论文部分内容阅读
本文在最小卡方估计方法基础上研究了高斯仿射利率模型的参数识别和估计问题。以标准化高斯模型为起点,从结构化模型和简约化模型参数的函数关系出发研究高斯仿射模型的可识别性,最小卡方估计量继承了结构化模型极大似然估计量的所有渐进性质并保证了参数估计量的可靠性。以上交所2006-2013年隐含于国债价格月度数据的零息票收益率为样本采用最小卡方方法实证研究了高斯仿射期限模型,结论表明高斯仿射模型很好的拟合了观测的期限结构,并且整体上看简约型和结构型参数估计量的统计性质的优劣具有一致性。
In this paper, based on the minimum chi-squared estimation method, the parameter identification and estimation problems of the Gaussian affine interest rate model are studied. Starting from the normalized Gaussian model, the identifiability of the Gaussian affine model is studied based on the functional relationship between the structured model and the reduced model parameter. The least-squares estimator inherits all the asymptotic properties of the maximum likelihood estimator of the structured model And ensure the reliability of parameter estimation. Based on the zero-coupon yield implied by the Shanghai Stock Exchange’s 2006-2013 monthly data, we use the least-squares chi-squared test to study the Gaussian affine period model. The conclusion shows that the Gaussian affine model fits well the observed period Structure, and overall the consistency of the statistical properties of simple and structured parameter estimates.