- The report systematically reviews 80 new quantitative finance-related research papers in May 2025, covering areas such as equity research, fixed income, fund studies, asset allocation, machine learning applications, and ESG-related studies [1][2][3] - Equity research includes studies on fundamental factors, price-volume and alternative factors, factor research, active quantitative strategies, and other categories, exploring investor behavior biases, asset pricing models, market structure distortions, prediction model innovations, and corporate resilience mechanisms [2][10] - Fixed income research focuses on high-frequency inflation forecasting, sovereign risk premium decomposition, and stochastic interest rate model innovations, with findings such as weekly online inflation rates predicting yield curve slope factors and semi-Markov-modulated Hull-White/CIR models achieving semi-analytical pricing for zero-coupon bonds [22][23] - Fund studies investigate fund selection factors, fund style evaluation, and behavioral biases, revealing strategies like liquidity picking driving excess returns and public pension funds underperforming benchmarks due to alternative investment errors post-2008 [28][30] - Asset allocation research explores multi-asset portfolio management paradigm shifts, systematic currency management, and volatility connectedness constraints, demonstrating dynamic adaptation mechanisms and enhanced performance during crises [32][33][35] - Machine learning applications in finance include innovations in volatility forecasting, credit risk prediction using GraphSAGE models, and long-memory stochastic interval models, significantly improving prediction accuracy and economic value [36][38][40] - ESG-related studies analyze green innovation drivers, ESG evaluation distortions, and corporate environmental response strategies, highlighting mechanisms like family business constraints on green innovation and AI-driven manufacturing green transformation [42][43][45]

Group 1: Quantitative Finance Research Overview - A total of 80 new quantitative finance-related research papers were added this month, with the following distribution: 31 on equity research, 4 on fund research, 8 on bond research, 9 on asset allocation, 3 on machine learning applications in finance, and 22 on ESG-related research[1] - Equity research covers various topics including investor behavior biases, asset pricing models, and market structure distortions, impacting capital markets[2] - Bond research focuses on interest rate bonds, credit bonds, and other bond markets, analyzing high-frequency inflation forecasting and pricing distortion mechanisms[2] Group 2: Specific Findings in Research - High-frequency online inflation rates predict yield curve slope factors with a contribution rate of 61%[22] - The sovereign risk premium in the Eurozone is primarily driven by credit risk premiums, with Italy accounting for 78% of this effect[22] - Climate disasters lead to a temporary premium for green bonds over brown bonds, which diminishes within five months due to behavioral overreaction[24] Group 3: Machine Learning and Risk Management - Machine learning models significantly improve the prediction of implied volatility, showing economic value superior to traditional models[38] - The GraphSAGE model enhances credit risk prediction accuracy by 19% through integrating stock returns, risk spillovers, and trading networks[38] - Long Memory Stochastic Interval Models (LMSR) capture persistent characteristics in volatility, reducing out-of-sample prediction loss by 38%[38]

