以下文章来源于兴证全球基金 ，作者与您相伴的 兴证全球基金 . 从我们的经验来看，主动基金经理很多都是行业研究员出身，对行业的了解比我们深，他们做行业轮 动的成功概率都不高，我们作为 FOF 或者投顾的管理人，难度就更高了。相对来说，我们会在我们 的研究比市场平均认知水平更高一点的领域，去争取创造超额，要在一个我们不擅长并且又很卷的领 域去创造超额，胜率比较低。推而广之，我觉得可能大部分普通投资者也不太适合去做行业轮动，因 为我们面临的困境，大多数投资者也是类似的。 投资复盘：你在哪里赚到了钱，在哪里亏了钱？ 文子： 我最近做了一些复盘，发现我真正赚到钱的投资其实只有两笔。第一笔是FOF ，因为我 们公司的 FOF 管理理念是 "追求赚到比平均好一点点的钱"，这个理念特别打动我，于是从 2020 年开始，我每周定投 FOF ，金额不大，但坚持几年下来，已经默默积累成了一笔可观的资产，而且 在2021 年市场波动很大的时候，因为 FOF 相对波动比较小，我也没有想过去减仓。虽然收益率不 算惊人，但它反而成为了我投资收益里贡献很大的一部分。 第二笔是今年 4 月买入的一笔投资，当时市场主流指数跌幅都很大，我觉得这 ...

策略说明: · 恒泰证券研究所基于策略报告《行业轮动下的策略组合报告:基于行业风格延续和切 换视角下的定量分析》(20241007)和《股票型ETF市场概览与配置方法研究:以基于 行业轮动策略的ETF组合为例》（20241013）,构建基于行业和主题ETF的策略组合。 恒泰证券 HENGTAI SECURITIES 研究所 行业轮动ETF策略周报 (20250818-20250824) 证券研究报告·策略周报 证券分析师:张一 S0670524030001 010-83270999-97050 zhangy i@cnht. com. cn 2025年8月25日 证券分析师:李杜 S0670524040001 021-50800937 l idu@cnht. com. cn 图表3: 行业轮动ETF策略建仓以来累计收益率(20241014-20250822) 策略更新: | 基金代码 | ETF名称 | ETF市值 | 持有情况 | 重仓申万 行业 | | 周度择时信号 日度择时信号 | | --- | --- | --- | --- | --- | --- | --- | | | | (亿元) | | 及权重 ...

证券研究报告：金融工程报告 研究所 分析师:肖承志 SAC 登记编号:S1340524090001 Email:xiaochengzhi@cnpsec.com 研究助理:李子凯 SAC 登记编号:S1340124100014 Email:lizikai@cnpsec.com 近期研究报告 《OpenAI 发布 GPT-5，Claude Opus 4.1 上线——AI 动态汇总 20250811》 - 2025.08.12 《融资余额新高，创新药光通信调整， 指数预期仍将震荡上行挑战前高—— 行业轮动周报 20250810》 - 2025.8.11 《ETF 资金偏谨慎流入消费红利防守， 银行提前调整使指数回调空间可控— — 行 业 轮 动 周 报 20250803 》 - 2025.08.04 《ETF 资金持续净流出医药，雅下水电 站成短线情绪突破口——行业轮动周 报 20250727》 – 2025.07.28 《ETF 资金净流入红利流出高位医药， 指数与大金融回调有明显托底——行 业轮动周报 20250720》 – 2025.07.21 《大金融表现居前助指数突破，GRU 行 业轮动调入非银行金融—— ...

