1. Report Industry Investment Rating - Not provided in the report 2. Core Viewpoints of the Report - Direct factorization may be the optimal way to use the industry rotation model. The industry rotation model can complement existing price - volume factors and show good offensive ability in upward markets, which is an important idea to solve the excess drawdown of CSI 500 index enhancement. The industry momentum is the key incremental information of the industry rotation model and the basis for factorizing the industry model [4][62]. - The industry momentum factor has a low cross - sectional correlation with traditional factors, providing an additional source of excess returns. In the perspective of industry momentum domain, factors perform differently in high - momentum and medium - low - momentum groups, which provides a basis for factorizing the industry momentum [4]. - The performance of direct factorization in index enhancement is better than that of the model only controlling industry deviation. However, the discontinuity of factor values needs to be addressed in the future [4]. 3. Summary According to the Table of Contents 1. Re - discussing the Factorization of the Industry Rotation Model - **1.1 Disassembling the Incremental Information of the Industry Rotation Model in Stock Selection** - The traditional industry rotation model has overlapping information such as financial reports and fund flows with stock - selection factors. The technical aspects based on industry momentum and crowding are the incremental information different from stock - selection. The special treatment of industry momentum in the Shenwan Hongyuan industry rotation model is an important reason for its more stable excess returns [7][10]. - **1.2 Performance of the Industry Momentum Factor** - When only the industry momentum is factorized, its stability decreases compared to the overall industry rotation model, but it still maintains good monotonicity and offensive characteristics. The monthly IC is 3.03% and the ICIR is 0.22. The industry momentum factor has a low cross - sectional correlation with other factors, providing an additional source of excess returns and showing complementarity with price - volume factors [16][21]. 2. Factor Performance from the Perspective of Industry Momentum Domain - **2.1 Low - Volatility Factor under Industry Momentum Domain: Obvious Differences in Recent Years** - The low - volatility factor shows good monotonicity in all intervals, high - momentum group, and medium - low - momentum group. The low - volatility factor in the high - momentum group has better returns in recent years, but this characteristic is unstable [30][31]. - **2.2 Reversal Factor under Industry Momentum Domain: Obvious Differences** - The reversal factor shows good monotonicity in the full - interval and high - momentum group, but poor monotonicity in the medium - low - momentum group. The reversal factor in the high - momentum group performs better, while it hardly contributes any returns in the medium - low - momentum group [36][39]. - **2.3 Momentum Factor under Industry Momentum Domain: Obvious Differences** - The momentum factor has stable five - group monotonicity in the full - interval and high - momentum group, but poor monotonicity in the medium - low - momentum group. The momentum factor in the high - momentum group has stronger long - term performance, while the medium - low - momentum group shows continuous drawdown in recent years [40][41]. - **2.4 Low - Liquidity Factor under Industry Momentum Domain: Little Differentiation** - The low - liquidity factor has poor monotonicity in all three intervals. Its performance has little relation with the industry momentum domain [45][47]. 3. Performance of the Industry Momentum Factor in the Portfolio - **3.1 Performance Comparison of the CSI 500 Index Enhancement Simulation Portfolio** - Whether it is a five - factor or six - factor combination, the performance of direct factorization is better than that of using the industry model only to control industry deviation [53]. - **3.2 Comparison after Changing Constraint Conditions** - After adjusting the constraint conditions, the performance of the six - factor combination is still better than that of the five - factor combination with industry model control deviation. Factorizing the industry rotation model is an effective way to use it, but the problem of discontinuous factor values needs to be solved [57][59]. - **3.3 Factorization vs. Industry Deviation Control** - When the industry momentum factor participates in stock selection, it is easier to form resonance with effective factors in the domain, increasing the exposure of the selected stocks to effective factors. Factorization can better play the advantage of resonance [60][61]. 4. Summary - The industry rotation model can complement existing price - volume factors and show good offensive ability in upward markets. Direct factorization may be the optimal way to use the industry rotation model. The industry momentum is the key incremental information of the industry rotation model, and the differences in factor performance in different industry momentum groups provide a basis for factorization. The performance of direct factorization is better than that of only controlling industry deviation, but the discontinuity of factor values needs to be addressed [62][63].

2026 年 02 月 27 日 行业动量分域视角：再论行业轮动 模型的因子化 ——申万金工因子观察第 4 期 20260226 相关研究 证券分析师 邓虎 A0230520070003 denghu@swsresearch.com 本研究报告仅通过邮件提供给 博时基金 博时基金管理有限公司(researchreport@bosera.com) 使用。1 权 益 量 化 研 究 证 券 研 究 报 告 请务必仔细阅读正文之后的各项信息披露与声明 股 票 基 金 - ⚫ 直接因子化可能是使用行业轮动模型的较优方式。行业轮动模型因偏动量的逻辑可以和 现有的量价因子形成互补，也能在上行行情里体现出较好的进攻性，成为解决当前中证 500 指数增强超额回撤的重要思路之一。但行业轮动模型面临着如何与传统指数增强框 架去结合的问题，我们给出的答案是：直接因子化可能是使用行业轮动模型的较优方式。 ⚫ 行业动量是行业轮动模型能够贡献的重要增量信息，是行业模型因子化的基础。因此，我 们剥离了行业轮动模型里的信息增量，传统行业轮动模型框架里有一部分如财报、资金流 等信息与选股信息是重叠的，而基于行业动量和拥挤度的技术面属于与选股不 ...