丨证券研究报告丨 战略数据研究丨专题报告 [Table_Title] 成长指数拥挤度突破今年三季度高点—— W134 市场观察 报告要点 [Table_Summary] 当周科技成长继续领涨，长江创业板成长动量周度涨幅靠前，长江动量指数周度涨幅超 4%， 成长整体拥挤度已经越过今年三季度成长拥挤度高点。当周 A 股市场成交活跃度显著提升，周 度日均成交额达 1.95 万亿，较前周明显放量，12 月 12 日单日成交额达 2.12 万亿元，创近期 新高。各类风格和行业拥挤度比较上看，红利风格拥挤度普遍边际变化不大，港股波动下，港 股通红利成交热度有所回暖。电信业务、纸类及包装等交易活跃度回暖相对较多。当周中盘行 情修复相对靠前。 分析师及联系人 [Table_Author] 陈洁敏 SAC：S0490518120005 SFC：BUT348 请阅读最后评级说明和重要声明 %% %% %% %% research.95579.com 1 [Table_Title 成长指数拥挤度 2] 突破今年三季度高点—— W134 市场观察 [Table_Summary2] 度量市场情绪：沪深 300 成交热度持续下滑，成长拥 ...

丨证券研究报告丨 战略数据研究丨专题报告 [Table_Title] 成长指数拥挤度重回今年三季度高点 —— W133 市场观察 报告要点 [Table_Summary] 当周大盘修复反弹，科技成长领涨，长江创业板成长动量周度涨幅超 9%，长江动量指数周度 涨幅近 10%，今年三季度成长热点板块周度回归。不过成长修复背后，周度全 A 的成交额相较 于今年三季度仍有一定程度回落。各类风格和行业拥挤度比较上看，红利风格拥挤度普遍边际 变化不大，社会服务、农产品、传媒互联网等交易活跃度回暖相对较多。当周小微盘表现优于 大盘，成交活跃度也呈现同步回升。 分析师及联系人 [Table_Author] 陈洁敏 SAC：S0490518120005 SFC：BUT348 请阅读最后评级说明和重要声明 %% %% %% %% research.95579.com 1 [Table_Title 成长指数拥挤度重回今年三季度高点 2] —— W133 市场观察 [Table_Summary2] 度量市场情绪：小微盘指数成交热度持续回暖，成长指数成交拥挤度边际修复 机构赚钱效应：成长回暖下，基金重仓悉数反弹 行业板块：通讯业务、信 ...

Quantitative Models and Construction Methods Dividend Enhancement Portfolio - **Model Name**: Dividend Enhancement Portfolio - **Model Construction Idea**: The model focuses on selecting high-quality dividend stocks and constructing portfolios that outperform the benchmark dividend index by leveraging dividend quality and growth factors[7][15][21] - **Model Construction Process**: - The portfolio is constructed by selecting stocks with high dividend yields and strong dividend growth potential - The selection process incorporates a combination of fundamental analysis and quantitative screening to identify stocks with stable and growing dividends - The portfolio is rebalanced periodically to ensure alignment with the dividend quality and growth criteria[13][14][15] - **Model Evaluation**: The model demonstrates strong performance in capturing excess returns over the benchmark dividend index, particularly in volatile market conditions[7][15][21] Electronic Balanced Allocation Enhancement Portfolio - **Model Name**: Electronic Balanced Allocation Enhancement Portfolio - **Model Construction Idea**: This model aims to achieve excess returns by balancing exposure across various sub-sectors within the electronics industry, focusing on growth and stability[7][24][29] - **Model Construction Process**: - The portfolio is constructed by allocating weights to sub-sectors such as LED chips and semiconductor distribution, which exhibit strong growth potential - Quantitative screening is applied to identify leading companies within these sub-sectors - The portfolio is periodically rebalanced to maintain a balanced exposure across the electronics industry[13][14][24] - **Model Evaluation**: The model effectively captures excess returns and ranks in the top 15% of active technology-themed products in terms of weekly performance[7][24][29] --- Model Backtesting Results Dividend Enhancement Portfolio - **Weekly Excess Return**: Approximately 0.23% for the "Central SOE High Dividend 30 Portfolio" and 1.45% for the "Balanced Dividend 50 Portfolio"[7][15][21] - **Year-to-Date Excess Return**: The "Balanced Dividend 50 Portfolio" achieved an excess return of approximately 7.91% relative to the benchmark dividend index, ranking in the top 35% of all dividend-themed funds[21] Electronic Balanced Allocation Enhancement Portfolio - **Weekly Excess Return**: Approximately 0.89%, ranking in the top 15% of active technology-themed products[7][24][29]

