硕士学位论文中国国债利率期限结构研究——动态模型及预测.doc

硕士研究生学位论文
中国国债利率期限结构研究——动态模型及预测
摘要
本文以Nelson-Siegel类利率期限结构静态模型为基础构建了利率期限结构的动态模型,并选取中国国债即期收益率数据进行样本内实证以及样本外预测。目前对国内利率的研究主要集中在对期限结构静态模型和短期利率动态模型的研究。对于整条利率期限结构曲线的动态行为的研究以及动态模型预测能力的研究还处于空白。不同于传统的动态模型只关注瞬时利率的动态行为,本文所构建的动态模型很好的拟合了整条利率曲线的动态过程,并且模型中的隐含变量可以很好的解释为利率期限结构的水平,斜率和曲度。在对数据进行动态建模后,使用样本外数据对模型的预测能力进行了全面的分析和比较。我们发现,在短期1个交易日的预测中,本文所建立的利率期限结构动态模型的预测效果显著的差于三个对比模型(随机游走模型,利率水平的AR(1)过程和利率变化的AR(1)过程),而在短期5个交易日的预测中,我们的利率期限结构动态模型与三个对比模型没有显著的差别。在长期预测中,利率期限结构动态模型要显著优于三个对比模型。
关键词:利率模型;Nelson-Siegel模型;国债即期收益率;模型预测
The Chinese treasury yield term structure: dynamic models and their forecasting ability
Han Haifeng (Major in Economics)
Directed by Professor Jin E. Zhang
Abstract
We use dynamic Nelson-Siegel class models to model and forecast the Chinese treasury yield term structure, using in-sample data and out-of-sample data separately. Most of existing researches on the Chinese interest rate focus on the statistic model of term structure and the dynamics of short rate. As far as we know, there is no published research on the dynamics of the whole term structure and the forecasting ability of various dynamic models, using Chinese data. As we showed in the paper, the dynamics of the Chinese treasury yield term structure is fitted well by the dynamic Nelson-Siegel class models. We estimate the statistic Nelson-Siegel class models day by day, and find that the time-varying factors may be interpreted as factors corresponding to level, slope and curvature. Then we build dynamic models for the time-varying factors, using five different dynamic processes. We produce forecasts of the term structure by dynamic Nelson-Siegel class models at both short and long horizons, pare to the forecasts produced by three benchmark models. In particular, our forecasts appear more accurate than the benchmark models at long horizons forecast. But at short horizons, the benchmark models seem produce better forecast.
Keywords: Nelson-Siegel class models; Term structures; Dynamic model;
Chinese Tr