国君晨报0207|机械、交运、有色、金融工程

Quantitative Models and Factor Construction Quantitative Models and Construction Process 1. Model Name: A-share Style Factor System Model Construction Idea: The model is built to analyze and predict style factor performance in the A-share market, covering eight major categories and 20 style factors. It is also used for risk prediction and portfolio attribution. [19] Model Construction Process: - The model incorporates eight major categories of style factors, including value, volatility, liquidity, and size. - A covariance matrix is constructed using a multi-factor model, which decomposes the stock covariance matrix into a combination of factor covariance and idiosyncratic risk matrices. - The model is updated periodically to reflect the latest market data, such as the factor covariance matrix updated on 2025/01/24. [20] Model Evaluation: The model provides a comprehensive framework for risk prediction and portfolio attribution, enabling investors to analyze factor contributions to returns and risks. [19][20] 2. Model Name: Monthly Scoring Quantitative Model for Size Rotation Model Construction Idea: This model evaluates the relative performance of large-cap and small-cap styles on a monthly basis, combining quantitative scoring with subjective analysis to form monthly style rotation views. [19] Model Construction Process: - Monthly scoring is conducted based on quantitative factors, including historical performance and market conditions. - The model incorporates monthly seasonality effects and subjective analysis to finalize the style rotation view. - For February 2025, the model indicates a preference for small-cap style, leading to an overweight allocation recommendation for small-cap stocks. [20] Model Evaluation: The model effectively integrates quantitative and qualitative insights, providing actionable style rotation strategies. [20] Quantitative Factors and Construction Process 1. Factor Name: Volatility Factor Factor Construction Idea: Measures the sensitivity of stock returns to market volatility, capturing the risk-return tradeoff. [19] Factor Construction Process: - Historical stock price data is used to calculate the standard deviation of returns over a specific period. - The factor is normalized across the universe to ensure comparability. Factor Evaluation: The volatility factor has shown positive returns in 2024, indicating its effectiveness in capturing market risk dynamics. [19] 2. Factor Name: Value Factor Factor Construction Idea: Identifies undervalued stocks based on fundamental metrics such as price-to-earnings (P/E) and price-to-book (P/B) ratios. [19] Factor Construction Process: - Fundamental data is collected for all stocks in the universe. - Stocks are ranked based on their P/E and P/B ratios, with lower values indicating higher attractiveness. - The factor is standardized to ensure consistency across the universe. Factor Evaluation: The value factor has delivered strong positive returns both in the week before the Chinese New Year and throughout 2024, demonstrating its robustness. [19][21] 3. Factor Name: Liquidity Factor Factor Construction Idea: Captures the ease of trading a stock, with higher liquidity stocks expected to have lower transaction costs. [19] Factor Construction Process: - Average daily trading volume and turnover ratios are calculated for each stock. - Stocks are ranked based on their liquidity metrics, with higher values indicating better liquidity. - The factor is normalized for cross-sectional analysis. Factor Evaluation: The liquidity factor has shown negative returns in 2024, suggesting a preference for less liquid stocks during this period. [19] Backtesting Results for Models 1. A-share Style Factor System: - Fund Heavyweight Index: - Weekly excess return contribution: -0.07% (momentum factor -0.09%, growth factor 0.08%) - Weekly excess risk contribution: 68.61% (momentum factor 27.32%, size factor 22.53%) - 2024 excess return contribution: 0.72% (size factor 1.94%, volatility factor 1.70%) - 2024 excess risk contribution: 85.78% (size factor 52.04%, volatility factor 12.14%) [20] - China Dividend Index: - Weekly excess return contribution: 0.32% (value factor 1.01%, volatility factor -0.64%) - Weekly excess risk contribution: 90.99% (volatility factor 72.07%, value factor 8.01%) - 2024 excess return contribution: 3.44% (value factor 8.10%, volatility factor -8.22%) - 2024 excess risk contribution: 87.49% (volatility factor 65.21%, value factor 8.42%) [21] - Micro-Cap Index: - Weekly excess return contribution: -0.49% (value factor -0.47%, momentum factor -0.37%) - Weekly excess risk contribution: 98.03% (size factor 79.20%, volatility factor 11.20%) - 2024 excess return contribution: -0.66% (size factor -10.19%, volatility factor 9.87%) - 2024 excess risk contribution: 98.79% (size factor 89.31%, volatility factor 5.67%) [21] Backtesting Results for Factors 1. Volatility Factor: - Weekly excess return: -0.64% (China Dividend Index) - 2024 excess return: 1.70% (Fund Heavyweight Index) [20][21] 2. Value Factor: - Weekly excess return: 1.01% (China Dividend Index) - 2024 excess return: 8.10% (China Dividend Index) [20][21] 3. Liquidity Factor: - Negative returns observed in 2024, indicating underperformance. [19]