基于联合估计法的债券信用利差期限结构拟合及应用

基于2016~2018年月度数据,通过独立估计的单曲线样条模型和SV 模型、联合估计的多曲线样条模型和SV模型拟合公司债信用利差期限结构,进而对模型拟合效果进行比较,讨论模型在宏观经济预测中的应用,得到以下结论:(1)拟合模型的函数形式是导致理论信用利差期限结构曲线翻折的原因。样条模型和SV模型拟合的信用利差曲线形状完全不同,且模型函数变动引起的误差变动大于曲线变动引起的误差变动。(2)联合估计模型可以修正独立估计模型的人为扭曲形式。多曲线模型的结果更接近实际信用利差,误差波动性明显减小,曲线更为平滑,且联合估计的多曲线样条模型优于独立估计的单曲线样条模型、独立估计的SV 模型和联合估计的SV模型。(3)公司债信用利差期限结构在一定程度上蕴含了市场对未来宏观经济的预期信息,且在短期内预测结果随先行期限延长而改善。因此,宏观经济政策制定者需关注信用利差和期限结构模型拟合研究,重视对信用利差期限结构的深度信息挖掘,从而提高中国宏观政策制定者调控手段的前瞻性和有效性。

Fitting Term Structure of Bond Credit Spread Based on Joint Estimation Method and Its Application

Based onmonthly frequency data from 2016 to 2018, this paper fits the term structure of credit spreads by four methods: single-curve spline model and SV model of independent estimation, multi-curve spline model and SV model of joint estimation,and compares the applicability of the model, anddiscusses the practical application of fitting model in macroeconomic forecasting. Three empirical conclusions are drawn. Firstly, the function form of model is the cause of the reversal of the term structure curve of theoretical credit spreads. The shape of the credit spreads curve fitted by spline model and SV model is totally different. The fitting curve is largely determined by the function of models. Secondly, the joint estimation model can modify the artificial distortion of the independent estimation model. The results of the multi-curve model are closer to the actual credit spreads, with a significantly reduced error volatility. Thirdly, the term structure of corporate bond credit spreads implies some information about the future macro-economy. And in the short term, the forecast results become better with the extension of the advance period. Therefore, macroeconomic policymakers should pay attention to the credit spreads and the information implied by it.