Quantitative Models and Construction Diffusion Index Model - **Model Name**: Diffusion Index Model - **Model Construction Idea**: The model is used to monitor the critical points of future diffusion index changes, predicting potential turning points in the market[6][43] - **Model Construction Process**: - The horizontal axis represents the relative price change of stocks in the micro-cap index components over a future period, ranging from +10% to -10% - The vertical axis represents the length of the review or forecast window, ranging from 20 days to 10 days - For example, a value of 0.16 at the intersection of a -5% price change (horizontal axis) and a 15-day window (vertical axis) indicates the diffusion index value under these conditions - The model uses historical data to calculate the diffusion index for different scenarios and predicts the likelihood of market turning points based on these values[43][45] - **Model Evaluation**: The model provides a systematic way to identify potential market turning points, but its accuracy depends on the stability of the index components and market conditions[6][43] - **Model Testing Results**: - Current diffusion index value: 0.91 - Historical signals: - Left-side threshold method triggered a sell signal on May 8, 2025, with a value of 0.9850[47] - Right-side threshold method triggered a sell signal on May 15, 2025, with a value of 0.8975[51] - Dual moving average method triggered a buy signal on April 30, 2025[52] --- Quantitative Factors and Construction Leverage Factor - **Factor Name**: Leverage Factor - **Factor Construction Idea**: Measures the financial leverage of companies, indicating their risk and potential return[5][38] - **Factor Construction Process**: Calculated as the ratio of total debt to equity or assets, normalized for comparison across companies[5][38] - **Factor Evaluation**: Demonstrated strong performance in the current week, with a rank IC of 0.143, significantly above its historical average of -0.006[5][38] Turnover Factor - **Factor Name**: Turnover Factor - **Factor Construction Idea**: Reflects the liquidity of stocks, with higher turnover indicating more active trading[5][38] - **Factor Construction Process**: Calculated as the ratio of trading volume to free float market capitalization over a specific period[5][38] - **Factor Evaluation**: Rank IC of 0.051 this week, outperforming its historical average of -0.08[5][38] PB Inverse Factor - **Factor Name**: PB Inverse Factor - **Factor Construction Idea**: Represents the inverse of the price-to-book ratio, identifying undervalued stocks[5][38] - **Factor Construction Process**: Calculated as 1 divided by the price-to-book ratio, normalized for comparison[5][38] - **Factor Evaluation**: Rank IC of 0.042 this week, slightly above its historical average of 0.034[5][38] Free Float Ratio Factor - **Factor Name**: Free Float Ratio Factor - **Factor Construction Idea**: Measures the proportion of shares available for public trading, indicating potential liquidity[5][38] - **Factor Construction Process**: Calculated as the ratio of free float shares to total shares outstanding[5][38] - **Factor Evaluation**: Rank IC of 0.04 this week, outperforming its historical average of -0.012[5][38] 10-Day Return Factor - **Factor Name**: 10-Day Return Factor - **Factor Construction Idea**: Captures short-term momentum by analyzing recent stock performance[5][38] - **Factor Construction Process**: Calculated as the percentage change in stock price over the past 10 trading days[5][38] - **Factor Evaluation**: Rank IC of 0.025 this week, significantly above its historical average of -0.061[5][38] Non-Adjusted Stock Price Factor - **Factor Name**: Non-Adjusted Stock Price Factor - **Factor Construction Idea**: Reflects the raw stock price without adjustments for splits or dividends[5][38] - **Factor Construction Process**: Directly uses the stock's current market price[5][38] - **Factor Evaluation**: Rank IC of -0.19 this week, underperforming its historical average of -0.017[5][38] PE_TTM Inverse Factor - **Factor Name**: PE_TTM Inverse Factor - **Factor Construction Idea**: Represents the inverse of the price-to-earnings ratio based on trailing twelve months, identifying undervalued stocks[5][38] - **Factor Construction Process**: Calculated as 1 divided by the PE_TTM ratio, normalized for comparison[5][38] - **Factor Evaluation**: Rank IC of -0.143 this week, underperforming its historical average of 0.018[5][38] ROE (Single Quarter) Factor - **Factor Name**: ROE (Single Quarter) Factor - **Factor Construction Idea**: Measures the profitability of companies based on their return on equity for a single quarter[5][38] - **Factor Construction Process**: Calculated as net income divided by shareholders' equity for the most recent quarter[5][38] - **Factor Evaluation**: