关键词: 高频宏观因子, 利率期限结构, 非规则混频

Abstract: Timely reflection of the dynamic mechanism of financial market and macro economy is the key to making forward-looking monetary policies and investment decision analysis in the era of big data. Major emergencies have increased the uncertainty of dynamic mechanism analysis, so that it is necessary to mine more high-frequency information for effective identification. Therefore, an Irregular Mixed-Frequency Macro-Factor Nelson-Siegel (IR-MF-NS) Model is proposed and tested based on Chinese Treasury bond yields and macroeconomic information. As the results show, firstly, compared with the traditional model, the new model shows better sample fitting performance. Secondly, we find a bidirectional mechanism between the term structure of treasury bond yields and the macro economy. There is a negative correlation between the real economic activity and the slope factor, which is consistent with the goal of counter-cyclical regulation in China; and the high-frequency curve factor has a significant positive and persistent impact on inflation, which verifies the conclusion that the curve factor plays an important role in high-frequency data. Thirdly, the mechanism is affected by the information frequency. The “smoothing effect” of low-frequency data makes the low-frequency mechanism show the stable effect. The “reversal effect” caused by information fluctuation in high-frequency data makes the low-frequency mechanism and high-frequency mechanism diverge. Finally, the correlation between treasury bond yields and macro fundamentals weakens systematically after the major emergency, and the contribution of term structure factors to forecasting variance is also weakened, while the macro fundamentals play a more important role in predicting the variance of the treasury bond yields.

Key words: High-Frequency Macro-Factor, Term Structure of Interest Rates, Irregular Mixed-Frequency