东北固收行业轮动系列专题二：行业关联网络下的二级行业轮动策略

Core Insights - The report emphasizes the importance of industry selection and timing in achieving relative returns and outperforming indices, especially in the context of public fund assessment mechanism reforms [2][13] - A dual-layer framework combining "basic layer + relational layer" is proposed to enhance the coverage of industry rotation strategies, moving from 15 to a broader range of industries [2][14] - The "low position + marginal improvement" strategy has shown significant excess returns, with an annualized return of 11.65%, maximum drawdown of 11.25%, and a Sharpe ratio of 1.86, indicating strong practical value [2][18] Industry Rotation Strategy - The report constructs a framework for industry rotation based on macroeconomic data, which has demonstrated significant excess returns, validating the effectiveness of a fundamental perspective [2][13] - Five strategies were developed: TOP20, long-short, marginal improvement, low position recovery, and low position + marginal improvement, with the latter two strategies particularly excelling in performance [2][3] - The model's effectiveness has been verified through backtesting since 2017, showcasing good fund utilization efficiency and stability in portfolio construction [3][18] Methodology Enhancements - The report details the breakdown of 31 first-level industries into 124 second-level sub-industries, allowing for a more granular approach to indicator selection and economic index construction [14][32] - A relational layer was introduced to capture inter-industry correlations, enhancing the model's ability to reflect market dynamics and improve predictive power [14][15] - The model incorporates dynamic adjustments to valuation logic, aligning better with real-time market conditions and improving its responsiveness to economic changes [15][20] Performance Analysis - The "low position + marginal improvement" strategy has consistently outperformed, with notable monthly performance in various sectors such as automotive and electronics in early 2025 [19][22] - Despite some limitations in predicting short-term policy changes, the model remains valuable for horizontal comparisons of industry fundamentals and constructing valuation expectations based on historical data [20][22] - The report suggests combining model signals with macroeconomic insights for a more comprehensive investment strategy, enhancing overall portfolio efficiency and risk management [20][22]