Rank IC of -0.124 this week, underperforming its historical average of 0.023[5][38] Nonlinear Market Cap Factor - **Factor Name**: Nonlinear Market Cap Factor - **Factor Construction Idea**: Captures the nonlinear relationship between market capitalization and stock performance[5][38] - **Factor Construction Process**: Applies a nonlinear transformation to market capitalization data, such as logarithmic or polynomial adjustments[5][38] - **Factor Evaluation**: Rank IC of -0.116 this week, underperforming its historical average of -0.033[5][38] Log Market Cap Factor - **Factor Name**: Log Market Cap Factor - **Factor Construction Idea**: Measures the logarithmic transformation of market capitalization to reduce skewness[5][38] - **Factor Construction Process**: Calculated as the natural logarithm of market capitalization[5][38] - **Factor Evaluation**: Rank IC of -0.116 this week, underperforming its historical average of -0.033[5][38] --- Factor Backtesting Results - **Leverage Factor**: Rank IC 0.143[5][38] - **Turnover Factor**: Rank IC 0.051[5][38] - **PB Inverse Factor**: Rank IC 0.042[5][38] - **Free Float Ratio Factor**: Rank IC 0.04[5][38] - **10-Day Return Factor**: Rank IC 0.025[5][38] - **Non-Adjusted Stock Price Factor**: Rank IC -0.19[5][38] - **PE_TTM Inverse Factor**: Rank IC -0.143[5][38] - **ROE (Single Quarter) Factor**: Rank IC -0.124[5][38] - **Nonlinear Market Cap Factor**: Rank IC -0.116[5][38] - **Log Market Cap Factor**: Rank IC -0.116[5][38]

Group 1 - The A-share market experienced a contraction with major indices declining, indicating a cautious market sentiment amid reduced trading volume [3] - The REITs market showed an upward trend in secondary market prices, with the weighted REITs index closing at 139.74 and a weekly return of 1.36%, outperforming other major asset classes [4] - The copper industry is facing pressure from trade conflicts and rising domestic inventory, but prices may gradually increase with potential domestic stimulus policies and U.S. interest rate cuts [5] Group 2 - In the livestock sector, the average weight of slaughtered pigs has decreased, and the price of pigs has seen a larger decline, indicating a potential turning point in inventory levels and a long-term upward profit cycle post-deinventory [6] - Nobon Co., a leading player in the spunlace non-woven fabric industry, has shown strong performance in 2024 and Q1 2025, with advanced production lines and a focus on high-margin clients [7] - The small-cap style is currently favored in the market, with private equity research strategies showing significant excess returns [8]

Group 1 - The article discusses various industry reports and investment outlooks scheduled for late May 2023, focusing on sectors such as sugar-free tea, electronics, transportation, media, engineering machinery, and raw pharmaceutical materials [1][2][3][4][5][6][7][8][9][10][11][12][26][27][28][29] - Each report is led by a chief analyst specializing in the respective field, indicating a structured approach to industry analysis and investment recommendations [1][2][3][4][5][6][7][8][9][10][11][12][26][27][28][29] - The meetings are set to provide insights into current market trends, potential investment opportunities, and sector-specific developments, which are crucial for investors looking to make informed decisions [1][2][3][4][5][6][7][8][9][10][11][12][26][27][28][29] Group 2 - The schedule includes a deep dive into the sugar-free tea market, highlighting its growth potential and consumer trends [1] - The electronics industry report aims to provide an investment perspective amidst technological advancements and market shifts [2] - The transportation sector analysis will focus on the performance of companies like 安徽皖通高速, assessing their financial health and market position [3] - The media industry report will explore investment opportunities in the evolving landscape of digital content and advertising [4] - The engineering machinery update will cover recent developments and market dynamics affecting the sector [5] - The raw pharmaceutical materials report will summarize the market outlook for 2024 and the first quarter of 2025, focusing on supply chain and regulatory factors [6][7]

