[Table_Title] 金 金融工程点评 [Table_Message]2025-12-24 金工 ETF 点评：宽基 ETF 单日净流入 51.19 亿元；家电、环保拥挤变幅较大 [Table_Author] 证券分析师：刘晓锋 电话：13401163428 E-MAIL：liuxf@tpyzq.com 执业资格证书编码：S1190522090001 证券分析师：孙弋轩 电话：18910596766 E-MAIL：sunyixuan@tpyzq.com 执业资格证书编码：S1190525080001 一、资金流向 二、行业拥挤度监测 ◼ 通过构建行业拥挤度监测模型，对申万一级行业指数的拥挤度进行每日监测， 前一交易日军工、有色、建材靠前，相比较而言，银行、计算机、传媒的拥挤 度水平较低，建议关注。此外，家电、环保拥挤度变动较大。从主力资金流 动来看，前一交易日主力资金流入化工、电力设备；流出军工、电子。近三 个交易日主力资金减配电子、军工、计算机；增配电力设备、化工。 三、ETF 产品关注信号 ◼ 根据溢价率 Z-score 模型搭建相关 ETF 产品筛选信号模型，通过滚动测算提 供存在潜在套利机会的标 ...

Quantitative Models and Construction Methods 1. Model Name: Industry Crowding Monitoring Model - **Model Construction Idea**: This model is designed to monitor the crowding levels of Shenwan First-Level Industry Indices on a daily basis, identifying industries with high or low crowding levels and significant changes in crowding dynamics[3] - **Model Construction Process**: The model calculates the crowding levels of various industries based on daily data. It identifies industries with the highest and lowest crowding levels and highlights industries with significant changes in crowding dynamics. Specific calculation methods or formulas are not provided in the report[3] - **Model Evaluation**: The model provides actionable insights into industry crowding trends, helping investors identify potential opportunities or risks in crowded or undercrowded sectors[3] 2. Model Name: Premium Rate Z-Score Model - **Model Construction Idea**: This model is used to screen ETF products for potential arbitrage opportunities by calculating the Z-score of premium rates on a rolling basis[4] - **Model Construction Process**: The Z-score of the premium rate is calculated for each ETF product over a rolling window. The Z-score helps identify ETFs with significant deviations from their historical premium rates, signaling potential arbitrage opportunities. Specific formulas or parameters are not detailed in the report[4] - **Model Evaluation**: The model effectively identifies ETFs with potential arbitrage opportunities, but it also warns of potential risks of price corrections in the identified ETFs[4] --- Model Backtesting Results 1. Industry Crowding Monitoring Model - No specific backtesting results or quantitative metrics are provided for this model in the report[3] 2. Premium Rate Z-Score Model - No specific backtesting results or quantitative metrics are provided for this model in the report[4] --- Quantitative Factors and Construction Methods No specific quantitative factors are mentioned or constructed in the report --- Factor Backtesting Results No specific backtesting results for factors are mentioned in the report

