摘要
采用NS混合模型动态估计中国利率期限结构,考察动态NS模型,无套利NS模型及广义无套利NS模型等NS混合模型对我国利率期限结构的动态估计效率,比较NS混合模型的样本外预测能力,检验无套利约束对混合模型动态估计的影响.本文的经验分析结果表明:无套利条件的引入增强了NS混合模型的样本内动态估计能力和样本外预测能力;五因素的广义无套利NS模型(AFGNS)无论在利率期限结构样本内动态估计还是在总体预测效率上都要高于其他模型,可将其作为利率期限结构研究的基础模型:
We used Hybrid Nelson-Siegel models,such as DNS,AFNS and AFGNS,to dynamically estimate the term structure of interest rates in China.We investigated the dynamic estimation efficiencies of these Hybrid Nelson-Siege models,compared the prediction capabilities of the Hybrid models,and also tested the impact of the Arbitrage-Free restriction for NS models.We found that,by adding an Arbitrage-Free restriction,the out of sample forecasting and in sample estimating capability of the Hybrid Nelson-Siegel model can be enhanced.The five-factor AFGNS model has a better performance on both in-sample and out-of-sample dynamic estimation than the other models,and it can be applied for the term structure research as a basic model.
出处
《数学的实践与认识》
CSCD
北大核心
2013年第20期1-11,共11页
Mathematics in Practice and Theory
基金
国家自然科学基金(71173030)
教育部人文社会科学重点研究基地重大项目(2009JJD790004)
辽宁省教育厅高等学校创新团队研究项目(WT2010009)
关键词
利率期限结构
NS混合模型
动态估计
卡尔曼滤波
term structure of interest rate
hybrid nelson-siegel model
dynamic estimation
kalman filter