Stockholm (HedgeNordic) – There is a new wave of hedge funds that have their analysis, security selection and trading processes controlled and steered by machine learning algorithms. But will the latest rush to embrace artificial intelligence (AI) disrupt the hedge fund industry? Whereas some believe that AI will unsettle the asset management industry, others consider that the impact of AI-powered automation will be limited to certain areas of the industry that rely on big data and systematic, fast-paced trading.

The use of AI in the Nordic hedge fund industry has been accelerating quite dramatically in recent months and years. Calculo Evolution Fund, Minastir Currency Fund, Lynx, Volt Diversified Alpha Fund, and Innolab Capital are just several vehicles in the industry that make use of AI as part of their strategies. And more AI-assisted hedge funds are set to come on the market later this year, at least one from Lynx and another one from Innolab.

HedgeNordic talked to Michael Halling, a Professor of Finance who also holds a PhD in Computer Science, and several players in the Nordic hedge fund industry in an attempt to formulate a conclusion on the impact of AI on the asset management industry.

Thoughts on How AI Will Change the Industry

Michael Halling states that machine learning algorithms are “a new kind of quantitative models the industry applies,” adding that artificial intelligence “will not fundamentally change the industry in the sense that asset managers will be gone and robots will run everything.” Halling, who is not new to the frontier between finance and technology, argues that machine learning algorithms are “a new technique to look at data and analyse data.”

Cliff Asness, the co-founder of quantitative investment firm AQR, corroborates Halling’s opinion, having previously said that “machine learning is basically a way to find more and more patterns” in a more efficient manner by processing large amounts of data. Halling argues that “AI is not going to be a revolution,” but he reckons that machine learning algorithms “are pretty good at pattern recognition and finding deviations from standard patterns.” This implies jobs that “process a lot of data to find patterns are obviously at risk due to the emergence of these machine learning techniques.”

According to Patrik Säfvenblad, chief investment officer at Volt Capital Management, “machine learning is a tool that provides two main benefits: scalability and adaptability.” Explaining the notion of scalability, Säfvenblad says that machine learning allows the investment process to be “scaled to cover many inputs, signals, and instruments.” At the same time, adaptability stemming from the use of machine learning enables the investment process to react more quickly to market regime changes over time. “This is particularly important in macro trading, where market regimes shift regularly,” explains Säfvenblad, adding that “these shifts are often difficult to handle for discretionary managers as previously-successful models might fail.”

The just-mentioned benefits are particularly relevant for systematic strategies relying on big data. “In my expectation, machine learning will be used to improve the scalability and adaptability primarily of systematic strategies,” Säfvenblad tells HedgeNordic. He expects trading-oriented strategies such as global macro, short- term trading, statistical arbitrage, and quantitative equity market-neutral to benefit from the use of artificial intelligence. Machine learning algorithms may have “some applications for discretionary managers,” reckons Säfvenblad, but “the benefit is mostly time-saving in nature. It will not really impact the underlying strategy.”

The Implementation of AI: Gradual Process

Although the discussions around artificial intelligence have been enjoying a significant resurgence in recent months, Säfvenblad considers that “any change to the industry will be gradual as market participants learn where machine learning adds value and where it does not.” Thomas Jacobson, the founder and portfolio manager of AI-assisted Minastir AI Currency Fund, believes that “moving forward, there will be a natural process where companies start experimenting with AI and try to implement AI in their existing systems to see how that can work out.” Whereas AI helps the currency- focused trading approach at Minastir, “for others, AI does not really work” according to Jacobson. “Others may not find how AI could add value to them.”

Peter Smedegaard, who is planning the launch of a fully autonomous AI market-neutral fund in the second half of 2019, believes artificial intelligence has more transformative potential in the asset management industry. “I see AI transforming most of the hedge fund and asset management industry gradually but steadily,” Smedegaard tells HedgeNordic. “We are aware of the rise of artificial intelligence systems that meet and exceed human abilities,” says the CEO of Innolab Capital, arguing that the rise of AI stems from “the exponential increases in data storage and computing power over the past couple of decades.”

Michael Halling, on the other hand, does not completely understand why artificial intelligence receives so much attention these days. “Currently there is a lot of attention to AI, but in the end, it is not so obvious why the attention is coming now,” says Halling, arguing that “most of these techniques in machine learning have been around for at least 20 years.” Although he acknowledges the argument that the discussions around AI “are booming now because data is more available; that could be the case,” but equally important, “many text archives, news archives, balance sheet archives have been around for many years.”

The Role of Traditional Analysts at Risk

Leaving aside the question of why AI is receiving so much attention now, Smedegaard says “artificial intelligence has already created waves in the industry, as assets managers find out that the ability to extract value from big data is going to be a key differentiator.” He also reckons that “AI and machine learning will remain and will undoubtedly replace the roles of many traditional analysts,” arguing that AI-powered strategies “adapt quickly to rapidly changing market conditions, whereas humans do not.” Whereas machine learning and artificial intelligence systems may indeed be able to respond quickly to changing market conditions, Halling argues that “for standard ML and AI approaches, it is actually doubtful whether they are able to respond quickly to changes in the underlying data structures.”

On the debate about the role of traditional analysts, Halling suggests that the role of most analysts may have been under question even without machine learning. “Whether you need artificial intelligence or machine learning to make that claim, the question surrounding the role of analysts has already been researched in the academia,” says Halling. “I am not an expert on this literature, but the bottom line is that only some star analysts appear to have some skill in predicting, but not the majority.”

Smedegaard, meanwhile, already notices the effects of AI playing out in the asset management industry. “The rise of robo advisors is the biggest example of how AI is enabling changes in investment,” he tells HedgeNordic. “Robo advisors and fully autonomous funds now perform all the functions of a financial advisor and enable passive investing at different price rates.”

Smedegaard also considers that investment advisors who learn to integrate AI into their decision-making are in a better position than competitors to succeed in this new investment landscape. “As AI continues to grow, it will also start moving into active investing because of the cheaper computational power and advances in software, hardware, and data storage,” argues Smedegaard.

Photo by Hitesh Choudhary on Unsplash