Do algorithms make stock markets wilder or wiser?

Many worry that lightning-speed machine processing makes the market more volatile and overreactive, but research finds that the reality may be quite the opposite

Featured faculty: Young Danqing
Written by Meg Chan

Trading floors are no longer exclusive to humans. Instead of shouting and yelling like a classic movie scene, silent computers have run in nearly all trading pits worldwide. From stocks, bonds, futures, commodities, foreign exchange, and even crypto, algorithmic trading now automates transactions faster than a blink of an eye.

According to Nasdaq and a 2025 Bloomberg report, algorithmic trading drives half of US stock volume and more than 60 per cent of European orders. Algorithms monitor live market data and execute buy or sell orders instantly when pre-programmed rules are met. However, critics argue that such advanced trading undermines market stability and transparency.

Some algorithms track the moves of large institutional investors. These investors have the resources and direct access to company executives to analyse stock markets deeply. Algorithms anticipating a massive stock trade from a major investor can then be programmed to place orders ahead of it.

To stay on top, institutional and large individual investors also deploy algorithms to track each other’s trading activity. Most algorithm-driven trading strategies focus on short-term price movements rather than fundamentals, making the market more volatile and harder for investors to value stocks accurately.

Nevertheless, properly monitored and regulated algorithmic trading can help new information be reflected in stock prices faster and improve overall market efficiency. The key is to find the true market signal and read between the lines.

Young Danqing, Professor (Teaching) in the School of Accountancy at the Chinese University of Hong Kong (CUHK) Business School, has identified a source of important industry-level information, and algorithmic trading can speed up the transfer of these valuable market signals.

“When a listed firm announces its earnings, algorithms quickly interpret the news as a hint about how other firms in the same industry might be doing,” says Professor Young. “Algorithmic trading helps information from firms that announce their earnings earlier to be incorporated into the stock prices of other firms, even before these firms release their own earnings.”

How algorithms help information transfer

The US securities regulator requires companies to report earnings publicly within 45 days after the fiscal quarter ends. In her new study, Algorithmic trading and intra‑industry information transfer, together with Zhang Xiaori of Guangdong University of Technology and Christine Jiang of Fudan University, Professor Young examines earnings announcements of more than 3,200 US-listed firms from 2012 to 2021.

They find that earnings announced earlier can affect stock prices of companies in the same industry that have yet to report, amplified by algorithmic discovery. When algorithmic trading activity increases, affected stock prices react more strongly to earlier earnings reports, moving up or down by about 39 to 66 per cent.

However, this isn’t an overreaction. Early earnings reports provide industry-specific information about changes in demand, supply chain pressures, and industry trends, all factors that are relevant for firms in the same sector. The key is sector exchange-traded funds (ETFs).

“When a listed firm announces its earnings, algorithms quickly interpret the news as a hint about how other firms in the same industry might be doing.”

Professor Young Danqing

An ETF is an investment fund that trades like a single share but holds a basket of companies. A sector ETF holds stocks of firms from one industry or a closely related group. For example, the Technology Select Sector SPDR Fund focuses on tech giants, software creators, and hardware makers, while the Financial Select Sector SPDR Fund tracks banks, investment firms, and insurance companies.

When a company in a sector ETF announces earnings, algorithms quickly respond by trading the sector ETF. The stock prices of other companies within the basket then also move, even if they have not yet announced their own earnings.

“Sector ETFs provide a natural channel for transmitting industry information as their constituent firms are relatively homogeneous,” Professor Young says. “Firms in the same sector are likely to be exposed to similar demand conditions, supply-chain developments, and other industry-wide shocks. For algorithms, that structure is efficient.”

Conversely, ETFs that track broad groups of securities, e.g., broad-market and total-market ETFs, dilute industry-specific signals as useful information gets averaged across many industries. The algorithms are not efficient in transferring information in these ETFs. However, this finding is only part of the picture.

The value of high-quality reports

Algorithmic trading does more when firms announcing early reports have historically similar profits and costs, and sell overlapping or comparable products. The benefits are even greater when earlier earnings reports are of higher quality.

This point has a practical edge since algorithms are data-sensitive. Overall, high-quality earnings reports have lower absolute accruals or estimated earnings amounts. More direct financial reports that reflect the true performance are easier for algorithms to interpret.

“A high-quality report is crucial not only for the reporting company but also for economically related firms. It helps investors assess other firms in the same industry more efficiently,” Professor Young adds. “The quality of corporate disclosure can shape how effectively algorithmic trading can find the true value of a company based on new information.”

Breaking the myth of volatility

Another criticism is that algorithms may exaggerate short-term volatility. Lightning-fast automated transactions amplify responses to early earnings reports and trigger sudden price jumps across related stocks in the same industry. Such overreaction gets corrected when the peer companies report their own earnings later.

Professor Young finds the opposite pattern. When algorithmic trading is more active after a firm announces its earnings, the stock prices of other firms in the same industry that have yet to announce earnings tend to reflect market information more accurately. When these firms eventually report their own earnings later, their stock prices are less likely to swing wildly.

In this case, more algorithmic trading helps the market process industry information more precisely, leading to more stable and reliable reactions. “Algorithmic trading can improve how accurately markets process information,” Professor Young says.

“Stocks traded by algorithms see greater price stability following earnings reports announced by other firms in the same industry. This suggests that algorithms help stock prices reflect the true value of industry news right away, rather than just causing big, temporary market swings.”

Finally, in more automated markets, disclosure quality becomes even more important because algorithms depend heavily on reliable, processable information. Regulators and policymakers should also look beyond trading speed or market disruption, and emphasise improving financial reporting, which allows algorithms to incorporate industry news more accurately.