For stock investors, the Golden Kirin analyst research reports offer authoritative, professional, timely and comprehensive insights to help uncover potential thematic opportunities. In the investment strategy summary from Chen Guo, A-shares continued their correction this week, but a substantial net inflow into broad-based indices on Friday drove a notable market rally. We believe this shift could break the negative feedback loop between sentiment and capital flows, signaling the start of a rebound phase. Furthermore, the magnitude of the subsequent rebound is also tied to the external macro environment and requires dynamic response, with key focus on the Middle East situation and oil price trends around the U.S. midterm elections, as well as marginal changes in domestic demand policy guidance for next year's tone. Rebound clues: balance thematic growth directions with low chip resistance, oversold rebounds in technology hardware, and domestic demand directions that logically benefit from falling oil prices and domestic policy expectations. The ChiNext Index and STAR 50 have fallen about 41% from their late-June and early-July highs this round, not reaching the decline speed of the ChiNext in 2015. Referring to the A-share performance after two rounds of declines in late August 2015: after policy support (rate and reserve requirement cuts) and a sharp deleveraging, the market experienced a considerable recovery. The ultimate space for the rebound phase also requires dynamic tracking of subsequent macro and industry changes. Currently, the core overseas macro issue is whether oil prices and interest rates will decline, while the core domestic macro issue is policy to expand domestic demand, and the core industry issue is AI commercialization progress. At the market micro level, net inflows into broad-based ETFs strengthened further this week. On the 23rd, broad-based ETFs across the entire market saw single-day net inflows of approximately 55.5 billion yuan. Additionally, since October, 88 companies have announced buyback proposals or plans, with the electronics sector having the highest number. Meanwhile, our constructed market sentiment indicator has fallen back to near extremely low territory this week. Historically, in this zone, short-term downside risk is smaller than rebound opportunities, and structural trading crowding has also declined significantly. The total A-share margin balance has decreased by about 591.6 billion yuan from its June high, the proportion of stocks in the top 5% by turnover has fallen from 51.6% at the end of July to around 42%, and the turnover share of electronics and communications has dropped to about 33%. The trading structure has improved, and these factors are conducive to a short-term bottoming and rebound. Rebound clues: prefer thematic growth with low chip resistance and industries that can be prioritized for allocation based on high certainty of third-quarter earnings delivery and order and prosperity trends still pointing upward. Allocation directions: computers, media, defense, non-ferrous metals, chemicals, real estate, consumer, non-bank finance, etc. Thematic investments: AI applications (physical AI, AI pharmaceuticals, robotics, etc.). Risk warnings: Overseas interest rates and cloud厂商 capital expenditures falling short of expectations; capital flow pressure exceeding expectations; third-quarter earnings falling short of expectations; historical analogies have limitations. 1. This week the market fluctuated and then rebounded, with risk mitigation signals emerging. This week the A-share market was mainly characterized by narrow-range fluctuations and corrections in the early part, followed by a clear rebound on Friday with synchronized recovery in volume. Most major broad-based indices closed lower, but the declines were significantly narrower than last week, with small- and mid-cap styles still dominant. At the sector level, trading crowding in the TMT sector decreased notably from earlier levels, with the turnover share continuously falling to around 41%, indicating an improved trading structure. The computer sector was the first to rebound and led the industry gains. Moreover, from the perspective of capital flows and sentiment positioning, the market has shown risk mitigation signals this week: first, broad-based ETFs continued to see net inflows, with daily net inflows significantly increasing and remaining positive for multiple consecutive days, providing support for the index; the margin balance has continued to decline after falling from previous highs, with the absolute scale of leveraged funds continuously depleting, mitigating the risk of negative chip feedback. Second, market sentiment is approaching historical lows. Our constructed market sentiment indicator has fallen this week to a zone near extremely low values, where further significant declines are usually limited, and short-term market downside risk has converged in stages. 