1. Report Industry Investment Rating - Not provided in the content 2. Core Viewpoints of the Report - The A-share market showed a continuous upward trend this week, with the Shanghai Composite Index stable above 3,600 points. Although the weekly average daily trading volume decreased by 6.26% compared to last week, the margin trading balance exceeded 2 trillion and continued to rise, indicating that investors' risk appetite remained relatively high in the short term [1][10]. - Most industry sectors' trading volume proportions reached new lows in the past four weeks, suggesting that the market trading focus was concentrating on a small number of sectors. Investors should pay attention to the congestion risk of industry sectors and focus on capital flows in the market with rapid rotation of industry themes [10]. - In terms of market style, small-cap stocks had significant excess returns. The cyclical style led the gains among the five major CITIC style indices, while the consumer style had the smallest increase [12]. - It is recommended to focus on three main investment lines: the domestic computing power industry chain, the AI application end, and the consumption recovery sector. These sectors have relatively reasonable valuations and strong potential for supplementary growth under the background of loose liquidity [13]. 3. Summary According to Relevant Catalogs 3.1 This Week's Market Review 3.1.1 Industry Index - This week, sectors such as non-ferrous metals, machinery, and national defense and military industry led the gains. The pharmaceutical sector, which had performed well last week, corrected significantly, while the coal and non-ferrous metals sectors, which had large declines last week, rebounded sharply [10]. - The trading volume proportions of most industry sectors reached new lows in the past four weeks, and the trading activity of the comprehensive finance and non-bank finance sectors decreased significantly [10]. 3.1.2 Market Style - All five major CITIC style indices rose this week, with the cyclical style leading the gains at 3.49%. The growth style rose 1.87%, and its trading volume proportion reached a four-week high. The consumer style had the smallest increase at 0.77%, and its trading volume proportion decreased slightly [12]. - Small-cap stocks had significant excess returns. The CSI 1000 and CSI 2000 rose 2.51% and 3.54% respectively, and their trading volume proportions reached four-week highs [12]. 3.2 Active Equity Funds 3.2.1 Funds with Excellent Performance This Week in Different Theme Tracks - The report selected single-track and double-track funds based on six sectors: TMT, finance and real estate, consumption, medicine, manufacturing, and cyclical sectors, and listed the top five funds in each sector [17][18]. 3.2.2 Funds with Excellent Performance in Different Strategy Categories - The report classified funds into different types such as deep undervaluation, high growth, high quality, quality growth, quality undervaluation, GARP, and balanced cost-effectiveness, and listed the top-ranked funds in each type [19][20] 3.3 Index Enhanced Funds 3.3.1 This Week's Excess Return Distribution of Index Enhanced Funds - The average and median excess returns of CSI 300 index enhanced funds were 0.22% and 0.20% respectively; those of CSI 500 index enhanced funds were 0.05% and 0.07% respectively; those of CSI 1000 index enhanced funds were -0.15% and -0.14% respectively; those of CSI 2000 index enhanced funds were -0.09% and 0.04% respectively; those of CSI A500 index enhanced funds were 0.24% and 0.26% respectively; those of ChiNext index enhanced funds were 0.45% and 0.39% respectively; and those of STAR Market and ChiNext 50 index enhanced funds were 0.18% and 0.21% respectively [23][24]. - The average and median absolute returns of neutral hedge funds were 0.29% and 0.27% respectively; those of quantitative long funds were 1.75% and 1.83% respectively [24]. 3.4 This Issue's Bond Fund Selection - The report comprehensively screened the fund pools of medium- and long-term bond funds and short-term bond funds based on indicators such as fund scale, return-risk indicators, the latest fund scale, Wind fund secondary classification, rolling returns in the past three years, and maximum drawdowns in the past three years [38] 3.5 This Week's High-Frequency Position Detection of Funds - Active equity funds significantly increased their positions in the machinery and computer industries this week and significantly reduced their positions in the electronics, banking, and automobile industries [3]. - From a one-month perspective, the positions in the communication, banking, and non-bank finance industries increased significantly, while the position in the food and beverage industry decreased significantly [3] 3.6 This Week's Weekly Tracking of US Dollar Bond Funds - Not provided in the content

Investment Insights - The market sentiment remains strong with margin trading exceeding 2 trillion, indicating a potential overheating risk [1][4] - The cyclical sector is gaining strength driven by expectations from projects like the Xinjiang-Tibet Railway, while the rotation between growth and cyclical stocks continues [1][4] Equity Market Overview - Small-cap growth stocks significantly outperformed last week, while the volatility of both large and small-cap styles increased [6] - The dispersion of excess returns among industry indices is at a near one-year low, indicating a slowdown in industry rotation [6] - The trading concentration has increased, with the top 100 stocks and top 5 industries seeing a rise in transaction value share [6] Commodity Market Analysis - Precious metals and agricultural products showed increased trend strength, while other sectors remained stable or declined [15][16] - The volatility in black and energy chemical sectors remained stable, with a slight decrease in the volatility of non-ferrous metals [15][16] Options Market Insights - Implied volatility for the Shanghai Stock Exchange 50 and CSI 1000 indices continues to decline, reflecting a market that is both strong and cautious [24] Convertible Bond Market Trends - The premium rate for convertible bonds is approaching a one-year high, while the proportion of bonds with low conversion premiums is increasing, indicating structural growth characteristics [26]

