Artificial intelligence and web scanning to uncover corporate secrets: Robert Gottliebsen
Originally printed in The Australian Newspaper – 9th August 2022
Sweeping the world is an investment revolution based on using artificial intelligence and web scanning to determine when chief executives are concealing the truth about current trading and other issues.
When about 10 years ago the so called “quantitative” or “alternative” computer driven investment strategies emerged on a small scale, they were loudly scoffed by traditional fundamental analysts. But now managers of hundreds of billions of dollars are using these methods to enhance fundamental analysis.
“Quantitative” computer driven investment strategies are now taking off in Australia and our local CEOs need to understand just how they will be caught out when they fail to reveal, in their market briefings, what is really happening to their enterprises.
One method now being used to get to the facts sees artificial intelligence being harnessed to analyse the real meaning of particular words a CEO uses, particularly in question-and-answer sessions with fundamental analysts.
Late last week, I was introduced to these new evaluation methods by Tarun Gupta who is Invesco’s Global Head of Investment Technology for Quantitative Strategies. He focuses on innovative research to enhance these new methods.
Some $20bn of the global giant’s investment portfolios have now been specifically devoted to this method of research. I was spellbound as Gupta explained how the “alternative” research methods are used.
But first let’s remind ourselves of the techniques fundamental investment managers have been using for countless decades.
The analysts listen carefully to the briefings from the company and prepare estimates of the likely earnings in the coming year or two plus other factors impacting the company.
They normally show their profit estimates to the corporate chief financial officer who tells them if they too high or too low.
Most companies try to keep the estimates on the low side so they can pleasantly surprise the market with better-than-expected results. It’s an antiquated system that gets around the insider trading rules and the reluctance of boards and CEOs to share with shareholders exactly what is currently happening in the company.
The methods that are being incorporated into the alternative information gathering can now be used on a wide scale because of artificial intelligence and today’s computer systems.
The computers analyse the words used by CEOs in question-and-answer sessions with analysts. By comparing current and past remarks and later profit performances the computers discover that certain phrases and words used by a CEO indicate future good or bad times. This profiling of remarks of CEOs has proved very successful in discovering the current trends in trading that are being concealed.
But the alternative methods cover a much wider canvass so many strands can be linked.
For example, in the US and UK it is possible to discover up to date credit card sales of particular enterprises which can reveal sales trends well before directors announce them.
Similarly the web can be scanned for data on a company‘s performance. Sometimes social media can reveal a particular product is either not being well received by the market or has become a market darling.
Other areas that can produce information that is not available in shareholder briefings are:
- Export activity from analysing bills of lading (shipping) data
- Foot traffic data to predict economic activities and macro trends
- Job posting data for predicting company growth
One area that is proving particularly difficult to gain real facts is in evaluating Environmental Social and Governance performance. The majority of companies make long ESG statements that please the analysts and markets but those statements are not always duplicated in practice.
In due course, the alternative strategists and their computers will devise ways to monitor a company’s ESG actual performance with what it tells shareholders, customers and staff.
A number of investment houses are now specialising in this “alternative” or “quantitative” method of corporate analysis which they sell to big investment houses who use data from the outside specialists as an adjunct to (or substitute for) their own “alternative” data.
Of course the data needs to be matched against the “fundamental” analysis. The corporate share price and the “fundamental” analysts may have already discovered the trends revealed by the “alternative” data. The alternative methods become very powerful when they discover a trend (either up or down) that neither the fundamental analysts nor the market has appreciated.
At some point the penny will drop with global boards that their inner secrets about what is actually happening in their business are in fact being made public by these new investment tools. They may even begin to share with shareholders the real facts about current trading in their enterprise.
But directors don’t like bad news and are frightened to be too optimistic in case there is a sudden reverse. But the computer systems and artificial intelligence are changing the game. Directors will need to be much more open and frank with their shareholders.
In the meantime, CEOs will need to choose their words more carefully in question-and-answer sessions with analysts.
ROBERT GOTTLIEBSEN, BUSINESS COLUMNIST
Robert Gottliebsen has spent more than 50 years writing and commentating about business and investment in Australia. He has won the Walkley award and Australian Journalist of the Year award.
The use of this work has been licensed by Copyright Agency except as permitted by the Copyright Act, you must not re-use this work without the permission of the copyright owner or Copyright Agency.