Market Overview - The market experienced significant volatility, with the technology growth sector continuing to decline while low beta and dividend styles showed relative resilience[2] - The Shanghai Composite Index closed below 3900 on the last trading day of the week, with trading volume dropping to below 1.8 trillion CNY mid-week, indicating a prevailing risk-averse sentiment[2] Sector Performance - Weekly trading activity improved notably in the consumer sector, particularly in agricultural products, social services, and commercial trade[6] - Communication services and essential consumer sectors led the weekly gains, with excess returns of 2.82% and 2.48% respectively[28] Investment Trends - The small-cap stocks and the CSI 2000 index showed signs of recovery in trading activity, while institutional heavy positions across various indices experienced declines[6] - The fund overlap index reported a weekly loss of 5.01%, indicating a general downturn in institutional investment performance[22] Style Analysis - Small-cap stocks faced a weekly pullback, while low beta stocks demonstrated relative strength against market declines[29] - The performance of major style indices showed that small-cap and mid-cap growth indices experienced declines of 5.93% and 6.10% respectively, highlighting a challenging environment for growth-oriented investments[32] Thematic Highlights - The Yangtze Energy Efficiency Star and the Beijing-Tianjin-Hebei Regional Development Leading Index performed relatively well during the week, with the latter showing a loss of only 1.72%[34]

- The report highlights the launch of multiple active quantitative strategies since July 2023, including Dividend Selection Strategy and High Winning Rate Industry Strategy, aimed at tracking market trends and selecting industry-specific stocks [6][14][15] - Active quantitative strategies follow a top-down stock selection logic, leveraging industry and thematic insights to refine factor selection from a large fundamental factor pool, enabling precise identification of potential stocks within specific sectors [14] - The Dividend Series includes two products: "Central SOE High Dividend 30 Portfolio" and "Balanced Growth Dividend 50 Portfolio," focusing on stable and growth-oriented dividend strategies [15] - The Electronics Series includes two products: "Electronics Balanced Allocation Enhanced Portfolio" and "Electronics Sector Preferred Enhanced Portfolio," targeting mature sub-sector leaders within the electronics industry [15] - Weekly performance tracking shows that the Electronics Balanced Allocation Enhanced Portfolio and Electronics Sector Preferred Enhanced Portfolio achieved positive excess returns of approximately 1.86% and 1.62%, respectively, outperforming the benchmark [7][25][32] - The Dividend Series underperformed the benchmark this week, with the Central SOE High Dividend 30 Portfolio and Balanced Growth Dividend 50 Portfolio failing to exceed the returns of the CSI Dividend Total Return Index [7][16][22] - The CSI Dividend Index achieved a weekly return of 0.25%, while sub-indices like CSI Dividend Growth and CSI Dividend Low Volatility outperformed with average weekly returns of approximately 1.31% and 1.08%, respectively [7][16][19] - The Electronics Series portfolios ranked in the top 26%-28% among active technology-themed funds based on weekly returns [32] - The report emphasizes the differentiation of active quantitative strategies from traditional ones, highlighting their ability to integrate thematic and industry logic for enhanced stock selection and strategy validation [14]