Core Viewpoint - The Chinese wealth management industry is entering a high-growth cycle, with investment advisors playing a crucial role in guiding asset allocation for clients [1] Group 1: Market Trends and Strategies - The current market is characterized by a "structural bull market," with significant differences in returns across industries, making rotation strategies effective for capturing excess returns [2][3] - The second quarter GDP growth rate was 5.2%, indicating a recovery in corporate earnings and providing a solid foundation for market support [3] - Northbound capital saw a net increase of $10.1 billion in the first half of the year, while financing balances increased by 75 billion yuan since April, reflecting a trend of retail savings entering the market through public funds [3] Group 2: Investment Advisor Challenges and Opportunities - Investment advisors face challenges in transforming service models from "sell-side sales" to "buy-side advisory," requiring a restructuring of income sources and balancing short-term gains with long-term asset allocation [4][5] - The integration of technology is essential, as AI can replace basic analysis tasks, but advisors must enhance their skills in human-machine collaboration to meet clients' emotional needs [4][5] - The demand for cross-disciplinary knowledge is increasing, particularly in areas like retirement, taxation, and cross-border assets, highlighting the need for composite talent in the advisory field [4] Group 3: Future Development of Investment Advisory Services - The core path for enhancing service capabilities involves shifting to a client-centric approach, focusing on account-level returns and satisfaction, and building deep trust with clients [5] - The future of advisory services will rely on "human-machine collaboration," where AI handles standardized processes, allowing advisors to focus on emotional support and client relationships [5] - The goal is to enhance both the financial and emotional value of client accounts, addressing the issue of market gains not translating into client profits, and moving towards a new stage of inclusive finance [5]

- Model Name: Diffusion Index Model; Model Construction Idea: The model is based on the principle of price momentum; Model Construction Process: The model tracks the weekly and monthly changes in the diffusion index of various industries, ranking them accordingly. The formula used is $ \text{Diffusion Index} = \frac{\text{Number of Upward Trends}}{\text{Total Number of Trends}} $; Model Evaluation: The model has shown varying performance over the years, with significant returns in some periods and notable drawdowns in others[27][28][31] - Model Name: GRU Factor Model; Model Construction Idea: The model utilizes GRU deep learning networks to analyze minute-level volume and price data; Model Construction Process: The model ranks industries based on GRU factors, which are derived from deep learning algorithms processing historical trading data. The formula used is $ \text{GRU Factor} = \text{GRU Network Output} $; Model Evaluation: The model performs well in short cycles but has mixed results in longer cycles[33][34][36] - Diffusion Index Model, Average Weekly Return: 2.06%, Excess Return: -0.00%, August Excess Return: -0.45%, Year-to-Date Excess Return: -0.41%[31] - GRU Factor Model, Average Weekly Return: 2.71%, Excess Return: 0.65%, August Excess Return: 0.32%, Year-to-Date Excess Return: -4.35%[36] - Factor Name: GRU Industry Factor; Factor Construction Idea: The factor is derived from GRU deep learning networks analyzing minute-level trading data; Factor Construction Process: The factor ranks industries based on GRU network outputs, which are calculated from historical volume and price data. The formula used is $ \text{GRU Factor} = \text{GRU Network Output} $; Factor Evaluation: The factor has shown significant changes in rankings, indicating its sensitivity to market conditions[6][14][34] - GRU Industry Factor, Steel: 2.82, Building Materials: 1.72, Transportation: 1.3, Oil & Petrochemicals: 0.27, Construction: -0.46, Comprehensive: -1.87[6][14][34]