2. Comparison with 2015: similar adjustment magnitude. In terms of decline from highs, the current adjustment magnitude is already close to that of 2015, but the pace is noticeably slower. In 2015, the ChiNext Index fell about 41% within 23 trading days from its June 3 high, and by September 15 had fallen about 55% cumulatively; this round, the ChiNext Index fell about 41% from June 25 to October 22, and the STAR 50 fell about 41% from its June 30 high, over about 76 trading days. From the perspective of rebounding sectors, the second wave of rebound in 2015 was mainly concentrated in non-bank finance, real estate, beauty care, computers and other directions with lower institutional allocation and relatively good fundamentals. 3. Trading structure pressure may gradually ease, and market stabilization demonstrates confidence. The main reason for the A-share adjustment over the past two weeks was capital flows, with leveraged and crowded chips concentrated in a clearing-out process, exhibiting certain stampede characteristics. From October 9 to 22, the ChiNext Index and STAR 50 fell 15.1% and 10.3% respectively, while the CSI 300 and East Money All A fell 6.4% and 5.7% respectively; on October 22, the ChiNext Index fell 4.37% in a single day, with the Shanghai Composite Index, CSI 300, ChiNext Index and East Money All A all closing at new lows for the year. On October 23, the market bottomed and rebounded, with the ChiNext Index rising 4.16%. Turnover in the past two weeks did not increase significantly. From October 12 to 23, the average daily turnover of All A was about 1.56 trillion yuan, lower than the third quarter's average daily level of about 2.26 trillion yuan. The adjustment was more a reflection of concentrated selling under insufficient absorption; on the rebound day of October 23, turnover recovered to about 1.96 trillion yuan. The continuous deleveraging of leveraged funds was an important driver of this stampede. The total A-share margin balance fell from a high of about 2.89 trillion yuan on June 25 to about 2.30 trillion yuan on October 22, a cumulative decrease of about 591.6 billion yuan, or about 20%, of which about 108.3 billion yuan was reduced since October 9. The proportion of margin purchases in total A-share turnover fell from a monthly average high of about 10.6% in September 2025 to 6.6% since October, and was 6.0% on October 22, the lowest level since June 2025. Stop-loss selling by absolute-return funds has largely been realized. Using bond-biased funds with equity positions between 5% and 35% at the end of the second quarter as a fixed-income-plus sample (1,275 funds, with scale of about 2.62 trillion yuan at the end of Q2), as of October 23, about 60% of the scale had retraced more than 3% from their year-to-date net value highs, about 36% had retraced more than 5%, and about 28% of the scale had negative year-to-date returns, with scale-weighted year-to-date returns of about 1.2%. The proportion of scale with drawdowns exceeding 3% and 5% is already relatively high, and position reductions related to drawdown control have already occurred extensively during the previous decline. The space for further active stop-loss by active equity funds has narrowed. Among equity-biased funds (4,356 funds, with scale of about 4.66 trillion yuan at the end of Q2), about 79% of the scale had retraced more than 20% from year-to-date highs, about 57% had retraced more than 30%, and about 55% of the scale had negative year-to-date returns. From the perspective of holding costs, for the top 20 A-shares by market value in active equity-biased fund heavy holdings at the end of Q2, weighted by public fund holding market value, current prices are about 30% lower than the third-quarter average trading price and about 25% lower than the second-quarter average trading price; 19 of the 20 are below the third-quarter average price, and about 88% of holding market value is more than 20% below the third-quarter average price. Subsequent pressure will come more from passive position reductions caused by redemptions. Trading crowding has fallen back to a mid-to-low level for the year. Calculated by weekly turnover, the proportion of stocks in the top 5% by turnover in total A-share turnover has fallen from a high of 51.6% at the end of July to 41% to 43% in the past two weeks, already close to the first quarter's average level of about 41.4%; the combined turnover share of Shenwan Electronics and Communications in total A-shares (20-day average) fell from a high of about 40.9% in late July to 32.8% on October 23, with the 5-day average dropping to 29.5% on October 16, the lowest since July. Crowding in the technology direction, where funds were previously highly concentrated, has been significantly digested. Overall, the trading structure cleanup may have entered its latter half. The margin balance has fallen about 20% from its high, about 60% of fixed-income-plus scale has retraced more than 3%, institutional heavy holdings have generally fallen below quarterly costs, and trading crowding has dropped to a mid-to-low level for the year. The crowded chips previously concentrated in technology leaders have been largely cleared. The remaining pressure mainly comes from two aspects: first, the relatively deep net value drawdown of equity-biased funds may bring sustained redemption pressure; second, fixed-income-plus products' desire to preserve positive full-year returns before year-end assessment. Market stabilization funds have clearly