Algorithmic trading has become a significant force in the A-Share market.
At a forum with investors chaired by the securities regulator on July 20, one of the key suggestions was to "standardize the development of quantitative trading and AI applications." Subsequent meetings with listed companies and industry professionals also saw calls to "further regulate algorithmic trading practices."
For a long time, investors and researchers have struggled to form a comprehensive view of how quantitative trading operates and its specific impact on the A-Share market. Many retail investors believe algorithmic trading "sells in a coordinated manner" to amplify market volatility, expressing feelings of "unfairness" and being "taken advantage of."
Investors also have many questions. For example, is it only algorithmic trading that can execute "T+0" trades while retail investors are limited to "T+1"? Does algorithmic trading frequently place and cancel orders to mislead ordinary investors, or use securities lending and stock index futures to short the market, exacerbating declines? Does algorithmic trading ignore fundamentals, simply executing trades based on keywords from news sentiment?
Recently, the Private Securities Investment Fund Committee of the Asset Management Association of China (AMAC) provided a comprehensive analysis of these issues surrounding algorithmic trading in the A-Share market, including its fairness, trading volume, market impact, trading rules, and regulatory measures.
Does Algorithmic Trading Lead to Unfair Treatment of Investors?
For many retail and even some institutional investors, algorithmic trading feels like a "black box." Although they don't know the specific mechanisms, they believe algorithmic traders have more sophisticated tools, faster execution, higher frequency, and consequently, greater profits.
For instance, many investors think algorithmic trading can execute "T+0" trades, while retail investors can only do "T+1."
"The current exchange trading rules are fair to all types of investors. Regardless of the investor type, only T+1 trading is allowed in the stock market, and this applies to algorithmic traders as well," the committee stated. It noted that the regulator has already restricted the use of securities lending for disguised T+0 trades since 2024. While some investors holding a stock on T-1 day can buy more on T day and sell their T-1 holdings, achieving a so-called "T+0," this is a normal trade available to all investors and is not true T+0.
In recent years, the apparent outperformance of algorithmic trading has led to a sense of "unfairness" and "being exploited" among retail investors. This is primarily due to the information and technology advantages of algorithmic trading, not differences in trading rules or regulatory frameworks.
The committee explained that the current advantage of algorithmic trading lies in its ability to collect vast amounts of market information and data, summarise statistical patterns into automated strategies, and execute orders automatically. In contrast, retail investors mainly rely on manual monitoring, subjective decision-making based on limited personal information and price changes, and manual order placement. This creates a gap in information processing, decision-making efficiency, and execution speed, especially for short-term or multi-stock trades.
However, recognising the domestic market where retail investors are the majority, regulators have actively promoted the oversight of algorithmic trading from the perspective of fairness and standardisation. For example, by strengthening transaction monitoring, they have set specific indicators for "abnormally high instantaneous order submission rates" and "frequent instantaneous order cancellations." Excessive order submissions per second and frequent rapid cancellations are now classified as abnormal trading, establishing a regulatory "red line" to encourage lower frequency and speed.
Furthermore, since the start of the year, regulators have also weakened the speed advantage of algorithmic trading by regulating hosting services and trading units (gateways). For instance, exchanges have required brokers to cancel dedicated independent trading gateways for single clients and move client hosting servers out of exchange data centres.
Regarding high-frequency trading, the A-Share market defines it as "a single account submitting or cancelling a combined total of 300 or more orders per second, or a total of 20,000 or more orders per day." The committee noted that in actual trading, the vast majority of algorithmic trading peak submissions are significantly below 300 orders per second, and there has been no deliberate attempt to submit or cancel orders at near-threshold rates like 299 orders per second.
"There have been rumours online that the US SEC imposes specific limits on the frequency or speed of algorithmic or high-frequency trading, such as 'no more than 15 orders per second,' 'a daily cancellation rate not exceeding 15%,' or 'orders must sit for at least 50 milliseconds,'" the committee said. After verification, no such rules exist from the US SEC or exchanges. In the US market, hedge funds, index funds, and arbitrage firms often have order submission and cancellation rates reaching tens of thousands per second during active trading without regulatory restriction.
Additionally, the US market does not require quantitative firms to disclose their order books or specific trading details in real-time. The publicly available order book is a summary of overall market information and does not distinguish orders by investor type. Large traders, including major quantitative firms, are also not required to submit core strategy logic or risk control details to regulators.
Does Algorithmic Trading "Sell Off" and Act as a Short Seller?
Regarding the market's performance since July, a key question is the role of algorithmic trading. Did it execute a "coordinated sell-off"? Did it short the market?
"After checking with leading brokers and quantitative private funds, on several days of significant market decline last week, many quantitative firms were net buyers, not net sellers. Some even had substantial net purchases. There was no 'coordinated sell-off,'" the committee stated. Looking at the operating models of mainstream quantitative products like index enhancement, long-only quantitative, and market-neutral strategies, they are typically fully or near-fully invested, with trades being adjustments between different stocks. A "coordinated sell-off" in a short period would not enhance returns but would likely cause losses, so quantitative firms themselves lack the incentive for such an action.
Domestic quantitative traders are mainly institutional investors, including foreign quantitative firms, domestic quantitative private funds, and broker proprietary trading desks, along with some individual investors. The primary profit model for quantitative trading involves analysing historical data to develop strategy models and exploit a slight "winning edge" across numerous trades.
