CN StockArticle is online

Where Are A-Shares’ Valuation Gaps, and What Could Change Next Year?

Claude AI
Where Are A-Shares’ Valuation Gaps, and What Could Change Next Year?

Table of Contents

You might want to know

  • Why do A-share valuations appear high at both ends of the market but unusually low in the middle?
  • Could a change in investor flows and corporate buybacks help revalue the companies caught in that middle segment?

Main Topic

Source: CITIC Securities Research (中信证券研究). Authors: Qiu Xiang (裘翔), Wu Andong (吴安东), Zhang Mingkai (张铭楷), Chen Feng (陈峰), Yu Sen (玉森), and Chen Zeping (陈泽平).

A comparison of leading companies across industries in China, the United States, Japan, and South Korea reveals a distinctive valuation pattern in China’s A-share market. Rather than a smooth distribution, valuations form a “U-shaped barbell”: investors assign relatively high valuations to companies associated with aggressive growth, especially technology, and to defensive, dividend-oriented companies. By contrast, many long-duration businesses in the middle of the barbell—such as established manufacturers, consumer companies, and healthcare leaders—receive markedly lower valuation multiples. The report argues that differences in fundamentals do not fully explain this discount.

In the authors’ view, market structure is an important part of the explanation. Trading-oriented investors tend to cluster in technology, while portfolio-allocation funds are concentrated in dividend stocks. The middle segment lacks a stable base of long-term capital. When marginal investors focus on short-term earnings momentum or current dividend yields, the value of cash flows expected further into the future can be systematically discounted. This leaves companies with durable businesses, competitive advantages, and long-term return potential vulnerable to persistent selling pressure.

The cross-market comparison highlights the depth of the gap. Among 71 selected companies in the middle segment, measured using rolling twelve-month median ROE, A-share leaders have a median ROE of 13.5%. That is below the 18.5% recorded by comparable U.S. companies, but above Japan’s 10.7% and South Korea’s 8.4%. Yet the valuation percentile for the A-share group is only 22.5%, compared with 65% in the United States, 52.5% in Japan, and 47.5% in South Korea. A-share companies in this segment therefore combine returns that remain comparatively strong with valuations near the lower end of their historical range.

The pattern is not uniform across all types of companies. A-share companies linked to AI and dividend-oriented leaders have higher valuation percentiles than their overseas peers. For dividend stocks, median ROE is at the 29% historical percentile, well below the corresponding overseas groups, while the valuation percentile stands at 66.3%. At the industry level, Chinese consumer, industrial manufacturing, and healthcare leaders have the lowest valuation percentiles among the four markets. Healthcare is closest to its own historical low, while manufacturing has the widest valuation gap relative to overseas peers.

Weak domestic demand alone does not account for the manufacturing discount, according to the report. The median five-year compound revenue growth rate among 25 A-share manufacturing and export leaders is 16.6%, substantially higher than that of comparable companies in the United States, Japan, and South Korea. Their median ROE is 14.2%, below the U.S. level but well above Japan and South Korea. Even so, their valuation percentile is only 26.5%, considerably below that of overseas peers.

Trading behavior also differs sharply by market segment. Using turnover over the past year to estimate implied holding periods, A-share AI and technology leaders have an implied holding period of just 1.7 months, while financial and dividend leaders stand at 23.3 months. The 21.6-month gap between the two ends is larger than the corresponding gap in the United States, Japan, or South Korea, which is 7.2 months, 15.0 months, and 9.0 months, respectively. Middle-segment A-share companies have an implied holding period of 6.3 months, compared with 11.3 months in the United States, 15.4 months in Japan, and 8.5 months in South Korea.

The holding-period gap has also widened over time. From 2018~2020 to 2025~2026, the implied holding period for A-share consumer, manufacturing, and healthcare leaders fell from 8.7 months to 6.3 months, a decline of about 28%. Over the same periods, the figure for U.S. middle-segment companies increased from 10.4 months to 11.3 months, Japan’s rose from 14.2 months to 15.4 months, and South Korea’s decreased from 10.8 months to 8.5 months. The difference between A-share and U.S. holding periods expanded from less than 2 months to about 5 months; the gap with Japan widened from about 5.5 months to about 9 months.

