- The report includes a quantitative timing model that generates signals for various financial instruments and indices, such as government bonds, stock indices, and commodities. The model uses original signals and a 5-day moving average (MA5) to determine market views like "bullish," "adjustment," or "neutral" [2][6][7] - The construction process of the model involves calculating original signals and MA5 values for each asset or index. For example, the original signal for the 30-year government bond is 86.06%, while its MA5 is 78.05%. These values are used to derive the model's view, which is "adjustment" in this case. Similar calculations are applied to other assets like the 2-year government bond (original signal: 8.02%, MA5: 14.70, view: "bullish") and COMEX gold (original signal: 39.61%, MA5: 42.29, view: "neutral") [2][6][7] - The model's evaluation is based on its ability to provide actionable market views for different financial instruments. It categorizes assets into "bullish," "adjustment," or "neutral" based on signal persistence and alignment with market trends. For example, the 2-year government bond has maintained a "bullish" signal for over 10 trading days, indicating strong consistency [2][6][7] - The backtesting results of the model include specific signal values for each asset or index. For instance: - 30-year government bond: Original signal 86.06%, MA5 78.05%, view "adjustment" [2][6] - 2-year government bond: Original signal 8.02%, MA5 14.70%, view "bullish" [2][6] - COMEX gold: Original signal 39.61%, MA5 42.29%, view "neutral" [2][7] - IPE crude oil: Original signal 46.07%, MA5 40.68%, view "neutral" [2][7] - Various stock indices like the CSI Red Dividend Total Return Index and the Hang Seng Technology Index also have their respective signal values and views [2][6][7]

- The report includes a quantitative model that provides timing signals for various financial instruments, such as government bonds, stock indices, and commodities. The model outputs probabilities representing the likelihood of short-term upward movements in yields or indices[3][7][8] - The model uses a moving average (MA5) to smooth the daily timing signals. For example, the original signal for the 30-year government bond is 82.96%, while its MA5 value is 74.29%. This approach helps identify trends and reduces noise in the data[3][7] - The model evaluates multiple instruments, including 2-year and 10-year government bonds, stock indices like the CSI All A Index and the Hang Seng Tech Index, and commodities such as COMEX gold and IPE crude oil. Each instrument is assigned a viewpoint, such as "bullish," "adjustment," or "neutral," based on the signal probabilities and thresholds[3][7][8] - The thresholds for the model's viewpoints are defined as follows: probabilities above 60% indicate a bullish stance, below 40% suggest bearishness, and values in between are considered neutral[8] - The model's performance is tracked over a 10-day period, with daily updates on signal probabilities and viewpoints for each instrument. For example, the 2-year government bond has maintained a "bullish" viewpoint for over 10 days, while the CSI All A Index has been in "adjustment" for 3 days[3][7][8] - The report provides detailed signal probabilities and MA5 values for each instrument, such as the CSI All A Index (83.02% original signal, 64.68% MA5) and COMEX gold (57.89% original signal, 42.29% MA5)[3][7][8]

Core Viewpoint - The article emphasizes the importance of understanding the relationship between short-selling tools and their pricing anchors in constructing neutral strategies in the bond market [1]. Group 1: Neutral Strategy Tools - Common hedging tools include government bond futures and bond lending, with government bond futures prices closely tied to their cheapest-to-deliver (CTD) bonds [1]. - The TL contract typically anchors to 30-year old bonds, while the T contract anchors to 7-year government bonds, necessitating a focus on the stability of the spread between current bonds and CTD bonds when employing neutral strategies [1]. - Bond lending directly targets the most liquid active bonds, but investors must be cautious of high lending rates and concentration risks that can lead to negative carry costs [1]. Group 2: Types of Neutral Strategies - The article identifies three main types of neutral strategy instruments: 1. **30-Year Old Government Bonds**: Investors need to clarify whether they are betting on mean reversion or a trend-based recovery. The spread between the old bonds and TL contract CTD bonds remains stable around 0 basis points, making it suitable for narrow spread fluctuations [2]. 2. **Ultra-Long Local Government Bonds**: These should prioritize bond lending for precise alignment, as their pricing logic closely mirrors that of 30-year government bonds, providing a solid basis for constructing neutral strategies [3]. 3. **Policy Bank Bonds**: The 7-year bonds are best hedged with government bond futures, while the 10-year bonds are more suited for bond lending. Using the T contract for 10-year policy bank bonds may lead to misalignment due to the influence of the 10Y-7Y spread [3].