Quantitative Models and Construction Methods 1. Model Name: Multivariate CNN-LSTM - **Model Construction Idea**: The model leverages the advantages of CNN and LSTM in different scenarios to predict multiple parallel financial time series by considering the correlation between them[12][14]. - **Detailed Construction Process**: - **General Structure**: The model consists of an input layer, a one-dimensional convolutional layer, a pooling layer, an LSTM hidden layer, and a fully connected layer to produce the final prediction results[14]. - **Formula**: $$ {\hat{x}}_{k,t+h}=f_{k}(x_{1,t},\dots,x_{k,t},\dots,x_{1,t-1},\dots,x_{k,t-1},\dots) $$ This formula indicates that each variable depends not only on its past values but also on the past values of other variables[11]. - **Hyperparameters**: - Number of convolution filters: 64 - Convolution kernel size: 2 - Use of padding: Yes - Pooling layer window size: (2,2) - Number of hidden units in the first LSTM layer: 128 - Number of hidden units in the second LSTM layer: 128 - Activation method between LSTM layers: ReLU - Time series look-back window: 10 - Number of training epochs: 100[20] - **Evaluation Metric**: Root Mean Square Error (RMSE) $$ RMSE={\sqrt{\frac{1}{n}\sum_{i}({\hat{y_{i}}}-y_{i}\,)^{2}}} $$ where \( y_i \) represents the standardized index price, and \( \hat{y_i} \) represents the CNN-LSTM prediction value[21]. - **Model Evaluation**: The model achieved good tracking and high accuracy in predicting multiple parallel financial time series, similar to the performance in predicting stock indices in the Asia-Pacific market[14][17]. 2. Model Name: Grouped Multivariate CNN-LSTM - **Model Construction Idea**: To improve prediction accuracy, the industry indices are grouped based on investment attributes, and a separate prediction model is constructed for each group[26][27]. - **Detailed Construction Process**: - **Grouping**: The industry indices are divided into six groups: Consumer and Medicine, Upstream Resources and Materials, High-end Manufacturing, Real Estate and Infrastructure, Big Tech, and Big Finance[27]. - **Model Structure**: Each group of industry indices is predicted using a separate CNN-LSTM model, as shown in the general structure diagram[28]. - **Evaluation Metric**: The prediction accuracy is evaluated using RMSE, similar to the original model[33]. - **Model Evaluation**: Grouping and training different CNN-LSTM sub-models for each industry group improved the prediction accuracy, especially for industries with previously low prediction accuracy[30][32]. Model Backtesting Results 1. Multivariate CNN-LSTM Model - **Prediction Error (Training Phase)**: 1.52% to 3.18%[23] - **Prediction Error (Testing Phase)**: 1.56% to 3.30%[23][25] 2. Grouped Multivariate CNN-LSTM Model - **Prediction Error (Training Phase)**: 1.49% to 2.60%[33] - **Prediction Error (Testing Phase)**: 1.61% to 2.82%[33] Quantitative Factors and Construction Methods 1. Factor Name: Weekly Industry Rotation Signal - **Factor Construction Idea**: Use the predicted values from the multivariate CNN-LSTM model to estimate the future weekly returns of industry indices and select the top five industries with the highest expected returns for equal-weight allocation[3]. - **Detailed Construction Process**: - **Prediction**: Predict the future weekly returns of industry indices using the multivariate CNN-LSTM model[34]. - **Allocation**: Every five trading days, select the top five industries with the highest expected returns for equal-weight allocation[35]. - **Training**: Retrain the model at the beginning of each quarter using an extended window of historical data from March 2014 to the training point[35]. - **Factor Evaluation**: The annualized return of the industry rotation portfolio reached 15.6%, with an annualized excess return of approximately 11.6%, and the risk-return characteristics significantly improved compared to the benchmark[3][35]. Factor Backtesting Results 1. Weekly Industry Rotation Signal - **Annualized Return**: 15.6%[38] - **Annualized Volatility**: 25.6%[38] - **Maximum Drawdown**: -27.1%[38] - **Sharpe Ratio**: 0.7[38] - **Longest Drawdown Recovery Time**: 248 days[38]

