GF Securities: A-Share Pessimism Fully Priced In, Conditions Ripe for Oversold Rebound
nashnova research
GF Securities argues that A-share valuation compression has exceeded the historical average by nearly 10 percentage points, with multiple short-term sentiment gauges hitting near-zero extremes simultaneously — an oversold bounce window is opening, but how far it runs depends on signal checks in November and December.
How severe is the valuation squeeze?
A-share 2026 net-profit growth is forecast above 15%, yet the valuation contribution stands at -24% — nearly double the historical average of about -14%.
This means → share prices are not just failing to reflect earnings growth; they are actively *discounting* it. The market has crammed all its fears about trade friction, interest rates, and risk appetite into valuation.
In plain terms = companies are earning more money, but investors' "pricing confidence" has dropped to a historically rare low.
AI stocks rallied hard — where did the valuation go?
The valuation premium AI-focused companies built up in 2024–2025 has been entirely erased by 2026. Full-year net-profit growth is expected at 85%, yet back-calculating from share prices shows a valuation drag of over -40%.
This means → three consecutive years of share-price gains were driven almost entirely by earnings. The market awarded AI zero "imagination premium."
This reflects expectations for 2027–2028 AI-industry growth that have already fallen to very low levels — current prices are essentially paying only for profits already booked.
Where do the technical gauges stand?
As of October 16, short-term sentiment indicators for the Wind All-A, ChiNext, and STAR Market indices have all dropped to near 0% — a historical extreme. GF Securities notes that past instances of the gauge hitting 0% corresponded to high win rates for short-term rebounds.
The ChiNext this week deviated more than 10% below its 20-day moving average — only the fifth time since the 2016 circuit-breaker episode. All four prior signals triggered a bounce: the shortest lasted 15 days (+9.8%), the longest 94 days (+27.6%).
In plain terms = several "panic thermometers" are reading near their physical limits at the same time. History says a technical bounce from these levels is more likely than not.
Lock-up expiries are coming — should A-shares or Hong Kong stocks worry more?
A-shares: December brings a lock-up expiry peak, but data since 2010 show the All-A rose 56% of the time in high-expiry months; the following month's win rate dips to 38% — the transmission chain is weak.
Hong Kong: A wave of expiries concentrated from late 2026 into early 2027 covers large-model, GPU, optical-module, and autonomous-driving companies, creating sentiment pressure in Q4.
This means → A-share lock-up expiries are mostly psychological noise; Hong Kong is the market that genuinely needs to watch for selling pressure. Historically, Hong Kong stocks fall the most before large expiries and often sit at a cyclical floor afterward; dividend stocks are relatively insulated.
How long can the bounce last?
GF Securities sees an oversold bounce as likely to begin, but its duration and magnitude hinge on three signals in November–December: the U.S. midterm elections' effect on major asset classes, global AI-industry catalysts, and the U.S. earnings-season window.
The report flags three risks: geopolitical conflict driving global inflation higher, U.S. economic resilience pulling global liquidity into faster tightening, and China's growth-support measures falling short of expectations.
In plain terms = the technical setup is flashing "bounce possible," but whether fundamentals can sustain it won't be clear for another month or so.
市场有风险，内容仅供研究参考，不构成投资建议。