Group 1: Overview of QDII ETF - QDII ETFs are primarily focused on Hong Kong stocks, experiencing rapid growth since 2021, with a significant increase in non-Hong Kong ETFs starting in 2023, reflecting a shift in investor interest towards overseas markets [3][12] - As of November 28, 2025, the cumulative scale of QDII ETFs reached 185.86 billion, indicating a growing trend in overseas investment opportunities [12] - The top 15 QDII ETFs by scale include products like Huaxia Hang Seng Technology ETF and Guotai Junan Nasdaq 100 ETF, with scales of 47.64 billion and 16.90 billion respectively [9][13] Group 2: Mechanism of Premium and Discount Formation - The premium formation mechanism for QDII ETFs relies on cash subscription and redemption, where the return to premium depends on the ability to arbitrage through "subscription + sale" [3][21] - High premiums often arise from disruptions in the arbitrage chain, particularly when subscription limits are imposed due to insufficient QDII quotas [27] - The expected arbitrage returns vary based on market conditions, with overlapping trading hours leading to different calculations of the indicative optimized price value (IOPV) [31][36] Group 3: Evaluation of QDII ETF Premiums - The premium rate is negatively correlated with the cost-effectiveness of QDII ETFs; for instance, when the premium rate exceeds 8%, holding the ETF for a month typically results in significant negative returns [3] - Daily subscription limits and market sentiment are significant factors influencing premium rates, with emotional factors providing long-term explanations for premium levels [3][27] - Some QDII ETFs exhibit additional premiums that cannot be easily explained by market sentiment, indicating potential inefficiencies in pricing [3][27] Group 4: Recent Developments in QDII ETFs - The number of QDII ETFs tracking new indices has been increasing, particularly those focusing on non-Hong Kong indices such as the Dow Jones Industrial Average and the S&P 500 Consumer Select Index [17] - Newly launched QDII ETFs often experience initial premiums, with the Huatai Baichuan South China Arabian ETF showing a first-day premium of 6.31% [17][18] - The overall market enthusiasm for QDII ETFs remains high, as evidenced by the sustained premium levels of newly listed products [17][18]

金 金融工程点评 [Table_Title] [Table_Message]2025-12-01 金工 ETF 点评：宽基 ETF 单日净流出 14.37 亿元；建装、交运、家电拥挤变幅较大 [Table_Author] 证券分析师：刘晓锋 电话：13401163428 E-MAIL：liuxf@tpyzq.com 执业资格证书编码：S1190522090001 证券分析师：孙弋轩 电话：18910596766 E-MAIL：sunyixuan@tpyzq.com 执业资格证书编码：S1190525080001 一、资金流向 二、行业拥挤度监测 ◼ 通过构建行业拥挤度监测模型，对申万一级行业指数的拥挤度进行每日监测， 前一交易日通信、房地产靠前，相比较而言，家电、汽车、非银的拥挤度水 平较低，建议关注。此外，建筑装饰、交运、家电拥挤度变动较大。从主力资 金流动来看，前一交易日主力资金流入电子、有色、汽车；流出医药、传媒。 近三个交易日主力资金减配传媒、计算机；增配电子、汽车。 三、ETF 产品关注信号 ◼ 根据溢价率 Z-score 模型搭建相关 ETF 产品筛选信号模型，通过滚动测算提 供存在潜在套利机会的 ...

- The report introduces an **industry crowding monitoring model** to track the crowding levels of Shenwan primary industry indices on a daily basis. The model identifies industries with high crowding levels (e.g., communication and electronics) and low crowding levels (e.g., automotive and non-bank financials). It also highlights significant changes in crowding levels for industries like construction decoration and military industries[3] - A **Z-score premium model** is constructed to screen ETF products for potential arbitrage opportunities. The model uses rolling calculations to identify ETFs with significant deviations in premium rates, which may indicate arbitrage opportunities or potential risks of price corrections[4] - The report provides detailed data on **ETF fund flows**, categorizing them into broad-based ETFs, industry-themed ETFs, style-strategy ETFs, and cross-border ETFs. For example, the top three net inflows for broad-based ETFs include the SSE 50 ETF (+6.60 billion yuan), A500 ETF (+5.84 billion yuan), and ChiNext 50 ETF (+2.75 billion yuan), while the top three net outflows include the ChiNext ETF (-7.26 billion yuan), CSI 500 ETF (-5.56 billion yuan), and STAR 50 ETF (-5.10 billion yuan)[5]

