基于动量与风险优化双重视角的ETF行业轮动策略研究

Research on ETF Sector Rotation Strategy Based on Dual Perspectives of Momentum and Risk Optimization

摘要: 本文构建了一种基于动量因子与协方差矩阵收缩优化的ETF行业轮动策略。策略首先通过线性回归模型计算各ETF的复合动量得分，筛选趋势强度与稳定性兼具的投资标的；随后采用协方差矩阵收缩方法优化组合权重，以降低参数估计误差并实现风险控制；最终通过月度调仓机制实现动态资产配置。基于2021年6月至2025年12月的回测结果显示，策略累计收益达74.81%，相较沪深300指数获得101.30%的超额收益，最大回撤为25.97%，夏普比率为0.567，索提诺比率为0.825。进一步的稳健性检验表明，在不同动量窗口、风险厌恶系数及协方差收缩强度设定下，策略风险收益特征保持稳定，验证了模型并非依赖特定参数设定。研究结果表明，动量信号与风险优化机制的有效结合能够为行业ETF轮动配置提供具有实践价值的量化投资框架。

Abstract: This paper proposes an ETF sector rotation strategy based on a momentum factor and covariance matrix shrinkage optimization. The strategy first constructs a composite momentum score using linear regression to identify ETFs with both strong and stable price trends. Portfolio weights are then optimized through covariance matrix shrinkage to reduce estimation errors and enhance risk control, followed by a monthly rebalancing mechanism for dynamic allocation. Backtesting from June 2021 to December 2025 shows a cumulative return of 74.81% and an excess return of 101.30% over the CSI 300 Index, with a maximum drawdown of 25.97%, a Sharpe ratio of 0.567, and a Sortino ratio of 0.825. Robustness tests further demonstrate that the strategy maintains stable risk-return characteristics under different momentum windows, risk-aversion parameters, and shrinkage intensities, indicating that performance does not rely on specific parameter settings. The results suggest that integrating momentum signals with risk optimization provides a practical quantitative framework for ETF sector rotation allocation.