The provided content does not contain any specific quantitative models or factors, nor does it include detailed construction processes, formulas, or backtesting results for such models or factors. The report primarily focuses on market trends, style performance, valuation metrics, and other financial indicators. Therefore, no summary of quantitative models or factors can be generated from this content.

金融工程 | 证券研究报告 —周报 2025 年 5 月 18 日 杠杆资金持续回升，大盘及成长风格占优 中银量化大类资产跟踪 小盘 vs大盘：小盘风格超额净值及拥挤度持续处于历史低位；大盘风格 拥挤度下降至历史低位。 微盘股 vs基金重仓：近期微盘股拥挤度快速下降至历史低位；基金重仓 拥挤度及超额累计净值处于历史低位。 A 股行情及成交热度 ◼ 本周领涨的行业为国防军工、房地产、纺织服装；领跌的行业为电力设 备及新能源、计算机、综合金融；成交热度最高的行业为家电、纺织服 装、电力设备及新能源。 A 股估值与股债性价比 股票市场概览 ◼ 本周 A 股上涨，港股上涨，美股上涨，其他海外权益市场普遍上涨。 A 股风格与拥挤度 成长 vs 红利：成长风格拥挤度及超额净值持续处于历史低位；红利风 格拥挤度当前仍处于历史较低位置，近期快速下降。 A 股资金面 机构调研活跃度 ◼ 当前机构调研活跃度历史分位居前的行业为纺织服装、商贸零售、房地 产，居后的行业为食品饮料、医药、电力设备及新能源。 利率市场 ◼ 本周中国国债利率上涨，美国国债利率上涨，中美利差处于历史高位。 汇率市场 ◼ 近一周在岸人民币较美元升值，离岸人民 ...

Investment Rating - The report maintains an "Overweight" rating for the sector [1] Core Viewpoints - The market is expected to experience a period of volatility, with a focus on relative opportunities in the pharmaceutical and dividend sectors [19][21] - The model predicts a potential shift to a downward trend for the Wande All A Index, indicating a possible adjustment phase in May [19][26] - The pharmaceutical sector is highlighted for its relative stability and potential for returns, while the dividend sector is also expected to perform well after a short-term adjustment [19][21] Summary by Sections A-share Market Overview (May 12-16, 2025) - The top three broad indices were: North Certificate 50 (3.13%), Wande Micro-Pan Daily Equal Weight Index (1.58%), and ChiNext Index (1.38%) [10] - The bottom three indices were: Sci-Tech Innovation 100 (-1.29%), Sci-Tech Innovation 50 (-1.10%), and Sci-Tech Comprehensive Index (-1.00%) [10] A-share Market Outlook (May 19-23, 2025) - The Wande All A Index's daily model shifted from a positive to a negative signal on May 15, suggesting a potential adjustment phase [19][26] - The monthly model for May scored -2.5, indicating a slight adjustment in the A-share market [19][26] - The report anticipates a "V-shaped" market movement, with ongoing pressure on trading volumes [19] Fund Allocation Recommendations - The report suggests a defensive ETF allocation strategy, focusing on the pharmaceutical and dividend sectors for relative returns [19][21]

Group 1 - The report introduces a quantitative timing strategy for style indices based on valuation, liquidity, and crowding models, emphasizing that "efficient markets" are dynamic processes rather than static states [1][6] - The quantitative timing model effectively captures the characteristics of style index bottoms and tops while mitigating risks associated with crowded trades, achieving an average annual return of 18.54% and an average excess annual return of 16.46% since 2011 [1][6] - The report highlights the performance of the mixed style index model, which has achieved an annual return of 20.10% and an excess annual return of 16.24% since December 2013 [1][6] Group 2 - The style index valuation model includes factors such as PB, PE, PBPE, and equity risk premium, with an average annual return of 10.38% and an average excess annual return of 8.30% since 2011 [1][6][17] - The market liquidity model incorporates factors like buy and sell impact costs and liquidity indices, showing a significant accuracy in bottom timing with an average rebound return of 6.86% [1][6][19] - The trading crowding model serves as a top-timing hedge factor, effectively complementing the valuation and liquidity models, achieving an excess annual return of 4.87% since 2011 [1][6][19] Group 3 - The report outlines a quantitative timing research framework that includes data processing, model factor calculation, model testing, and composite model synthesis [1][6][19] - The valuation factors are constructed by calculating the historical percentile levels of the style index valuation factors, which are then compared against set thresholds to trigger buy or sell signals [1][6][21] - The report emphasizes the need for timing factors to be logical and mean-reverting, with specific thresholds established for different style indices to determine market conditions [1][6][20]