- The report involves several quantitative models and factors, including the 30-year Treasury bond, 3-year AAA medium-term notes, 10-year Treasury bond, 2-year Treasury bond, Wind All A Index, CSI Dividend Total Return Index, Hang Seng Technology Index, STAR 50 Index, Wind Micro Index, and CNI 2000 Index[2][3][6] - The construction idea of these models is based on the original signal and the 5-day moving average (MA5) to generate model viewpoints such as "adjustment," "bullish," or "shock"[2][6] - The specific construction process involves calculating the original signal and the 5-day moving average (MA5) for each asset. For example, the original signal for the 30-year Treasury bond is 73.10%, and its MA5 is 73.46%. The model viewpoint is "adjustment" if the original signal is below the MA5 and "bullish" if it is above the MA5[2][6] - The evaluation of these models is qualitative, indicating whether the market is in a state of adjustment, bullish, or shock based on the comparison between the original signal and the MA5[2][6] - The specific test results for these models are as follows: - 30-year Treasury bond: original signal 73.10%, MA5 73.46%, model viewpoint "adjustment," signal duration 10 days[2][6] - 3-year AAA medium-term notes: original signal 73.81%, MA5 80.07%, model viewpoint "adjustment," signal duration 2 days[2][6] - 10-year Treasury bond: original signal 69.40%, MA5 74.91%, model viewpoint "adjustment," signal duration 7 days[2][6] - 2-year Treasury bond: original signal 17.95%, MA5 10.80%, model viewpoint "bullish," signal duration over 10 days[2][6] - Wind All A Index: original signal 64.74%, MA5 65.43%, model viewpoint "adjustment," signal duration 1 day[2][6] - CSI Dividend Total Return Index: original signal 76.94%, MA5 72.05%, model viewpoint "adjustment," signal duration 5 days[2][6] - Hang Seng Technology Index: original signal 79.89%, MA5 80.42%, model viewpoint "adjustment," signal duration 6 days[2][6] - STAR 50 Index: original signal 78.50%, MA5 78.98%, model viewpoint "adjustment," signal duration over 10 days[2][6] - Wind Micro Index: original signal 66.86%, MA5 62.01%, model viewpoint "adjustment," signal duration 1 day[2][6] - CNI 2000 Index: original signal 76.44%, MA5 76.88%, model viewpoint "adjustment," signal duration 9 days[2][6] - COMEX Gold: original signal 31.26%, MA5 51.98%, model viewpoint "shock," signal duration 1 day[7] - IPE Brent Oil: original signal 57.93%, MA5 48.82%, model viewpoint "shock," signal duration 8 days[7]

Group 1 - The report emphasizes the importance of distinguishing between different hedging strategies based on the type and maturity of bonds, specifically focusing on 30-year government bonds, local government bonds, and policy bank bonds for implementing neutral strategies [1][4]. - The core trading strategy involves identifying the most cost-effective spread opportunities in the bond market and utilizing government bond futures to construct neutral strategies that mitigate directional risks while stabilizing the spread convergence [7][8]. - The report outlines that common neutral strategy varieties include: 1) 30-year government bonds, where the focus is on mean reversion versus trend-based trading; 2) ultra-long local government bonds, which require precise alignment; and 3) policy bank bonds, where different maturities necessitate different hedging tools [4][18][25]. Group 2 - The report details that government bond futures are closely tied to the cheapest-to-deliver (CTD) bonds, which may not always align with the movements of actively traded bonds of the same maturity, necessitating careful consideration of the spread between the target bonds and CTD bonds [8][12]. - It highlights that the TL contract is primarily anchored to 30-year government bonds, and the spread between old and new bonds should be monitored for effective arbitrage [9][19]. - The report indicates that the T contract is linked to 7-year government bonds, and the strategy must account for the 10-7 year spread, which has shown significant fluctuations in recent years [13][25]. Group 3 - The analysis of neutral strategies reveals that the 30-year government bonds exhibit a stable spread around 0 basis points, making them suitable for high-certainty narrow spread trading [19][28]. - For ultra-long local government bonds, the report suggests that bond lending is a more robust choice due to the stable mean-reversion characteristics of the spread with 30-year government bonds, while also allowing for precise targeting of liquidity [22][30]. - The report notes that for policy bank bonds, the 7-year bonds are better suited for hedging with government bond futures, while the 10-year bonds require careful consideration of the 10-7 year spread to avoid hedging mismatches [25][35].

