微盘股指数周报：本周市场持续缩量，微盘表现欠佳-20251208

Quantitative Models and Construction Methods 1. Model Name: Diffusion Index Model - Model Construction Idea: The diffusion index is used to monitor the critical points of market trend changes and predict potential reversals in the micro-cap stock index[40][41] - Model Construction Process: The diffusion index is calculated based on the relative price changes of micro-cap stock index constituents over a specific time window. The horizontal axis represents the percentage change in stock prices (e.g., from +10% to -10%), while the vertical axis represents the length of the review or forecast window (e.g., 10 to 20 trading days). For example, a horizontal value of 0.95 and a vertical value of 15 days indicates that if all constituent stocks drop by 5% after 5 days, the diffusion index value is 0.03[40][42] - Key Methods: - Left Threshold Method: Triggered a short signal on November 14, 2025, when the index closed at 0.925[43] - Right Threshold Method: Triggered a short signal on November 17, 2025, when the index closed at 0.8975[46] - Dual Moving Average Method: Triggered a short signal on November 25, 2025[48] - Model Evaluation: The diffusion index is effective in identifying critical points for market reversals and provides actionable signals for trading strategies[40][41] 2. Model Name: Small-Cap Low-Volatility 50 Strategy - Model Construction Idea: This strategy selects 50 stocks with small market capitalization and low volatility from the micro-cap stock index constituents, aiming to achieve stable returns with reduced risk[8][35] - Model Construction Process: - Stock Selection: Identify stocks with the smallest market capitalization and lowest volatility within the micro-cap stock index[35] - Rebalancing Frequency: Rebalance the portfolio bi-weekly[35] - Transaction Costs: Assume a bidirectional transaction cost of 0.3%[35] - Benchmark: Use the Wind Micro-Cap Stock Index (8841431.WI) as the benchmark[35] - Model Evaluation: The strategy demonstrates strong performance in capturing excess returns, particularly in 2025, with a significant year-to-date (YTD) return[35] --- Model Backtesting Results 1. Diffusion Index Model - Left Threshold Method: Short signal triggered at 0.925 on November 14, 2025[43] - Right Threshold Method: Short signal triggered at 0.8975 on November 17, 2025[46] - Dual Moving Average Method: Short signal triggered on November 25, 2025[48] 2. Small-Cap Low-Volatility 50 Strategy - 2024 Return: 7.07%, with an excess return of -2.93%[35] - 2025 YTD Return: 72.07%, with a weekly excess return of 0.001%[35] --- Quantitative Factors and Construction Methods 1. Factor Name: PE_TTM Inverse Factor - Factor Construction Idea: This factor uses the inverse of the price-to-earnings ratio (trailing twelve months) to identify undervalued stocks[5][33] - Factor Construction Process: - Calculate the trailing twelve months (TTM) PE ratio for each stock - Take the reciprocal of the PE ratio to construct the factor[33] - Factor Evaluation: Demonstrates strong predictive power with a weekly rank IC of 0.237, significantly above its historical average of 0.016[5][33] 2. Factor Name: PB Inverse Factor - Factor Construction Idea: This factor uses the inverse of the price-to-book ratio to identify undervalued stocks[5][33] - Factor Construction Process: - Calculate the price-to-book (PB) ratio for each stock - Take the reciprocal of the PB ratio to construct the factor[33] - Factor Evaluation: Shows strong performance with a weekly rank IC of 0.213, well above its historical average of 0.034[5][33] 3. Factor Name: Profitability Factor - Factor Construction Idea: Measures the profitability of companies to identify fundamentally strong stocks[5][33] - Factor Construction Process: - Use metrics such as return on equity (ROE) and net profit margin to construct the factor[33] - Factor Evaluation: Exhibits robust predictive power with a weekly rank IC of 0.185, outperforming its historical average of 0.022[5][33] 4. Factor Name: Dividend Yield Factor - Factor Construction Idea: Focuses on stocks with high dividend yields to capture stable income[5][33] - Factor Construction Process: - Calculate the dividend yield for each stock - Rank stocks based on their dividend yield to construct the factor[33] - Factor Evaluation: Performs well with a weekly rank IC of 0.179, exceeding its historical average of 0.022[5][33] 5. Factor Name: Single-Quarter ROE Factor - Factor Construction Idea: Evaluates the return on equity for a single quarter to identify high-performing stocks[5][33] - Factor Construction Process: - Calculate the ROE for the most recent quarter - Rank stocks based on their single-quarter ROE to construct the factor[33] - Factor Evaluation: Delivers strong results with a weekly rank IC of 0.163, above its historical average of 0.021[5][33] --- Factor Backtesting Results 1. PE_TTM Inverse Factor - Weekly Rank IC: 0.237 - Historical Average Rank IC: 0.016[5][33] 2. PB Inverse Factor - Weekly Rank IC: 0.213 - Historical Average Rank IC: 0.034[5][33] 3. Profitability Factor - Weekly Rank IC: 0.185 - Historical Average Rank IC: 0.022[5][33] 4. Dividend Yield Factor - Weekly Rank IC: 0.179 - Historical Average Rank IC: 0.022[5][33] 5. Single-Quarter ROE Factor - Weekly Rank IC: 0.163 - Historical Average Rank IC: 0.021[5][33]