型与LSC模型对利率期限结构的拟合与预测能力。结果表明:动态Svensson(LSCC)模型具有更加显著的拟合能力和预测能力,能够更好地反映国债利率期限结构的动态特征。

The term structure of treasure bond yield is a very important part of benchmark interest rate and its dynamic characteristic has positive implication for interest rate liberalization and asset pricing.The paper extends NSS model to Dynamic Svensson Model,which nests the dynamic Nelson-Siegel model as a special case.Then we use state space form model and Kalman filtering to estimate two different models with the original data obtained by non-smooth Fama-Bliss method.Comparing with dynamic Nelson-Siegel model,Dynamic Svensson model is shown to have more outstanding fitting and forecasting ability,which can capture more dynamic information of treasure bond yield.