Summary of Conference Call Records Industry Overview - The records primarily discuss the bond market, focusing on convertible bonds and government bonds, with insights into the broader financial market dynamics as of March 2026. Key Points and Arguments Market Performance - The China Convertible Bond Index fell by 7.1% in March, reversing all gains for the year, indicating high volatility in the market [1][2] - The Shanghai Composite Index dropped by 3.6% on March 23, 2026, marking a significant decline, with historical comparisons showing similar drops in previous years due to external factors [2] Institutional Fund Flows - There was a notable net redemption of fixed-income products, particularly from wealth management subsidiaries, which significantly impacted high-volatility products [2] - Institutions primarily sold off sectors such as electronics, power equipment, banking, non-ferrous metals, and pharmaceuticals, while showing interest in automotive, basic chemicals, and machinery sectors [2] Future Market Scenarios for Convertible Bonds - Three potential scenarios for the convertible bond market were outlined: 1. **Optimistic Scenario**: A short-term rebound leading to price increases in equity-type and mid-to-low priced convertible bonds [3] 2. **Volatile Scenario**: A period of market fluctuation with limited new capital inflow, delaying overall valuation increases [4] 3. **Pessimistic Scenario**: A shift in market sentiment leading to significant risks, particularly for lower-quality convertible bonds [4] Investment Strategies - Emphasis on identifying convertible bonds with strong debt support and a high likelihood of conversion success, targeting a price of 130 yuan to trigger strong redemption [5] - Recommendations to maintain cautious positions and focus on structural allocations, particularly in mid-to-low priced convertible bonds with favorable conversion prospects [5] Government Bond Market Insights - The yield on 10-year government bonds is expected to fluctuate between 1.78% and 1.85%, while 30-year bonds may reach up to 2.3% [6] - The market is anticipated to experience a directional choice soon, with significant attention on the upcoming special government bond issuance plan [6][10] Monetary Policy Outlook - The likelihood of the central bank actively withdrawing MLF (Medium-term Lending Facility) is low, with a supportive monetary policy stance expected to continue [7] - Key signals to watch for include the wording in MLF announcements and the potential impact of external geopolitical factors on domestic monetary policy [7] Credit Bond Market Dynamics - Funds are the primary players in the credit bond market, with significant net purchases observed, particularly in the 1-3 year maturity range [9] - The trend of funds heavily investing in credit bonds is expected to persist into the second quarter, despite potential risks associated with concentrated positions [10] Key Variables to Monitor - Upcoming special government bond issuance plans and their market impact [8] - Economic data releases in early April to assess ongoing economic trends [10] - Fluctuations in oil prices and their implications for monetary policy [10] Additional Important Content - The records highlight the importance of structural investment strategies in a volatile market environment, emphasizing the need for careful monitoring of both domestic and international economic indicators [5][10]

