Quantitative Models and Construction Methods 1. Model Name: ERP and DRP Standardized Equal-Weight Model for A-Share Odds - **Model Construction Idea**: The model calculates A-share odds using standardized values of ERP (Equity Risk Premium) and DRP (Default Risk Premium) with equal weighting[12] - **Model Construction Process**: - Standardized values of ERP and DRP are calculated - These values are equally weighted to derive the A-share odds - As of the end of November, the A-share odds declined to near the zero axis, indicating a neutral level[12] - **Model Evaluation**: The model reflects a neutral positioning for A-shares, with odds returning to a balanced state[12] 2. Model Name: Bond Odds Indicator - **Model Construction Idea**: The model uses the expected return difference between long-term and short-term bonds to construct a bond odds indicator[19] - **Model Construction Process**: - The expected return difference between long-term and short-term bonds is calculated - This difference is standardized to derive the bond odds indicator - Recently, the bond odds indicator has significantly rebounded but remains at a low level of -0.9 standard deviations[19] - **Model Evaluation**: The model effectively captures the rebound in bond odds, though it remains at a relatively low level[19] 3. Model Name: AIAE Indicator for US Stocks - **Model Construction Idea**: The AIAE (Asset Implied Allocation Efficiency) indicator measures the historical positioning of US stocks to assess risk and return[20] - **Model Construction Process**: - The AIAE indicator is calculated based on historical data - Currently, the AIAE indicator is at 55%, the highest point in its history, corresponding to 2.4 standard deviations above the mean - Historical analysis shows that when the AIAE indicator exceeded 50% in 2000 and 2022, the S&P 500 experienced significant corrections of 46% and 25%, respectively[20] - **Model Evaluation**: The model highlights elevated risks for US stocks, with the AIAE indicator at a historically high level[20] 4. Model Name: Federal Reserve Liquidity Index - **Model Construction Idea**: Combines quantity and price dimensions to construct a liquidity index for asset allocation[20] - **Model Construction Process**: - The index incorporates multiple factors, including net liquidity, Federal Reserve credit support, market expectations, and announcement surprises - After the October FOMC meeting, the announcement surprise signal turned negative, and net liquidity continued to decline - Other indicators showed easing signals, and the liquidity index returned to a moderately high level of 20%[20] - **Model Evaluation**: The model provides a comprehensive view of liquidity conditions, highlighting mixed signals in the current environment[20] --- Model Backtesting Results 1. ERP and DRP Standardized Equal-Weight Model for A-Share Odds - Odds: Neutral (near zero axis)[12] - Win Rate: 16%[12] 2. Bond Odds Indicator - Odds: -0.9 standard deviations (low level)[19] - Win Rate: -4% (medium level)[19] 3. AIAE Indicator for US Stocks - Odds: 2.4 standard deviations (historically high level)[20] - Win Rate: Not explicitly mentioned[20] 4. Federal Reserve Liquidity Index - Liquidity Index: 20% (moderately high level)[20] --- Quantitative Factors and Construction Methods 1. Factor Name: Small-Cap Factor - **Factor Construction Idea**: Evaluates small-cap stocks based on odds, trends, and crowding levels[21] - **Factor Construction Process**: - Odds: 0.2 standard deviations (neutral level) - Trend: 1.2 standard deviations (high level) - Crowding: -1.4 standard deviations (low level) - Comprehensive score: 4[21] - **Factor Evaluation**: The factor shows strong trends and low crowding, making it attractive for allocation[21] 2. Factor Name: Value Factor - **Factor Construction Idea**: Assesses value stocks using odds, trends, and crowding levels[23] - **Factor Construction Process**: - Odds: 0.8 standard deviations (moderately high level) - Trend: 0.1 standard deviations (neutral level) - Crowding: -1.3 standard deviations (low level) - Comprehensive score: 3[23] - **Factor Evaluation**: The factor ranks high among others, suggesting it is worth focusing on[23] 3. Factor Name: Quality Factor - **Factor Construction Idea**: Evaluates quality stocks based on odds, trends, and crowding levels[26] - **Factor Construction Process**: - Odds: 1.2 standard deviations (high level) - Trend: -0.6 standard deviations (low level) - Crowding: 0.1 standard deviations (medium level) - Comprehensive score: 0[26] - **Factor Evaluation**: The factor's weak trend reduces its allocation value, requiring confirmation of a right-side trend[26] 4. Factor Name: Growth Factor - **Factor Construction Idea**: Analyzes growth stocks using odds, trends, and crowding levels[29] - **Factor Construction Process**: - Odds: 0.1 standard deviations (neutral level) - Trend: 0.5 standard deviations (moderately high level) - Crowding: 1.5 standard deviations (high level) - Comprehensive score: -1.2[29] - **Factor Evaluation**: The factor's high crowding level indicates elevated trading risks[29] --- Factor Backtesting Results 1. Small-Cap Factor - Odds: 0.2 standard deviations[21] - Trend: 1.2 standard deviations[21] - Crowding: -1.4 standard deviations[21] - Comprehensive Score: 4[21] 2. Value Factor - Odds: 0.8 standard deviations[23] - Trend: 0.1 standard deviations[23] - Crowding: -1.3 standard deviations[23] - Comprehensive Score: 3[23] 3. Quality Factor - Odds: 1.2 standard deviations[26] - Trend: -0.6 standard deviations[26] - Crowding: 0.1 standard deviations[26] - Comprehensive Score: 0[26] 4. Growth Factor - Odds: 0.1 standard deviations[29] - Trend: 0.5 standard deviations[29] - Crowding: 1.5 standard deviations[29] - Comprehensive Score: -1.2[29]