Quantitative Models and Construction Methods Quantitative Factors and Their Construction 1. **Factor Name**: Intraday Skewness Factor **Construction Idea**: Captures the skewness of intraday stock returns to identify potential return asymmetry[3][6] **Construction Process**: Referenced in the report "Stock Selection Factor Series Research (19) - High-Frequency Factors on Stock Return Distribution Characteristics"[11] **Evaluation**: Demonstrates mixed performance with positive returns in some periods but underperformance in others[3][6] 2. **Factor Name**: Downside Volatility Proportion Factor **Construction Idea**: Measures the proportion of downside volatility in intraday price movements to assess risk[3][6] **Construction Process**: Referenced in the report "Stock Selection Factor Series Research (25) - High-Frequency Factors on Realized Volatility Decomposition"[16] **Evaluation**: Shows consistent positive returns in certain periods but limited robustness in others[3][6] 3. **Factor Name**: Post-Open Buy Intention Proportion Factor **Construction Idea**: Quantifies the proportion of buy orders after market open to gauge investor sentiment[3][6] **Construction Process**: Referenced in the report "Stock Selection Factor Series Research (64) - Low-Frequency Applications of High-Frequency Data Using Intuitive Logic and Machine Learning"[20] **Evaluation**: Exhibits moderate performance with occasional strong returns[3][6] 4. **Factor Name**: Post-Open Buy Intention Intensity Factor **Construction Idea**: Measures the intensity of buy orders after market open to reflect market momentum[3][6] **Construction Process**: Referenced in the report "Stock Selection Factor Series Research (64) - Low-Frequency Applications of High-Frequency Data Using Intuitive Logic and Machine Learning"[24] **Evaluation**: Performance is inconsistent, with periods of underperformance[3][6] 5. **Factor Name**: Post-Open Large Order Net Buy Proportion Factor **Construction Idea**: Tracks the proportion of large net buy orders after market open to identify institutional activity[3][6] **Construction Process**: Derived from high-frequency trading data[30] **Evaluation**: Generally positive performance with strong returns in specific periods[3][6] 6. **Factor Name**: Post-Open Large Order Net Buy Intensity Factor **Construction Idea**: Measures the intensity of large net buy orders after market open to capture market trends[3][6] **Construction Process**: Derived from high-frequency trading data[35] **Evaluation**: Mixed results with moderate returns in some periods[3][6] 7. **Factor Name**: Improved Reversal Factor **Construction Idea**: Enhances traditional reversal factors by incorporating high-frequency data[3][6] **Construction Process**: Derived from intraday price reversals[40] **Evaluation**: Limited performance improvement over traditional reversal factors[3][6] 8. **Factor Name**: Tail-End Trading Proportion Factor **Construction Idea**: Measures the proportion of trading activity near market close to capture end-of-day effects[3][6] **Construction Process**: Derived from high-frequency trading data[45] **Evaluation**: Underperformance in most periods[3][6] 9. **Factor Name**: Average Single Transaction Outflow Proportion Factor **Construction Idea**: Tracks the proportion of outflows in single transactions to assess liquidity[3][6] **Construction Process**: Derived from high-frequency trading data[50] **Evaluation**: Limited effectiveness in predicting returns[3][6] 10. **Factor Name**: Large Order Push-Up Factor **Construction Idea**: Measures the impact of large orders on price increases to identify market movers[3][6] **Construction Process**: Derived from high-frequency trading data[55] **Evaluation**: Moderate performance with occasional strong returns[3][6] 11. **Factor Name**: Deep Learning High-Frequency Factor (Improved GRU(50,2)+NN(10)) **Construction Idea**: Combines GRU and neural networks to capture complex patterns in high-frequency data[3][6] **Construction Process**: Utilizes GRU(50,2) and NN(10) architectures for feature extraction and prediction[59] **Evaluation**: Strong performance in certain periods but underperformance in others[3][6] 