Group 1: Market Overview - The domestic macro liquidity is expected to remain loose, with the central bank injecting 300 billion yuan through MLF in August, marking six consecutive months of increased operations [2] - There are two major positive factors for September: the 93rd National Day parade, which is believed to be closely related to market sentiment and risk appetite, and the anticipated interest rate cut by the Federal Reserve [2][3] - The market's expectation for a rate cut in September has risen to over 91%, with a possibility of a significant cut of 50 basis points similar to last September [3] Group 2: Sector Focus - Focus on sectors that may benefit from liquidity easing, prioritizing financials, healthcare, and real estate, followed by technology growth, fintech, and biotechnology [5] - Historical data indicates that cyclical sectors like non-ferrous metals and chemicals tend to perform well in the month of a Fed rate cut [5] Group 3: Company Recommendations - Meitu (3690.HK) has a current P/E ratio of 19.58x, close to its lowest level in a year, indicating high cost-effectiveness for long-term holding [6] - Laopuhuangjin (6181.HK) reported a revenue of 12.354 billion yuan in the first half of the year, a significant increase of 251% year-on-year, with a net profit of 2.268 billion yuan, up 285.8% [8] - ZTE Corporation (0763.HK) achieved a revenue of 71.553 billion yuan in the first half of the year, a year-on-year increase of 14.51%, with a notable growth in its government and enterprise business [13] - Lens Technology (6613.HK) reported a revenue of 32.96 billion yuan in the first half of the year, a 14.18% increase year-on-year, with a net profit of 1.143 billion yuan, up 32.68% [15] - Duolingo (DUOL.US) saw a revenue of $252 million in Q2 2025, a 41.6% increase year-on-year, despite a slowdown in user growth [17][18] - Snap (SNAP.US) reported Q2 revenue of $1.34 billion, slightly below expectations, but the company anticipates a revenue increase in Q3 [20][21] - DoorDash (DASH) achieved a revenue of $3.28 billion in Q2 2025, a 25% year-on-year increase, significantly exceeding analyst expectations [22]

Quantitative Models and Construction Methods - **Model Name**: "Offense and Defense Dividend 50 Portfolio" **Model Construction Idea**: This model aims to enhance returns by selecting high-dividend stocks with a balance of growth and stability, outperforming the benchmark dividend indices[6][15] **Model Construction Process**: The portfolio is constructed by combining stocks with high dividend yields, growth potential, and low volatility. The selection process involves filtering stocks based on dividend-related factors and optimizing the portfolio to achieve a balance between growth and defensive characteristics[6][15] **Model Evaluation**: The model demonstrates strong performance, consistently outperforming the benchmark dividend indices and ranking in the top percentile among dividend-focused funds[6][21] - **Model Name**: "Central SOE High Dividend 30 Portfolio" **Model Construction Idea**: This model focuses on central state-owned enterprises (SOEs) with high dividend payouts, aiming to capture stable returns from these entities[15] **Model Construction Process**: The portfolio is constructed by selecting 30 central SOEs with the highest dividend yields. The selection criteria emphasize stability and consistent dividend payouts[15] **Model Evaluation**: The model shows stable performance, delivering excess returns over the benchmark dividend indices[15][21] - **Model Name**: "Electronic Sector Enhanced Portfolios" **Model Construction Idea**: These models aim to enhance returns within the electronic sector by focusing on high-growth sub-sectors and leading companies in mature sub-sectors[15][31] **Model Construction Process**: 1. **Balanced Allocation Enhanced Portfolio**: This portfolio is constructed by evenly allocating weights across various electronic sub-sectors to achieve diversification[15] 2. **Sector Leader Enhanced Portfolio**: This portfolio focuses on leading companies in mature sub-sectors, emphasizing their growth potential and market dominance[15][31] **Model Evaluation**: Both portfolios demonstrate positive returns, with the Sector Leader Enhanced Portfolio delivering higher excess returns relative to the electronic sector index[31] Model Backtesting Results - **Offense and Defense Dividend 50 Portfolio**: - Weekly excess return: ~1.41% over the CSI Dividend Total Return Index[6][21] - Year-to-date excess return: ~3.52% over the CSI Dividend Total Return Index[21] - Weekly performance percentile: ~11% among dividend-focused funds[6][21] - **Central SOE High Dividend 30 Portfolio**: - Weekly excess return: ~0.35% over the CSI Dividend Total Return Index[6][21] - **Electronic Sector Enhanced Portfolios**: - **Balanced Allocation Enhanced Portfolio**: Weekly excess return: ~0.89% over the electronic sector index[31] - **Sector Leader Enhanced Portfolio**: Weekly excess return: ~0.89% over the electronic sector index[31] Quantitative Factors and Construction Methods - **Factor Name**: Dividend Quality **Factor Construction Idea**: This factor evaluates the stability and sustainability of a company's dividend payouts[16][18] **Factor Construction Process**: The factor is calculated using metrics such as dividend payout ratio, historical dividend growth rate, and earnings stability. Companies with higher scores on these metrics are ranked higher[16][18] **Factor Evaluation**: The factor demonstrates strong predictive power for identifying high-performing dividend stocks[16][18] - **Factor Name**: Dividend Growth **Factor Construction Idea**: This factor focuses on the growth potential of a company's dividends over time[16][18] **Factor Construction Process**: The factor is derived from the historical growth rate of dividends and projected earnings growth. Companies with consistent and high dividend growth rates are ranked higher[16][18] **Factor Evaluation**: The factor shows significant excess returns compared to pure dividend yield factors[16][18] - **Factor Name**: Low Volatility Dividend **Factor Construction Idea**: This factor targets stocks with high dividend yields and low price volatility[16][18] **Factor Construction Process**: The factor is constructed by combining dividend yield with a volatility measure (e.g., standard deviation of returns). Stocks with high yields and low volatility are ranked higher[16][18] **Factor Evaluation**: The factor provides a defensive characteristic, outperforming during market downturns[16][18] Factor Backtesting Results - **Dividend Quality Factor**: - Weekly excess return: ~1.94% over the CSI Dividend Index[18] - **Dividend Growth Factor**: - Weekly excess return: ~0.92% over the CSI Dividend Index[18] - **Low Volatility Dividend Factor**: - Weekly excess return: ~0.69% over the CSI Dividend Index[18]

