最优多因子仿射利率期限结构模型选择

The Optimal Multi-factor Affine Term Structure Model Selecting

在利用NS模型估计出市场即期利率的基础上,采用卡尔曼滤波方法对多因子Vasieck和CIR模型进行参数估计,最后运用蒙特卡罗模拟方法对交易所国债价格进行模拟,并与实际价格进行比较,进而确定了符合我们国债市场的最优多因子仿射利率期限结构模型。研究结果表明:多因子CIR模型对数据的拟合效果及对国债价格模拟效果要明显优于多因子Vasicek模型;对于多因子CIR模型而言,因子个数增加并没有提高模型的价格模拟效果;两因子CIR模型具有最优的国债价格模拟效果。

The article firstly uses the Kalman filtering method to estimate the multi-factor Vasieck and CIR models based on the spot rate fitted by NS model,and then uses the method of Monte Carlo simulation to simulate the bond price to select the optimal multi-factor affine term structure model which fits our debt market.The results show that the effect of multi-factor CIR models fitting data and simulating bond prices is much better than the multi-factor Vasicek models;the increase in the number of factors does not mean the increase of effect of simulation effect for the multi-factor CIR models;two-factor CIR model has the best effect of bond price simulation.