This round of structural market movement has been primarily driven by portfolio reallocation and margin trading inflows. Following a peak in the trading volume share of the TMT sector, the turnover rate for the entire A-share market is expected to gradually decline below 1.8%. The recent visit by former U.S. President Trump to China has re-established the framework of "competition without conflict" in Sino-U.S. relations. However, market sentiment may be at a temporary high. Once sentiment cools, the expectation for a global liquidity tightening will gain greater influence in investor decision-making. As the market cools, its structural diversity is likely to return, increasing the necessity of re-establishing a barbell investment strategy. It is anticipated that a combination of AI and energy/chemicals will be a more suitable structure for the year.
The primary drivers of this structural market movement are portfolio reallocation and margin trading inflows. After a market-wide sharp adjustment involving significant position reductions, the market typically undergoes three phases: a broad-based position replenishment, a differentiated replenishment structure, and a rotation of existing funds. The current phase of position replenishment is nearing its end, and the market is gradually entering a stage of rotation among existing funds. The rapid surge in margin trading has been a key incremental factor in this structural trend. Leveraged funds demonstrated resilience during the March correction and have continued to rise since. As of May 14, the five-day moving average of the proportion of margin purchases in the total trading volume of the Shanghai and Shenzhen markets reached 10.3%, at the 92.2nd percentile level since 2021. Furthermore, leveraged funds have been primarily concentrated in the broad AI sector, represented by communications and electronics. Over the past four weeks, net margin purchases in the communications and electronics sectors accounted for 51%, 55%, 42%, and 61% of the total, respectively. Beyond leveraged funds, other fund flow indicators suggest limited market inflows, highlighting the characteristic of a structural market movement driven by reallocation of existing capital. Since April 9, existing A-share ETFs have seen cumulative net redemptions of 459.1 billion yuan. During this period, broad-based ETFs experienced cumulative net redemptions of 398.6 billion yuan, with the major broad-based ETFs held by Central Huijin Investment Ltd. seeing cumulative net redemptions of 286.1 billion yuan by the end of 2025. Sector/theme ETFs saw cumulative net redemptions of 60.5 billion yuan, with technology, cyclical, and manufacturing sectors recording net redemptions of 45.7 billion, 17.5 billion, and 16 billion yuan, respectively, while dividend-focused ETFs saw net subscriptions of 14.6 billion yuan. The increase in positions by small and medium-sized active private funds has been limited, with their surveyed positioning still below the January high. Retail investors have also not entered the market on a large scale; the number of new accounts in April failed to maintain March levels, dropping 46% month-on-month. According to channel surveys, the scale of settlement funds in May did not exceed the January level despite the market rise. Search indices for A-share related keywords on WeChat also remained significantly lower than in January this year. Overall, the concentration of leveraged funds in the tech sector, coupled with the shift of other existing funds into tech, constitutes the main funding characteristic of this structural trend. Based on calculations, over the past ten weeks, the profit-making effect in the tech sector relative to the other six major sectors reached its highest level since 2025.
Since September 2024, the proportion of trading volume in the A-share TMT sector relative to total market turnover has shown a high degree of synchronization with the daily turnover rate of the entire A-share market. From a microstructure perspective, reviewing the seven periods since 2019 where the TMT sector's trading share exceeded 40%, the daily turnover rate of the entire A-share market exhibited specific patterns within 120 trading days after the peak of the main theme: initial volatile decline, mid-period compression of fluctuations, and later-stage restructuring for a breakthrough. In the first phase (T to T+25 days), market trading sentiment peaked on the day of the high (median daily turnover rate of 1.47%, with an upper limit of 2.22%). Subsequently, as core TMT assets cooled, the market entered a painful period of "main theme stagnation," with the median A-share turnover rate declining to a range of 0.6% to 1.8% (median 1.11%). The second phase (T+25 to T+60 days) saw the market enter a prolonged "quiet period without a clear main theme" lasting about a month, during which the upper limit of the turnover rate was around 2%. The third phase (after T+60 days) involved a spontaneous repair of the market's trading structure. Overall, based on historical experience, during a market sentiment cooling phase, the A-share turnover rate needs to fall below 1.8% to be considered essentially concluded. Based on the current total market capitalization, this corresponds to a daily trading volume of approximately 2.5 trillion yuan.
The visit by former U.S. President Trump to China has re-established the framework of "competition without conflict" in Sino-U.S. relations. However, market sentiment may be at a temporary high. Overall, the core purpose of this high-level meeting was to establish guardrails and maintain stability, with its constructive role being relatively limited. Following strategic setbacks in Iran, the U.S. is expected to enter a state of "strategic calm." Within this context, Sino-U.S. relations will focus more on establishing guardrails, maintaining stability, engaging in limited transactions in controllable areas, and managing competition in key areas to avoid loss of control, essentially returning to the Biden administration's framework of "competition without conflict." Looking ahead, it is estimated that the relatively stable pattern in Sino-U.S. relations may not be limited to this year. Given the ongoing fluctuations and reconstruction in strategic thinking, it could potentially persist throughout the tenure. From a market perspective, significant contacts between China and the U.S., whether reaching trade agreements or leader meetings, are often seen as signals of peak optimistic sentiment. The reason is that such meetings typically occur when differences are relatively minimal and consensus is easiest to reach. In the six months or even longer following past meetings, opportunities for further improvement in Sino-U.S. relations have been rare, with marginal negative disturbances being more common. Therefore, viewing such events as temporary points of minimal bilateral differences and peak market optimism is a principle some investors use to gauge medium-term market sentiment. Historical data supports this view. An analysis of eight major Sino-U.S. contact events since 2018, including summits and significant agreement signings, shows the following pattern: in the 20 trading days before an event, a constructed A-share sentiment index rose from an average of 62 to 66, and the CSI 300 Index gained an average of 1.6%. In the 20 trading days after the event, the sentiment index declined from an average of 66 to 45, and the CSI 300 Index fell an average of 2.2% (with two instances of gains and six of declines, maximum drawdown of -11.9%, minimum drawdown of -2.1%). The adjustment was most pronounced in the first five trading days, gradually normalizing thereafter.
