Quantitative Models and Construction Methods 1. Model Name: Short-term Quantitative Timing Model - **Model Construction Idea**: The model integrates macroeconomic fundamentals, valuation, sentiment, and liquidity indicators to generate short-term market timing signals[24][25][26] - **Model Construction Process**: - **Macroeconomic Fundamentals**: - Manufacturing PMI: Current value at 49.70, 44.92% percentile over the past 5 years, indicating neutral sentiment[24] - Long-term RMB loan growth: 0.00% percentile, indicating weak credit growth and cautious signals[24] - M1 growth rate: 94.92% percentile, indicating strong growth and optimistic signals[24] - **Valuation**: - PE median: 43.18, 97.19% percentile, indicating high valuation and neutral signals[25] - PB median: 2.85, 86.77% percentile, indicating high valuation and neutral signals[25] - **Sentiment**: - Beta dispersion: -0.59%, 40.68% percentile, indicating neutral sentiment[25] - Volume sentiment score: 0.98, 99.59% percentile, indicating strong sentiment and optimism[25] - Volatility: 7.53% (annualized), 0.17% percentile, indicating optimism[25] - **Liquidity**: - Monetary rate: -0.10, 33.90% percentile, indicating relative ease and optimism[26] - Exchange rate expectations: -0.09%, 40.68% percentile, indicating neutrality[26] - 5-day average financing: 50.66 billion RMB, 95.53% percentile, indicating neutral leverage signals[26] - **Model Evaluation**: The model demonstrates strong performance with significant excess returns and reduced drawdowns compared to benchmarks[26][30] 2. Model Name: Growth-Value Style Rotation Model - **Model Construction Idea**: The model evaluates growth and value styles based on macroeconomic fundamentals, valuation, and sentiment indicators to recommend allocation[35] - **Model Construction Process**: - **Macroeconomic Fundamentals**: - Profit cycle slope: High, favoring growth[37] - Interest rate cycle: High, favoring value[37] - Credit cycle: Weak, favoring value[37] - **Valuation**: - PE difference: 19.57% percentile, favoring growth[37] - PB difference: 38.03% percentile, favoring growth[37] - **Sentiment**: - Turnover difference: 38.13% percentile, favoring value[37] - Volatility difference: 17.73% percentile, favoring balanced allocation[37] - **Model Evaluation**: The model has historically delivered significant excess returns over benchmarks, though recent performance has been mixed[36][39] 3. Model Name: Small-Cap vs. Large-Cap Style Rotation Model - **Model Construction Idea**: The model assesses small-cap and large-cap styles using macroeconomic fundamentals, valuation, and sentiment indicators to suggest balanced allocation[40] - **Model Construction Process**: - **Macroeconomic Fundamentals**: - Profit cycle slope: High, favoring small-cap[42] - Interest rate cycle: High, favoring large-cap[42] - Credit cycle: Weak, favoring large-cap[42] - **Valuation**: - PE difference: 78.86% percentile, favoring large-cap[42] - PB difference: 96.59% percentile, favoring large-cap[42] - **Sentiment**: - Turnover difference: 72.56% percentile, favoring small-cap[42] - Volatility difference: 62.60% percentile, favoring large-cap[42] - **Model Evaluation**: The model has consistently outperformed benchmarks, delivering significant excess returns over time[41][44] 4. Model Name: Four-Style Rotation Model - **Model Construction Idea**: Combines insights from growth-value and small-cap-large-cap models to allocate across four styles: small-cap growth, small-cap value, large-cap growth, and large-cap value[45] - **Model Construction Process**: - Allocation recommendation: Small-cap growth (12.5%), small-cap value (37.5%), large-cap growth (12.5%), large-cap value (37.5%)[45] - **Model Evaluation**: The model has historically generated significant excess returns, though recent performance has been slightly below benchmarks[45][46] --- Model Backtest Results 1. Short-term Quantitative Timing Model - Annualized return: 16.98% - Annualized volatility: 14.55% - Maximum drawdown: 27.70% - Sharpe ratio: 1.0138 - Excess return (2024 onwards): 2.26%[26][30][33] 2. Growth-Value Style Rotation Model - Annualized return: 11.82% - Annualized volatility: 20.79% - Maximum drawdown: 43.07% - Sharpe ratio: 0.5457 - Excess return (2025 YTD): -2.32%[36][39] 3. Small-Cap vs. Large-Cap Style Rotation Model - Annualized return: 12.38% - Annualized volatility: 22.69% - Maximum drawdown: 50.65% - Sharpe ratio: 0.5408 - Excess return (2025 YTD): -5.11%[41][44] 4. Four-Style Rotation Model - Annualized return: 13.29% - Annualized volatility: 21.53% - Maximum drawdown: 47.91% - Sharpe ratio: 0.6001 - Excess return (2025 YTD): -3.25%[45][46]

证券研究报告 | 金融工程 2025 年 6 月 22 日 地缘风险仍压制市场表现 ——A 股趋势与风格定量观察 20250622 1. 当前市场观察 风险提示：择时和风格轮动模型结论基于合理假设前提下结合历史数据统计规 律推导而出，市场环境变化下可能导致出现模型失效风险。 任瞳 S1090519080004 rentong@cmschina.com.cn 王武蕾 S1090519080001 wangwulei@cmschina.com.cn 敬请阅读末页的重要说明 定期报告 ❑ 本周市场整体下跌，大盘价值风格维持上涨，小盘成长风格跌幅较大。具体 来看，万得全 A 指数下跌 1.07%，上证 50 上涨约 0.15%，沪深 300、中证 1000 分别下跌约 0.33%、1.63%，国证成长下跌约 0.92%，国证价值上涨约 0.14%。 ❑ 短期来看，地缘风险未见缓解，市场情绪转为谨慎，建议仍持半仓观望。在 上周周报中，我们认为虽然国内经济数据以及投资者情绪均有所回暖，但鉴 于中东地缘风险存在持续发酵可能，短期建议维持半仓。此外，我们也回顾 了俄乌冲突期间全球权益市场表现。从结果上来看，在局势尚未明朗前， ...