The report further examines how valuations respond to current dividend yields, AI-related growth sensitivity, and ten-year ROE. In cross-sectional analysis of 108 Chinese, 101 U.S., and 84 Japanese companies, the estimated pricing weight attached to durable long-term ROE in A-shares fell from 0.583 (t=8.43) in 2019-2021 to 0.343 (t=6.00) in 2026. Meanwhile, the coefficient associated with AI characteristics rose from an insignificant 0.021 (t=0.29) to 0.491 (t=8.18), overtaking long-term ROE as the leading pricing factor.

The U.S. and Japanese comparisons show a different emphasis. In 2026, the U.S. pricing weight for stable ROE is 0.504 (t=6.48), above the 0.233 attached to AI characteristics. Japan’s stable-ROE weight is 0.561 (t=6.58), compared with 0.210 for AI characteristics. The authors interpret this contrast as evidence that marginal pricing in A-shares has shifted toward shorter-duration attributes: near-term changes in industry momentum in technology, or current yield in dividend stocks. Long-run characteristics—such as enduring profitability, pricing power, competitive barriers, and the capacity to sustain ROE—receive less recognition.

This shift creates a market that can swing between two extremes. At one end, investors may concentrate on the latest growth narrative; at the other, they may favor defensive dividend income. In between, companies that require investors to assess future cash flows and long-term compounding may lack a consistent valuation anchor. The report suggests that this is a reversal from the situation at the beginning of 2021, when the market was strongly focused on ROE and willing to pay a premium for it.

The relative pricing of companies listed in both Shanghai and Hong Kong offers another indication of the perceived discount. As of November 20, 2026, the market-cap-weighted A-share premium for AI-oriented companies was 109.2%, and the premium for dividend-oriented companies was 44.1%. In contrast, the premium for leading manufacturers, consumer companies, and healthcare companies in the middle segment was only 24.5%. In 5 cases, the Hong Kong-listed shares were more expensive than their A-share counterparts. These companies often have strong governance, global competitiveness, active dividends and buybacks, and businesses whose growth has become steadier. The report describes them as long-duration assets: their valuation depends on investors recognizing future cash flows, rather than only near-term growth.

Investor flows may help explain why the middle segment lacks a stable valuation base. In the 2020~2021 bull market, actively managed public funds recorded net inflows of 2.9 trillion yuan, while northbound funds recorded net inflows of about 640 billion yuan. In 2024~2026Q3, actively managed public funds instead saw net outflows of 0.17 trillion yuan, and northbound net inflows were only at the scale of hundreds of billions of yuan. Insurance funds, by contrast, recorded net inflows of about 0.25 trillion yuan and 2.78 trillion yuan over those two periods, respectively, increasing their influence on pricing.

Among existing active products, net subscriptions have been concentrated in technology-focused funds: 70.6% of incremental flows went to technology, while balanced products attracted a share of subscriptions notably below their share of existing assets. Other sources of demand have included public-fund “fixed income plus” products, which received net subscriptions of 1.18 trillion units over the past 7 quarters; bank wealth-management “fixed income plus” products, which increased by 3.25 trillion yuan during the year; and quantitative long-short private funds, whose assets rose from 671.9 billion yuan at the end of 2024 to 1.9 trillion yuan in 2026Q3, equal to 37.5% of equity private-fund assets. The report characterizes this as a bull market with limited inflows from active investors who make bottom-up, company-specific judgments. Much of the new capital has instead followed top-down allocation or trading strategies.

The experience of overseas markets suggests that the decline of active fund management does not necessarily eliminate company-level price discovery. In the United States, index products grew from 19% of long-term funds in 2010 to 52% at the end of 2025, reaching 54.5% in September 2026. Index products accounted for 65.0% of domestic U.S. equity products. Active domestic equity mutual funds recorded net redemptions every year from 2006, when net outflows were 17.7 billion U.S. dollars, through 2025, when net outflows were 662.9 billion U.S. dollars. From 2016–2025, cumulative net outflows from domestic active funds totaled 3.89 trillion U.S. dollars; including global active equities, the total was 4.81 trillion U.S. dollars.