Quantitative Models and Construction Methods Genetic Programming Industry Rotation Model - **Model Name**: Genetic Programming Industry Rotation Model - **Model Construction Idea**: Directly extract factors from industry index data such as volume, price, and valuation, and update the factor library at the end of each quarter[30] - **Model Construction Process**: The model adopts weekly frequency rebalancing, selecting the top five industries with the highest composite multi-factor scores for equal-weight allocation every weekend[30] - **Model Evaluation**: The model has achieved an absolute return of 28.79% this year, outperforming the industry equal-weight benchmark by 17.68 percentage points[30] - **Model Testing Results**: - Annualized Return: 31.39% - Annualized Volatility: 18.12% - Sharpe Ratio: 1.73 - Maximum Drawdown: -19.63% - Calmar Ratio: 1.60 - Last Week Performance: 3.15% - Year-to-Date (YTD): 28.79%[32] Absolute Return ETF Simulation Portfolio - **Model Name**: Absolute Return ETF Simulation Portfolio - **Model Construction Idea**: The asset allocation weights are mainly calculated based on the recent trends of various assets, with stronger trend assets assigned higher weights. The internal equity asset allocation weights directly adopt the monthly views of the monthly frequency industry rotation model[34] - **Model Construction Process**: The model's latest holdings include dividend style ETFs and ETFs related to pharmaceuticals, non-ferrous metals, media, steel, and energy chemicals[36] - **Model Evaluation**: The model has risen by 0.34% last week and has accumulated a 5.69% return this year[34] - **Model Testing Results**: - Annualized Return: 6.52% - Annualized Volatility: 3.81% - Maximum Drawdown: 4.65% - Sharpe Ratio: 1.71 - Calmar Ratio: 1.40 - Year-to-Date (YTD): 5.69% - Last Week Performance: 0.34%[39] Global Asset Allocation Simulation Portfolio - **Model Name**: Global Asset Allocation Simulation Portfolio - **Model Construction Idea**: Predict future returns of global major assets using a cycle three-factor pricing model, and construct the portfolio using a "momentum selects assets, cycle adjusts weights" risk budgeting framework[40] - **Model Construction Process**: The strategy currently overweights bonds and foreign exchange, with higher risk budgets assigned to assets such as Chinese bonds and US bonds[40] - **Model Evaluation**: The strategy has achieved an annualized return of 7.22% in the backtest period, with a Sharpe ratio of 1.50[40] - **Model Testing Results**: - Annualized Return: 7.22% - Annualized Volatility: 4.82% - Maximum Drawdown: -6.44% - Sharpe Ratio: 1.50 - Calmar Ratio: 1.12 - Year-to-Date (YTD): -3.04% - Last Week Performance: 0.61%[41] Quantitative Factors and Construction Methods Sentiment Indicators - **Factor Name**: Sentiment Indicators - **Factor Construction Idea**: Construct sentiment indicators from the perspectives of the put-call ratio, implied volatility, and basis in the options and futures markets[2] - **Factor Construction Process**: - **Put-Call Ratio**: Observe the ratio of the trading volume of call options to put options in the 50ETF and 500ETF options markets[17] - **Implied Volatility**: Construct the implied volatility ratio series of call and put options[20] - **Basis**: Construct the annualized basis rate weighted by the open interest for the four major stock index futures products[26] - **Factor Evaluation**: The sentiment indicators show that the previous overheating sentiment in the A-share market has continued to ease[2] Factor Backtesting Results Sentiment Indicators - **Put-Call Ratio**: The ratio has significantly fallen from the high levels observed on July 23, indicating a more rational market sentiment[17] - **Implied Volatility Ratio**: Despite the stock market rebound last week, the implied volatility ratio of call options to put options has been trending downward, further reflecting rational investor sentiment[20] - **Annualized Basis Rate**: The basis rate has been fluctuating downward, indicating rational sentiment in the futures market[26]

Core Conclusions - The speed of market rotation has significantly accelerated, with the rotation index dropping to 61.95% in 2025, and the duration of hot themes shortening, with most themes lasting less than or equal to 20 days [3][4] - The relationship between rotation speed and funding structure indicates that during accelerated rotation, financing balances are highly synchronized with the index, while during slower rotations, financing responses lag [3][14] - Based on the alpha158 factor, derived strategies were constructed for wind hot rotation, industry rotation, and hot index mapping leading stocks. The index rotation strategy achieved an annualized return of 20.25%, outperforming industry rotation at 16.03% [3][4] Industry Rotation Effective Factors - Quantile factors (QTLU/QTUD) are identified as effective for industry rotation, with support momentum (QTUD) being more effective in bear markets and resistance momentum (QTLU) in bull markets [3][4] - The proportion of positive volatility (SUMN) indicates stronger industry strength, while extreme value factors (RSV/MAX) are sensitive to hot themes [3][4] Hot Index Rotation Optimization - The analysis utilized 68 Wind hot indices, focusing on core factors such as quantile factors (QTLU_20_95) and residual ranking factors (RESI30, RANK20) which have shown high win rates in recent years [4][6] - The adjustment strategy involves T+1 closing for rebalancing to mitigate factor decay, with the top 5 components of hot indices yielding an annualized return of 15.79%, significantly outperforming the CSI 300 [4][6] Strategy Application - For industry rotation holdings in 2025, high-frequency positions include banking, automotive, and non-ferrous metals, with recent additions in coal and basic chemicals [4][6] - Hot index holdings for July 2025 included semiconductor, lithium mining, and energy equipment, while automotive parts and liquor indices were removed [4][6] Market Rotation Dynamics - The analysis indicates that the speed of rotation is influenced by the structure of market participation funds, with rapid rotation correlating with high retail participation and financing balance synchronization [14][18] - In contrast, slower rotation reflects a dominance of institutional funds, leading to a significant lag in financing balances compared to index gains [14][18] Performance of Hot Rotation Strategies - The report suggests that in recent years of rapid hot rotation, short-term trend strategies are more likely to achieve excess returns [21][27] - The effectiveness of the index rotation has been higher than that of industry rotation in the past three years, indicating a shift in alpha generation from broader industry to more granular segments [27][28]