Quantitative Models and Construction Methods 1. **Model Name**: Industry Crowding Monitoring Model **Model Construction Idea**: This model is designed to monitor the crowding levels of Shenwan First-Level Industry Indices on a daily basis, identifying industries with high or low crowding levels and tracking changes in crowding over time[3] **Model Construction Process**: The model calculates crowding levels for each industry index based on specific metrics (not detailed in the report) and ranks them accordingly. It provides daily updates on crowding levels and highlights industries with significant changes in crowding[3] **Model Evaluation**: The model effectively identifies industries with high or low crowding levels and tracks significant changes, providing actionable insights for investors[3] 2. **Model Name**: Premium Rate Z-Score Model **Model Construction Idea**: This model identifies potential arbitrage opportunities in ETF products by calculating the Z-score of premium rates for ETFs over a rolling window[4] **Model Construction Process**: - The premium rate for each ETF is calculated as the difference between the ETF's market price and its net asset value (NAV), divided by the NAV - The Z-score is then computed as: $ Z = \frac{(Premium\ Rate - Mean\ Premium\ Rate)}{Standard\ Deviation\ of\ Premium\ Rate} $ where the mean and standard deviation are calculated over a rolling window[4] **Model Evaluation**: The model is useful for identifying ETFs with significant deviations from their historical premium rates, which may indicate potential arbitrage opportunities. However, it also highlights the need to be cautious of potential price corrections[4] --- Model Backtesting Results 1. **Industry Crowding Monitoring Model**: No specific numerical backtesting results were provided in the report[3] 2. **Premium Rate Z-Score Model**: No specific numerical backtesting results were provided in the report[4] --- Quantitative Factors and Construction Methods No specific quantitative factors were detailed in the report --- Factor Backtesting Results No specific quantitative factor backtesting results were provided in the report

Quantitative Models and Construction Methods 1. **Model Name**: Industry Crowding Monitoring Model **Model Construction Idea**: This model is designed to monitor the crowding levels of industries on a daily basis, specifically for the CSI Level-1 Industry Index. It identifies industries with high or low crowding levels to provide actionable insights. [3] **Model Construction Process**: The model calculates the crowding levels of industries by analyzing daily fund flows and other relevant metrics. It ranks industries based on their crowding levels, highlighting those with significant changes or extreme values. For example, the previous trading day showed high crowding levels in power equipment, basic chemicals, and environmental protection, while industries like computers, automobiles, and non-bank financials had lower crowding levels. [3] **Model Evaluation**: The model effectively identifies industries with significant crowding changes, providing valuable insights for fund allocation and risk management. [3] 2. **Model Name**: Premium Rate Z-Score Model **Model Construction Idea**: This model is used to screen ETF products for potential arbitrage opportunities by calculating the Z-score of their premium rates. [4] **Model Construction Process**: The model employs a rolling calculation of the Z-score for the premium rates of ETF products. A high Z-score indicates a potential arbitrage opportunity, while a low Z-score may signal a risk of price correction. [4] **Model Evaluation**: The model provides a systematic approach to identify ETFs with potential arbitrage opportunities, but it also warns of potential price correction risks. [4] --- Model Backtesting Results 1. **Industry Crowding Monitoring Model**: No specific numerical backtesting results were provided in the report. [3] 2. **Premium Rate Z-Score Model**: No specific numerical backtesting results were provided in the report. [4] --- Quantitative Factors and Construction Methods No specific quantitative factors were explicitly mentioned or constructed in the report. --- Factor Backtesting Results No specific backtesting results for factors were provided in the report.

- The report constructs an industry congestion monitoring model to monitor the congestion of Shenwan first-level industry indices on a daily basis[3] - The ETF product screening signal model is built based on the premium rate Z-score model, which provides potential arbitrage opportunities through rolling calculations[4] - The industry congestion monitoring model indicates that the congestion levels of the power equipment, basic chemicals, and environmental protection industries were high, while the congestion levels of the computer, automotive, and machinery equipment industries were low[3] - The premium rate Z-score model is used to identify potential arbitrage opportunities in ETF products, while also warning of potential pullback risks[4]