Report Industry Investment Rating No information provided in the text. Core Viewpoints - The economic data from January to February was positive, with the industrial added - value in the production sector rising to 6.3%, the year - on - year growth of fixed asset investment turning positive, the year - on - year growth of total retail sales of consumer goods rising to 2.8%, and exports growing by 19.2%. The domestic economic growth momentum improved. With the volatile Middle East situation causing oil prices to rise and the A - share market plummeting, the bond market mainly declined, and ultra - long bonds continued to fall [1][4][10][33]. - The probability of a recent bond market correction is high. Factors include the surge in crude oil due to the escalation of the Middle East geopolitical conflict, the good economic growth rate at the beginning of the year reducing the need for significant monetary policy easing, and the reduction in the central bank's bond - buying scale due to the low absolute level of interest rates [2][3][11][12]. Summary by Directory 1. Ultra - long Bond Review - Economic data from January to February was positive, the bond market mainly declined, and ultra - long bonds continued to fall. The trading activity of ultra - long bonds decreased slightly last week but remained very active. The term spread of ultra - long bonds widened, and the variety spread narrowed [1][4][10]. 2. Ultra - long Bond Investment Outlook 30 - year Treasury Bonds - As of March 20, the spread between 30 - year and 10 - year Treasury bonds was 47BP, at a historically low level. The recent bond market correction probability is high, and the 30 - 10 spread is expected to fluctuate at a high level in the short term [2][11]. 20 - year China Development Bank Bonds - As of March 20, the spread between 20 - year China Development Bank bonds and 20 - year Treasury bonds was 14BP, at a historically low position. The recent bond market correction probability is high, and the variety spread of 20 - year China Development Bank bonds is expected to continue to fluctuate within a narrow range [3][12]. 3. Ultra - long Bond Basic Overview - The balance of outstanding ultra - long bonds is 25.3 trillion. As of February 28, ultra - long bonds with a remaining term of over 14 years totaled 1,655,081 billion, accounting for 15.3% of the total bond balance. Local government bonds and Treasury bonds are the main sub - varieties. The 30 - year variety has the highest proportion [13]. 4. Primary Market Weekly Issuance - Last week (March 16 - 22, 2026), the issuance of ultra - long bonds increased. A total of 1,532 billion yuan of ultra - long bonds were issued. By variety, local government bonds accounted for 1,461 billion yuan, and government - supported institutional bonds accounted for 50 billion yuan. By term, 210 billion yuan had a 15 - year term, 836 billion yuan had a 20 - year term, and 486 billion yuan had a 30 - year term [18]. This Week's Scheduled Issuance - The announced ultra - long bond issuance plan for this week totals 1,288 billion yuan, all of which are ultra - long local government bonds [22]. 5. Secondary Market Trading Volume - Last week, ultra - long bonds were very actively traded, with a trading volume of 9,826 billion yuan, accounting for 9.9% of the total bond trading volume. The trading activity decreased slightly compared with the previous week [24]. Yield - Due to positive economic data and other factors, the bond market mainly declined, and ultra - long bonds continued to fall. The yields of 15 - year, 20 - year, 30 - year, and 50 - year Treasury bonds, China Development Bank bonds, local bonds, and railway bonds changed to varying degrees [33]. Spread Analysis - The term spread of ultra - long bonds widened last week, with an absolute low level. The variety spread of ultra - long bonds narrowed, with an absolute low level [43][44]. 6. 30 - year Treasury Bond Futures - Last week, the main 30 - year Treasury bond futures contract TL2606 closed at 110.67 yuan, with a decline of 0.35%. The total trading volume was 384,400 lots (a decrease of 86,770 lots), and the open interest was 134,400 lots (an increase of 3,697 lots) [49].