Once sentiment cools, the expectation for a global liquidity tightening will gain greater influence in investor decision-making. The inflationary impact of disruptions in the Strait of Hormuz has begun to materialize. China recently released its April PPI, which recorded a month-on-month increase of 1.7%, the second-highest value since 2021, with upstream components like raw materials showing significant strength while downstream components remained weak. The recently released U.S. April inflation data (CPI, core CPI, PPI, core PPI) all exceeded expectations. High inflation continues to erode the real purchasing power of American households; in April, real average hourly earnings turned negative year-on-year for the first time in three years, declining by 0.3%. The CME interest rate futures-implied year-end rate has been rising continuously since April 17, approaching its highest level this year at 3.7% (implying no rate cuts within the year). Global bond markets have already begun pricing in inflation. U.S. 10-year and 30-year bond yields have been rising, with the 10-year Treasury yield nearing 4.5%, close to its high since June 2025. Japan's long-term bond yields have also been rising, recently climbing to 2.6% in a one-sided move. European government bond markets have also seen widespread selling. Previously, the stock market primarily focused on the high-growth segments of the AI-related industrial trend, temporarily overlooking the impact of inflation and interest rates. However, excessively high interest rates may ultimately affect the level and pace of spending on debt-financed AI infrastructure. This analytical logic has always existed but could be selectively ignored when market enthusiasm was high. As the market cools marginally, liquidity expectations will carry greater weight in investor decisions, and negative narratives may temporarily dominate.
The necessity of re-establishing a barbell investment strategy is increasing. A combination of AI and energy/chemicals is likely a more suitable structure for this year. 1) Stable funds are temporarily absent, but the demand for stable allocation has never disappeared. The recent new highs in large Hong Kong-listed bank stocks demonstrate that stable funds are not absent from the market. However, the extreme structure and excessive focus in the A-share market have obscured the movements of these stable funds. If stable return sectors in the A-share market decline alongside a broader market adjustment, entering cheaper valuation ranges, they are likely to attract these stable funds to expand their purchasing scope. Of course, even when constructing a barbell strategy, it is necessary to focus on areas with certain supply-demand gaps or stable cash flows. The energy/chemicals chain is expected to be the most suitable choice for one end of the barbell this year. The only short-term issue is the significant volatility in commodity prices ahead of potential U.S.-Iran negotiations, making it difficult for the market to assess the real supply-demand gap and its sustainability after the Strait resumes navigation, thereby constraining participation from stable funds. Once a U.S.-Iran agreement is reached, commodity price volatility is expected to stabilize, significantly improving the investment experience. Even if short-term commodity price volatility persists, energy/chemical varieties are expected to exhibit some defensive attributes during the market sentiment cooling process. 2) A combination of AI and energy/chemicals may be a more suitable barbell structure for this year. Similar to the AI + dividends theme in 2023-2024 and the AI + resources theme in 2025, the AI + energy/chemicals combination has the potential to become a primary source of supply-demand gaps and excess returns on an annual basis. In terms of allocation, the underlying logic remains the re-rating of pricing power in China's advantaged manufacturing sectors, with the most representative industries being new energy, chemicals, non-ferrous metals, and power equipment. Continued close attention should be paid to the progress of domestic AI. On the hardware side, the logic of "volume" expansion remains the direction with relatively larger expectation gaps within the AI chain, while advancements in domestic models are expected to drive both volume and price increases in cloud services. Domestic computing power and cloud platforms are viewed favorably. Additionally, it is recommended to continue increasing allocations to some low-valuation varieties, with a focus on securities and insurance. For cyclical products experiencing price increases, the景气度 of周期成长品 like the AIDC chain and lithium battery chain remains持续, but the expectation gap is now very limited. It is advisable to focus on the tightest supply-demand links, reflected in recent price increase frequency, mainly including copper-clad laminates, fiberglass, high-speed silicon, electronic special gases, optical fiber, MLCCs, chromium, lithium carbonate, rare earths, and carbon fiber. For traditional cyclical products, the focus should be on varieties that have undergone genuine systematic capacity elimination or have absolute supply constraints, such as phosphorus chemicals, MDI, spandex, dyes, glyphosate, urea, rubber, and refrigerants.
Risk factors include intensified Sino-U.S. friction in technology, trade, and finance; domestic policy strength, implementation effects, or economic recovery falling short of expectations; unexpected tightening of domestic and international macro liquidity; further escalation of regional conflicts such as in Russia-Ukraine or the Middle East; and slower-than-expected digestion of China's real estate inventory.
Comments