债券投资新探索：基于机器学习的利率择时策略实践
洪钰,李燕婷,万淑珊,朱隼,陈奕安,鲁炎
摘要(Abstract)：
利率择时是债券资产配置的重要基石之一。传统利率预测方法一般是以宏观经济指标或动量指标为出发点,预测未来结果,但并没对宏观环境及其变迁进行建模识别和预测。本文创新地提出一种基于平安独有的MEC债券投资时钟路由器的模块,通过动态识别当前市场状态,将预测样本“路由”至最优的子模型,充分利用不同模型在不同经济环境下的优势,提升模型预测精度,产出有效信号。本文以2010—2022年的10年期国债收益率作为预测目标,在7-10年国债财富指数上回测交易。引入模块后,年化收益从2.97%提升到4.03%,夏普比从0.33提升至0.72,回撤从4.18%降至2.15%,胜率由64.52%升到68.82%,策略表现有显著提升。
关键词(KeyWords)： 利率择时;机器学习;投资时钟;债券投资
基金项目(Foundation):
作者(Author): 洪钰,李燕婷,万淑珊,朱隼,陈奕安,鲁炎
DOI: 10.16453/j.2096-5036.2023.02.005
参考文献(References)：
- [1]邓燊,杨朝军.基于经济周期的资产配置研究[J].金融理论与实践,2011(1):5.
- [2]陈健恒,范阳阳,东旭,等.经济周期、金融周期与利率周期的轮动[R].中金研究,2023.
- [3]宋唯实,周萧潇.量化配置系列(12)宏观数据建模应用手册[R].中金研究,2023
- [4]LIN H,ZHOU D,LIU W,et al.Learning multiple stock trading patterns with temporal routing adaptor and optimal transport[C]//Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery&Data Mining,August 14-18,2021,Singapore.New York:Association for Computing Machinery,2021.
- [5]周金涛.涛动周期论:经济周期决定人生财富命运[M].北京:机械工业出版社,2017.
- [6]高志刚.产业经济学[M].北京:中国人民大学出版社,2016.
- [7]拉斯·特维德.逃不开的经济周期:历史,理论与投资现实[M].北京:中信出版社,2012.
- [8]董德志.投资交易笔记:续:2011-2015年中国债券市场研究回眸[M].北京:经济科学出版社,2016.
- [9]亚伯,伯南克,克鲁肖,等.中级宏观经济学:Macroeconomics[M].任曙明,马强,译.北京:机械工业出版社,2009:220-222.
- [10]燕翔,许茹纯,朱成成,等.宏观周期时钟类投资框架综述[R].方正证券,2023.
- [11]TREVOR G,MICHEAL H.The investment clock-special report:making money from macro[R].Merrill Lynch,2014.
- [12]周岳.资产配置深度报告:以普林格时钟为鉴，我们如何搭建资产配置框架?[R].中泰证券,2022.
- [13]王灿华.基于改进型美林时钟分析框架的理财产品设计方案与启示[J].债券,2019(9):7.
- [14]董德志.基于"货币+信用"风火轮方法的我国大类资产配置研究[J].金融发展,2019(1):21.
- [15]COX J C,INGERSOLL J E,ROSS S A.A theory of the term structure of interest rates[J].Econometrica,1985,53(2):385-407.
- [16]刘郁,田乐蒙.利率预测，历史致敬未来[R/OL].广发证券.2021.https://mp.weixin.qq.com/s/Lpe Thky5rj3pt W3tw3Lqq A.
- [17]CHRISTENSEN J,DIEBOLD F X,RUDEBUSCH G D.An arbitrage‐free generalized Nelson-Siegel term structure model[J].Econometrics Journal,2009,12(3):C33-C64.
- [18]高宇.“债券定价逻辑”下的宏观数据库-利率系列(货币、财政、汇率、资金面)[R].浙商证券,2023.
- [19]张超,罗军.技术分析框架下的利率择时和久期轮动策略[R].广发证券,2021.
- [20]LEIPPOLD M,WANG Q,ZHOU W.Machine learning in the chinese stock market[J].Journal of Financial Economics,2021(1):64-82.
- [21]KANEVSKI M F,TIMONIN V.Machine learning analysis and modeling of interest rate curves[C]//Proceedings of the 18th European Symposium on Artificial Neural Networks(ESANN 2010),April 28-30,2010,Bruges,Belgium.ESANN,2010.
- [22]NUNES M,GERDING E H,MCGROARTY F,et al.Artificial Neural Networks in Fixed Income Markets for Yield Curve Forecasting[J/OL].SSRN Electronic Journal,2018.https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3144622.DOI:10.2139/ssrn.3144622.
- [23]NUNES M,GERDING E H,MCGROARTY F,et al.The Memory Advantage of Long Short-Term Memory Networks for Bond Yield Forecasting[J/OL].SSRN Electronic Journal,2019.https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3415219.DOI:10.2139/ssrn.3415219.
- [24]SHU L C,CHOU J K.Using Deep Learning Techniques to Predict 10-Year US Treasury Yield[C]//2021 11th International Conference on Information Science and Technology (ICIST).IEEE,2021.
- [25]任瞳,周靖明,罗星辰.引入宏观维度的改进dtw算法在择时策略中的应用[R/OL].招商证券,2023.http://www.yanbaohui.com/doc-9f54371720a842757d7eb9e863800a6e.html.
- [26]ZHANG R,YI C,CHEN Y.Explainable Machine Learning for Regime-Based Asset Allocation[C]//2020 IEEE International Conference on Big Data (Big Data).IEEE,2020.
- [27]AGRESTI A.Categorical Data Analysis[J].Journal of the American Statistical Association,1991.