Sinolink: In Q3, the scale of "fixed income plus" funds continued to rebound, with increased positions in pro-cyclical sectors and reduced holdings in technology growth.

Sinolink released a research report stating that in 2026Q3, the scale of "fixed income plus" funds continued to rebound, while positions edged down but remained at a high level since 2023Q4. On the liability side, "fixed income plus" funds still received relatively notable incremental capital, but both net subscription scale and new issuance scale declined. In terms of allocation, "fixed income plus" funds mainly increased positions in sectors such as non-ferrous metals, defense, petroleum and petrochemicals, coal, and transportation, while reducing positions in electronics, electrical equipment and new energy, non-bank financials, home appliances, and telecommunications, showing similarities with active equity funds in Q3. Reversal after extreme divergence in active equity funds In 2026Q3, the equity positions of active equity funds fell markedly to 83.00% (A-shares 74.84%, Hong Kong stocks 8.16%), and the decline in positions was entirely due to the decline in A-share positions, while the allocation ratio to Hong Kong stocks edged up. Combined with southbound flows, public fund holdings as a proportion of southbound flows continued to decline to 11.76% in Q3, and both active/passive fund allocation ratios continued to fall. In terms of performance, the median performance of active funds in Q3 was approximately -13.2%, of which only 37.11% of active funds outperformed their benchmarks, comparable to the central level from 2022Q4 to 2024Q3, but performance divergence among active funds remained evident. The performance gap between the top/bottom 10% of active funds in Q3 remained at the fourth-highest level in history, mainly because the performance of the bottom 10% was at the second-worst level in history. In terms of factor exposure, top-performing active equity funds in 2026Q3 mainly pursued assets with strong profitability, high cash flow capability, and dividend attributes, while also featuring low valuation and low trading, which differs significantly from previous top-performing funds that pursued high growth/above-expectation. On the liability side, equity funds (active + passive) regained net capital inflows on the liability side in 2026Q3, driven on one hand by passive funds again receiving large net subscriptions, and on the other by a decline in net redemptions from active funds: 1) active funds' net outflow narrowed from RMB91.666 billion in 2026Q2 to RMB52.979 billion; 2) passive funds regained net inflows of RMB458 billion. It is worth noting that as the market retreated markedly in Q3, the profitability of liability-side holders of active funds dropped from overall "significant profit" to "only slight profit," but the net redemption scale on the liability side of active equity funds did not expand; instead, a higher proportion of active funds received net subscriptions, with active funds with higher exposure to strong profitability/dividends/low valuation receiving more favor. 2026Q3 active equity funds: consensus has loosened In 2026Q3, the holdings concentration of active equity funds declined somewhat. By industry, active funds mainly increased positions in sectors such as pharmaceuticals, non-ferrous metals, autos, defense, and coal, while reducing positions in electronics, non-bank financials, telecommunications, banks, and building materials. From a track perspective, active equity funds' allocation/overweight ratio to AI hardware (optical modules, PCB, semiconductors, etc.) fell rapidly to 45.2%/26.4% in 2026Q3, but the allocation/overweight ratio to AI hardware remained at a historical high, with the overweight ratio in AI hardware even still higher than the peak overweight ratios of most historical mainstream tracks. That is, although AI hardware suffered a large drawdown in Q3, active equity funds' reduction in AI hardware was relatively limited, and overall holdings exposure remained high. Combined with funds heavily positioned in different sectors and incremental market capital, the consensus among various market participants on the AI hardware sector in 2026Q3 showed fairly obvious cracks, which is also an important reason why top-performing funds in Q2 significantly underperformed in Q3. It is worth noting that funds heavily positioned in electronics and telecommunications tend to also be heavily positioned in electrical equipment and new energy, pharmaceuticals, machinery, and other sectors. As the liability side of active equity funds continues to shrink, this holding correlation may amplify volatility in the aforementioned industries, stemming on one hand from potential common redemption pressure on the liability side, and on the other hand from the "seesaw" effect on the asset side. Against this backdrop, active equity funds have temporarily faced a dilemma similar to 2022-2024, when large amounts of capital accumulated in tracks but the wealth effect gradually disappeared, incremental capital was relatively limited, and the briefly returned "pricing power" gradually faded. In addition, funds heavily positioned in different sectors also faced different divergence in 2026Q3, and this divergence continued to breed "fear of losing" and "fear of missing out" on the liability side of active funds: funds heavily positioned in sectors such as electronics and telecommunications began to see net redemptions as floating profits on the liability side fell sharply; funds heavily positioned in sectors such as coal, pharmaceuticals, and non-ferrous metals saw varying degrees of net subscriptions due to improvement in liability-side floating profit levels. However, in absolute scale terms, the liability-side impact brought by this divergence may temporarily be relatively limited; nevertheless, if the divergence pattern continues in the future, the impact on the liability side may gradually strengthen, further affecting active funds' "pricing power." Since October 2026, the market has again undergone "deleveraging," with ETFs still the main supporting force Since October, leverage funds represented by margin financing have not returned after the long holiday; instead, a new round of rapid deleveraging and chip clearing began: margin financing activity and margin financing balances both fell rapidly, active equity funds and northbound capital both first showed varying degrees of net selling, and new fund issuance scale also fell markedly compared with September. Subsequently, supported by large-scale stock ETF buying, the market shifted to range-bound trading. By industry, leverage funds represented by margin financing again concentrated net selling in sectors such as electronics and telecommunications since October, which Sinolink believes actually continues the "de-crowding" trade since Q3. On the other hand, precisely because active equity funds' reduction in sectors such as electronics and telecommunications in Q3 was relatively limited and their absolute holding ratios remain at historical highs, while new issuance scale continues to decline, AI hardware may in the future rely more on incremental capital from other participants. However, considering that both new fund issuance and margin financing behavior depend on the market's wealth effect, and that other investors including "fixed income plus" funds also still have allocation ratios to AI hardware at highs, this means the overall holding crowding in AI hardware may still be at a relatively high level, and the market may still face repeated trading-level fluctuations and volatility in the future.