Quantitative Models and Construction Methods 1. Model Name: "攻守兼备红利50组合" (Balanced Dividend 50 Portfolio) - **Model Construction Idea**: This model adopts an active quantitative strategy with a "top-down" stock selection logic. It aims to identify high-potential stocks within dividend-related themes by leveraging a refined factor library and aligning with industry and thematic insights[14][15] - **Model Construction Process**: 1. The model focuses on dividend-related stocks, emphasizing a balance between "stability" and "growth" factors[15] 2. It selects stocks from a universe of high-dividend companies, incorporating both fundamental and thematic factors to refine the portfolio[14][15] 3. The portfolio is actively managed and periodically rebalanced to maintain alignment with the strategy's objectives[15] - **Model Evaluation**: The model demonstrates a strong ability to generate excess returns over the benchmark, showcasing its robustness in balancing defensive and growth-oriented attributes[22] 2. Model Name: "央国企高分红30组合" (Central SOE High Dividend 30 Portfolio) - **Model Construction Idea**: This model targets high-dividend stocks within central state-owned enterprises (SOEs), emphasizing stability and defensive characteristics[15] - **Model Construction Process**: 1. The portfolio is constructed by selecting 30 high-dividend stocks from central SOEs[15] 2. It prioritizes companies with consistent dividend payouts and strong financial health[15] 3. The portfolio is actively managed to ensure alignment with its defensive strategy[15] - **Model Evaluation**: The model is effective in capturing stable returns, making it suitable for risk-averse investors seeking consistent income[22] 3. Model Name: "电子均衡配置增强组合" (Electronics Balanced Allocation Enhanced Portfolio) - **Model Construction Idea**: This model aims to achieve balanced exposure within the electronics sector, focusing on diversification and stability[15] - **Model Construction Process**: 1. The portfolio selects stocks across various sub-sectors within the electronics industry[15] 2. It employs a balanced allocation strategy to mitigate concentration risks[15] 3. The portfolio is periodically rebalanced to maintain its diversified structure[15] - **Model Evaluation**: The model has demonstrated positive excess returns over the electronics sector index, indicating its effectiveness in achieving balanced growth[32] 4. Model Name: "电子板块优选增强组合" (Electronics Sector Preferred Enhanced Portfolio) - **Model Construction Idea**: This model focuses on identifying leading companies within mature sub-sectors of the electronics industry, emphasizing growth potential[15] - **Model Construction Process**: 1. The portfolio targets mature sub-sector leaders with strong fundamentals and growth prospects[15] 2. It employs a factor-based approach to refine stock selection within the electronics sector[15] 3. The portfolio is actively managed to capitalize on emerging opportunities within the sector[15] - **Model Evaluation**: The model has achieved significant excess returns over the electronics sector index, highlighting its ability to capture growth opportunities[32] --- Model Backtesting Results 1. "攻守兼备红利50组合" - Excess return over the CSI Dividend Total Return Index: 4.76% (YTD 2025)[22] - Weekly excess return: 0.85%[22] 2. "央国企高分红30组合" - Weekly excess return over the CSI Dividend Total Return Index: 0.11%[22] 3. "电子均衡配置增强组合" - Excess return over the electronics sector index (YTD 2025): 2.40%[32] 4. "电子板块优选增强组合" - Excess return over the electronics sector index (YTD 2025): 5.83%[32]

