国内流动性延续宽松，欧洲央行如期降息 ooo[Table_ReportInfo] 2025 年 06 月 06 日 证券研究报告•宏观定期报告 宏观周报（6.2-6.6） 摘要 [Table_Summary] 一周大事记 国内：央行发布买断式逆回购，AI 赋能新型工业化。6月 3日，经文化和旅游 部数据中心测算，端午节全国国内出游 1.19亿人次，同比增长 5.7%，今年端 午出行整体或受部分地区降雨天气影响，但新消费成为增长亮点；同日，中国 5 月财新制造业 PMI 录得 48.3，较 4月下降 2.1个百分点，5日，5月财新服 务业 PMI 录得 51.1，较 4 月上升 0.4 个百分点，随着未来更多财政政策相继 落地，国内内需有望提振，制造业景气度也有望逐步回升；同日，工信部党组 书记、部长李乐成主持召开会议，系统部署人工智能与新型工业化融合发展路 径，有望改变我国工业发展范式，为新型工业化带来新的活力；4 日，中证报 头版文章报道，业内人士认为，货币政策在"适度宽松"的方向上还有发力空 间，6日，中国人民银行开展期限 3个月的一万亿元买断式逆回购操作，保持 流动性合理充裕；同日，香港财经事务及库务局 ...

Group 1 - The report highlights that China's economy is entering a high elasticity period, with structural growth opportunities in high-end manufacturing, urban renewal, and service consumption, supported by domestic investment policies [6][8][29] - Manufacturing investment is expected to maintain an annual growth rate of over 8% due to domestic demand expansion policies, while infrastructure investment growth may exceed 9% [6][8][20] - The report suggests that the real estate market is shifting towards "quality over quantity," which will help stabilize prices amid reduced supply [6][29] Group 2 - The report identifies short, medium, and long-term industry selection strategies, emphasizing sectors such as intelligent manufacturing, beverage and dairy, and chemical pharmaceuticals for short-term focus [6][7] - In the medium term, the report notes that trade conflicts have limited impact on China's industrial development, with technology, services, and education sectors showing strong growth [6][7] - Long-term investment opportunities are seen in high-end manufacturing, pharmaceutical biotechnology, and new discretionary consumption as interest rates decline [6][7] Group 3 - The report indicates that the broad infrastructure investment growth rate is expected to decline slightly in the third quarter, with a projected annual growth rate of over 9% [20][22] - Specific sectors such as electricity, heat, gas, and water supply are experiencing significant investment growth, with fixed asset investment in these areas increasing by 25.5% [20][22] - The report also notes that the approval of central and local projects is accelerating, which will support infrastructure investment in the second half of the year [20][22] Group 4 - The report emphasizes that consumer confidence remains weak, with retail sales growth driven primarily by "trade-in" policies, which are expected to support a modest recovery in consumption [41][43] - The service retail sector is outpacing goods retail, with service retail sales growing by 5.1% compared to 4.7% for goods retail [48] - The report highlights the potential of the "谷子经济" (Guzi Economy), which focuses on emotional value through cultural IP, as a new consumption driver [48]

Domestic Insights - Industrial enterprises' profits increased by 1.4% year-on-year in the first four months, with a slight acceleration of 0.6 percentage points compared to Q1[1] - State-owned enterprises' profits decreased by 1.7% year-on-year, indicating pressure from volume growth but price decline[14] - The manufacturing sector showed strong performance, particularly in equipment manufacturing, which saw a profit increase of 11.2%[13] International Developments - The U.S. tariff policy experienced reversals, with a federal court temporarily halting the implementation of tariffs announced by the Trump administration[16] - The U.S. Treasury Department announced a reduction in short-term debt issuance, reflecting ongoing political negotiations over the debt ceiling[21] - Japan's central bank signaled a cautious approach to monetary policy, with low short-term interest rate hike probabilities amid rising inflation pressures[19] Market Trends - Brent crude oil prices fell by 1.03% week-on-week and decreased by 21.69% year-on-year, indicating a significant decline in energy prices[26] - The price of rebar dropped by 1.81% week-on-week and 14.22% year-on-year, reflecting ongoing challenges in the construction materials market[32] - The average collection period for accounts receivable in state-owned enterprises extended to 70.3 days, indicating increased financial pressure[15]

