KAIYUAN SECURITIES·2025-09-01 06:16Quantitative Models and Construction Methods 1. Model Name: "All Stocks Portfolio" - Model Construction Idea: This model aggregates all broker-recommended stocks ("golden stocks") and evaluates their performance as a portfolio[18][21] - Model Construction Process: 1. Collect all broker-recommended stocks for the month 2. Weight the stocks within the portfolio based on the number of recommendations by brokers 3. Exclude non-A-share stocks and Hong Kong-listed stocks to focus solely on A-shares[18] - Model Evaluation: The model demonstrates strong performance, significantly outperforming benchmark indices such as CSI 300 and CSI 500[18][21] 2. Model Name: "Newly Added Stocks Portfolio" - Model Construction Idea: Focuses on stocks newly added to the broker-recommended list, as they tend to exhibit better performance compared to repeated recommendations[18][23] - Model Construction Process: 1. Identify stocks newly added to the broker-recommended list for the month 2. Construct a portfolio weighted by the number of broker recommendations 3. Exclude non-A-share stocks and Hong Kong-listed stocks[18] - Model Evaluation: Newly added stocks outperform repeated recommendations, showcasing their superior return potential[18][23] 3. Model Name: "Repeated Stocks Portfolio" - Model Construction Idea: Focuses on stocks that have been repeatedly recommended by brokers across multiple months[18] - Model Construction Process: 1. Identify stocks that were recommended in the previous month and continue to be recommended in the current month 2. Construct a portfolio weighted by the number of broker recommendations 3. Exclude non-A-share stocks and Hong Kong-listed stocks[18] - Model Evaluation: While the performance is positive, it is generally weaker compared to newly added stocks[18] 4. Model Name: "Optimized Golden Stocks Portfolio" - Model Construction Idea: Selects the top 30 newly added stocks with the highest earnings surprise factor (SUE factor) to construct an optimized portfolio[23] - Model Construction Process: 1. Filter newly added stocks based on their earnings surprise factor (SUE factor) 2. Select the top 30 stocks with the highest SUE factor 3. Weight the portfolio based on the number of broker recommendations[23] - Model Evaluation: This optimized portfolio demonstrates superior performance compared to the "All Stocks Portfolio" and benchmark indices[23][25] --- Model Backtesting Results 1. All Stocks Portfolio - August Return: 13.6%[21] - 2025 YTD Return: 33.5%[21] - Annualized Return: 13.7%[21] - Annualized Volatility: 23.6%[21] - Sharpe Ratio: 0.58[21] - Maximum Drawdown: 42.6%[21] 2. Newly Added Stocks Portfolio - August Return: 11.8%[21] - 2025 YTD Return: 37.9%[21] - Annualized Return: 16.5%[21] - Annualized Volatility: 24.3%[21] - Sharpe Ratio: 0.68[21] - Maximum Drawdown: 38.5%[21] 3. Repeated Stocks Portfolio - August Return: 15.6%[21] - 2025 YTD Return: 30.2%[21] - Annualized Return: 11.3%[21] - Annualized Volatility: 23.7%[21] - Sharpe Ratio: 0.48[21] - Maximum Drawdown: 45.0%[21] 4. Optimized Golden Stocks Portfolio - August Return: 19.6%[25] - 2025 YTD Return: 37.6%[25] - Annualized Return: 22.3%[25] - Annualized Volatility: 25.5%[25] - Sharpe Ratio: 0.88[25] - Maximum Drawdown: 24.6%[25] --- Quantitative Factors and Construction Methods 1. Factor Name: Earnings Surprise Factor (SUE Factor) - Factor Construction Idea: Measures the degree to which a company's earnings exceed or fall short of market expectations, serving as a key indicator for stock selection[23] - Factor Construction Process: 1. Calculate the earnings surprise for each stock as the difference between reported earnings and consensus estimates 2. Normalize the earnings surprise to account for variations across stocks and industries 3. Rank stocks based on their normalized earnings surprise values[23] - Factor Evaluation: The SUE factor demonstrates strong predictive power, particularly in identifying high-performing newly added stocks[23] --- Factor Backtesting Results 1. SUE Factor - Performance: The SUE factor is highly effective in selecting top-performing stocks within the newly added category, contributing to the superior returns of the Optimized Golden Stocks Portfolio[23]