量化组合跟踪周报：动量因子占上风，公募调研选股组合表现佳-20250915

Quantitative Models and Construction Methods 1. Model Name: PB-ROE-50 Combination - Model Construction Idea: This model focuses on selecting stocks with low Price-to-Book (PB) ratios and high Return on Equity (ROE) to construct a portfolio that aims to achieve excess returns[24] - Model Construction Process: The portfolio is constructed by screening stocks based on their PB and ROE metrics. Stocks with the lowest PB ratios and highest ROE values are selected to form the top 50 stocks in the portfolio. The portfolio is rebalanced periodically to maintain the selection criteria[24] - Model Evaluation: The model demonstrates significant excess returns in the all-market stock pool, though it underperforms in specific indices like the CSI 500 and CSI 800[24][25] 2. Model Name: Public and Private Institutional Research Combination - Model Construction Idea: This model leverages the stock selection strategies of public and private institutional research to identify stocks with potential for excess returns[27] - Model Construction Process: The portfolio is constructed by tracking the stocks that public and private institutions have recently researched. Stocks with higher research frequency or positive sentiment are included in the portfolio. The portfolio is rebalanced periodically to reflect updated research data[27] - Model Evaluation: The public institutional research strategy shows significant excess returns compared to the CSI 800 index, while the private institutional research strategy also achieves positive but smaller excess returns[27][28] 3. Model Name: Block Trade Combination - Model Construction Idea: This model identifies stocks with high block trade activity and low volatility, as these characteristics are associated with better subsequent performance[31] - Model Construction Process: The portfolio is constructed based on two key metrics: "block trade transaction amount ratio" and "6-day transaction amount volatility." Stocks with higher transaction ratios and lower volatility are selected. The portfolio is rebalanced monthly[31] - Model Evaluation: The model experienced a drawdown in the past week, with negative excess returns relative to the CSI All Share Index[31][32] 4. Model Name: Directed Issuance Combination - Model Construction Idea: This model focuses on stocks involved in directed issuance events, which are analyzed for their potential investment value based on event-driven factors[37] - Model Construction Process: The portfolio is constructed by identifying stocks with directed issuance announcements. Factors such as market capitalization, rebalancing cycles, and position control are considered. The portfolio is rebalanced periodically to reflect new issuance events[37] - Model Evaluation: The model experienced a drawdown in the past week, with negative excess returns relative to the CSI All Share Index[37][38] --- Model Backtesting Results 1. PB-ROE-50 Combination - Weekly Excess Return: All-market stock pool: +0.79%; CSI 500: -0.57%; CSI 800: -0.02%[24][25] - Year-to-Date Excess Return: All-market stock pool: +22.30%; CSI 500: +3.00%; CSI 800: +16.16%[25] - Weekly Absolute Return: All-market stock pool: +2.87%; CSI 500: +2.79%; CSI 800: +1.89%[25] - Year-to-Date Absolute Return: All-market stock pool: +48.27%; CSI 500: +28.59%; CSI 800: +36.42%[25] 2. Public and Private Institutional Research Combination - Weekly Excess Return: Public research: +3.82%; Private research: +0.51%[27][28] - Year-to-Date Excess Return: Public research: +8.10%; Private research: +12.02%[28] - Weekly Absolute Return: Public research: +5.81%; Private research: +2.44%[28] - Year-to-Date Absolute Return: Public research: +26.96%; Private research: +31.56%[28] 3. Block Trade Combination - Weekly Excess Return: -1.77%[31][32] - Year-to-Date Excess Return: +0.26%[32] - Weekly Absolute Return: Not explicitly stated - Year-to-Date Absolute Return: +62.65%[32] 4. Directed Issuance Combination - Weekly Excess Return: -1.71%[37][38] - Year-to-Date Excess Return: -0.77%[38] - Weekly Absolute Return: Not explicitly stated - Year-to-Date Absolute Return: +20.29%[38] --- Quantitative Factors and Construction Methods 1. Factor Name: Beta Factor - Factor Construction Idea: Measures the sensitivity of a stock's returns to market movements, capturing systematic risk[20] - Factor Construction Process: Beta is calculated using regression analysis of a stock's returns against the market index over a specified period[20] - Factor Evaluation: Demonstrated significant positive returns in the past week, indicating a preference for high-beta stocks[20] 2. Factor Name: Momentum Factor - Factor Construction Idea: Captures the tendency of stocks with strong past performance to continue performing well in the short term[20] - Factor Construction Process: Momentum is calculated based on the cumulative returns of a stock over a specific lookback period, such as 1 month or 5 days[20][22] - Factor Evaluation: Significant positive returns were observed, with notable momentum effects in sectors like media, real estate, and agriculture[20][22] 3. Factor Name: Scale Factor - Factor Construction Idea: Reflects the size effect, where larger-cap stocks tend to outperform smaller-cap stocks in certain market conditions[20] - Factor Construction Process: Scale is measured using market capitalization, with adjustments for sector and industry effects[20] - Factor Evaluation: Demonstrated positive returns, indicating a preference for large-cap stocks in the past week[20] --- Factor Backtesting Results 1. Beta Factor - Weekly Return: +0.70%[20] 2. Momentum Factor - Weekly Return: +0.46%[20] 3. Scale Factor - Weekly Return: +0.16%[20]