Workflow
小盘大盘风格轮动模型
icon
Search documents
A股趋势与风格定量观察20260329:关注交易逻辑由“胀”转“滞”
CMS· 2026-03-29 08:40
Quantitative Models and Construction Methods 1. Timing Model - **Model Name**: Short-term Timing Model - **Model Construction Idea**: The model aims to provide signals for short-term market timing based on various indicators such as macro fundamentals, valuation, sentiment, and liquidity[17][18][19]. - **Model Construction Process**: - **Macro Fundamentals**: Indicators include the latest manufacturing PMI (49.00), long-term loan pulse growth rate (91.53% percentile), and M1 growth rate (91.53% percentile)[17]. - **Valuation**: Indicators include the overall PE median (44.99, 91.73% percentile) and PB median (2.98, 87.59% percentile)[18]. - **Sentiment**: Indicators include overall Beta dispersion (-3.66%, 16.95% percentile), volume sentiment score (-0.42, 25.72% percentile), and volatility (25.09%, 90.74% percentile)[18]. - **Liquidity**: Indicators include money market rate (0.00, 33.90% percentile), exchange rate expectation (-1.36%, 8.47% percentile), and average new financing amount (-25.40 billion, 11.17% percentile)[19]. - **Model Evaluation**: The model provides a comprehensive view of market conditions, balancing multiple factors to generate a neutral signal for the current week[17][18][19]. 2. Growth-Value Rotation Model - **Model Name**: Growth-Value Rotation Model - **Model Construction Idea**: The model aims to balance growth and value stocks based on dynamic macro signals, valuation regression signals, short-term momentum signals, style breadth signals, and style crowding signals[27]. - **Model Construction Process**: - **Dynamic Macro Signals**: Currently at 0%[27]. - **Valuation Regression Signals**: Currently at 100%[27]. - **Short-term Momentum Signals**: Currently at 100%[27]. - **Style Breadth Signals**: Currently at 0%[27]. - **Style Crowding Signals**: Currently at 0%[27]. - **Model Evaluation**: The model suggests a balanced allocation between growth and value stocks due to the frequent switching of signals in a volatile market[27]. 3. Small-Cap vs. Large-Cap Rotation Model - **Model Name**: Small-Cap vs. Large-Cap Rotation Model - **Model Construction Idea**: The model aims to rotate between small-cap and large-cap stocks based on 11 effective rotation indicators[30]. - **Model Construction Process**: - **Indicators**: Include A-share Dragon Tiger List buying intensity, R007, financing buy balance change, theme investment trading sentiment, grade spread, option volatility risk premium, beta dispersion, PB differentiation, block trading discount rate, MACD (10,20,10) for CSI 1000, and trading volume for CSI 1000[31]. - **Current Signals**: All indicators currently suggest a preference for large-cap stocks[31]. - **Model Evaluation**: The model has consistently generated positive excess returns since 2014, with a current signal favoring large-cap stocks[30][31]. Model Backtesting Results 1. Short-term Timing Model - **Annualized Return**: 15.84%[19] - **Benchmark Annualized Return**: 4.78%[19] - **Annualized Excess Return**: 11.06%[19] - **Maximum Drawdown**: 15.05%[19] - **Sharpe Ratio**: 0.9295[23] - **Monthly Win Rate**: 65.63%[23] - **Quarterly Win Rate**: 59.26%[23] - **Annual Win Rate**: 73.33%[23] 2. Growth-Value Rotation Model - **Annualized Return**: 14.28%[27] - **Benchmark Annualized Return**: 6.38%[27] - **Annualized Excess Return**: 7.89%[27] - **Maximum Drawdown**: 40.08%[27] - **Sharpe Ratio**: 0.64[27] - **Monthly Excess Win Rate**: 66.32%[27] - **Annualized Tracking Error**: 5.88%[27] - **Annualized Information Ratio**: 1.34[27] 3. Small-Cap vs. Large-Cap Rotation Model - **Annualized Return**: 19.49%[31] - **Annualized Excess Return**: 12.43%[31] - **Maximum Drawdown**: 40.70%[31] - **Average Turnover Interval**: 19 trading days[31] - **Win Rate (by trade)**: 50.00%[31] Quantitative Factors and Construction Methods 1. Timing Factors - **Factor Name**: Manufacturing PMI, Long-term Loan Pulse Growth Rate, M1 Growth Rate, PE Median, PB Median, Beta Dispersion, Volume Sentiment Score, Volatility, Money Market Rate, Exchange Rate Expectation, Average New Financing Amount[17][18][19]. - **Factor Construction Idea**: These factors are used to assess macro fundamentals, valuation, sentiment, and liquidity to generate timing signals[17][18][19]. - **Factor Construction Process**: - **Manufacturing PMI**: 49.00[17] - **Long-term Loan Pulse Growth Rate**: 91.53% percentile[17] - **M1 Growth Rate**: 91.53% percentile[17] - **PE Median**: 44.99, 91.73% percentile[18] - **PB Median**: 2.98, 87.59% percentile[18] - **Beta Dispersion**: -3.66%, 16.95% percentile[18] - **Volume Sentiment Score**: -0.42, 25.72% percentile[18] - **Volatility**: 25.09%, 90.74% percentile[18] - **Money Market Rate**: 0.00, 33.90% percentile[19] - **Exchange Rate Expectation**: -1.36%, 8.47% percentile[19] - **Average New Financing Amount**: -25.40 billion, 11.17% percentile[19] - **Factor Evaluation**: These factors provide a comprehensive view of market conditions, balancing multiple aspects to generate a neutral signal for the current week[17][18][19]. 2. Growth-Value Rotation Factors - **Factor Name**: Dynamic Macro Signals, Valuation Regression Signals, Short-term Momentum Signals, Style Breadth Signals, Style Crowding Signals[27]. - **Factor Construction Idea**: These factors are used to balance growth and value stocks based on various market signals[27]. - **Factor Construction Process**: - **Dynamic Macro Signals**: 0%[27] - **Valuation Regression Signals**: 100%[27] - **Short-term Momentum Signals**: 100%[27] - **Style Breadth Signals**: 0%[27] - **Style Crowding Signals**: 0%[27] - **Factor Evaluation**: The factors suggest a balanced allocation between growth and value stocks due to the frequent switching of signals in a volatile market[27]. 3. Small-Cap vs. Large-Cap Rotation Factors - **Factor Name**: A-share Dragon Tiger List Buying Intensity, R007, Financing Buy Balance Change, Theme Investment Trading Sentiment, Grade Spread, Option Volatility Risk Premium, Beta Dispersion, PB Differentiation, Block Trading Discount Rate, MACD (10,20,10) for CSI 1000, Trading Volume for CSI 1000[31]. - **Factor Construction Idea**: These factors are used to rotate between small-cap and large-cap stocks based on various market signals[31]. - **Factor Construction Process**: - **A-share Dragon Tiger List Buying Intensity**: 100%[31] - **R007**: 100%[31] - **Financing Buy Balance Change**: 0%[31] - **Theme Investment Trading Sentiment**: 0%[31] - **Grade Spread**: 0%[31] - **Option Volatility Risk Premium**: 100%[31] - **Beta Dispersion**: 100%[31] - **PB Differentiation**: 0%[31] - **Block Trading Discount Rate**: 0%[31] - **MACD (10,20,10) for CSI 1000**: 0%[31] - **Trading Volume for CSI 1000**: 0%[31] - **Factor Evaluation**: The factors currently suggest a preference for large-cap stocks, with consistent positive excess returns since 2014[30][31].