The fluctuations in China’s A-share market have significantly detached from changes in economic fundamentals, evolving into a "policy-driven market" of speculation around policy expectations. The continuous and violent volatility of the capital market, alongside the concurrent occurrence of concentrated sell-offs by national-level industrial funds, batch cash-outs by listed company shareholders, floating investment losses on insurance capital equities, and intensified cross-border capital fluctuations, clearly indicates the expectation disorders caused by repeated policy adjustments.

This round of volatility began with the semiconductor industry support cycle. As industrial policies stepped up support for the domestic chip industry chain, massive amounts of capital flooded into the semiconductor track, rapidly elevating sector valuations. Market data showed that the CSI Semiconductor Industry Index rose continuously from April to June 2026, with a single-month gain of 24.73% in April, an increase of 14.83% in May, and a surge of 26.57% in June. By the end of June, the sector's price-to-earnings ratio (TTM) reached 161.68, placing it at 99.72% of its historical range since 2019, while the STAR 50 Index concurrently touched a P/E (TTM) of 258.97, resting at 99.93%. Valuations thus significantly detached from current earnings fundamentals.

Although policy-driven incentives generated a short-term market rally, fundamental issues such as supply-demand imbalances across the industry and technological bottlenecks remained largely unresolved. Between January and May 2026, China’s National Integrated Circuit Industry Investment Fund (the “Big Fund”) significantly reduced its holdings in semiconductor stocks, with cumulative disposals exceeding RMB 10 billion. As a state-backed investment vehicle entrusted with guiding the development of the semiconductor industry, its decision to take profits and exit ahead of the market essentially reflected the industry capital's assessment that short-term valuations had become severely overstretched, prompting it to capitalize on the window of policy-driven market enthusiasm. This move not only intensified direct selling pressure across the sector but also sent a clear signal that valuations may have peaked, prompting controlling shareholders of listed companies and institutional investors to follow suit with further share disposals. Among them, a number of insurance funds had entered the semiconductor sector at elevated prices, driven by policy expectations, only to suffer substantial unrealized losses and subsequently be forced to reduce their positions and cut losses, further accelerating the market's downward trajectory. As the market euphoria faded, sector valuations rapidly declined, dragging the broader stock market indices into a downward trend. Meanwhile, risk governance in the capital market continued to tighten, with a batch of major cases exposed centrally, further suppressing market risk appetite and continuously increasing downward pressure on the index. This was because investors and the general market universally anticipated that the existence of similar problems might involve a much broader scope, rendering the risks uncontrollable. In July, the CSI Semiconductor Industry Index fell by 27.25% in a single month, and a large volume of funds that had chased the highs earlier exited at a loss. This adjustment also coincided with external factors such as the downturn of the overseas semiconductor sector and peripheral macroeconomic disturbances. The rapid pullback of the entire sector precisely corroborated that the expectation attribute of this market wave was far stronger than its performance attribute.

Following the cooling of the semiconductor rally, speculative capital rapidly shifted its focus to the artificial intelligence (AI) sector in an attempt to capture the next wave of policy-driven gains. At the industrial level, the government introduced a series of supportive policies to encourage AI technology research and development as well as its commercial and industrial adoption. Market capital subsequently flowed heavily into hardware segments such as computing power infrastructure, optical modules and memory chips, driving a rapid rally in AI-related stocks that became increasingly detached from underlying fundamentals. In the first half of the year, total trading value across the market reached a record RMB 317.53 trillion for a six-month period. Large volumes of short-term quantitative trading funds, speculative capital and trend-following investors poured into computing hardware stocks, while long-term institutional investors, particularly insurance funds, were also compelled to increase their positions amid the prevailing policy-driven market enthusiasm, further pushing up sector valuations.

Just as market sentiment was fully ignited by policy dividends, China’s policy direction experienced a rapid shift. Regulatory documents were promptly issued, with the Interim Measures for the Management of Anthropomorphic AI Interactive Services taking effect on July 15, 2026. At the same time, public consultations were launched on the labelling of AI-generated content and controls on technology exports. This marked a shift in policy priorities from accelerating development towards a more balanced approach that places equal emphasis on development and security, with greater attention to mitigating employment disruption, content-related risks and the potential spillover of critical technologies. Although these measures were intended to strengthen industry governance, market participants interpreted them as signals of diminishing policy support and increasingly restrictive growth prospects for the sector. This was compounded by a wave of shareholding reduction plans announced by major shareholders of listed AI companies, as industrial capital took advantage of elevated valuations to cash out on a large scale, directly triggering a concentrated sell-off in high-flying AI and computing-power stocks. Cross-border capital, including northbound funds, also accelerated its withdrawal from technology and growth-oriented sectors amidst policy uncertainty, causing overall market risk preference to deteriorate rapidly and pushing the technology and growth sectors into another deep correction.

Within the short span of half a year, the market successively experienced two rounds of industrial theme hype in semiconductors and AI. The initiation of each market wave was ignited by industrial encouragement policies, while the collapse of each wave was the result of the resonance of multiple factors, including changes in policy expectations, reductions by industrial capital, and external market disturbances. The entire market constantly chased new policy trends, and the capital pricing logic no longer anchored on corporate earnings or long-term industrial prosperity, but instead bet on the landing points of the next round of policy-favorable conditions. Each round of stimulus spawned a short-term market wave, while each round of policy tightening combined with other disturbances triggered a rapid decline. Amidst repeated policy switches, investors grew exhausted coping with policy changes, long-term capital dared not make long-term allocations, and the market was left with only short-term speculative capital trading.