Report Industry Investment Rating No information provided in the given content. Core Viewpoints - The special treasury bond issuance plan this year may be earlier than last year. The issuance plan in 2024 was announced in May, and in 2025 it was advanced to April 16. With the earlier release of the list of 'two - heavy' construction projects and the statement of accelerating fund allocation, it is judged that this year's plan may be even earlier [8][14]. - The current 30 - year treasury bond active bond (2500006) is 'unreasonable' as its balance is much less than 2500002. If the active bond expectation changes, the spread between 2500006 and 2500002 may widen [8][15]. - Market prices active bonds based on the special treasury bond issuance plan. In 2024, the active bond did not change as the scale of special treasury bonds was similar to 230023. In 2025, due to the earlier issuance plan and larger expected scale of new bonds, the active bond expectation of 250002 weakened, and the spread between 250002 and 230023 widened [8][15][16]. - For this year, if the estimated balance based on the special treasury bond issuance plan is higher than the current active bond, 2500006 may weaken, and funds may flow to 2500002 or 230023, with the spreads of 2500006 - 2500002 and 2500006 - 230023 likely to widen. It is advisable to appropriately allocate some 30 - year old bonds before the special treasury bond issuance plan [8][16]. - Currently, 2500006 has a high bond - lending concentration. If it weakens, short - sellers can close positions for profit. However, if there are significant positive events in the bond market before the issuance plan, 2500006 may perform relatively stronger [9][17]. Summary by Directory 1. This Week's Bond Market Review - There are many 30 - year treasury bond varieties with certain liquidity. As the second quarter approaches, investors are advised to pay attention to the impact of the special treasury bond issuance plan on the active bond switch [14]. - The special treasury bond issuance plan this year may be earlier than last year. The earlier release of the 'two - heavy' project list and the call for accelerating fund allocation support this judgment [8][14]. - If the issuance is advanced, it will affect the long - term bond trend. The current active bond 2500006 is 'unreasonable', and its spread with 2500002 may widen if the expectation changes [8][15]. - By comparing history, it is found that the market prices active bonds according to the issuance plan. In 2024, the active bond did not change, while in 2025, the active bond expectation of 250002 weakened, and the spread with 230023 widened [8][15][16]. - For this year, 2500006 may weaken, and the spreads with 2500002 and 230023 may widen. It is recommended to allocate some 30 - year old bonds before the issuance plan [8][16]. - 2500006 has a high bond - lending concentration. Its performance may be affected by short - selling and positive market events [9][17].

Report Industry Investment Rating - The report does not provide a specific industry investment rating. Core Views - The report analyzes various financial products, including bonds, stock indices, and commodities, and provides model - based views on their trends. The views are classified as "看多" (bullish), "调整" (adjustment), and "震荡" (sideways). Bullish products include 3 - year AAA medium - short - term notes, 2 - year treasury bonds, and the Wind Micro - cap Index. Products in adjustment include 30 - year treasury bonds, 10 - year treasury bonds, the Wind All - A Index, the CSI Dividend Total Return Index, the Hang Seng Tech Index, the STAR 50 Index, and the China Securities 2000 Index. Sideways products include COMEX gold and IPE Brent crude oil [2][3][6]. Summary by Related Content Bond Analysis - **30 - year Treasury Bond**: The original signal is 72.23%, MA5 is 71.18%, the model view is "调整" (adjustment), and the signal has lasted for 6 trading days [2][6][7]. - **3 - year AAA Medium - short - term Note**: The original signal is 93.88%, MA5 is 27.75%, the model view is "看多" (bullish), and the signal has lasted for 4 trading days [2][6][7]. - **10 - year Treasury Bond**: The original signal is 77.63%, MA5 is 66.70%, the model view is "调整" (adjustment), and the signal has lasted for 3 trading days [2][6][7]. - **2 - year Treasury Bond**: The original signal is 5.82%, MA5 is 27.35%, the model view is "看多" (bullish), and the signal has lasted for over 10 trading days [2][6][7]. Stock Index Analysis - **Wind All - A Index**: The original signal is 82.26%, MA5 is 60.51%, the model view has changed from "震荡" (sideways) to "调整" (adjustment), and the signal has lasted for 1 trading day [2][6][7]. - **CSI Dividend Total Return Index**: The original signal is 46.79%, MA5 is 61.58%, the model view has changed from "震荡" (sideways) to "调整" (adjustment), and the signal has lasted for 1 trading day [2][6][7]. - **Hang Seng Tech Index**: The original signal is 87.97%, MA5 is 71.03%, the model view is "调整" (adjustment), and the signal has lasted for 2 trading days [2][6][7]. - **STAR 50 Index**: The original signal is 77.75%, MA5 is 73.52%, the model view is "调整" (adjustment), and the signal has lasted for over 10 trading days [2][6][7]. - **Wind Micro - cap Index**: The original signal is 60.71%, MA5 is 30.48%, the model view is "看多" (bullish), and the signal has lasted for 7 trading days [2][6][7]. - **China Securities 2000 Index**: The original signal is 62.11%, MA5 is 61.27%, the model view is "调整" (adjustment), and the signal has lasted for 4 trading days [2][6][7]. Commodity Analysis - **COMEX Gold**: The original signal is 53.33%, MA5 is 58.47%, the model view is "震荡" (sideways), and the signal has lasted for 2 trading days (not yet closed, postponed by one trading day) [2][6][7]. - **IPE Brent Crude Oil**: The original signal is 28.07%, MA5 is 41.07%, the model view is "震荡" (sideways), and the signal has lasted for 4 trading days (not yet closed, postponed by one trading day) [2][6][7].

