Quantitative Models and Factor Analysis Quantitative Models and Construction - **Model Name**: Bank Stock Momentum Strategy **Model Construction Idea**: The strategy leverages the momentum of market styles to identify outperforming sub-segments within the banking sector. It assumes that the internal style rotation of bank stocks mirrors the broader market's style rotation, albeit with a lag [28][48]. **Model Construction Process**: 1. Use market-wide style momentum to determine the dominant style (e.g., growth vs. value, large-cap vs. small-cap) based on a two-month relative momentum difference exceeding 5% [48]. 2. Map the dominant market style to corresponding banking sub-segments (e.g., large-cap value, small-cap growth) [28][29]. 3. Select stocks within the identified banking sub-segment for the next month [48]. **Model Evaluation**: The model effectively captures value-style dominance in the banking sector, particularly during market-wide value-oriented phases [48]. - **Model Name**: Fundamental-Adjusted Growth Selection **Model Construction Idea**: To mitigate the downside risk of high-PB (price-to-book) stocks, the model incorporates ROE (return on equity) adjustments to prioritize high-quality fundamentals when selecting growth-oriented bank stocks [58]. **Model Construction Process**: 1. Combine PB and ROE_TTM (trailing twelve months) values into an equal-weighted scoring system [58]. 2. When the signal indicates a growth-oriented market, select stocks with high PB and strong fundamentals [58]. **Model Evaluation**: This adjustment improves annualized returns by 1.82% compared to using PB alone, demonstrating its effectiveness in balancing growth potential and risk [58]. Model Backtesting Results - **Bank Stock Momentum Strategy**: - Annualized Return: 14.45% - Annualized Volatility: 22.62% - Sharpe Ratio: 0.64 - Maximum Drawdown: 33.08% - Excess Annualized Return (vs. CSI Bank Index): 5.65% - Information Ratio (IR): 0.57 [48][49][54] - **Fundamental-Adjusted Growth Selection**: - Annualized Return: 16.27% - Excess Annualized Return (vs. CSI All Share Index): 13.20% - Excess Annualized Return (vs. CSI Bank Index): 7.33% [58] --- Quantitative Factors and Construction - **Factor Name**: Price-to-Book Ratio (PB) **Factor Construction Idea**: Low PB stocks are expected to outperform due to their value-oriented characteristics [32]. **Factor Construction Process**: 1. Divide bank stocks into quintiles based on PB values. 2. Track the performance of the lowest PB quintile over time [32]. **Factor Evaluation**: The low PB group consistently outperformed the CSI All Share Index, particularly during value-dominant market phases [32]. - **Factor Name**: Dividend Yield **Factor Construction Idea**: High dividend yield stocks are expected to provide stable returns, especially in defensive market conditions [32]. **Factor Construction Process**: 1. Divide bank stocks into quintiles based on dividend yield. 2. Track the performance of the highest dividend yield quintile over time [32]. **Factor Evaluation**: While generally effective, the factor underperformed low PB stocks in certain years, indicating weaker pricing power [32]. - **Factor Name**: Market Capitalization **Factor Construction Idea**: Large-cap stocks are expected to outperform during periods of market uncertainty due to their defensive nature [39]. **Factor Construction Process**: 1. Divide bank stocks into quintiles based on market capitalization. 2. Track the performance of the largest market-cap quintile over time [39]. **Factor Evaluation**: Large-cap stocks demonstrated strong performance, particularly after 2023, with a clear upward trend [39]. - **Factor Name**: Return on Equity (ROE) **Factor Construction Idea**: High ROE stocks are expected to outperform due to their superior profitability [39]. **Factor Construction Process**: 1. Divide bank stocks into quintiles based on ROE. 2. Track the performance of the highest ROE quintile over time [39]. **Factor Evaluation**: High ROE stocks outperformed in most years but showed limitations in certain market environments, such as 2014 [39]. Factor Backtesting Results - **PB Factor**: - Outperformed the CSI All Share Index across the entire observation period [32]. - **Dividend Yield Factor**: - Generally outperformed the benchmark but lagged behind the PB factor in specific years [32]. - **Market Capitalization Factor**: - Demonstrated a strong upward trend post-2023, consistently outperforming the benchmark [39]. - **ROE Factor**: - Achieved superior performance in years like 2012, 2017, and 2019, but underperformed in 2014 [39]. --- Composite Factor Analysis - **Composite Model**: Four-Quadrant Style Model **Construction Idea**: Combine market capitalization and PB factors to create four distinct style groups: large-cap value, large-cap growth, small-cap value, and small-cap growth [43]. **Construction Process**: 1. Divide stocks into large-cap and small-cap groups based on the top 20% and bottom 80% of market capitalization. 2. Further divide each group into high-PB and low-PB subgroups. 3. Analyze the performance of the resulting four style groups [43]. **Evaluation**: The model revealed clear style rotations within the banking sector, with small-cap value dominating in 2014-2015 and large-cap value leading post-2022 [43]. Composite Model Backtesting Results - **Four-Quadrant Style Model**: - Small-Cap Value: Outperformed during 2014-2015 - Large-Cap Value: Consistently outperformed post-2022 [43]