- Model Name: Short-term Quantitative Timing Model; Model Construction Idea: The model aims to provide short-term market timing signals based on various market indicators; Model Construction Process: The model evaluates four main aspects: fundamentals, valuation, sentiment, and liquidity. Each aspect is assessed using specific indicators such as PMI, loan growth, M1 growth, PE and PB ratios, beta dispersion, trading volume sentiment, volatility, interest rates, exchange rate expectations, and financing amounts. The signals from these indicators are combined to generate an overall market timing signal. For example, the formula for the fundamental signal is based on the PMI and loan growth: $$ \text{Fundamental Signal} = \text{PMI} \times \text{Loan Growth} $$ where PMI represents the manufacturing PMI index and Loan Growth represents the year-on-year growth rate of medium and long-term loans in RMB. Model Evaluation: The model has shown significant improvement over the benchmark in terms of annualized returns and maximum drawdown reduction[19][22][23] - Model Name: Growth-Value Style Rotation Model; Model Construction Idea: The model aims to rotate between growth and value styles based on economic cycles and market conditions; Model Construction Process: The model uses a quantitative economic cycle analysis framework to assess the profitability cycle, interest rate cycle, and credit cycle. For example, the profitability cycle slope is calculated as: $$ \text{Profitability Cycle Slope} = \frac{\text{Current Profitability} - \text{Previous Profitability}}{\text{Time Period}} $$ The model also considers valuation differences (PE and PB ratios) and sentiment differences (turnover and volatility). The signals from these indicators are combined to generate a style rotation recommendation. Model Evaluation: The model has shown significant improvement over the benchmark in terms of annualized returns and maximum drawdown reduction[31][32][33] - Model Name: Small-Cap vs. Large-Cap Style Rotation Model; Model Construction Idea: The model aims to rotate between small-cap and large-cap styles based on economic cycles and market conditions; Model Construction Process: Similar to the Growth-Value Style Rotation Model, this model uses a quantitative economic cycle analysis framework to assess the profitability cycle, interest rate cycle, and credit cycle. It also considers valuation differences (PE and PB ratios) and sentiment differences (turnover and volatility). The signals from these indicators are combined to generate a style rotation recommendation. Model Evaluation: The model has shown significant improvement over the benchmark in terms of annualized returns and maximum drawdown reduction[35][36][37] - Model Name: Four-Style Rotation Model; Model Construction Idea: The model combines the Growth-Value and Small-Cap vs. Large-Cap Style Rotation Models to provide a comprehensive style rotation strategy; Model Construction Process: The model integrates the signals from the Growth-Value and Small-Cap vs. Large-Cap Style Rotation Models to recommend allocations across four styles: small-cap growth, small-cap value, large-cap growth, and large-cap value. The recommended allocation is based on the combined signals from the underlying models. Model Evaluation: The model has shown significant improvement over the benchmark in terms of annualized returns and maximum drawdown reduction[39][40][41] Model Backtest Results - Short-term Quantitative Timing Model: Annualized Return 16.27%, Annualized Volatility 14.73%, Maximum Drawdown 27.70%, Sharpe Ratio 0.9620, IR 0.5875[22][27] - Growth-Value Style Rotation Model: Annualized Return 11.35%, Annualized Volatility 20.89%, Maximum Drawdown 43.07%, Sharpe Ratio 0.5239, IR 0.2634[32][34] - Small-Cap vs. Large-Cap Style Rotation Model: Annualized Return 11.99%, Annualized Volatility 22.79%, Maximum Drawdown 50.65%, Sharpe Ratio 0.5241, IR 0.2367[36][38] - Four-Style Rotation Model: Annualized Return 12.90%, Annualized Volatility 21.64%, Maximum Drawdown 47.91%, Sharpe Ratio 0.5776, IR 0.2693[40][41]