Group 1 - The final share confirmation ratio for Huaxia Brazil ETF is 11.54% [1] - The final share confirmation ratio for E Fund Brazil ETF is 11.82% [2] - The timing for selling these ETFs is crucial, with reference to the performance of the Saudi ETF on its listing day [2][3] Group 2 - Both Huaxia and E Fund Brazil ETFs track the Brazil IBOVESPA Index, which covers major listed companies in Brazil [10] - The IBOVESPA Index accounts for approximately 80% of the trading volume in the Brazilian stock market [12] - The index is a total return index, meaning dividends are included in its performance [14] Group 3 - The annualized return of the IBOVESPA Index is approximately 8.9%, which is lower than the annualized returns of the S&P 500 and CSI 800 when dividends are considered [17] - The Brazilian real has depreciated significantly against the US dollar, impacting investment returns [19] - From January 2016 to May 2021, the IBOVESPA Index increased by 191.16%, but currency depreciation reduced the effective returns [20] Group 4 - The largest company by market capitalization in the IBOVESPA Index is Vale S.A., a global leader in iron ore production [27][29] - The index is diversified across various sectors, including materials, utilities, finance, and energy, with no single sector dominating [31][32] Group 5 - The Brazilian stock market is currently experiencing high volatility, with significant trading activity around the 4000-point mark [34] - There is a noticeable lack of clear market leadership among sectors, leading to a cautious investment sentiment [37] - Recent trends show that while some sectors have performed well, investor confidence remains fragile, with many choosing to sell during price increases [46]

- The industry crowding monitoring model was constructed to monitor the crowding level of Shenwan primary industry indices daily. The model identified that the crowding levels of power equipment and environmental protection were high, while non-bank and home appliances had lower crowding levels. Additionally, significant changes in crowding levels were observed in banking and agriculture, forestry, animal husbandry, and fishery industries[3] - The Z-score model for premium rate was developed to screen ETF products with potential arbitrage opportunities. The model uses rolling calculations to identify signals and warns of potential risks of price corrections for the identified ETFs[4] - Daily net capital inflow for broad-based ETFs was 24.71 billion yuan, with top inflows observed in the following ETFs: China Securities A500ETF (+7.83 billion yuan), A500ETF (+5.14 billion yuan), and SSE 50ETF (+2.61 billion yuan). Conversely, the top outflows were seen in CSI 300ETF (-7.13 billion yuan), CSI 300ETF E Fund (-2.21 billion yuan), and ChiNext ETF (-0.43 billion yuan)[5] - Daily net capital inflow for industry-themed ETFs was 41.72 billion yuan, with top inflows observed in the following ETFs: Securities ETF (+7.78 billion yuan), Banking ETF (+6.03 billion yuan), and Power Grid Equipment ETF (+3.98 billion yuan). Conversely, the top outflows were seen in Wine ETF (-2.71 billion yuan), Robotics ETF E Fund (-2.23 billion yuan), and Battery ETF (-1.26 billion yuan)[5] - Daily net capital inflow for style-strategy ETFs was 7.92 billion yuan, with top inflows observed in the following ETFs: Dividend ETF E Fund (+3.44 billion yuan), Dividend Low Volatility ETF (+1.75 billion yuan), and Dividend Low Volatility ETF (+1.01 billion yuan). Conversely, the top outflows were seen in Dividend ETF (-0.36 billion yuan), Dividend State-Owned Enterprise ETF (-0.27 billion yuan), and Dividend Low Volatility 50ETF (-0.20 billion yuan)[5] - Daily net capital inflow for cross-border ETFs was 67.28 billion yuan, with top inflows observed in the following ETFs: Hang Seng Technology ETF (+12.00 billion yuan), Hang Seng Technology Index ETF (+9.20 billion yuan), and Hong Kong Non-Bank ETF (+6.53 billion yuan). Conversely, the top outflows were seen in Saudi ETF (-0.19 billion yuan), H-Share ETF (-0.18 billion yuan), and Hong Kong Stock Connect 100ETF (-0.08 billion yuan)[5]