Quantitative Models and Construction Methods GAN_GRU Model - **Model Name**: GAN_GRU - **Model Construction Idea**: The GAN_GRU model utilizes Generative Adversarial Networks (GAN) for processing volume-price time series features and then employs the GRU model for time series feature encoding to derive the stock selection factor[2][9]. - **Model Construction Process**: 1. **GRU Model**: - **Volume-Price Features**: Includes 18 volume-price features such as closing price, opening price, trading volume, turnover rate, etc.[10][13][15]. - **Training Data and Input Features**: Uses past 400 days of 18 volume-price features for all stocks, sampling every 5 trading days. The feature sampling shape is 40*18, predicting cumulative returns for the next 20 trading days[14]. - **Training and Validation Set Ratio**: 80% training set, 20% validation set[14]. - **Data Processing**: Extreme value removal and standardization in time series for each feature, cross-sectional standardization at the stock level[14]. - **Model Training Method**: Semi-annual rolling training, training points are June 30 and December 31 each year[14]. - **Stock Screening Method**: Excludes ST and stocks listed for less than half a year[14]. - **Training Sample Screening Method**: Excludes samples with empty labels[14]. - **Hyperparameters**: batch_size is the number of stocks in the cross-section, optimizer Adam, learning rate 1e-4, loss function IC, early stopping rounds 10, maximum training rounds 50[14]. - **Model Structure**: Two GRU layers (GRU(128, 128)) followed by MLP layers (256, 64, 64), with the final output pRet as the stock selection factor[18]. 2. **GAN Model**: - **Generator**: Learns the real distribution of data and generates samples that look like real data. The loss function is: $$L_{G}\,=\,-\mathbb{E}_{z\sim P_{z}(z)}[\log(D(G(z)))]$$ where \( z \) is random noise, \( G(z) \) is the data generated by the generator, and \( D(G(z)) \) is the probability output by the discriminator[20][21][22]. - **Discriminator**: Distinguishes real data from generated data. The loss function is: $$L_{D}=-\mathbb{E}_{x\sim P_{d a t a}(x)}[\log\!D(x)]-\mathbb{E}_{z\sim P_{z}(z)}[\log(1-D(G(z)))]$$ where \( x \) is real data, \( D(x) \) is the probability output by the discriminator for real data, and \( D(G(z)) \) is the probability output by the discriminator for generated data[23][24][25]. - **Training Process**: Alternating training of generator and discriminator until convergence[25][26]. - **Model Structure**: Uses LSTM as the generator to retain the time series nature of the input features and CNN as the discriminator to match the two-dimensional volume-price time series features[29][30][31]. - **Feature Generation**: The generator part of the trained GAN model is used for feature generation, inputting original volume-price time series features and outputting processed features[33][34]. Model Evaluation - **Evaluation**: The GAN_GRU model effectively combines GAN and GRU to process and encode volume-price time series features, showing promising results in stock selection[2][9]. Model Backtest Results - **GAN_GRU Model**: - **IC Mean**: 11.57%[37][38] - **ICIR**: 0.89[38] - **Turnover Rate**: 0.83[38] - **Recent IC**: -0.28%[37][38] - **One-Year IC Mean**: 11.54%[37][38] - **Annualized Return**: 36.60%[38] - **Annualized Volatility**: 24.02%[38] - **IR**: 1.66[38] - **Maximum Drawdown**: 27.29%[38] - **Annualized Excess Return**: 24.89%[38] Quantitative Factors and Construction Methods GAN_GRU Factor - **Factor Name**: GAN_GRU Factor - **Factor Construction Idea**: The GAN_GRU factor is derived from the GAN_GRU model, which processes volume-price time series features using GAN and encodes them using GRU[2][9]. - **Factor Construction Process**: Same as the GAN_GRU model construction process described above[2][9][10][14][18][19][20][21][22][23][24][25][26][29][30][31][33][34]. Factor Backtest Results - **GAN_GRU Factor**: - **IC Mean**: 11.57%[37][38] - **ICIR**: 0.89[38] - **Turnover Rate**: 0.83[38] - **Recent IC**: -0.28%[37][38] - **One-Year IC Mean**: 11.54%[37][38] - **Annualized Return**: 36.60%[38] - **Annualized Volatility**: 24.02%[38] - **IR**: 1.66[38] - **Maximum Drawdown**: 27.29%[38] - **Annualized Excess Return**: 24.89%[38]