"For quantitative private funds, they typically maintain a high position of over 90% regardless of bull or bear markets, and market declines often trigger automatic buying programs." Regarding the accusation of short-selling to depress the market, the committee stated that based on the use of related instruments, quantitative private funds have not used securities lending or stock index futures to suppress the market.
"Furthermore, since February 2024, regulators have clearly prohibited securities firms from providing securities lending to investors using it for disguised T+0 trades," the committee added.
Regulators also closely monitor extreme situations. The committee noted that the Shanghai, Shenzhen, and Beijing stock exchanges have set specific monitoring indicators for "large-volume trades in a short time," classifying actions that trigger significant index fluctuations through rapid, massive trading as abnormal. Once triggered, exchanges promptly take self-regulatory measures to mitigate the negative impact of algorithmic trading on market stability.
What Should Ordinary Investors Do About the Trend of Algorithmic Trading?
Algorithmic trading is a product of the combination of modern information technology and capital markets, and the recent widespread application of AI technology has further empowered it.
From a global perspective, the standardised development of algorithmic trading is a trend in the professional upgrade of capital markets. As technologies like AI rapidly advance, investment techniques will inevitably evolve.
For ordinary investors, returning to the fundamentals of investing, and cultivating a rational, long-term, and professional investment philosophy, may be a better choice.
Rational investing often requires reducing frequent trading, chasing rises and selling on dips, and panicking. It involves managing one's investment pace and cultivating the right mindset to enhance core stock market investment skills.
Long-term investing focuses on "leveraging strengths and avoiding weaknesses." Ordinary investors can use their natural advantages of flexible fund cycles and no short-term performance pressure to focus on the growth value of high-quality listed companies. Relying on a company's long-term growth to generate stable returns can also help avoid short-term volatility and navigate market cycles.
Professional investing involves leveraging the power of professional institutions for scientific asset allocation. Compared to individuals, professional institutions have advantages in research, modelling, and risk management. Investors can use this expertise to improve the scientific nature of their asset allocation and diversify risk through multiple strategies, such as investing in public funds with low entry barriers, high transparency, and standardised regulations.
Long-term value investing and short-term trading are two common market approaches, and they are not mutually exclusive.
Algorithmic trading, by submitting orders based on market conditions, can objectively improve market depth, narrow bid-ask spreads, and enhance liquidity. This benefits other investors, including long-term capital, by allowing them to trade at lower costs and facilitating smoother entry and exit channels. When the market experiences significant volatility, algorithmic trading can often trigger counter-trend mechanisms, which can help to some extent in curbing excessive price swings.
Research has also found no significant correlation between algorithmic trading and the medium- to long-term trajectory of the stock market. A considerable portion of quantitative strategy models includes factors like fundamentals and dividends, and some strategies are primarily fundamental-based. Products like quantitative private fund index enhancement and active quantitative stock selection typically maintain high positions, participating in the market continuously. By holding a basket of assets over the long term, dynamically optimising the portfolio structure, and managing risk exposure, they aim to achieve relatively stable long-term returns.
Regulatory Oversight is Tight, with Continuous Crackdowns on Manipulation and Unfair Trading
Currently, both algorithmic and non-algorithmic trading are subject to strict legal and regulatory requirements. The Securities Law explicitly prohibits market manipulation through "submitting and cancelling orders frequently or in large quantities without the intent to execute."
Recently, the securities regulator stated it would continue to refine the regulatory framework for programmatic trading, with a greater emphasis on fairness and standardisation, strengthening targeted supervision, and effectively preventing the abuse of technological advantages. It is expected that regulators will gradually introduce more policies to optimise regulatory arrangements and further enhance the fairness of market trading.
To address investor concerns about algorithmic trading using frequent order placements and cancellations to mislead ordinary investors or engage in market manipulation, exchanges have set up specific abnormal trading monitoring indicators to limit the submission frequency and cancellation rate of algorithmic trading.
According to reports, in actual A-Share market trading, algorithmic trading is typically very diversified, often involving over 1,000 or even thousands of individual stocks. The trading volume, number of order submissions, and single order sizes in any individual stock are relatively small. Frequent order cancellations and reverse trades are not common practices for algorithmic trading in the A-Share market. Compared to the average cancellation rate of over 90% in the US market and around 70% in Japan, the A-Share market has explicitly set a 50% cancellation rate as a monitoring threshold for abnormal trading behaviour, making its regulatory standards significantly stricter than those in overseas markets. Exchanges will continue to study and optimise monitoring indicators to improve regulatory effectiveness and maintain market stability.
"If any individual investor uses programmatic tools to place and cancel a large number of orders, misleading other investors, and then engages in market manipulation through reverse trades, they will face severe legal penalties," the committee stated. Regulators are cracking down on illegal activities involving algorithmic trading, including market manipulation, disrupting fair and orderly trading, and harming investor rights. They are also continuously optimising screening mechanisms based on market conditions to crack down on serious manipulation cases. In May 2025, the regulator fined an individual investor, Mr. Peng, 400 million yuan for using a programmatic trading tool to manipulate stock prices through false orders, intraday price ramping, and limit-up order placement.
Furthermore, regulators maintain a high-pressure stance on new types of illegal stock recommendations disguised as "quantitative investment." Recently, the regulator, in conjunction with public security authorities, launched a series of targeted crackdowns on "quantitative investment" related illegal stock recommendation activities, investigating several cases using "quantitative investment" as a gimmick.
Comments