The immediate consequence has been the persistence of two defining characteristics in China's A-share market， namely increasingly short-term, speculative capital flows and sustained capital outflows. Margin financing and securities lending funds have experienced substantial inflows and outflows, fluctuations in northbound capital flows have become significantly more pronounced, and share disposals by industrial capital have remained elevated. Meanwhile, the proportion of long-term equity capital from insurance companies, pension funds and other institutional investors has increased only gradually, with these investors even recording periodic net outflows amid repeated market volatility. Against the backdrop of successive rounds of policy-driven market fluctuations, major stock indices have continued to weaken. Rather than fully fulfilling its role in facilitating corporate financing and supporting the real economy, the capital market has instead faced mounting capital outflow pressures exacerbated by policy uncertainty and destabilized market expectations. Moreover, this recurring policy cycle of “encouragement, reversal and renewed encouragement” has not only been unable to reverse the economic downturn but has also inflicted a dual negative impact on the real economy and the capital market. Each round of policy stimulus has generated little more than a short-lived market rally, without establishing a solid foundation for sustainable industrial development. Conversely, every abrupt policy reversal has disrupted industrial investment planning. Companies that had just expanded production capacity in response to supportive policies subsequently found themselves facing new regulatory constraints, forcing them to suspend or shelve earlier investments and further weakening business confidence and operating expectations. In the capital market, this has translated into declining confidence among industrial capital in the sustainability of policy incentives, with companies increasingly inclined to capitalize on policy-driven market peaks to reduce holdings and realize gains rather than commit resources to long-term research and development or capacity expansion. This repeated cycle of policy reversals has done little to restore the economy's endogenous growth momentum. Instead, it has continuously eroded market confidence, ultimately creating a vicious cycle in which increasingly frequent policy interventions coincide with deteriorating economic fundamentals and persistent capital outflows from the market.

At a deeper level, similar divergence between short-term and long-term objectives can be observed not only in technology industry policies but also in social governance initiatives such as Common Prosperity. Long-term objectives require gradual, steady, and systematic implementation. However, short-term pressures to stabilize economic growth and safeguard people's livelihoods can create incentives for policymakers to pursue results too hastily, forcing long-term reform objectives into an accelerated implementation timetable. When policy measures are introduced too aggressively and at an excessive pace, they disrupt the normal functioning of markets and society, amplify short-term volatility and, paradoxically, prolong the overall reform implementation cycle. The experience of China's private tutoring industry over the past few years provides a clear lesson in this regard. During the initial phase of industry regulation in 2021, in an effort to rapidly advance the long-term objectives of promoting educational equality and reducing the financial burden of household education expenditure, policymakers introduced comprehensive regulatory requirements across the entire sector in a single policy push. This effectively dismantled the industry's existing commercial ecosystem, leaving hundreds of thousands of industry workers facing career transitions within a short period. Subsequent regulatory adjustments were then required, with the initial blanket restrictions gradually giving way to a more differentiated regulatory framework. As a result, the overall reform implementation cycle was extended by nearly five years compared with the original expectations, while the associated social costs far exceeded initial assessments. A similar pattern can be observed in Zhejiang, China's designated demonstration province for Common Prosperity. Adjustments to the provincial Party Committee's No. 1 Document in 2026 clearly reflected an effort to recalibrate the pace of implementation. Rather than continuing to pursue several short-term quantitative catch-up targets proposed during the demonstration zone's initial launch, the document shifted its policy focus towards urban-rural integration, emphasizing gradual progress in addressing three structural disparities, which are urban-rural, regional, and income inequalities. In essence, this represented a correction to the campaign-style implementation practices adopted by certain counties and districts in previous years under pressure from short-term performance assessments, including, in some cases, the implicit imposition of targets for village collective income and charitable donations. The underlying problem, however, remains unchanged. When short-term policy interventions pursue immediate results too aggressively, long-term strategic objectives that require gradual implementation are forced into annual policy task lists. This inevitably leads to repeated shifts and inconsistencies in policy implementation.

ANBOUND has long maintained that the implementation of economic policies must adhere to the principle of prudence. The essence of this principle lies in clearly defining the boundaries between government policy and market forces, as well as distinguishing long-term strategic objectives from short-term economic regulation. Kung Chan emphasizes that policymakers must abandon the conception that government policies can control everything and fully substitute for the functions of the market. Administrative intervention should only serve to address market failures, not replace the market's role in allocating resources. Short-term countercyclical adjustments must remain strictly subordinate to long-term development strategies. Long-term policy orientations should avoid being arbitrarily altered in pursuit of short-term economic stabilization or growth targets. This is essential to preventing frequent and abrupt policy reversals and maintaining stable long-term market expectations.

Final analysis conclusion:

The market has its own cycles and operating laws. In particular, the rise and fall of China’s industrial prosperity and the alternation of investment heat and cold are themselves the process of spontaneous market clearance. The duty of policy is to improve institutional rules, remedy infrastructure shortcomings, and prevent systemic risks, rather than frequently intervening in industrial prosperity, artificially manufacturing industrial trends, and forcibly driving up investment heat. Looking back at the recent fluctuations of the tech sector, from national-level industrial funds taking the lead in reducing holdings to realize gains, to batch cash-outs by listed company shareholders, and further to insurance capital exiting with floating losses and cross-border capital fluctuating continuously, the problem behind these various issues lies not in the wrong direction of industrial upgrade, but in excessive policy intervention intensity, overly rapid pacing, and external market disturbances amplifying fluctuations. Between relaxation and tightening, it is easy to bypass the stage of spontaneous market adjustment and alienate the industrial cycle into a policy expectation cycle, which ultimately will only overdraft market confidence and exacerbate capital outflow.

_____________

He Yan is a researcher at ANBOUND, an independent think tank.