Quantitative Models and Construction Methods 1. Model Name: GAN_GRU - **Model Construction Idea**: The GAN_GRU model combines Generative Adversarial Networks (GAN) for processing volume-price time-series features and Gated Recurrent Unit (GRU) for encoding time-series features to create a stock selection factor[2][9] - **Model Construction Process**: 1. **GAN Component**: - **Generator (G)**: Generates realistic data from random noise (e.g., Gaussian distribution). The generator's loss function is: $$ L_{G} = -\mathbb{E}_{z\sim P_{z}(z)}[\log(D(G(z)))] $$ where \( z \) represents random noise, \( G(z) \) is the generated data, and \( D(G(z)) \) is the discriminator's output probability that the generated data is real[20][21] - **Discriminator (D)**: Distinguishes real data from generated data. The discriminator's loss function is: $$ L_{D} = -\mathbb{E}_{x\sim P_{data}(x)}[\log D(x)] - \mathbb{E}_{z\sim P_{z}(z)}[\log(1-D(G(z)))] $$ where \( x \) is real data, \( D(x) \) is the discriminator's output probability for real data, and \( D(G(z)) \) is the output probability for generated data[23][25] - Training alternates between updating the generator and discriminator to improve feature generation and discrimination capabilities[26] 2. **GRU Component**: - Two GRU layers (GRU(128, 128)) are used to encode time-series features, followed by a Multi-Layer Perceptron (MLP) with layers (256, 64, 64) to output predicted returns (\( pRet \))[18] 3. **Feature Input and Processing**: - Input features include 18 volume-price characteristics (e.g., closing price, turnover rate) sampled over the past 40 days to predict cumulative returns for the next 20 days[10][14] - Data preprocessing includes outlier removal, standardization, and cross-sectional normalization[14] - Training is conducted semi-annually with rolling updates[14] 4. **GAN_GRU Integration**: - The GAN generator processes raw volume-price time-series features (Input_Shape=(40,18)) and outputs features encoded by LSTM. These features are then passed to the GRU model for further processing[33][34] - **Model Evaluation**: The GAN_GRU model effectively captures time-series and cross-sectional features, demonstrating strong predictive power for stock selection[2][9] --- Model Backtesting Results 1. GAN_GRU Model - **IC Mean**: 11.57%[37][38] - **ICIR**: 0.89[38] - **Turnover Rate**: 0.83[38] - **Recent IC**: -0.28%[37][38] - **1-Year IC Mean**: 11.54%[37][38] - **Annualized Return**: 36.60%[38] - **Annualized Volatility**: 24.02%[38] - **IR**: 1.66[38] - **Maximum Drawdown**: 27.29%[38] - **Annualized Excess Return**: 24.89%[38] --- Quantitative Factors and Construction Methods 1. Factor Name: GAN_GRU Factor - **Factor Construction Idea**: Derived from the GAN_GRU model, this factor leverages GAN for feature generation and GRU for time-series encoding to predict stock returns[2][9] - **Factor Construction Process**: 1. **Input Features**: 18 volume-price characteristics (e.g., closing price, turnover rate) sampled over the past 40 days[10][14] 2. **GAN Feature Generation**: - LSTM-based generator processes raw time-series features to retain temporal properties[29][33] - CNN-based discriminator identifies realistic features from generated ones[29] 3. **GRU Encoding**: Encodes GAN-generated features using two GRU layers and an MLP to output predicted returns[18][33] 4. **Factor Normalization**: Industry and market capitalization neutralization, followed by standardization[18] - **Factor Evaluation**: The GAN_GRU factor demonstrates robust