Quantitative Models and Construction 1. Model Name: CDS Timing Strategy - **Model Construction Idea**: The model uses the 20-day difference signal of China's sovereign CDS as a timing indicator for Hong Kong stocks, leveraging the negative correlation between CDS and stock performance[12][13] - **Model Construction Process**: 1. Calculate the 20-day difference of China's sovereign CDS 2. Use the signal to time Hong Kong stock investments 3. Evaluate the annualized excess return relative to the benchmark - **Model Evaluation**: The model demonstrates a strong fit with Hong Kong stock returns and provides a reliable timing signal[12][13] 2. Model Name: Duration Timing Strategy - **Model Construction Idea**: The model estimates the expected return of government bonds for any duration and holding period using a multi-step process[17] - **Model Construction Process**: 1. Decompose government bond yields 2. Predict interest rates using modeling techniques 3. Simulate scenarios via Monte Carlo methods 4. Calculate expected returns for different durations and holding periods - **Model Evaluation**: The strategy is effective for short-term bond timing and provides actionable insights for duration allocation[17][19] 3. Model Name: Equity Index Return Prediction Model - **Model Construction Idea**: Predict the future returns of broad-based indices using a combination of macroeconomic and valuation factors[22][27] - **Model Construction Process**: 1. Use macroeconomic indicators and valuation metrics 2. Apply the model to predict returns for indices like CSI 300, CSI 500, etc. 3. Compare predicted returns to historical benchmarks - **Model Evaluation**: The model shows strong predictive power for large-cap indices like CSI 300, while small-cap indices like CSI 500 exhibit lower reliability[22][27] 4. Model Name: Industry Rotation Strategy - **Model Construction Idea**: Evaluate industries based on momentum, turnover, volatility, and beta to identify rotation opportunities[60] - **Model Construction Process**: 1. Calculate 12-month information ratios for industry momentum 2. Assess turnover, volatility, and beta for crowding metrics 3. Combine these dimensions to rank industries - **Model Evaluation**: The strategy effectively identifies high-potential industries and provides actionable rotation insights[60][63] 5. Model Name: Odds + Win Rate Strategy - **Model Construction Idea**: Combine odds and win rate metrics to allocate assets dynamically[65][70] - **Model Construction Process**: 1. Construct odds and win rate indicators for each asset 2. Combine the two metrics into a unified score 3. Allocate assets based on the combined score - **Model Evaluation**: The strategy balances risk and return effectively, achieving stable performance over time[65][70] --- Model Backtest Results 1. CDS Timing Strategy - Annualized Return: 11.8% - Annualized Volatility: 13.9% - Maximum Drawdown: 19.1% - Sharpe Ratio: 0.851[15] 2. Duration Timing Strategy - Annualized Return: 6.8% - Annualized Volatility: 2.1% - Maximum Drawdown: 2.3% - Calmar Ratio: 2.94[19] 3. Equity Index Return Prediction Model - CSI 300: Predicted Return 19.7% - CSI 500: Predicted Return -27.8%[22][26] 4. Industry Rotation Strategy - Annualized Excess Return: 12.2% (since 2011) - Tracking Error: 10.9% - Maximum Drawdown: 25.4% - IR: 1.12[61] 5. Odds + Win Rate Strategy - Annualized Return: 6.9% (since 2011) - Annualized Volatility: 2.3% - Maximum Drawdown: 2.8% - Sharpe Ratio: 3.03[72] --- Quantitative Factors and Construction 1. Factor Name: Quality Factor - **Factor Construction Idea**: Combines odds, trend, and crowding metrics to evaluate quality stocks[46] - **Factor Construction Process**: 1. Calculate odds (valuation) at 1.3 standard deviations 2. Assess trend at -0.1 standard deviations 3. Measure crowding at -1.1 standard deviations 4. Combine metrics into a composite score - **Factor Evaluation**: High composite score indicates strong potential for long-term allocation[46] 2. Factor Name: Growth Factor - **Factor Construction Idea**: Evaluates growth stocks based on trend, odds, and crowding metrics[47] - **Factor Construction Process**: 1. Calculate trend at 0.5 standard deviations 2. Assess odds at -1.1 standard deviations 3. Measure crowding at 0.2 standard deviations 4. Combine metrics into a composite score - **Factor Evaluation**: Low composite score suggests limited allocation value[47] 3. Factor Name: Dividend Factor - **Factor Construction Idea**: Focuses on dividend-paying stocks with moderate odds and low crowding[50] - **Factor Construction Process**: 1. Calculate trend at -1.7 standard deviations 2. Assess odds at -0.2 standard deviations 3. Measure crowding at -1.6 standard deviations 4. Combine metrics into a composite score - **Factor Evaluation**: Low composite score indicates limited allocation potential[50] 4. Factor Name: Small-Cap Factor - **Factor Construction Idea**: Evaluates small-cap stocks based on trend, odds, and crowding metrics[53] - **Factor Construction Process**: 1. Calculate trend at -0.06 standard deviations 2. Assess odds at -0.05 standard deviations 3. Measure crowding at 0.3 standard deviations 4. Combine metrics into a composite score - **Factor Evaluation**: High uncertainty and low composite score suggest caution[53] --- Factor Backtest Results 1. Quality Factor - Odds: 1.3 SD - Trend: -0.1 SD - Crowding: -1.1 SD - Composite Score: 3[46] 2. Growth Factor - Odds: -1.1 SD - Trend: 0.5 SD - Crowding: 0.2 SD - Composite Score: 0[47] 3. Dividend Factor - Odds: -0.2 SD - Trend: -1.7 SD - Crowding: -1.6 SD - Composite Score: 0[50] 4. Small-Cap Factor - Odds: -0.05 SD - Trend: -0.06 SD - Crowding: 0.3 SD - Composite Score: 0[53]