本文构建了具有平方根扩散特征的三因子仿射利率期限结构模型,给出了基于卡尔曼滤波法的模型参数估计过程,利用蒙特卡罗模拟对我国国债进行定价预测,并与Longstaff-Schwartz模型、Vasicek模型、Cox-Inger-soll-Ross模型的定价效果进行实证比较。结果表明多因子模型要优于单因子模型,双因子模型要略优于三因子模型,从而为我国国债合理定价提供技术支持。
The three-factor affine term structure models with square-root diffusion process are developed in this paper,and the Kalman filter method to estimate the parameters of the model is given.So the prices of treasury bonds are analyzed by Monte Carlo stimulation.The pricing results of Chinese treasury bonds are compared with Longstaff-Schwartz model,Vasicek model and Cox-Ingersoll-Ross model.The results show that multi-factor models are superior to the single factor,and the two-factor affine model has the higher precision than three-factor model.It can provide the technology support for effectively pricing the treasury bonds in China.