Revolutionary markets research focuses on the other AI

0
71
Revolutionary markets research focuses on the other AI

Take a short walk from the trendy Italian restaurant scene in Lygon Street, Melbourne, and you will find a small laboratory conducting revolutionary research into financial market trading.

The work being done inside the Brain, Mind and Markets Lab at the University of Melbourne has the potential to fundamentally change the way financial institutions think about risk management.

It could well result in a lucrative new line of export business for the university as the world’s leading fund managers seek to gain an edge through the use of intellectual property developed by a team of post-doctoral students and professors.

In financial markets competition is intense. In some segments of trading the difference between winning and losing is defined in milliseconds thanks to the widespread use of high frequency trading, algorithms that learn from their mistakes and machines that attempt to remove human errors.

Leading the work in Melbourne is a maverick, Belgian-born social scientist Peter Bossaerts, who is Professor of Experimental Finance and Decision Neuroscience and Honorary Fellow, Florey Institute for Neuroscience and Mental Health.

Advertisement

He pioneered the use of controlled experimentation with human subjects in the study of asset pricing. He initially focused on the Capital Asset Pricing Model, which is used throughout the financial industry to evaluate risk and return trade-offs and portfolio performance evaluation.

This caused enormous controversy in economics and finance academia, which found the idea of using laboratory science techniques hard to swallow.

“I am trying to revolutionise methods in finance and economics and that’s extremely hard in a field that does not have lab science,” Bossaerts says.

Maverick seizes opportunity

So how did world leading research in financial markets end up in Melbourne?

“If you look at the top finance and economic departments at Stanford, Boston, Chicago nobody there does experiments,” he says.

“I am the only one in the world doing that.

“Once you start doing market experiments you need to work with hundreds of people and you need to have software that is commercial grade, you need your own markets, and you need robots as well.

“It becomes extremely complicated and relative to social science its extremely expensive.

“We are talking about $400,000 a year and that’s peanuts for medicine or something like that but for me it was almost impossible to get that anywhere. University of Melbourne has given me the opportunity to do this work.”

Bossaerts is regarded as a maverick because he has pushed back against the key findings from the populist research revolving around behavioural science.

This work peaked with the publication of the book Nudge about 10 years ago by Richard Thaler and Cass Sunstein. Their work focused on the propensity for people in all walks of life, whether it is financial markets or buying a mobile phone, to make irrational decisions.

Mind you, all that work and the subsequent public debates about the cognitive biases driving poor decisions has not stopped companies from exploiting consumer irrationality.

Behavioural finance

Chanticleer attended a conference in Sydney this week hosted by one of the world’s leading consultants in price management, Simon-Kucher & Partners. It advises companies on the best way to deliver top-line revenue growth through lifting prices.

Companies attending the Simon-Kucher seminar included NRMA, KFC, News Corp and Westpac Banking Corp. A prime case study delivered at the conference showed the benefits of the “anchor effect”.

The strategy is to subconsciously influence consumers by offering three price bands. Most consumers tend to pick the price in the middle. Simon-Kucher restructured a mobile phone plan in a way that boosted average revenue per user by 36 per cent and monthly recurring revenue by 29 per cent.

In financial markets the commonly found cognitive bias is the disposition effect, which is the tendency to sell shares that have risen in price and keep those that have dropped in value.

Bossaerts says the problem with behavioural finance is that it is all about telling people what’s wrong with their decision making.

“Certainly in the context of finance we do make an enormous amount of mistakes but behavioural finance is very short on prescription and what we should do about it,” he says.

“For a very long time the idea was you just had to teach people. But we know that doesn’t work because it’s much deeper than that.

“Then came the idea of nudging, which is that you just have to give people nudges so they don’t make mistakes.

“We came at it from a completely different angle. These are cognitive biases that have a biological basis and you have to look at them by conditional biases.”

Theory of mind

Bossaerts says his work starts from a similar position to the aviation industry, which trains pilots how to deal with situations that can cause total confusion such as flying in clouds.

“We start from the position we are not as good as birds which can fly through clouds for instance,” he says.

“You cannot fly through clouds for instance without losing the sense of what’s up or down.”

He says pilots are trained to look at instrument panels when in clouds and they have to do regular sessions to keep their skills.

“So it’s not enough to be aware of it. It’s not enough to be nudging people to do that. In fact you have to coach them you have to teach them you have to train them.

“Then, we actually look at how can you use something in people’s biology in order to set up a training scheme that is hopefully effective.

“That’s what we did and we found one that is effective.”

The Brain, Mind and Markets Lab has developed a training scheme which exploits theory of mind, which is the ability to form correct beliefs of others’ intentions by abstracting from one’s own situation.

Forward-thinking approach

Another professor who worked on the scheme, Petko Kalev from La Trobe University, says the project exploited a human skill that facilitated transfer away from instinct.

He says correct investment requires a forward-thinking approach, while ignoring personal history, especially the purchase price of assets.

“Humans do have the skill to abstract from their own circumstances,” Kalev says.

“It is this skill that forms the basis of our training scheme. Not surprisingly, our evidence shows that its success requires one to have well-developed theory of mind in the first place.

“Our study presents the first attempt to develop rigorous training schemes for financial decision-making that are based on a bio-centric analysis of human behaviour and opens a promising research avenue associated with designing and testing theory of mind-based training strategies for effective reduction of disposition effect.”

The paper, which is being released at next week’s eighth Behavioural Finance and Capital Markets Conference (BFCM) at La Trobe University in Melbourne, was written by Bossaerts, Kalev, Kristian Rotaru from Monash University and Nitin Yadav from the University of Melbourne.

‘Concerning situation’

When Bossaerts spoke to Chanticleer this week he was in Amsterdam visiting the global headquarters of Optiva, one of the most innovative trading companies in financial markets. Bossaerts came away from the Optiva visit convinced that machines should never be allowed to take over all financial market trading.

“There are a whole host of robots there that are black boxes that have been trained using machine learning and nobody can tell me what they have learnt,” he says. “That is a very concerning situation.”

Bossaerts argues that financial institutions have a responsibility to at least know what robots have been trained to do in financial markets. At least then you know when to turn them off and they “wont’ turn on you”.

It is somewhat ironic that Bossaerts is concerned about artificial intelligence (AI) given that much of his work in the lab is related to neuroscience and the focus on the anterior insula (AI), which is a cortical structure in humans thought to be responsible for translating emotions.

Tony Boyd

Read More

LEAVE A REPLY

Please enter your comment!
Please enter your name here