Guosen: Model differences and proprietary trading divergence; maintain "Outperform" rating on the securities industry.

Guosen released a research report maintaining an "Outperform" rating on the industry. From the perspective of marginal industry changes, the securities industry is undergoing a dual iteration of performance and business models. In the short term, it benefits from the rally driven by rising market trading activity, while in the medium to long term, the logic stemming from endogenous transformation continues to strengthen. The bank focuses on recommending leading brokerages with strong capital strength, balanced business layouts, and leadership in AI and cross-border businesses. Guosen's main views are as follows: Third-quarter proprietary trading performance dragged on growth momentum In the first three quarters of 2026, 43 listed brokerages collectively achieved operating revenue of RMB499.231 billion, up 18.64% year-on-year, and net profit attributable to parent of RMB209.473 billion, up 23.74% year-on-year, with a weighted average ROE of 6.15%, up 0.60 percentage points year-on-year. However, cumulative growth rates have fallen significantly from the first half (revenue +45.9%, net profit +49.1%). Third-quarter single-quarter revenue was RMB134.52 billion, -21% year-on-year and -37% quarter-on-quarter; net profit attributable to parent was RMB54.103 billion, -17% year-on-year and -43% quarter-on-quarter, with only 11 companies achieving positive year-on-year growth in net profit attributable to parent. The problem lies mainly in proprietary investment: investment business revenue in the first three quarters was RMB203.59 billion, up only 8.66% year-on-year, accounting for 40.8% of operating revenue, down 3.7 percentage points year-on-year, ranking last in growth among the five major business lines (brokerage +27.74%, investment banking +26.11%, asset management +20.31%, interest +37.58%). Single-quarter proprietary income was RMB34.857 billion, turning from the largest contributor in the first half into the largest drag. The simultaneous emergence of slowing aggregate growth and structural divergence means that the relative performance of proprietary trading directly determines changes in rankings. The divergence in proprietary trading is essentially a difference among three types of investment models First, directional trading positions (mainly trading financial assets, i.e., TPL): against the backdrop of a 12.49% decline in the CSI 300 and a 27.80% decline in the ChiNext Index in Q3 alone, institutions with large equity exposure and strong directional bias mainly bore fair value drawdowns, with 35 (over 80%) posting year-on-year declines in single-quarter proprietary income and 9 recording negative values. Second, OCI accounts (with fluctuations recorded in other comprehensive income): at the end of the third quarter, the 43 brokerages had combined OCI equity of RMB832.262 billion, up 18% from the beginning of the year, accounting for 10.7% of financial investment assets, a share hitting a new high in recent years. Among them, CITIC (RMB123.9 billion, +21% quarter-on-quarter), Huatai (RMB32.1 billion, +63% quarter-on-quarter), and GF (RMB44.7 billion, +46.4% from the beginning of the year) were all accelerating allocations; Guotai Haitong's OCI equity was stable at RMB84.4 billion, with trading financial assets down 2% quarter-on-quarter. This OCI-dominated allocation structure enabled its proprietary income to still record RMB24.771 billion, up 21.61% year-on-year. Third, direct equity investment and follow-on investment: returns from this type of business depend on the pace of primary market exits and have weaker correlation with secondary market . In the third quarter, hard-tech projects such as CXMT Corporation, Unitree Robotics, and Enflame Technology were listed in concentrated fashion, and with brokerages simultaneously playing direct investment and sponsorship roles, Huaan's proprietary income in the first three quarters was RMB6.003 billion (+203%), with single-quarter RMB3.910 billion (+336%); CMSC's RMB20.959 billion (+186%), with single-quarter proprietary income of RMB8.044 billion (+151%), rose to the top, with fair value change gains and losses of RMB15.149 billion as the main source of realization, and CICC's RMB15.719 billion (+43.35%) was also among them. Differences in proprietary trading models will amplify into operating performance volatility In the first three quarters, the CR5 (top five brokerages) share of net profit attributable to parent rose to 50%, up 4 percentage points from the end of 2025; Huaan (+169.56%), CICC (+101.71%), China Merchants (+94.81%), and GF (+46.58%) ranked at the top. Investment income dominating brokerage performance will have two effects. First, most high-growth performance carries a "realization of gains" attribute (such as fair value change gains and losses, equity exit gains, etc.), whose sustainability is weaker than fee-based income and should not be treated on par with businesses such as brokerage and asset management; on the other hand, OCI unrealized gains are not included in current-period profit, which may mean that some proprietary investment returns are underestimated, and institutions with larger directional equity exposure will have greater elasticity in either the rebound or pullback in the fourth quarter. Therefore, measuring proprietary trading should focus more on the matching among "allocation structureperformance volatilitycapital occupation" rather than simply looking at revenue growth, which is also the micro-foundation for the industry's shift from scale-driven to asset-capability pricing. Risk warnings: risk of significant secondary market volatility; risk that the comprehensive deepening of fundamental institutional reform in the capital market progresses less than expected; risk of stricter financial regulation; risk that innovative business progress falls short of expectations, etc.