Core Viewpoint - The adjustment of QDII fund's regional allocation strategy, particularly the increased allocation to Hong Kong stocks, has drawn market attention as the second quarter reports are released [1][2]. Group 1: QDII Fund Allocation Changes - As of July 20, 37 QDII funds have allocated assets to Hong Kong stocks, with many increasing their holdings [1]. - The Nasdaq index's volatility in the first quarter prompted QDII funds to reduce their U.S. stock positions and shift towards the more favorably valued Hong Kong market [2]. - For instance, the allocation of the Chuangjin Hexin Global Pharmaceutical Biotechnology Stock Fund to U.S. stocks decreased from 39.04% at the end of last year to 29.33% by the end of the first quarter, while its Hong Kong stock allocation rose from 13.97% to 24.07% [2]. Group 2: Sector Focus - The technology sector remains a core focus for QDII funds, with 38 funds investing a total of 33.147 billion yuan in this sector [4]. - The information technology industry, which includes sub-sectors like artificial intelligence, semiconductors, and software services, is the largest allocation area for QDII funds [4]. - Sixteen QDII funds maintain over 30% allocation to the information technology sector, with some, like the Chuangjin Hexin Global Chip Industry Stock Fund, increasing their allocation from 74.65% to 75.47% [4]. Group 3: Future Investment Trends - Many QDII funds are focusing on the AI industry chain as a key investment theme for the second half of the year, emphasizing the importance of AI's development from foundational infrastructure to application scenarios [5]. - The trend of increasing allocation to Hong Kong stocks and focusing on technology reflects a rational response to global market valuation differences and an active grasp of industrial transformation opportunities [5]. - Investors are advised to pay attention to the technological barriers, global competitiveness, and policy adaptability of leading technology stocks in Hong Kong [5].

Group 1: Market Performance - Fund holdings outperformed northbound holdings, with non-fund holdings leading the gains[2] - The real estate sector showed significant growth, with a 6.07% increase, exceeding the overall A-share market by 4.37%[22] - The healthcare sector's leading stocks significantly outperformed the industry benchmark[4] Group 2: Market Trends - Market rotation speed across industries and styles remains high[4] - Small-cap and high-valuation stocks are currently favored, with the reversal index showing strong performance[4] - Digital currency and specialized innovation themes are leading the market[4] Group 3: Investment Insights - Non-fund holdings index gained 2.88%, while fund holdings index only increased by 1.16%[14] - The reversal index has shown a notable increase, indicating a shift in market sentiment towards recovery[26] - The specialized innovation 100 index rose by 3.61%, highlighting strong interest in niche sectors[28]