performance across various industries and time periods, indicating its effectiveness in stock selection[2][9] --- Factor Backtesting Results 1. GAN_GRU Factor - **IC Mean**: 11.57%[37][38] - **ICIR**: 0.89[38] - **Turnover Rate**: 0.83[38] - **Recent IC**: -0.28%[37][38] - **1-Year IC Mean**: 11.54%[37][38] - **Annualized Return**: 36.60%[38] - **Annualized Volatility**: 24.02%[38] - **IR**: 1.66[38] - **Maximum Drawdown**: 27.29%[38] - **Annualized Excess Return**: 24.89%[38] Industry-Specific Performance - **Top 5 Industries by Recent IC (May 2025)**: - Social Services: 30.15% - Defense & Military: 28.07% - Banking: 25.31% - Computers: 24.86% - Real Estate: 12.07%[39] - **Top 5 Industries by 1-Year IC Mean**: - Construction & Decoration: 18.54% - Utilities: 18.14% - Communication: 17.37% - Non-Banking Finance: 16.76% - Defense & Military: 16.53%[39] - **Top 5 Industries by Recent Excess Return (May 2025)**: - Retail: 8.22% - Defense & Military: 7.15% - Social Services: 4.58% - Construction & Decoration: 3.91% - Electronics: 3.64%[42] - **Top 5 Industries by 1-Year Average Monthly Excess Return**: - Oil & Petrochemicals: 5.60% - Building Materials: 5.29% - Home Appliances: 5.06% - Non-Ferrous Metals: 4.57% - Communication: 4.29%[42]

Fund Manager Profile - Fund manager Qu Shaojie has extensive experience in overseas market investments, managing a total of 1.154 billion CNY across three funds as of Q1 2025[1] - Qu's investment philosophy focuses on long-term holdings of fundamentally strong companies with matching valuations and earnings[1] Fund Performance - The Changcheng Hong Kong Stock Connect Value Selection Multi-Strategy A fund has achieved cumulative returns of 19.39%, 48.59%, and 51.88% over the past three, two, and one years, respectively, ranking in the top 6.8%, 1%, and 1.7% of its peers[2] - Year-to-date return for 2025 is 33.52%, placing it in the top 1.5% of its category[2] Market Resilience - During the market turbulence from October 2022 to June 2023, the fund returned 11.14%, significantly outperforming the peer average of -1.44%[2] - From February 2024 to September 2024, the fund achieved a return of 16.89%, compared to the peer average of 3.19%[2] Portfolio Composition - The fund maintains a high concentration in its top three and five sectors, averaging 73.90% and 85.69% respectively since Qu's tenure began[2] - As of Q1 2025, the fund's stock allocation reached 93.11%, an increase of 4.58% from Q4 2024[2] Stock Selection - Notable heavy positions include Xiaomi Group (9.66%) and Pop Mart (11.70%), reflecting Qu's focus on companies with strong fundamentals and growth potential[4] - The weighted average excess return of the fund's heavy positions is 12.66%, with an excess win rate of 82.61%[4] Risk Management - The fund's maximum drawdown was 20.35% in 2024, with a recovery time of only 51 days[2] - The fund's average turnover rate has decreased to 1.09, indicating a more stable investment approach[2]

Investment Rating - The report assigns a "Buy" rating for the company with a target price of 115.75 CNY over the next six months, based on a current price of 91.45 CNY [1]. Core Insights - The company is a leading domestic cosmetics brand that has surpassed 10 billion CNY in revenue, demonstrating strong brand power and high penetration rates, indicating robust long-term growth potential [6][8]. - The domestic cosmetics market continues to grow, with a projected CAGR of 6.2% from 2022 to 2025, expected to reach 579.1 billion CNY by 2025, highlighting the ongoing trend of domestic brands replacing international ones [6][8]. - The company's makeup brand, 彩棠, has shown impressive growth, with a CAGR of 77.5% from 2021 to 2024, and has consistently ranked among the top ten in Tmall's beauty GMV [6][8]. Summary by Sections 1. Company Overview - The company was established in 2006 and became a publicly traded company in 2017, focusing on the research, production, and sales of cosmetics, including brands like 珀莱雅 and 彩棠 [13][16]. - The company achieved a revenue of 10.78 billion CNY in 2024, marking it as the first domestic beauty brand to surpass the 10 billion CNY threshold [16]. 2. Market Dynamics - The domestic cosmetics market is experiencing a slowdown in growth, with retail sales growth fluctuating significantly, but local brands are gaining market share [38][42]. - The company has shifted its focus to online sales, with online revenue growing from 6.4 billion CNY in 2017 to 102.3 billion CNY in 2024, accounting for over 90% of total revenue [21][56]. 3. Brand Performance - The main brand, 珀莱雅, has seen revenue grow from 2.09 billion CNY in 2018 to 8.58 billion CNY in 2024, while 彩棠 has rapidly increased its revenue from 250 million CNY in 2021 to 1.19 billion CNY in 2024 [24][67]. - The company has successfully implemented a "big product" strategy, enhancing customer loyalty and product lifecycle through continuous upgrades and marketing efforts [74][78]. 4. Financial Projections - The company is expected to maintain a compound annual growth rate (CAGR) of 17.2% in net profit over the next three years, with a projected net profit of 1.55 billion CNY in 2024 [2][6]. - The report estimates a price-to-earnings (PE) ratio of 25 for 2025, supporting the target price of 115.75 CNY [6][8].

Market Performance - The overall performance of the Beijing Stock Exchange (BSE) was weak during the period from May 12 to May 23, 2025, with a closing market value of 780.92 billion yuan on May 23[5] - The BSE 50 Index fell by 3.68% compared to the opening market value on May 12, underperforming the ChiNext by approximately 2.8 percentage points[5] - Among 266 stocks on the BSE, 88 stocks rose, while 170 stocks declined during this period[20] Trading and Valuation Metrics - The total trading volume for the BSE during this period was 353.16 billion yuan, with an average trading amount of 1.33 billion yuan per stock[3] - The turnover rate reached 90.46%, indicating an increase in liquidity compared to the previous period[12] - The median price-to-earnings (PE) ratio for the BSE was 51.76 times, which is at a historical high level[12] New Listings and Upcoming IPOs - Tian Gong Co., Ltd. (920068.BJ) was the only new stock listed during this period, debuting on May 13, 2025, with a first-day increase of 411.9%[28] - The upcoming IPO for Jiao Da Tie Fa (920027.BJ) is set to issue 19.09 million shares at a price of 8.81 yuan per share, with an expected total fundraising of approximately 170 million yuan[30] Sector Highlights - Strong performance was noted in sectors such as shipping, controllable nuclear fusion, and mergers and acquisitions, with stocks like Guo Hang Yuan Yang and Chang Fu Co., Ltd. leading the gains[5] - The top five gainers included Tian Gong Co. (up 422.3%) and Guo Hang Yuan Yang (up 47.2%), while Lin Tai New Materials saw a decline of 32.8%[20][24] Regulatory Updates - The China Securities Regulatory Commission revised the "Major Asset Restructuring Management Measures," which aims to simplify review processes and enhance regulatory inclusiveness, potentially boosting M&A activities on the BSE[35]

Investment Rating - The report does not explicitly provide an investment rating for the industry Core Insights - Historical stock returns are a significant reference for future stock performance, and the report explores effective K-line patterns and their corresponding volume-price states without time and stock constraints [1][16] - The K-line investment framework combines K-line shape recognition with volume-price state recognition to enhance prediction accuracy and investment success rates [2][19] - Effective K-line patterns can assist in constructing investment portfolios and optimizing quantitative investments by providing low-correlated pricing information [5][16] Summary by Sections K-line Pattern Investment Framework - K-line charts represent the trading trajectory formed by the interaction of buyers and sellers, and relying solely on K-line shapes for stock selection is considered one-dimensional [2][19] - The report emphasizes the importance of integrating K-line shapes with volume-price state information to improve predictive accuracy [2][19] K-line Shape Recognition and Volume-Price State Recognition - The report utilizes various K-line characteristics such as Yin & Yang, body ratio, upper shadow ratio, lower shadow ratio, returns, open, high, low, and close prices for shape recognition [3][22] - The most commonly used volume-price states are volume expansion and contraction, along with high and low price levels [3][24] Effective K-line Patterns - Strongly effective positive K-line patterns include the large bullish line and inverted hammer, while effective negative patterns include the hammer and gap-down large bearish line [4] - In multi-K patterns, effective positive shapes include the morning star and three white soldiers, while effective negative shapes include the evening star and three black crows [4] Application of Effective K-line Patterns - Effective K-line patterns can directly aid in portfolio construction and enhance subjective stock selection and position management when combined with macroeconomic and fundamental analysis [5] - K-line patterns can serve as event factors to optimize investment portfolios based on traditional fundamental and volume-price factors [5]

Investment Rating - The report maintains an "Outperform" rating for the automotive industry as of May 25, 2025 [1] Core Insights - Retail sales of passenger vehicles from May 1-18 reached 932,000 units, a year-on-year increase of 12% and a month-on-month increase of 18%. Cumulatively, 7.804 million units have been sold this year, reflecting an 8% year-on-year growth. The Passenger Car Association expects a stable and positive retail trend for May [6][61] - Xiaomi launched its first SUV, the YU7, equipped with a self-developed intelligent driving system that integrates deep learning and neural network algorithms for precise operation in complex road environments. Dongfeng Motor signed a comprehensive strategic cooperation agreement with Huawei, focusing on smart assisted driving and digital upgrades in automotive R&D and production [6][64] - The report highlights investment opportunities in companies involved in technology R&D, market expansion, and those benefiting from policy subsidies and improved infrastructure in the new energy vehicle sector [6][61] Summary by Sections Passenger Vehicles - From May 1-18, retail sales of passenger vehicles were 932,000 units, up 12% year-on-year and 18% month-on-month. Cumulative sales for the year reached 7.804 million units, an 8% increase year-on-year. The Passenger Car Association anticipates continued positive retail trends for May [6][62] - Key companies to watch include BYD (002594), Geely (0175.HK), Xpeng Motors (9868.HK), SAIC Motor (600104), Changan Automobile (000625), GAC Group (601238), and Leap Motor (9863.HK) [6][62] New Energy Vehicles - Retail sales of new energy passenger vehicles from May 1-18 reached 484,000 units, a 32% year-on-year increase and a 15% month-on-month increase, with a retail penetration rate of 52%. Cumulative sales for the year reached 3.808 million units, a 35% increase year-on-year [6][63] - Key companies to focus on include BYD (002594), Geely (0175.HK), Huayu Automotive (600741), Xinquan Co. (603179), Doli Technology (001311), Chuanhuan Technology (300547), and Wuxi Zhenhua (605319) [6][63] Smart Vehicles - Xiaomi's YU7 SUV features a self-developed intelligent driving system that utilizes deep learning and neural networks for enhanced operational precision. Dongfeng Motor's partnership with Huawei aims to advance smart driving and digital upgrades in automotive production [6][64] - Companies with technological advancements in smart driving algorithms, sensors, and smart cockpit systems are recommended for investment, including BYD (002594), Geely (0175.HK), and SAIC Motor (600104) [6][64][65]