Artificial intelligence – investment decisions without intuition

3 Min.

Artificial intelligence has become an integral part of our world. But can this technology also be used profitably in the financial sector, and how does machine decision-making differ from that of humans and rule-based approaches?

Stefan Fröhlich

In January 1848, the first gold nugget was discovered in California. The press in New York did not report this until seven months later. Different conditions prevail in today’s information age; important information is already available worldwide in the shortest possible time. Investors are flooded with information about companies, industrial trends, regions and economies. The key question today is how can we extract the really important information from the vast amount of readily available data and use it profitably.

Gut decisions – the unconscious power of intuition

When making decisions, investors consider company analyses and valuation models, but they usually listen to their own intuition for the final investment decision. A gut decision is not divine inspiration, but a rational, unconscious decision by the mind. It is based on less, but relevant information. Gerd Gigerenzer1, psychologist and leading researcher in decision-making and heuristics, found that people overestimate unimportant information when they have access to lots of decision-making parameters. According to his research, better results are achieved when decisions are based on only a few, truly relevant criteria and unimportant details are neglected, which is just what our gut feeling intuitively does. However, there is a risk of premature conclusions being drawn for complex issues that cannot be simplified with basic rules of thumb.

Rules-based decisions – the rational choice

Benjamin Franklin, one of the founding fathers of the United States, promoted the balance sheet method in the 18th century. In this systematic, rules-based decision-making process, the most important influencing factors are determined by empirical analysis and weighted according to their relevance. Similar to unconscious gut decisions, less important criteria are ignored. Conversely, however, the decision-making is transparent and the criteria and their weighting are known, while humans are not aware of how an intuition-based decision arises.

Artificial intelligence – the Swiss army knife for handling huge amounts of data

Andrew Ng2, one of the most renowned experts in the field of artificial intelligence, says: «Artificial intelligence is the new electricity.» Whether unlocking an iPhone using facial recognition, googling or driving around in a Tesla, machine learning has become an integral part of our world. These machine processes achieved a major breakthrough in the fields of «big data» around 2010, i.e. applications with very large amounts of data that people can no longer process themselves. The algorithms learn independently from data and are able to identify patterns in complex systems.

In recent years, the topic of artificial intelligence has also become increasingly important in the financial sector. The industry has recognised that this technology opens up new opportunities and business areas and that the know-how in this area needs to be expanded in order to remain competitive in the future.

For investment decisions, machine learning algorithms can be trained with all the quantifiable information that is also available to a human investor. The process then uncovers hidden correlations between company key figures and future returns without human intervention. Finally, return forecasts are generated on the basis of the prediction models.

At present, there are only a few public investment funds whose strategy is based on machine learning algorithms. This will change in the coming years. US financial institutions saw the potential of artificial intelligence earlier, and European banks are lagging behind.

Human versus machine

Which method delivers better investment decisions? All three approaches can be justified. Human intuition has proven itself over thousands of years and allows rapid decisions to be made even in unusual situations. Rules-based approaches are transparent and it is easy to assess how the model will behave in a scenario. With artificial intelligence, a powerful, data-driven decision-making method is now available. These machine learning algorithms provide promising return predictions because they have a perfect memory compared to humans, so they don't forget anything, can accurately filter for the relevant criteria and process huge amounts of data quickly. In addition, these models generate returns that are less correlated with traditional investment strategies. Strategies based on artificial intelligence are therefore suitable as additional investment components.

Financial markets are among the most complex systems in the world, comparable to chaotic weather systems. Return forecasts remain difficult. For this reason, a broad-based investment approach that takes advantage of the three decision-making methods is promising.

1Gerd Gigerenzer, Gut Feelings – The Intelligence of the Unconscious, 2008
2Andrew Ng, Chief Scientist at Google and Baidu, professor at Stanford University, co-founder of the Coursera learning platform

 

Categories

Equity