- The report includes a quantitative timing model that outputs probabilities for short-term movements in bond yields or stock indices. The model uses historical data to estimate probabilities and provides multi-directional views (bullish, bearish, or neutral) based on thresholds of 40% and 60% for probabilities. Probabilities above 60% indicate bullish sentiment, below 40% indicate bearish sentiment, and between 40%-60% indicate neutral sentiment [5][6]. - The model's specific construction involves calculating daily timing signals and a 5-day moving average (MA5) of these signals. The timing signals represent the probability of upward or downward movements in the target asset. For example, the 30-year government bond signal is calculated as 76.88% for the day, with an MA5 of 67.56%, indicating an "adjustment" view. Similarly, the 2-year government bond signal is 7.40%, with an MA5 of 30.40%, indicating a "bullish" view [5][6]. - The model evaluates multiple assets, including government bonds (2-year, 10-year, 30-year), corporate bonds (3-year AAA medium-term notes), stock indices (e.g., CSI All A Index, CSI Dividend Total Return Index, Hang Seng Technology Index, STAR 50 Index, Wind Micro Index, and CNI 2000 Index), and commodities (COMEX Gold and IPE Brent Oil). Each asset is assigned a daily signal, MA5, and a directional view (bullish, bearish, or neutral) [5][6]. - The model's performance is assessed through the persistence of signals over multiple trading days. For example, the bullish signal for COMEX Gold has persisted for over 10 trading days, while the adjustment signal for the CSI All A Index has persisted for 4 trading days. This persistence is used to validate the model's reliability in predicting short-term movements [5][6]. - The report provides detailed backtesting results for each asset, including daily timing signals, MA5 values, and directional views. For instance, the CSI Dividend Total Return Index has a daily signal of 65.65% and an MA5 of 55.78%, indicating a "neutral" view. Similarly, the STAR 50 Index has a daily signal of 86.91% and an MA5 of 77.16%, indicating an "adjustment" view [5][6]. - The model is qualitatively evaluated as a useful tool for short-term market timing, providing actionable insights for various asset classes. However, the report highlights potential risks, such as model failure due to changing market conditions, factor failure, and data quality issues [5][6][7]. - Backtesting results for the quantitative timing model: - 30-year government bond: daily signal 76.88%, MA5 67.56%, view "adjustment" [5][6] - 3-year AAA medium-term notes: daily signal 17.77%, MA5 26.20%, view "bullish" [5][6] - 10-year government bond: daily signal 74.05%, MA5 60.92%, view "adjustment" [5][6] - 2-year government bond: daily signal 7.40%, MA5 30.40%, view "bullish" [5][6] - CSI All A Index: daily signal 35.32%, MA5 64.67%, view "adjustment" [5][6] - CSI Dividend Total Return Index: daily signal 65.65%, MA5 55.78%, view "neutral" [5][6] - Hang Seng Technology Index: daily signal 62.49%, MA5 59.50%, view "neutral" [5][6] - STAR 50 Index: daily signal 86.91%, MA5 77.16%, view "adjustment" [5][6] - Wind Micro Index: daily signal 22.76%, MA5 16.86%, view "bullish" [5][6] - CNI 2000 Index: daily signal 49.34%, MA5 66.20%, view "adjustment" [5][6] - COMEX Gold: daily signal 74.78%, MA5 38.96%, view "bullish" [5][6] - IPE Brent Oil: daily signal 60.69%, MA5 49.96%, view "neutral" [5][6]