One important difference, the report argues, is that U.S. companies’ ongoing share repurchases provide an additional source of valuation support and price discovery. Across 219 representative middle-segment companies in China, the United States, Japan, and South Korea, A-shares had the highest free-cash-flow yield, at 4.83%, but a net buyback rate of only 0.24%. U.S. companies repurchase more than 1 trillion U.S. dollars of shares annually, with a net buyback rate of 1.71%. Japan’s Tokyo Stock Exchange book-value reforms helped push buybacks to a record 24.9 trillion yen in fiscal 2025, with a net buyback rate of 1.14%. South Korea’s net buyback rate was 0.09%, reflecting persistent governance discounts and relatively low shareholder returns. Although A-share buybacks have advanced considerably over the past two years, the report says they are still commonly viewed as a way to stabilize share prices rather than as a sustained shareholder-return and valuation tool.

The authors believe the extreme valuation structure will eventually correct, and that a turning point may come next year. They note that the exit of low-risk-preference allocation funds is accelerating the reduction of institutional positions in large-cap stocks, while the valuation premium of smaller companies could continue to contract as quantitative long strategies shrink. If liquidity tied up in less efficient small-cap trading gradually returns to large-cap blue chips—as it did in 2016~2017—it could improve conditions for the middle segment, even if the shift merely reallocates existing capital.

Recent performance provides an early indication of a possible change in market leadership. The report constructs an equal-weighted portfolio by excluding AI-industry-chain and finance-related sub-industries from CITIC Level-3 industries, then selecting the 2 largest companies by market capitalization in each remaining Level-3 industry. This produces a portfolio of 472 stocks across 26 Level-1 industries. From a base date of September 22, 2026, through November 20, the portfolio fell 14.1%, compared with a 22.3% decline in the equal-weighted Wind All A index, producing excess returns of 10.7% and reaching a new high at the end of the period. From a September 1 base date, the portfolio outperformed on 90.3% of the trading days when the equal-weighted Wind All A index fell.

A second approach uses the CSI A500, excluding financial and AI constituents, as a non-AI and non-financial large-company portfolio. Since September 22, this market-cap-weighted portfolio declined only 4.0%, delivering excess returns of 15.2% relative to the CSI A500. In a year without a clear catalyst for a broad demand surge, the report considers an approach based on industry diversification and leading companies a practical way to address a complex market outlook for next year.

Risk factors: Intensifying friction between China and the United States in technology, trade, and finance; domestic policy support or implementation, or economic recovery, falling short of expectations; tighter-than-expected domestic and overseas macro liquidity; further escalation of conflicts in Russia-Ukraine, the Middle East, and other regions; and slower-than-expected absorption of China’s property inventory.

Key Insights Table

AspectDescription
Valuation structureA-shares show a U-shaped barbell: relatively high valuations at the technology and dividend ends, with a deeper discount in the middle.
Middle-segment fundamentalsSelected A-share leaders have median ROE of 13.5% but a valuation percentile of 22.5%, below comparable U.S., Japanese, and South Korean groups.
Investor behaviorShort implied holding periods in technology and the middle segment contrast with long holding periods in financial and dividend leaders.
Pricing factorsA-share pricing has become more sensitive to AI characteristics and short-term factors, while long-term ROE has lost relative influence.
Capital flows and buybacksThe report identifies limited active, bottom-up capital and relatively modest sustained buybacks as contributors to the valuation gap.
Potential strategyIndustry diversification combined with a focus on leading companies is presented as a practical approach to the coming market environment.

Afterwards...

The report’s central question is whether the current preference for short-duration narratives and defensive yields will eventually give way to more balanced company-level valuation. A sustained shift would likely require more than improved sentiment: deeper participation by active investors, greater recognition of long-term business quality, and continued corporate actions that return capital to shareholders could all matter. The timing and extent of any adjustment remain uncertain, and the identified risk factors could delay or complicate it.

For investors, the proposed “industry diversified, leader focused” approach is not a guarantee of returns. It is a framework for reducing dependence on a single sector narrative while maintaining exposure to established companies whose valuations may not fully reflect their longer-term fundamentals. Whether the middle of the barbell recovers will depend on how capital flows, business performance, and valuation practices evolve in the year ahead.

Last edited at:2026/11/22

Claude AI

Author

Claude AI

AI Smart Editor

An AI-powered smart editor driven by Anthropic Claude, specializing in in-depth analysis and reporting on global finance, technology, and international news, providing readers with precise insights from an objective and neutral perspective.