利率择时信号

Search documents
债市机构行为周报(8月第2周):股份行机构行为触发做多信号-20250817
Huaan Securities· 2025-08-17 03:42
Group 1: Report Overview - The report is a fixed - income weekly report titled "Institutional Behavior of Joint - stock Banks Triggers Bullish Signals - Weekly Report on Bond Market Institutional Behavior (Week 2 of August)" dated August 17, 2025 [1][2] - The chief analyst is Yan Ziqi, and the analyst is Hong Ziyan [2] Group 2: Core Viewpoints - The bond market's bullish space has opened, and institutional behavior indicators have triggered bullish signals. A trading - following strategy based on joint - stock banks' transactions may have a high win - rate [2][3][12] - Although there are short - term bearish factors in the bond market, the medium - and long - term trend remains unchanged, and the bullish space has opened [4][13] Group 3: Weekly Institutional Behavior Review 3.1 General Comment - A rate - timing signal was developed based on joint - stock banks' trading behavior. In the past year, it gave 5 long - lasting large - wave signals with a 100% win - rate on large waves. On August 13, it triggered a bullish signal that lasted for 3 days [3][12] - Big banks were the main buyers of short - and medium - term bonds, while funds and securities firms sold long - term interest - rate bonds, and rural commercial banks, insurance companies, and city commercial banks were the main buyers. The bond market's capital supply remained loose, and the overall spread of the curve widened [4][13] 3.2 Yield Curve - The yields of treasury bonds and China Development Bank bonds generally increased. For treasury bonds, the 1Y yield rose 2bp, 3Y fell 1bp, 5Y rose about 4bp, 7Y rose 5bp, 10Y rose 6bp, 15Y rose about 7bp, and 30Y rose 9bp. For China Development Bank bonds, the 1Y yield rose 3bp, 3Y rose 4bp, 5Y rose 8bp, 7Y rose 7bp, 10Y rose 8bp, 15Y rose 7bp, and 30Y rose 9bp [15] 3.3 Term Spread - For treasury bonds, the interest spread increased, and the term spread generally widened. For China Development Bank bonds, the interest - spread inversion deepened, and the short - end spread widened [16][19] Group 4: Bond Market Leverage and Funding 4.1 Leverage Ratio - The leverage ratio dropped to 107.22%. From August 11 to 15, it first increased and then decreased [20] 4.2 Repurchase Transactions - The average daily trading volume of pledged repurchase this week was 8.2 trillion yuan, with an average overnight proportion of 89.82%. The overnight trading volume increased by 0.03 trillion yuan month - on - month, and the overnight proportion decreased by 0.05pct [26][30] 4.3 Funding - Banks' fund lending first increased and then decreased. Big banks and policy banks' net lending on August 15 was 4.83 trillion yuan. The main fund borrowers were funds. DR007 fluctuated upward, and R007 continued to rise [32] Group 5: Duration of Medium - and Long - Term Bond Funds 5.1 Median Duration - The median duration of medium - and long - term bond funds remained at 2.81 years (de - leveraged) and 3.11 years (leveraged). On August 15, the de - leveraged median duration changed less than 0.01 year, and the leveraged median duration decreased by 0.02 years [44] 5.2 Duration by Bond - Fund Type - The median duration of interest - rate bond funds (leveraged) rose to 3.94 years, an increase of 0.02 years from last Friday. The median duration of credit - bond funds (leveraged) dropped to 2.86 years, a decrease of 0.03 years from last Friday [47] Group 6: Category Strategy Comparison 6.1 Sino - US Spread - The Sino - US treasury bond spreads generally widened. The 1Y spread widened by 2bp, 2Y by about 1bp, 3Y narrowed by 4bp, 5Y widened by about 3bp, 7Y widened by 1bp, 10Y changed less than 1bp, and 30Y widened by 2bp [52] 6.2 Implied Tax Rate - The implied tax rate generally widened. As of August 15, the 1Y spread between China Development Bank bonds and treasury bonds widened by about 2bp, 3Y by 5bp, 5Y by 3bp, 7Y by about 3bp, 10Y by 2bp, 15Y narrowed by 1bp, and 30Y changed less than 1bp [53] Group 7: Bond Lending Balance Changes - On August 15, the lending concentration of active 10Y treasury bonds and active 10Y China Development Bank bonds increased, while that of less - active 10Y treasury bonds, less - active 10Y China Development Bank bonds, and active 30Y treasury bonds decreased. By institution, the lending concentration of securities firms and other institutions increased, while that of big banks and small - and medium - sized banks decreased [54]
