Workflow
港股优选20组合
icon
Search documents
金融工程定期:港股量化:南下资金累计流入达2024年91%,7月增配成长
KAIYUAN SECURITIES· 2025-07-02 12:15
- Model Name: Hong Kong Stock Selection 20 Portfolio; Model Construction Idea: The model selects the top 20 stocks with the highest scores based on four types of factors (technical, capital, fundamental, and analyst expectations) from the Hong Kong Stock Connect sample stocks and constructs an equal-weighted portfolio at the end of each month; Model Construction Process: The model uses the Hong Kong Composite Index (HKD) (930930.CSI) as the benchmark. The specific construction process involves selecting the top 20 stocks with the highest scores based on the four types of factors and constructing an equal-weighted portfolio at the end of each month[3][37][39]; Model Evaluation: The model has shown excellent performance in the Hong Kong Stock Connect sample stocks[37] - Factor Name: Technical Factors; Factor Construction Idea: Technical factors are constructed based on the price and volume data of stocks; Factor Construction Process: The specific construction process involves calculating various technical indicators such as moving averages, relative strength index (RSI), and moving average convergence divergence (MACD) to evaluate the stock's price trend and momentum[37][38]; Factor Evaluation: Technical factors have shown good performance in the Hong Kong Stock Connect sample stocks[37] - Factor Name: Capital Factors; Factor Construction Idea: Capital factors are constructed based on the capital flow data of stocks; Factor Construction Process: The specific construction process involves calculating the net capital inflow and outflow of stocks to evaluate the stock's capital flow trend[37][38]; Factor Evaluation: Capital factors have shown good performance in the Hong Kong Stock Connect sample stocks[37] - Factor Name: Fundamental Factors; Factor Construction Idea: Fundamental factors are constructed based on the financial data of stocks; Factor Construction Process: The specific construction process involves calculating various financial ratios such as price-to-earnings (P/E) ratio, return on equity (ROE), and total market value to evaluate the stock's financial health and valuation[37][38]; Factor Evaluation: Fundamental factors have shown good performance in the Hong Kong Stock Connect sample stocks[37] - Factor Name: Analyst Expectation Factors; Factor Construction Idea: Analyst expectation factors are constructed based on the analyst ratings and target prices of stocks; Factor Construction Process: The specific construction process involves collecting and analyzing the analyst ratings and target prices of stocks to evaluate the stock's future performance expectations[37][38]; Factor Evaluation: Analyst expectation factors have shown good performance in the Hong Kong Stock Connect sample stocks[37] Model Backtest Results - Hong Kong Stock Selection 20 Portfolio, Annualized Excess Return: 13.3%, Excess Return Volatility Ratio: 1.0[4][40][41] Factor Backtest Results - Southbound Capital, Average Return: 25.7%[32][36] - Foreign Capital, Average Return: 13.5%[32][36] - Domestic Capital, Average Return: 14.0%[32][36] - Hong Kong Capital and Others, Average Return: 10.2%[32][36]
金融工程定期:港股量化:5月南下资金净流入有所放缓,6月增配价值
KAIYUAN SECURITIES· 2025-06-04 06:13
Quantitative Models and Construction Methods 1. Model Name: Hong Kong Stock Multi-Factor Model - **Model Construction Idea**: The model integrates four categories of factors: technical, capital flow, fundamental, and analyst expectations, to evaluate Hong Kong Stock Connect constituent stocks[38][39] - **Model Construction Process**: 1. Select Hong Kong Stock Connect constituent stocks as the sample universe 2. Construct four categories of factors: - **Technical factors**: Indicators derived from price and volume data - **Capital flow factors**: Metrics based on fund flow data, such as net inflow - **Fundamental factors**: Metrics like valuation ratios (e.g., PE, PB) and profitability indicators (e.g., ROE) - **Analyst expectation factors**: Metrics based on analyst ratings and earnings forecasts 3. Combine these factors into a composite score for each stock 4. Rank stocks based on their composite scores and select the top 20 stocks to form the portfolio[38][39][40] - **Model Evaluation**: The model demonstrates strong performance in historical backtesting, with significant excess returns over the benchmark[38][40] 2. Model Name: Hong Kong Stock Preferred 20 Portfolio - **Model Construction Idea**: This portfolio is constructed monthly by selecting the top 20 stocks with the highest composite scores from the multi-factor model, using equal weighting[40] - **Model Construction Process**: 1. At the end of each month, rank stocks based on their composite scores from the multi-factor model 2. Select the top 20 stocks 3. Allocate equal weights to each stock in the portfolio 4. Use the Hong Kong Composite Index (HKD, 930930.CSI) as the benchmark for performance comparison[40] - **Model Evaluation**: The portfolio has shown robust performance over the long term, with a high excess annualized return and a stable risk-return profile[40][44] --- Model Backtesting Results 1. Hong Kong Stock Multi-Factor Model - **Excess Annualized Return**: 13.3% (2015.1–2025.5)[40][44] - **Excess Annualized Volatility**: 13.4%[44] - **Excess Sharpe Ratio**: 1.0[44] - **Maximum Drawdown**: 18.2%[44] 2. Hong Kong Stock Preferred 20 Portfolio - **May 2025 Monthly Return**: 2.44%[40] - **May 2025 Excess Return**: -2.36% (Benchmark return: 4.80%)[40] - **Excess Annualized Return**: 13.3% (2015.1–2025.5)[40][44] - **Excess Annualized Volatility**: 13.4%[44] - **Excess Sharpe Ratio**: 1.0[44] - **Maximum Drawdown**: 18.2%[44] --- Quantitative Factors and Construction Methods 1. Factor Name: Technical Factors - **Factor Construction Idea**: Derived from price and volume data to capture market trends and momentum[38][39] - **Factor Construction Process**: 1. Calculate indicators such as moving averages, RSI, and MACD 2. Normalize and rank the indicators across the stock universe 3. Combine the normalized scores into a composite technical factor score[38][39] 2. Factor Name: Capital Flow Factors - **Factor Construction Idea**: Based on fund flow data to identify stocks with strong capital inflows[38][39] - **Factor Construction Process**: 1. Measure net fund inflows for each stock 2. Normalize and rank the net inflow data 3. Combine the normalized scores into a composite capital flow factor score[38][39] 3. Factor Name: Fundamental Factors - **Factor Construction Idea**: Focused on valuation and profitability metrics to identify undervalued stocks with strong fundamentals[38][39] - **Factor Construction Process**: 1. Calculate valuation ratios (e.g., PE, PB) and profitability indicators (e.g., ROE) 2. Normalize and rank these metrics across the stock universe 3. Combine the normalized scores into a composite fundamental factor score[38][39] 4. Factor Name: Analyst Expectation Factors - **Factor Construction Idea**: Based on analyst ratings and earnings forecasts to capture market sentiment and expectations[38][39] - **Factor Construction Process**: 1. Collect analyst ratings and earnings forecast data 2. Normalize and rank the data 3. Combine the normalized scores into a composite analyst expectation factor score[38][39] --- Factor Backtesting Results 1. Technical Factors - **Performance**: Demonstrated strong predictive power in identifying stocks with upward momentum[38][39] 2. Capital Flow Factors - **Performance**: Effective in capturing stocks with significant fund inflows, indicating strong market interest[38][39] 3. Fundamental Factors - **Performance**: Successfully identified undervalued stocks with robust financial performance[38][39] 4. Analyst Expectation Factors - **Performance**: Provided valuable insights into market sentiment and future earnings potential[38][39]