12. **Factor Name**: Deep Learning High-Frequency Factor (Residual Attention LSTM(48,2)+NN(10)) **Construction Idea**: Incorporates residual attention mechanisms with LSTM and neural networks for enhanced prediction[3][6] **Construction Process**: Utilizes LSTM(48,2) and NN(10) architectures with residual attention layers[61] **Evaluation**: Consistently strong performance across multiple periods[3][6] 13. **Factor Name**: Deep Learning Factor (Multi-Granularity Model - 5-Day Label) **Construction Idea**: Uses multi-granularity modeling with 5-day labels for short-term predictions[3][6] **Construction Process**: Trained using bidirectional AGRU[64] **Evaluation**: Strong performance with high returns in most periods[3][6] 14. **Factor Name**: Deep Learning Factor (Multi-Granularity Model - 10-Day Label) **Construction Idea**: Uses multi-granularity modeling with 10-day labels for medium-term predictions[3][6] **Construction Process**: Trained using bidirectional AGRU[65] **Evaluation**: Consistently strong performance across multiple periods[3][6] AI-Enhanced Portfolio Construction 1. **Portfolio Name**: CSI 500 AI Enhanced Wide Constraint Portfolio **Construction Idea**: Maximizes expected returns under wide constraints using deep learning factors[69][70] **Construction Process**: - Weekly rebalancing - Constraints on individual stocks, industries, market cap, and other factors - Objective function: $$ max\sum\mu_{i}w_{i} $$ where \( w_i \) is the weight of stock \( i \) and \( \mu_i \) is its expected excess return[71] **Evaluation**: Strong cumulative excess returns since 2017[72] 2. **Portfolio Name**: CSI 500 AI Enhanced Strict Constraint Portfolio **Construction Idea**: Similar to the wide constraint portfolio but with stricter constraints[69][70] **Construction Process**: Same as above with stricter constraints on market cap, ROE, SUE, and volatility[71] **Evaluation**: Moderate cumulative excess returns since 2017[73] 3. **Portfolio Name**: CSI 1000 AI Enhanced Wide Constraint Portfolio **Construction Idea**: Maximizes expected returns under wide constraints using deep learning factors for smaller-cap stocks[69][70] **Construction Process**: Same as CSI 500 portfolios but applied to CSI 1000 index[71] **Evaluation**: Strong cumulative excess returns since 2017[76] 4. **Portfolio Name**: CSI 1000 AI Enhanced Strict Constraint Portfolio **Construction Idea**: Similar to the wide constraint portfolio but with stricter constraints for smaller-cap stocks[69][70] **Construction Process**: Same as above with stricter constraints on market cap, ROE, SUE, and volatility[71] **Evaluation**: Strong cumulative excess returns since 2017[79] Backtest Results for Factors 1. **Intraday Skewness Factor**: IC (2025): 0.057, Multi-Period Returns: 14.35% (2025)[3][6] 2. **Downside Volatility Proportion Factor**: IC (2025): 0.055, Multi-Period Returns: 11.77% (2025)[3][6] 3. **Post-Open Buy Intention Proportion Factor**: IC (2025): 0.033, Multi-Period Returns: 10.32% (2025)[3][6] 4. **Post-Open Buy Intention Intensity Factor**: IC (2025): 0.026, Multi-Period Returns: 11.19% (2025)[3][6] 5. **Post-Open Large Order Net Buy Proportion Factor**: IC (2025): 0.039, Multi-Period Returns: 12.32% (2025)[3][6] 6. **Post-Open Large Order Net Buy Intensity Factor**: IC (2025): 0.028, Multi-Period Returns: 6.78% (2025)[3][6] 7. **Improved Reversal Factor**: IC (2025): 0.003, Multi-Period Returns: 9.34% (2025)[3][6] 8. **Tail-End Trading Proportion Factor**: IC (2025): 0.022, Multi-Period Returns: 5.43% (2025)[3][6] 9. **Average Single Transaction Outflow Proportion Factor**: IC (2025): 0.012, Multi-Period Returns: 0.82% (2025)[3][6] 10. **Large Order Push-Up Factor

证券研究报告 | 金融工程 2025 年 5 月 25 日 利率择时信号维持看空 ——利率市场趋势定量跟踪 20250523 利率市场结构变化 - 10 年期国债到期收益率录得 1.72%，相对上周升高 4.15BP。当前 利率水平、期限和凸性结构读数分别为 1.58%、0.27%、-0.04%， 从均值回归视角看，目前处于水平结构点位较低、期限结构点位 较低、凸性结构点位较低的状态。 利率价量周期择时信号：看空 - 利率数据的多周期择时信号为：长周期无信号、中周期向上突破、 短周期向上突破。综合来看，当前合计下行突破 0 票、上行突破 2 票，由于 3 种周期下的同向突破总得票数达到 2/3，最终信号的综 合评分结果为看空。 公募债基行为跟踪：久期微降、分歧微降 利率价量多周期择时策略表现 - 自 2023 年底以来，策略的短期年化收益率为 7.5%，最大回撤为 1.61%，收益回撤比为 6.43，相对业绩基准的超额收益率为 2.35%。 过去的 18 年内，策略逐年绝对收益大于 0 的胜率为 100%，逐年超 额收益大于 0 的胜率为 100%。 - 公募基金最新久期（计入杠杆后）中位数读数为 3.09 ...

