摘要
本文研究了基于不同波动率预测模型预测收益误差的波动率,在交易成本影响下的动态最优策略应用到商品期货的表现。将等权重策略作为本文模型的基准模型,分别使用样本协方差、GARCH型波动率模型和精度矩阵模型进行对比,并且比较不同向前预测长度的波动率预测策略的表现。在本文通过夏普率、净夏普率、以及换手率等策略评估指标来评价策略的表现。研究发现:向前预测长度对最优策略的表现影响不大,不同的波动率预测模型的策略表现差异很大,选择合适的波动率模型能带来超额收益。具体而言,样本协方差因相对较低的换手率表现较为稳定;从夏普率和净夏普率表现来看,应用不同波动率估计模型的最优策略均由优于等权策略,Efron等(1976)[1]提出的对精度矩阵直接估计的估计量表现优于样本协方差矩阵估计和其它波动率估计模型。
This paper studies the performance of dynamic optimal strategies with transaction cost based on different covariance matrix estimators applied to commodity futures.The equal weight strategy is used as the benchmark in this paper.we compare the sample covariance matrix,GARCH volatility model and precision matrix model,meanwhile,we compare the strategy performance in different fixed rolling windows.In this paper,the strategy performance is evaluated by three indicators:Sharpe ratio,net Sharpe ratio,and turnover.We conclude that the effect of rolling length in the optimal strategy is not very large,the gap of different volatility forecasting model in optimal strategy is very large,it can gain excess return to choose proper volatility model.In particular,sample covariance matrix is stable because of relative low turnover;From the aspect of the Sharpe ratio and net Sharpe ratio,the optimal strategy applied different volatility forecasting model is better than the equal weight strategy,Efron,et al(1976) putting forward the best estimator of the precision matrix is better than other volatility forecasting model.
作者
崔翔宇
杨澜芝
CUI Xiang-yu;YANGLan-zhi(School of Statistics and Management,Shanghai University of Finance and Economics,Shanghai 200433,China)
出处
《数理统计与管理》
CSSCI
北大核心
2023年第5期937-950,共14页
Journal of Applied Statistics and Management
关键词
动态投资组合优化
交易成本
GARCH型波动率模型
精度矩阵模型
dynamic portfolio optimization
transaction cost
GARCH volatility model
precision matrix dynamic portfolio optimization
transaction cost
GARCH volatility model
precision matri model
