Nelson-Siegel模型中的时变载荷因子及其对提高经济形势预测的重要作用  

Time-varying Loading Factor in Nelson-Siegel Model and Its Important Role in Improving Economic Situation Prediction

作　　者：张靖泽 沈根祥[2] Zhang Jingze;Shen Genxiang

机构地区：[1]中泰证券股份有限公司博士后科研工作站 [2]上海财经大学经济学院

出　　处：《统计研究》2024年第2期77-86,共10页Statistical Research

基　　金：国家社会科学基金青年项目“关联图谱视角下资本市场系统性风险传导及预警研究”(19CJL041)。

摘　　要：现有文献中Nelson-Siegel模型大多将载荷因子λ提前固定为常数或作为待估静态参数,较少考虑载荷因子λ的时变化。本文基于得分驱动时变参数建模方法,在状态空间模型框架内考虑时变载荷因子λ,构建GAS-λ-DNS模型,并提取相应时变载荷因子。结果显示,时变载荷因子λ_(t)呈现出极强的波动性,同时与经济周期密切相关;将时变载荷因子λ_(t)应用于预测工业产出增速,发现载荷因子相比于传统宏观预测因子具有额外增量信息,引入λ_(t)可以显著提高模型预测精度。本文结论对Nelson-Siegel利率期限结构建模和宏观经济预测因子选取具有参考价值。In general, the relevant literature of Nelson-Siegel models makes the loading factor λ a constant in advance or a static parameter to be estimated, and the time variation of the loading factor is rarely considered. Based on the score-driven time-varying parameter modeling method, we give the loading factor time-varying under the state space model, then construct GAS-λ-DNS model. The results show that the time-varying loading factor λ_(t) has high volatility and is closely related to the economic cycle.When the time-varying loading factor λ_(t) is used to predict the growth rate of industrial output, it is found that the loading factor has additional incremental information compared with the traditional macro predictors and λ_(t) can significantly improve the prediction accuracy of the models. The conclusion of this paper has reference value for Nelson-Siegel interest rate term structure modeling and the selection of macroeconomic predictors.

关 键 词：动态Nelson-Siegel模型 GAS模型 时变载荷因子 宏观经济预测 

分 类 号：F064.1[经济管理—政治经济学]