基于P-样条方法短期利率模型非参估计

NO-PARAMETRIC ESTIMATION FOR SHORT TERM INTEREST RATE MODEL BY P-SPLINE

引入风险补偿因子,建立半参短期利率模型,使用P-样条方法估计漂移项,并证明了在合适参数的约束条件下,相应的短期利率动态过程是非负的平稳过程.实证研究结果表明考虑半参模型将增加似然函数的估计值,同时考虑风险补偿因子将进一步改善模型的拟合效果.此外具有弹性系数模型比均方根模型能够更好地刻画时间序列数据,而且也发现风险补偿因子对于漂移项非线性现象是比较显著的.

This paper presents the semi-parametric model of short-term interest rates with the risk compensation factor. We show a technique for nonparametrically estimating the drift function by using P-spline approximation. Under the appropriately parameter constraints, the interest rate dynamics process is a non-negative stationary process. Empirical results show that the likelihood function will be improved for the semi-parametric model. Fklrthermore, the model provides the better fitting results by incorporating the risk compensation. Besides, the CEV model could show the better fitting than the CIR model. Finally, there are some evidence of substantial non-linearity in the drift for the model with the risk compensation factor.