基于股份行交易的利率择时信号
Huaan Securities· 2025-07-02 12:03
Report Industry Investment Rating No relevant content provided. Core View of the Report Since 2024, the bond trading attributes of joint-stock banks, which were "neglected" in traditional institutional behavior research, have expanded, and high-frequency operations of interest rate bands can be observed. Based on their behavior, corresponding buy and sell timing signals can be constructed, and the follow-up strategy outperforms the benchmark return (10Y Treasury bond) by about 30% [2][4]. Summary by Relevant Catalog 1. Why Do We Focus on the Trading Behavior of Joint-Stock Banks? 1.1 A Consensus on the Institutional Behavior of Joint-Stock Banks in the Market Joint-stock banks are the last "puzzle piece" in institutional behavior research. The market consensus on the institutional behavior of joint-stock banks is that they usually conduct distribution in the secondary market. Joint-stock banks and city commercial banks typically have an obvious nature of "primary subscription and secondary distribution," so they generally show a net selling feature in secondary cash bond transactions. Historically, joint-stock banks have shown a net selling state in nearly 90% of the time, similar to city commercial banks [11]. 1.2 New Features of the Institutional Behavior of Joint-Stock Banks Since 2024 Since 2024, joint-stock banks have started to increase interest rate band trading. On the one hand, as the Ministry of Finance has increased the issuance scale of single Treasury bonds, the stock scale of active bonds with key tenors (such as 10Y) has reached a new high, reducing the impact of distribution from new bond issuance. On the other hand, since 2024, the overall liability side of the banking system has been squeezed, making them more reliant on the income from bond investment business on the asset side. Therefore, the trading volume of joint-stock banks for old 10-year Treasury bonds has increased significantly, and they are no longer in the previous mode of distribution and passive buying during bond market corrections [2][14]. 2. Judgment of Interest Rate Timing Signals Based on the Trading Behavior of Joint-Stock Banks 2.1 Construction of Low-Frequency Interest Rate Timing Signals Based on the trading behavior of joint-stock banks, a cash bond follow-up strategy signal is constructed. By studying the trading behavior of joint-stock banks on long-term bonds and eliminating the impact of the latest bond distribution, in the past year, the long-term interest rate band trading return following joint-stock banks under the 10MA caliber has exceeded the benchmark by nearly 30%. From September last year to late June this year, the 10-year Treasury bond yield declined by 49bp, while the interest rate band return following joint-stock banks was 62bp, with an excess return of 13bp over the benchmark and an overall excess return rate of 26%. Moreover, if the continuous days of the signal triggered by joint-stock banks exceed 5 trading days, the winning rate of their buying signals in the past year was 100%, capturing a total band of 59bp, and the winning rate of selling signals was 60%, avoiding a capital loss of 15bp [2][18][21]. 2.2 Construction of Medium-Frequency and Other Timing Signals Two other perspectives can be used to observe the interest rate timing signals of the trading behavior of joint-stock banks. First, observe the higher-frequency 5MA caliber buying signals. Since 2025 is a year of strong fiscal efforts, there may still be distribution impacts under the old bond caliber at some time points, so the buying signals of joint-stock banks may be more meaningful than the selling signals. Second, observe the extreme value of single-day demand. If it exceeds a certain threshold, it can be regarded as an overbought or oversold signal, and the historical backtesting shows a relatively high winning rate [3][24][27].