摘要
本文论证了双曲模型是描述中国货币市场利率动态变化的最佳单因子利率模型。由极大似然估计可以得到单因子利率模型的边际密度函数。双曲模型的边际密度和非参数估计得到的边际密度函数拟合较好,其表现远远优于几个常见的利率模型(CIR、CKLS和AG模型)。与较一般的At-Sahalia模型相比差别很小,但参数形式得到简化,似然比检验也支持这一点。双曲模型在刻画利率的均值回复特征方面还克服了AG模型的不足。
This article demonstrates that hyperbolic model is the most suitable single-factor interest rate model to describe the dynamics of Chinese money market interest rates. The marginal densities of single-factor interest rate models can be obtained by maximum likelihood estimation. The marginal density of hyperbolic model can fit the marginal density obtained by nonparametric estimation very well, outperforming some popular interest rate models, such as CIR, CKLS and AG model; the difference between hyperbolic model and Ait-Sahalia model is small, but the parametric form of hyperbolic model is reduced, and the likelihood ratio test also support this. Furthermore, hyperbolic model characterize the mean reversion of interest rates better than AG model.
出处
《数量经济技术经济研究》
CSSCI
北大核心
2006年第12期54-63,共10页
Journal of Quantitative & Technological Economics
关键词
极大似然估计
非参数估计
边际密度函数
双曲模型
Maximum Likelihood Estimation
Nonparametric Estimation
Marginal Density Function
Positive Hyperbolic Model