stepped up, with broad-based ETFs as the main channel. From October 12 to 22, broad-based ETFs saw total net inflows of about 183.4 billion yuan, with single-day net inflows of about 44.6 billion yuan and 51.7 billion yuan on October 15 and 22 respectively. The single-day net inflow on October 22 was the third highest since July, second only to July 17 and 20. Funds flowed concentratedly into CSI 300 ETFs, with net inflows of about 78.9 billion yuan in the past two weeks, accounting for 43% of broad-based ETF net inflows; ETFs related to the ChiNext Index, STAR 50 and SSE 50 saw net inflows of about 33 billion yuan, 28 billion yuan and 16.8 billion yuan respectively. On the 23rd, broad-based ETFs across the entire market saw single-day net inflows of about 55.5 billion yuan. Regulators guided listed companies to increase buybacks, with "key minorities" actively providing support. On October 10, the CSRC held a symposium, where participating experts suggested "promoting qualified listed companies to increase dividend and buyback efforts, and improving investor return and rights protection mechanisms." On October 19, CSRC Chairman Wu Qing stated at the 2026 Financial Street Forum Annual Meeting that "qualified listed companies should be encouraged to increase buybacks and holdings, and better reward investors." Buybacks were concentrated in technology leaders that had previously been heavily sold off. Since October (as of October 23), a total of 88 companies have announced buyback proposals or board buyback plans, with a total proposed buyback amount cap of about 14.8 billion yuan, of which 32 were directly proposed by chairmen, controlling shareholders or actual controllers. By industry, electronics had 30 companies, the highest number; Changxin Technology, Dongyangguang, New Easy-Tech, Hygon Information and others had the largest proposed buyback amounts. Increased buybacks by listed companies after deep stock price adjustments help stabilize the pricing of institutional heavy holdings and reduce the risk of further stampedes. 4. Earnings verification window opens, selectively seize the rebound. Starting next week, we enter a dense period of domestic and overseas earnings verification, and market pricing is expected to return from capital games to fundamentals. Overseas, Meta was the first to disclose its third-quarter report on October 21, Google and Microsoft will release quarterly results on October 28, and Amazon will release its third-quarter report on October 29. Cloud vendors' capital expenditure guidance is key to testing AI computing demand; the Federal Reserve will hold its interest rate meeting on October 28; and on November 3 the U.S. will hold midterm elections. Changes in pre-election policy and interest rate expectations may increase volatility in overseas assets. Domestically, A-share third-quarter reports will be disclosed intensively before the end of October. As of October 23, 510 companies have disclosed third-quarter reports, with total net profit attributable to parent up about 19.5% year-on-year; according to scheduled disclosure dates, about 5,070 more companies will disclose from October 24 to 31, accounting for more than 90% of all A-shares. Trading structure pressure may gradually ease, market stabilization funds provide support, and the space for further significant market declines is limited. For the rebound, allocate selectively, preferring thematic growth with low chip resistance and industries that can be prioritized for allocation based on high certainty of third-quarter earnings delivery and order and prosperity trends still pointing upward. Allocation directions: computers, media, defense, non-ferrous metals, chemicals, real estate, consumer, non-bank finance, etc. Thematic investments: AI applications (physical AI, AI pharmaceuticals, robotics, etc.). Risk analysis: 1) Overseas interest rates and technology capital expenditures falling short of expectations. If the Federal Reserve's policy path is tighter, U.S. Treasury yields continue to rise, or cloud vendors' capital expenditure guidance is below expectations, global technology asset valuations may continue to face pressure. 2) Capital flow pressure exceeding expectations. If equity-biased fund redemptions accelerate, fixed-income-plus products concentrate position reductions before year-end, or the margin balance continues to fall rapidly, the time needed for chip clearing may be longer than expected. 3) Third-quarter earnings falling short of expectations. If listed companies' earnings delivery is below market expectations, the sustainability and structure of the rebound may be weaker than our judgment. 4) Limitations of historical analogies. There are differences between 2015 and the present in macro environment, market structure and policy tools, and historical trends do not represent future performance. Sina statement: This message is reprinted from Sina's cooperative media. Sina publishes this article for the purpose of conveying more information and does not mean it agrees with its views or confirms its description. The article content is for reference only and does not constitute investment advice. Investors operate at their own risk based on this information.
Comments