CUHK Business School Research Looks on the Limitations of Using Artificial Intelligence to Pick Stocks | India Education,Education News India,Education News

It’s been referred to as the holy grail of finance. Is it doable to harness the promise of synthetic intelligence to earn a living buying and selling shares? Many have tried with various levels of success. For instance, BlackRock, the world’s largest cash supervisor, has mentioned its Artificial Intelligence (AI) algorithms have constantly crushed portfolios managed by human inventory pickers. However, a current analysis research by The Chinese University of Hong Kong (CUHK) reveals that the effectiveness of machine studying strategies might require a re-assessment.


The research, titled “Machine Learning versus Economic Restrictions: Evidence from Stock Return Predictability”, analysed a big pattern of U.S. shares between 1987 and 2017. Using three well-established deep-learning strategies, researchers have been in a position to generate a month-to-month value-weighted risk-adjusted return of as a lot as 0.75 % to 1.87 %, reflecting the success of machine studying in producing a superior payoff. However, the researchers discovered that this efficiency would attenuate if the machine studying algorithms have been restricted to working with shares that have been comparatively straightforward and low cost to commerce.


“We find that the return predictability of deep learning methods weakens considerably in the presence of standard economic restrictions in empirical finance, such as excluding microcaps or distressed firms,” says Si Cheng, Assistant Professor at CUHK Business School’s Department of Finance and one of many research’s authors.


Disappearing Returns

Prof. Cheng, alongside together with her collaborators Prof. Doron Avramov at IDC Herzliya and Lior Metzker, a analysis scholar at Hebrew University of Jerusalem, discovered the portfolio payoff declined by 62 % when excluding microcaps — shares which might be tough to commerce due to their small market capitalisations, 68 % decrease when excluding non-rated corporations — shares which don’t obtain Standard & Poor’s long-term issuer credit standing, and 80 % decrease excluding distressed corporations round credit standing downgrades.


According to the research, machine learning-based buying and selling methods are extra worthwhile during times when arbitrage turns into harder, corresponding to when there’s excessive investor sentiment, excessive market volatility, and low market liquidity.


One caveat of the machine-learning primarily based methods highlighted by the research is excessive transaction prices. “Machine learning methods require high turnover and taking extreme stock positions. An average investor would struggle to achieve meaningful alpha after taking transaction costs into account,” she says, including, nonetheless, that this discovering didn’t indicate that machine learning-based methods are unprofitable for all merchants.


“Instead, we show that machine learning methods studied here would struggle to achieve statistically and economically meaningful risk-adjusted performance in the presence of reasonable transaction costs. Investors thus should adjust their expectations of the potential net-of-fee performance,” says Prof. Cheng.


The Future of Machine Learning

“However, our findings should not be taken as evidence against applying machine learning techniques in quantitative investing,” Prof. Cheng explains. “On the contrary, machine learning-based trading strategies hold considerable promise for asset management.” For occasion, they’ve the aptitude to course of and mix a number of weak inventory buying and selling alerts into significant info that might kind the premise for a coherent buying and selling technique.


Machine learning-based methods show much less draw back threat and proceed to generate optimistic payoff throughout disaster intervals. The research discovered that in a number of main market downturns, such because the 1987 market crash, the Russian default, the burst of the tech bubble, and the current monetary disaster, the perfect machine-learning funding methodology generated a month-to-month value-weighted return of three.56 %, excluding microcaps, whereas the market return got here in at a destructive 6.91 % throughout the identical interval.


Prof. Cheng says that the profitability of buying and selling methods primarily based on figuring out particular person inventory market anomalies — shares whose behaviour run counter to traditional capital market pricing principle predictions — is primarily pushed by quick positions and is disappearing in recent times. However, machine-learning primarily based methods are extra worthwhile in lengthy positions and stay viable within the post-2001 interval.


“This could be particularly valuable for real-time trading, risk management, and long-only institutions. In addition, machine learning methods are more likely to specialise in stock picking than industry rotation,” Prof. Cheng provides, referring to technique which seeks to capitalise on the subsequent stage of financial cycles by shifting funds from one business to the subsequent.


The research is the primary to offer large-scale proof on the financial significance of machine studying strategies, she provides.


“The collective evidence shows that most machine learning techniques face the usual challenge of cross-sectional return predictability, and the anomalous return patterns are concentrated in difficult-to-arbitrage stocks and during episodes of high limits to arbitrage,” Prof. Cheng says. “Therefore, even though machine learning offers unprecedented opportunities to shape our understanding of asset pricing formulations, it is important to consider the common economic restrictions in assessing the success of newly developed methods, and confirm the external validity of machine learning models before applying them to different settings.”



Avramov, Doron and Cheng, Si and Metzker, Lior, Machine Learning versus Economic Restrictions: Evidence from Stock Return Predictability (April 5, 2020). Available at SSRN: or


This article was first printed within the China Business Knowledge (CBK) web site by CUHK Business School:


Please enter your comment!
Please enter your name here