金融工程 | 动态跟踪 分红对期指的影响 20250523 研究结论 | | 收盘价 | 分红点数 | 实际价差 | 含分红价差 | | --- | --- | --- | --- | --- | | IH2506 | 2693.00 | 17.28 | -18.85 | -1.58 | | IH2507 | 2664.80 | 54.43 | -47.05 | 7.38 | | IH2509 | 2655.60 | 58.77 | -56.25 | 2.52 | | IH2512 | 2653.80 | 58.77 | -58.05 | 0.72 | 沪深 300 股指期货: | | 收盘价 | 分红点数 | 实际价差 | 含分红价差 | | --- | --- | --- | --- | --- | | IF2506 | 3846.20 | 20.75 | -36.07 | -15.32 | | IF2507 | 3808.80 | 55.62 | -73.47 | -17.86 | | IF2509 | 3777.80 | 66.11 | -104.47 | -38.37 | | IF2512 | 37 ...

Core Insights - The report presents a novel investment portfolio construction method that combines stock price momentum with hierarchical clustering (HC) to address the instability and concentration issues of Markowitz mean-variance (MV) portfolio weights [2][22] - The proposed Hierarchical Momentum (HM) strategy shows potential applicability in various domains such as stock portfolio construction, ETF portfolio construction, and asset allocation in the domestic market [2][22] Hierarchical Momentum Strategy - The HM strategy derives a distance function from the Pearson correlation coefficients between asset returns, using a bottom-up recursive approach to cluster assets based on proximity, resulting in a dendrogram [3][24] - At a certain height in the dendrogram, a horizontal cut is made to divide the tree into n clusters, identifying the assets with the highest momentum scores within each cluster while assigning zero weight to those with negative momentum scores [3][24] Empirical Results - The backtesting period spans from June 1997 to August 2022, utilizing a high-dimensional dataset from the MSCI All Country World Index (ACWI), which includes large-cap and mid-cap stocks from 23 developed and 24 emerging markets [5][43] - After accounting for transaction costs, the HM strategy outperforms all other strategies in terms of cumulative returns, average returns, risk-adjusted returns (Sharpe and Sortino ratios), and risk metrics (volatility and maximum drawdown) [5][55] - The HM strategy demonstrates improved stability in industry allocation compared to the Maximum Momentum (MM) and Threshold Momentum (TM) strategies, which are known for their potential large drawdown issues [5][56] Methodology - The HM portfolio construction method does not require the inversion of the covariance matrix, instead relying on a hierarchical clustering approach to reduce dimensionality and ensure sparse diversification [24][68] - The method involves two main steps: applying hierarchical clustering to create a distance matrix and then constructing portfolio weights based on the hierarchical structure and momentum scores [24][38] Conclusion - The report emphasizes the importance of sparse diversification in constructing superior investment portfolios, particularly in high-dimensional environments where traditional methods may underperform [68][69] - The HM strategy effectively captures momentum premiums while mitigating risks associated with traditional momentum strategies, demonstrating its robustness across different economic conditions [68][69]

Group 1 - The report introduces a quantitative model solution for addressing the issue of value and growth style switching, based on the combination of odds and win rates [1][8] - The recent performance of the growth style portfolio was 0.82%, while the value style portfolio achieved a return of 1.15% [1][8] Group 2 - The estimated odds for the growth style is 1.09, and for the value style, it is 1.08, indicating a negative correlation between relative valuation levels and expected odds [2][14] - The current win rates indicate that 4 out of 7 indicators favor growth, resulting in a win rate of 58.26% for growth and 41.74% for value [3][18] Group 3 - The latest investment expectation for the growth style is calculated to be 0.22, while the value style has an investment expectation of -0.13, leading to a recommendation for the growth style [4][19] - Since 2013, the annualized return of the style rotation model based on investment expectations is 27.18%, with a Sharpe ratio of 0.98 [4][20]