关键词: 经济政策不确定性, 杠杆效应, 短期利率波动, BHK-L-MIDAS模型, MCS检验

Abstract: The short-term interest rate is a key variable in the pricing of fixed-income securities and derivatives. Meanwhile,the volatility of short-term interest rate has important impact on the asset allocation for investors. The short-term interest rate displays mean reversion,and its empirical distribution exhibits stylized facts such as leptokurtosis and fat tail. In addition,the short-term interest rate volatility changes over time and responds asymmetrically to good and bad news. If those stylized facts of short-term interest rate are ignored,it may affect the optimal asset allocation for investors as well as the accuracy of derivatives valuation. As a consequence,developing a rational model to model and forecast the short-term interest rate volatility is of great importance.
　This paper contributes to the literature on the modelling of short-term interest rate volatility in three aspects. Firstly,we develop a new model,namely the BHK-L-MIDAS model,which is able to capture the leverage effect (volatility asymmetry) as well as the impact of economic policy uncertainty (EPU) on the short-term interest rate volatility,to model and forecast the short-term interest rate volatility. Secondly,we employ China EPU as a proxy for the EPU,and incorporate it into the BHK-L-MIDAS model to examine the link between the China EPU and short-term interest rate volatility. By doing so,this study provides a new perspective for modelling and forecasting the short-term interest rate volatility. Meanwhile,this study highlights the importance of incorporating the leverage effect for modelling the short-term interest rate volatility by empirically comparing the performance of the BHK-L-MIDAS and BHK-MIDAS models. Finally,based on various loss functions and the model confidence set (MCS) test,this paper examines the predictive ability of the BHK-L-MIDAS model for the short-term interest rate volatility. Furthermore,a VaR analysis is conducted to assess the economic value of the BHK-L-MIDAS model for short-term interest rate market risk measurement.
An empirical application to Shanghai Interbank Offered Rate (SHIBOR) and Chinese EPU index data based on the BHK-L-MIDAS model shows that the short-term interest rate exhibits a reverse leverage effect,and the EPU has a significantly negative impact on the short-term interest rate volatility. The BHK-L-MIDAS model outperforms the BHK and BHK-MIDAS models in in-sample fitting. Furthermore,out-of-sample analysis based on three loss functions and the MCS test suggests that the BHK-L-MIDAS model yields more accurate out-of-sample volatility forecasts than the BHK and BHK-MIDAS models. In particular,the superior forecast ability of the BHK-L-MIDAS model is robust to different forecasting windows. Finally,an empirical application to VaR estimation suggests the economic value of the BHK-L-MIDAS model for short-term interest rate market risk measurement.
Our empirical findings provide useful insights for researchers and financial practitioners. In fact,our study is important to researchers who are trying to understand the dynamic nature of the short-term interest rate volatility. It is also of great significance for financial practitioners who are concerned about financial applications,such as the interest rate risk management. It is worth pointing out that our work could be extended to several directions. For example,we can extend the BHK-L-MIDAS model to incorporate the jump dynamics to capture the time-varying jump in the short-term interest rate. In addition,combing the BHK-L-MIDAS model and the Copula or VAR approach to study the co-movement (spillover effect) of short-term interest rate volatilities between two countries is also worth future research.

Key words: economic policy uncertainty, leverage effect, short-interest rate volatility, BHK-L-MIDAS model, MCS test