322 | 0 | 51 |
下载次数 | 被引频次 | 阅读次数 |
本文在融合了金融中介理论(信贷、货币周期)的美林时钟框架下,结合VAR向量自回归预测、LSTM宏观经济预测模型,采用Black-Litterman策略进行大类资产配置.本文以信贷货币周期理论基础,拓展了美林时钟经济周期划分判别准则,将2001年—2021年国内的宏观经济运行状况划分成不同的经济状态.紧接着,构建了基于LSTM深度学习宏观经济预测模型,模拟中国未来五年的经济增长、通胀、利率(反映货币政策松紧程度)等宏观经济环境,以及未来四类资产(股票、大宗商品、债券、现金及其等价物)的四个指数收益率.随后,本文使用单资产配置、避险配置以及Markowitz夏普比率最大进行资产配置.但是Markowitz策略缺点是对于预期收益率的输入比较敏感.根据前面未来宏观经济环境的预测,本文发现未来的经济状态的主旋律主要是滞涨和衰退经济轮动,本文认为由于经济环境的利率上升和信贷渠道的萎缩导致的流动性下降,现金和大宗商品相对于股票和债券相对而言比较强势.为了更好的融入本文对于未来经济周期的观点,本文采用Black-Litterman资产配置策略.最后本文在计算了不同资产配置策略下投资组合的风险收益特征,并结合各自策略特性进行比较与分析.
Abstract:In this paper, Black-Litterman strategy is adopted for asset allocation under the Merrill Lynch clock framework that integrates financial intermediation theory(credit and money cycle), combined with VAR vector autoregressive forecasting and LSTM macroeconomic forecasting model. Based on the theory of credit and money cycle, the classification criterion of Merlin clock economic cycle is expanded, so as to divide the domestic macroeconomic operation status into different economic states from 2001 to 2021.A deep learning macroeconomic forecasting model based on LSTM is constructed to simulate China's macroeconomic environment in the next five years, including economic growth, inflation, interest rate(reflecting the degree of monetary policy tightening), as well as four index yields of four types of assets(stocks, commodities, bonds, cash and their equivalents). Finally, single asset allocation, hedge allocation and Markowitz Sharpe ratio maximum is used for asset allocation. However, Markowitz strategy is sensitive to the input of expected rate of return. According to the previous prediction of the future macroeconomic environment, this paper finds that the main theme of the future economic state is stagflation and recession economic rotation. This paper believes that due to the rise of interest rates in the economic environment and the decline of liquidity caused by the contraction of credit channels, cash and commodities are relatively strong compared with stocks and bonds. In order to better fit into this paper 's views on future economic cycles, Black-Litterman asset allocation strategy is adopted. Finally, this paper calculates the risk-return characteristics of the portfolio under different asset allocation strategies, and combines them with their own strategy characteristics for comparison and analysis.
[1]次必聪,张品一.基于ARIMA-LSTM模型的金融时间序列预测[J].统计与决策,2022,38(11):145-149.
[2]杜清源,龚六堂.带“金融加速器”的RBC模型[J].金融研究,2005(4):16-30.
[3]郜哲.基于投资时钟原理的中国大类资产配置研究与实证[J].河北经贸大学学报,2015,36(3):49-54.
[4]胡永刚,刘方.劳动调整成本、流动性约束与中国经济波动[J].经济研究,2007,42(10):32-43.
[5]雷海波.数字征信是否改善了信贷使用?———基于中国县域数据的实证研究[J].南方金融,2021(10):37-48.
[6]吕朝凤,黄梅波.习惯形成、借贷约束与中国经济周期特征———基于RBC模型的实证分析[J].金融研究,2011(9):1-13.
[7]欧阳红兵,黄亢,闫洪举.基于LSTM神经网络的金融时间序列预测[J].中国管理科学,2020,28(4):27-35.
[8]任泽平,方思元,梁珣.改良的中国投资时钟:周期轮动和大类资产表现[EB/OL].[2023-06-05].https://mp.weixin.qq.com/s/Xp2k48XoACtWE18GHzfD-A.2021-03-04.
[9]王国静,田国强.金融冲击和中国经济波动[J].经济研究,2014,49(3):20-34.
[10]张燃.信用价差变化的决定因素———一个宏观视角[J].当代财经,2008(9):62-67+83.
[11] ANSON M J P. Strategic versus tactical asse t allocation[J]. The Journal of Portfolio Management,2004,30(2):8-22.
[12] BERNANKE B S,BOIVIN J,ELIASZ P. Measuring the effects of monetary policy:a factor-augmented vector autoregressive(FAVAR)approach[J]. The Quarterly Journal of Economics,2005,120(1):387-422.
[13] BERNANKE B S, GERTLER M. Inside the black box:The credit channel of monetary policy transmission[J]. The Journal of Economic Perspectives,1995,9(4):27-48.
[14] BLACK F,LITTERMAN R. Global portfolio optimization[J]. Financial Analysts Journal,1992,48(5):28-43.
[15] DIAMOND D W. Financial intermediation and delegated monitoring[J]. The Review of Economic Studies,1984,51(3):393-414.
[16] HE Y,LI J M,RUAN S M,et al. A hybrid model for financial time series forecasting-integration of EWT,ARIMA With the improved ABC optimized ELM[J]. IEEE Access,2020,8:84501-84518.
[17] HOCHREITER S,SCHMIDHUBER J. Long short-term memory[J]. Neural Computation,1997,9(8):1735-1780.
[18] IBBOTSON R G,KAPLAN P D. Does asset allocation policy explain 40,90,or 100 percent of performance?[J]. Financial Analysts Journal,2000,56(1):26-33.
[19] SIMS C A. Macroeconomics and reality[J]. Econometrica:Journal of the Econometric Society,1980:1-48.
[20] TAYLOR G W. Composable,distributed-state models for high-dimensional time series[M]. Toronto:University of Toronto,2009.
基本信息:
DOI:
中图分类号:F832.51
引用信息:
[1]缪智伟,黄国文.信用货币周期下大类资产配置[J].汕头大学学报(自然科学版),2024,39(02):17-30.
基金信息: