- The report highlights the performance of various market indices, noting that the CSI 300 Index performed well among scale indices, while the SSE Composite Index performed well among sector indices, and the CSI 500 Value Index performed well among style indices[2][6] - The market sentiment on 2026-03-02 showed 98 stocks hitting the upper limit and 24 stocks hitting the lower limit, with a sealing rate of 74% and a continuous board rate of 28%[2][13][16] - The financing balance as of 2026-02-27 was 26,519 billion yuan, and the securities lending balance was 173 billion yuan, with the financing balance accounting for 2.5% of the circulating market value and the financing transaction accounting for 9.7% of the market turnover[2][18][22] - The ETF with the highest premium on 2026-02-27 was the CSI 500 ETF by BOCOM, with a premium of 0.86%, while the ETF with the highest discount was the Photovoltaic ETF by Guotai, with a discount of 0.85%[3][23][25] - The median annualized discount rates for the main contracts of the SSE 50, CSI 300, CSI 500, and CSI 1000 stock index futures over the past year were 0.68%, 3.88%, 11.15%, and 13.55%, respectively[3][29][32] - The stocks with the highest net inflows from institutional seats on the Dragon and Tiger List on 2026-03-02 were Minbao Optoelectronics, Xianglu Tungsten Industry, Baiyun Electric, China Ruilin, Cangzhou Dahua, Chuanrun Co., Ltd., Hunan Gold, Aerospace Rainbow, Zhongrun Optics, and Guohang Ocean[4][37][39] - The stocks with the highest net inflows from Northbound Trading on the Dragon and Tiger List on 2026-03-02 were Xiaocheng Technology, China Unmanned Aerial Vehicle, Liande Co., Ltd., Sanfu Co., Ltd., Aerospace Rainbow, Roman Co., Ltd., Yujing Co., Ltd., Potential Hengxin, Tin Industry Co., Ltd., and Baiyun Electric[4][38][41]

Core Conclusions - In the fourth week of February, a total net inflow of 45.3 billion yuan was recorded, reversing the previous week's outflow of 72.8 billion yuan [1][7] - Short-term sentiment indicators are at a mid-high level since 2005, while long-term sentiment indicators are at a mid-low level [1][2] - From an industry perspective, the highest trading volume share in the past week was seen in defense and military, communication, and semiconductor sectors [1][2] Market Trends - The total net inflow of funds amounted to 45.3 billion yuan, with a financing balance increase of 79.5 billion yuan, public fund issuance up by 0.9 billion yuan, ETF net redemption of 18 billion yuan, and an estimated net inflow of 2.1 billion yuan from northbound funds [7] - On the outflow side, IPO financing was 1.7 billion yuan, industrial capital net reduction was 7.8 billion yuan, and transaction fees totaled 9.6 billion yuan [7] Short-term Sentiment Indicators - The recent weekly turnover rate (annualized) was 501%, placing it at the 84th percentile historically [11][14] - The recent weekly financing transaction ratio was 10.04%, currently at the 76th percentile historically [11][14] Long-term Sentiment Indicators - The recent A-share risk premium (the inverse of the overall A-share PE minus the yield of ten-year government bonds) was 2.43%, at the 46th percentile historically [2][13] - The recent dividend yield of the CSI 300 index (excluding financials) compared to the yield of ten-year government bonds was 1.21, at the 6th percentile historically [2][13] Industry Analysis - The top three industries by trading volume share in the past week were defense and military (99%), communication (98%), and semiconductor (98%), while the lowest were food processing (0%), pharmaceutical biology (0%), and banking (1%) [2][13] - The highest financing transaction ratio industries were machinery equipment (92%), electric power equipment (85%), and social services (82%), while the lowest were comprehensive (12%), coal (18%), and oil and petrochemicals (20%) [2][13]

Group 1 - The report indicates that the central market interest rates are expected to remain low in March, but volatility may increase due to multiple factors [5][66] - In February, the interbank repo rates showed minor fluctuations, with R001 and R007 averaging 1.40% and 1.55% respectively, reflecting a change of -1BP and 0BP [11][21] - The liquidity pressure in March is significant, with a total of 16,000 billion yuan in reverse repos, 4,500 billion yuan in MLF, and 35,900 billion yuan in interbank certificates maturing [57][66] Group 2 - The central bank's supportive stance is clear, with an increase in fiscal spending expected at the end of the quarter, which is likely to supplement liquidity [61][66] - Historical data shows that interbank repo rates typically rise in March, with an average increase of 5BP from 2022 to 2025 [62][66] - The report concludes that while the market interest rates are expected to remain low, the pressures from maturing financial instruments may lead to increased volatility during the month [66]

Group 1 - The report introduces a paradigm shift in investment research through the integration of Deep Research technology, enabling autonomous planning capabilities in AI agents for investment research tasks [4][14]. - The system transforms investment research from "assisted search" to "end-to-end autonomous research," allowing for multi-round broad searches and cross-validation of information [4][14]. - The architecture employs a standardized three-layer decoupled structure with asynchronous parallel strategies, enhancing the efficiency of ESG task processing [5][19]. Group 2 - The deployment is cloud-native with zero-code interaction, allowing business personnel to customize research directions simply by adjusting natural language instructions [6][52]. - The production cycle of a single research report is compressed from hours to minutes, ensuring high-quality, standardized reports that can be replicated at scale [7][67]. - The system's design ensures consistency in data depth and formatting style, overcoming the quality fluctuations associated with manual delivery [67]. Group 3 - The report outlines a comprehensive ESG investment research architecture driven by Agentic AI, featuring a three-layer technical structure and dual-engine support [18][19]. - The system's core is designed to enhance research standardization and traceability, starting from precise user input to the delivery of structured JSON data alongside standardized Word reports [25][27]. - The architecture allows for dynamic loading of prompts and employs a layered design to convert unstructured natural language requirements into precise parameterized instructions [22][23]. Group 4 - The report emphasizes the efficiency revolution achieved through the integration of Deep Research capabilities and asynchronous parallel architecture, significantly improving response times to emerging events [67]. - The system is capable of 24/7 operation, providing scalable information services that reduce marginal costs and enhance coverage frequency [67]. - The introduction of this system aims to reshape the human-machine collaboration in research, allowing professionals to focus on deeper insights and strategic decision-making [67][68]. Group 5 - The construction process validates the feasibility of zero-code development, enabling users without technical backgrounds to quickly build and maintain investment research tools [68]. - Future enhancements suggest adopting OpenRouter interfaces for flexible model integration, allowing easy switching between different AI models [69]. - The customization of report styles and research focuses can be achieved through simple adjustments to prompt files, streamlining the iterative process of business logic [70].

Investment Rating - The report maintains an "Outperform" rating for the social services sector, indicating expected performance above the market index by over 10% [4][32]. Core Insights - The Chinese Spring Festival consumer market shows strong recovery momentum, with significant increases in travel and domestic tourism orders, as well as a notable rise in duty-free shopping sales in Hainan [2][21]. - The restaurant sector, particularly Haidilao, reported receiving over 14 million customers during the Spring Festival, showcasing effective marketing strategies to meet diverse consumer needs [22][23]. - The education sector, represented by China Oriental Education, is experiencing a surge in enrollment inquiries, driven by positive employment prospects for skilled workers [26]. Summary by Sections Market Review - The consumer services sector declined by 4.02% during the reporting period, underperforming the market by 5.09 percentage points [13][14]. - Notable stock performances included SanTe Cableway (+9.95%), JinJiang Hotels (+6.24%), and China Oriental Education (+4.84%) [14][17]. Industry and Company Dynamics - The travel sector saw a significant increase in passenger flow, with domestic travel orders reaching historical highs during the Spring Festival [18][19]. - Duty-free sales in Hainan during the Spring Festival reached 27.2 billion CNY, a year-on-year increase of 30.8% [21]. - Haidilao's marketing strategies led to a record number of customers during the holiday, highlighting its strong market presence [23]. Stock Holdings Analysis - Key stocks in the sector, such as Haidilao and China Oriental Education, saw increases in shareholding percentages during the reporting period [30]. Investment Recommendations - The report suggests focusing on companies like China Duty Free Group, Huazhu Group, and Haidilao, among others, for potential investment opportunities [4][32].

- Model Name: Securities Firm Golden Stock Performance Enhancement Portfolio - Model Construction Idea: The model aims to optimize the selection of stocks from the securities firm golden stock pool to outperform the median of active equity funds[38] - Model Construction Process: The model uses the securities firm golden stock pool as the stock selection space and constraint benchmark, employs portfolio optimization to control deviations in individual stocks and styles from the golden stock pool, and uses the industry distribution of all public funds as the industry allocation benchmark[43] - Model Evaluation: The model has shown stable performance historically, consistently outperforming the equity hybrid fund index from 2018 to 2022, ranking in the top 30% of active equity funds each year[12][44] - Model Test Results: - Absolute return for the month (20260202-20260227): 2.32%[42] - Excess return relative to equity hybrid fund index for the month: 1.28%[42] - Absolute return for the year (20260105-20260227): 14.06%[42] - Excess return relative to equity hybrid fund index for the year: 5.62%[42] - Ranking in active equity funds for the year: 19.91% percentile (741/3721)[42] - Factor Name: Total Market Value - Factor Construction Idea: The factor aims to capture the performance of stocks with different market capitalizations[3] - Factor Construction Process: The factor is constructed by grouping stocks based on their total market value and tracking the performance of these groups[27] - Factor Evaluation: The total market value factor has performed well in the recent month and year-to-date[3][27] - Factor Name: SUE (Standardized Unexpected Earnings) - Factor Construction Idea: The factor aims to capture the performance of stocks with unexpected earnings[3] - Factor Construction Process: The factor is constructed by grouping stocks based on their SUE values and tracking the performance of these groups[27] - Factor Evaluation: The SUE factor has performed well in the recent month[3][27] - Factor Name: Expected Dividend Yield - Factor Construction Idea: The factor aims to capture the performance of stocks with different expected dividend yields[3] - Factor Construction Process: The factor is constructed by grouping stocks based on their expected dividend yields and tracking the performance of these groups[27] - Factor Evaluation: The expected dividend yield factor has performed well in the recent month[3][27] - Factor Name: EPTTM (Earnings Price to Trailing Twelve Months) - Factor Construction Idea: The factor aims to capture the performance of stocks based on their earnings price to trailing twelve months ratio[3] - Factor Construction Process: The factor is constructed by grouping stocks based on their EPTTM values and tracking the performance of these groups[27] - Factor Evaluation: The EPTTM factor has performed poorly in the recent month and year-to-date[3][27] - Factor Name: Analyst Net Upgrade Ratio - Factor Construction Idea: The factor aims to capture the performance of stocks based on the net upgrade ratio by analysts[3] - Factor Construction Process: The factor is constructed by grouping stocks based on their analyst net upgrade ratios and tracking the performance of these groups[27] - Factor Evaluation: The analyst net upgrade ratio factor has performed well year-to-date[3][27] - Factor Name: Analyst Net Upgrade Magnitude - Factor Construction Idea: The factor aims to capture the performance of stocks based on the net upgrade magnitude by analysts[3] - Factor Construction Process: The factor is constructed by grouping stocks based on their analyst net upgrade magnitudes and tracking the performance of these groups[27] - Factor Evaluation: The analyst net upgrade magnitude factor has performed well year-to-date[3][27] - Factor Name: Single Quarter ROE - Factor Construction Idea: The factor aims to capture the performance of stocks based on their single quarter return on equity[3] - Factor Construction Process: The factor is constructed by grouping stocks based on their single quarter ROE values and tracking the performance of these groups[27] - Factor Evaluation: The single quarter ROE factor has performed poorly year-to-date[3][27] - Factor Name: Intraday Return - Factor Construction Idea: The factor aims to capture the performance of stocks based on their intraday returns[3] - Factor Construction Process: The factor is constructed by grouping stocks based on their intraday return values and tracking the performance of these groups[27] - Factor Evaluation: The intraday return factor has performed poorly in the recent month and year-to-date[3][27] - Factor Name: Stripped Limit-Up Momentum - Factor Construction Idea: The factor aims to capture the performance of stocks based on their stripped limit-up momentum[3] - Factor Construction Process: The factor is constructed by grouping stocks based on their stripped limit-up momentum values and tracking the performance of these groups[27] - Factor Evaluation: The stripped limit-up momentum factor has performed poorly in the recent month[3][27] - Factor Test Results: - Total Market Value: Recent month performance - good, Year-to-date performance - good[3][27] - SUE: Recent month performance - good[3][27] - Expected Dividend Yield: Recent month performance - good[3][27] - EPTTM: Recent month performance - poor, Year-to-date performance - poor[3][27] - Analyst Net Upgrade Ratio: Year-to-date performance - good[3][27] - Analyst Net Upgrade Magnitude: Year-to-date performance - good[3][27] - Single Quarter ROE: Year-to-date performance - poor[3][27] - Intraday Return: Recent month performance - poor, Year-to-date performance - poor[3][27] - Stripped Limit-Up Momentum: Recent month performance - poor[3][27]

Investment Rating - The report maintains a "Weaker than Market" rating for the U.S. stock market [5]. Core Insights - The geopolitical tensions following the U.S.-Israel military action against Iran have heightened market risk aversion, leading to a decline in U.S. Treasury yields [2]. - The implementation of a new 10% global tariff by the U.S. has raised uncertainties in international trade, impacting corporate investment decisions [1]. - The unexpected rise in the U.S. core PPI in January has complicated the Federal Reserve's policy outlook, with inflation pressures remaining persistent [3]. - The report suggests a defensive investment strategy focusing on short to medium-term bonds to mitigate risks associated with interest rate volatility and inflation [4]. Summary by Sections U.S. Treasury Market - The U.S. Treasury market is currently experiencing a tug-of-war between geopolitical risk and inflation resurgence, with yields dropping below 4% due to recent military actions [4]. - The core PPI data indicates a significant inflationary pressure that could influence future monetary policy decisions [3]. Economic Indicators - January's core PPI increased by 0.8% month-over-month and 3.6% year-over-year, surpassing market expectations and highlighting ongoing production cost pressures [3]. - The market is reassessing the likelihood of interest rate cuts, with the probability of a March rate cut now at only 7.4% [3]. Investment Strategy - The report recommends a cautious approach, advising investors to limit duration exposure and focus on 2-5 year bonds to secure stable yields while waiting for clearer economic signals [4].

Macro and Strategy - The market is expected to perform well before and after the Two Sessions, with cyclical industries showing higher probabilities of gains [9][10][11] - The spring market rally has continued, with significant trading volume observed, particularly in the A-share market [9][10] - Historical data indicates that the market often rallies before the Two Sessions, with a tendency for cyclical sectors to outperform [10][11] Industry and Company - The public utility and environmental protection sector is responding positively to the national carbon emissions trading market developments [7] - The media industry anticipates growth driven by AI applications, with the 2026 Spring Festival box office reaching 5.166 billion [4] - The banking sector shows resilience despite localized pressures, as highlighted by the U.S. household debt report [4] - The real estate sector is experiencing improved sentiment due to policy measures aimed at boosting purchasing power [4] - Companies like New Oriental and Baidu are focusing on high-quality growth, with revenue guidance being adjusted upwards [7] Fixed Income Strategy - The bond market is expected to experience fluctuations, with a potential rebound following the Two Sessions as policies are implemented [22][23] - The convertible bond market is seeing increased divergence, with expectations of declining asset Sharpe ratios [27][28] - The issuance of public REITs is gaining traction, with the first port REITs application marking a significant milestone in asset securitization [30] Financial Engineering - Public funds are increasingly utilizing stock index futures for risk management, with a notable preference for long positions [24][25] - The participation of public funds in stock index futures has shown a strong inclination towards liquidity and risk management strategies [25][26]

Investment Rating - The investment rating for Baidu Group is "Outperform the Market" [5][27]. Core Insights - Baidu Group's total revenue for Q4 2025 was 32.7 billion yuan, a year-on-year decrease of 4%. The core revenue was 26.1 billion yuan, down 6% year-on-year, while iQIYI's revenue increased by 3% to 6.8 billion yuan [1][9]. - The adjusted net profit attributable to shareholders for Q4 2025 was 3.9 billion yuan, a decline of 42% year-on-year, with an adjusted net profit margin of 12% [1][9]. - AI new business revenue reached 11.3 billion yuan, with AI high-performance computing facility subscription revenue growing by 143% year-on-year [2][21]. - The company plans to increase capital expenditure significantly, maintaining a scale of 10 billion yuan over the next 12-24 months, while also initiating a stock repurchase plan of 5 billion dollars [3][27]. Financial Performance Summary - For Q4 2025, Baidu's core revenue decreased by 6%, while iQIYI's revenue increased by 3%. Online marketing revenue fell by 16%, while non-online marketing revenue rose by 13% [11][14]. - The operating expenses for Q4 2025 were 13 billion yuan, up 6% year-on-year, primarily due to increased expected credit losses and one-time employee severance costs [9]. - The forecast for adjusted net profit for 2026-2027 is 22.1 billion yuan and 22.8 billion yuan, respectively, with an additional forecast for 2028 of 27.3 billion yuan [27][29]. Business Segment Analysis - The AI business is divided into three segments: 1. Intelligent Cloud Infrastructure, which generated 5.8 billion yuan in revenue, up 34% year-on-year. 2. AI Applications, with revenue of 2.7 billion yuan, a 5% increase year-on-year. 3. AI Native Marketing Services, which saw revenue of 2.7 billion yuan, growing by 110% year-on-year, accounting for 18% of Baidu's core online marketing revenue [2][21]. Market Outlook - The report expresses optimism about the potential impact of AI technology on marketing and intelligent cloud business, despite challenges in advertising revenue due to macroeconomic conditions [3][27].

1. Report Industry Investment Rating No information provided in the content. 2. Core View of the Report - The bond market first declined and then rose last week, and ultra - long bonds fell again. The trading activity of ultra - long bonds decreased significantly, and both the term spread and the variety spread of ultra - long bonds narrowed [1][4][11]. - In March, the bond market is expected to first decline and then rise. The economic stability since Q4 2024 was mainly due to the central government's leverage - increasing support. In Q4 2025, there was no additional issuance of treasury bonds, and the short - term government support for the economy weakened. The GDP growth rate in Q4 2025 dropped to the lowest level in the post - pandemic era, and the Chinese economy is still under pressure. Meanwhile, the Party Central Committee attaches more importance to high - quality development in 2026, and the importance of "seeking progress while maintaining stability" in the total economic volume has been adjusted down. On the other hand, during the Two Sessions, policy expectations are optimistic, the A - share spring rally continues, and the current absolute level of interest rates is low, so the stock - bond seesaw effect is expected to strengthen [2][3][12][13]. 3. Summary According to the Directory 3.1 Ultra - long Bond Review - After the Spring Festival, with a record - high number of travelers during the holiday and the release of the "Shanghai Seven Measures" in the first week after the holiday, the A - share market continued to rebound. The bond market first declined and then rose, and ultra - long bonds fell again. The trading activity of ultra - long bonds decreased significantly last week, but it was still very active. The term spread and variety spread of ultra - long bonds narrowed [1][4][11]. 3.2 Ultra - long Bond Investment Outlook - **30 - year Treasury Bonds**: As of February 27, the spread between 30 - year and 10 - year treasury bonds was 44BP, at a relatively low level in history. The economic downward pressure in December eased, with the estimated GDP growth rate of about 4.5% year - on - year, a 0.4% increase from November. In January, CPI was 0.2% and PPI was - 1.4%, and the deflation risk continued to ease. It is expected that the bond market will first decline and then rise in March. The 30 - 10 spread is expected to fluctuate at a high level in the short term [2][12]. - **20 - year CDB Bonds**: As of February 27, the spread between 20 - year CDB bonds and 20 - year treasury bonds was 14BP, at an extremely low level in history. The economic situation and inflation data are similar to those of 30 - year treasury bonds. It is also expected that the bond market will first decline and then rise in March. Considering that the bond market is still in a volatile range, the variety spread of 20 - year CDB bonds is expected to continue to fluctuate in a narrow range [3][13]. 3.3 Ultra - long Bond Basic Overview - As of February 28, the balance of ultra - long bonds with a remaining term of more than 14 years was 1,655,081 billion (excluding asset - backed securities and project revenue notes), accounting for 15.3% of the total bond balance. Local government bonds and treasury bonds are the main sub - varieties. By variety, treasury bonds accounted for 27.5%, local government bonds accounted for 67.3%, etc. By remaining term, the 30 - year variety accounted for the highest proportion [14]. 3.4 Primary Market - **Weekly Issuance**: Last week (February 23 - March 1, 2026), the issuance volume of ultra - long bonds increased. A total of 1,902 billion yuan of ultra - long bonds were issued. Compared with the previous week, the total issuance volume increased significantly. By variety, only local government bonds were issued, with an amount of 1,902 billion yuan. By term, 315 billion yuan was issued with a 15 - year term, 271 billion yuan with a 20 - year term, and 1,316 billion yuan with a 30 - year term [19]. - **This Week's Planned Issuance**: The announced issuance plan for ultra - long bonds this week is 1,442 billion yuan, all of which are ultra - long local government bonds [25]. 3.5 Secondary Market - **Trading Volume**: Last week, the trading of ultra - long bonds was very active, with a trading volume of 6,189 billion yuan, accounting for 10.5% of the total bond trading volume. By variety, the trading volume of ultra - long treasury bonds was 5,083 billion yuan, accounting for 31.7% of the total treasury bond trading volume; the trading volume of ultra - long local bonds was 1,009 billion yuan, accounting for 42.1% of the total local bond trading volume, etc. Compared with the previous week, the trading activity of ultra - long bonds decreased significantly, with the trading volume decreasing by 4,736 billion yuan and the proportion decreasing by 2.5% [27][29]. - **Yield**: After the Spring Festival, the bond market first declined and then rose, and ultra - long bonds fell again. The yields of 15 - year, 20 - year, 30 - year, and 50 - year treasury bonds changed by 1BP, 4BP, 5BP, and 4BP respectively to 2.12%, 2.27%, 2.29%, and 2.48%. Similar changes occurred in CDB bonds, local bonds, and railway bonds. The yield of the 30 - year treasury bond active bond 25 ultra - long special treasury bond 02 changed by 1.3BP to 2.24%, and the yield of the 20 - year CDB bond active bond 21 CDB 20 changed by 3.02BP to 2.25% [40][41]. - **Spread Analysis**: - **Term Spread**: Last week, the term spread of ultra - long bonds narrowed, and the absolute level was low. The spread between 30 - year and 10 - year treasury bonds was 44BP, a - 1BP change from the previous week, at the 45% quantile since 2010 [46]. - **Variety Spread**: Last week, the variety spread of ultra - long bonds narrowed, and the absolute level was low. The spread between 20 - year CDB bonds and treasury bonds was 14BP, and the spread between 20 - year railway bonds and treasury bonds was 17BP, with changes of 0BP and - 3BP respectively from the previous week, at the 12% and 10% quantiles since 2010 [52]. 3.6 30 - year Treasury Bond Futures - Last week, the main contract TL2606 of 30 - year treasury bond futures closed at 1112.07 yuan, with an increase of - 0.66%. The total trading volume was 34.59 million lots (162,994 lots), and the open interest was 13.42 million lots (32,541 lots). The trading volume increased significantly compared with the previous week, and the open interest increased slightly [55].