Since the release of the 2026 Annual Strategy Report in December last year, we have consistently emphasized that in the face of major external transformations, especially innovations at the productivity level like AI, some historical patterns and experiences are likely to require re-evaluation.
Patterns such as "things should not happen more than three times," trading volume share, concentration and crowding metrics, and institutional holdings have been continuously broken during the past year's AI technological breakthroughs. However, up to now, these indicators still serve as factors in short-term market games. This article will further explore the boundaries of applying these patterns.
I. Market Outlook: ① Since the bull market began on September 24th, the Shanghai Composite Index has only fallen below its 100-day moving average twice. ② The first instance was in April 2025 due to reciprocal tariffs, and the second was in March 2026 due to US-Iran conflict, both triggered by significant external black swan events. Other corrections have generally found support near the 100-day moving average. ③ Excessive deviation from the 100-day line often leads to cooling measures, while proximity to the 100-day line typically sees supportive policies. ④ Over the past two weeks, the index has repeatedly tested the 100-day moving average. If no major external black swan factors emerge in the near future, systemic risks are expected to be limited. ⑤ Therefore, in the absence of negative market beta, a full-scale capital outflow is unlikely. On a quarterly basis, the focus can remain on corporate performance and industry trends.
II. Current Hot Topic: Trading Volume Share of Top 5% Stocks - US Market vs. A-Share Market From the US market perspective, the trading volume share of the top 5% stocks has exceeded 70% three times historically: ① 1995-2000, during the dot-com boom; ② 2007-2008, during the subprime mortgage crisis and safe-haven flows; ③ 2017 to present, with capital clustering in large tech companies. Currently, this share stands around 70%.
Conclusions from US experience: ① When major industry trends emerge, trading concentration can break through long-standing historical thresholds. Using the fluctuation range of the past 10-20 years as a threshold may lead to missing significant opportunities. ② Current US trading concentration is not at a high level, slightly below the September 2020 peak and far below the dot-com bubble period in 2000.
From the A-share market perspective, the trading volume share of the top 5% stocks has risen rapidly but has not yet exceeded 50% or reached historical highs. With major external transformations occurring, the trading concentration in related A-share sectors still has room to increase. ① Even compared to historical highs, current concentration remains lower. ② In the face of major industry trends, trading concentration itself is expected to reach new highs. ③ The accelerated selling of traditional assets in this cycle may be difficult to sustain, and the steep slope of subsequent concentration increases is expected to slow.
The 100-day moving average of the Shanghai Composite Index serves as a key support line in the current bull market. Apart from two short-term breaches caused by external shocks, the index has generally found support when approaching this line, and outflows from broad-based ETFs have also slowed. Currently, the Shanghai Composite Index, representing traditional assets, is again nearing its 100-day moving average. It is highly likely that the index will stabilize, and selling pressure on broad-based ETFs will ease. Therefore, we believe the accelerated selling of traditional assets may not last long, and the steep increase in trading concentration is expected to moderate.
III. Trading Volume/Market Cap Share of Tech Industries: US Market vs. A-Share Market From the US market perspective, during the dot-com era, the trading volume and market cap shares of the internet industry repeatedly hit new highs. For example, the trading volume share of the hardware equipment sector had a stable threshold (e.g., 17%) during most periods, but it significantly and persistently broke through this multi-decade threshold during the 1990s dot-com cycle and the post-2023 AI cycle.
From the A-share market perspective: ① A-shares have experienced four waves of tech industry trends, with the TMT sector's market cap share and trading volume share showing systematic increases. ② Sectors like new energy vehicles in 2021 and optical modules in 2025-26 have also seen systematic rises in their trading volume share.
Therefore, in the face of accelerating industry prosperity and earnings, crowding indicators such as trading concentration, trading volume share, and market cap share can easily become ineffective. These types of indicators are more applicable to thematic sectors where earnings have not yet materialized, such as humanoid robots, commercial aerospace, and AI applications.
IV. During the phase when industry trends translate into earnings, the indicative significance of public fund holding structures and the magnitude of position increases for stock prices is also diminishing. Historical patterns that have become ineffective in recent years include: ① Historically, a single sector's holding exceeding 20% might face short-term pressure, but the electronics sector has broken this pattern. ② From a "bull market mindset" perspective: the sector with the largest quarterly position increase does not necessarily underperform in the next quarter. The current AI industry trend has broken market patterns under存量 conditions.
V. For industry trends at major inflection points, the aforementioned timing indicators have become ineffective. For short-term judgments, what other indicators can be referenced? Referring to Chinese and US cases, for major inflection points brought by industrial revolutions, some timing indicators have become ineffective. In this context, fundamentals and industry trends remain the most fundamental indicators for identifying the sustainability of market moves.
If seeking better risk-reward ratios in smaller波段, some short-term indicators can be referenced: At the sector level, the moving average deviation (EMA20 deviation) is a more suitable indicator, as its historical center is more stable. At the index level, sentiment diffusion indicators show decent short-term timing效果.
Risk warnings: Geopolitical risks, overseas inflation risks, lower-than-expected domestic稳增长 policies, etc.
Comments