Behavioural science: Model behaviour

Nicholas Pratt looks at the growing use of behaviour science in the funds industry and its potential role as the saviour of active management.

Although it could be argued that behavioural science has been employed in the investment process for decades, it first gained significant commercial attention in 2002 when the Nobel Prize for Economics went to Israeli psychologist Daniel Kahneman, “for having integrated insights from psychological research into economic science, especially concerning human judgement and decision-making under uncertainty”.

But while it has gained popularity in the funds industry, it is by no means pervasive or deployed systematically, says Greg Davies, head of behavioural science at Oxford Risk, a research firm providing investor risk profiling.

When discussing behavioural science, it is important to separate bias and emotion, he says. “We definitely want to remove bias because it leads to an inaccurate assessment of the world, but we don’t want to remove emotion or intuition. The goal is manage that emotion. If there is no emotion, there is no reason to make a decision. You need that emotion to be able make trade-off decisions. Ultimately investment decisions are based on the desire to increase wealth versus the fear of losing wealth.”

Emotions are not necessarily irrational and can be helpful when they are associative, says Davies, using chess grandmasters as an example. “They have trained their emotional sense so that they intuitively know the next step to take. That can be very effective in investment, despite the differences. With chess, you get immediate feedback and a fixed set of rules, which is a very good environment for learning. The investment world is more complex, given that the rules constantly change and that feedback is rarely immediate, making it very difficult to train intuition.”

The economic theory behind behavioural finance has been in play for a long time but practical application in the funds industry is much newer, says Mick Dillon, portfolio manager of Brown Advisory’s Global Leaders strategy. “Brown Advisory started using it seriously four to five years ago to understand where our behavioural biases lie and to go back to analyse our decisions and see where they were making mistakes or using skills. The goal is to make the decisions better and more informed.”

One critical change that Brown Advisory made was to analyse not just the active decisions – the buying and selling of stocks – but also the passive decisions whereby no action was taken. “The majority of decisions made by a portfolio manager (nine out of ten) are non-active and involve no capital allocation,” says Dillon. “They play a huge role and until we started capturing those decisions, they were lost to the ether.”

The team also wanted to capture decisions beyond the buying and selling of stocks, or the so-called hit rate, and also look at sizing decisions – i.e. how much capital was put into certain trades and do you have the right capital behind the right stocks.

To achieve this involved a huge qualitative effort. It was also where the use of machine learning technology came in, says Dillon. “We need machine learning to contextualise and give meaning to all of these decisions and to help us learn from them. We have used this technology a lot, especially for those non-active decisions.”

He adds that in recent years, the technology available has improved immensely in terms of data capture, visualisation and data analysis. Although these tools still need to be customised, the machine learning elements have helped to capture some of the nuance and the context behind the many decisions.

Brown Advisory has also started employing a third party to analyse the data as opposed to doing it themselves. “Their job is to understand the data that they then hand back to us. And the results show that it more than pays for itself,” says Dillon.

System of rules
Technology has a huge role to play in behavioural science, agrees Davies. “If we want to improve human decision-making, we need to build a system of rules and be able to detect where emotion leads to either the right or wrong answer. We want to create a rules framework for decision-making, along the lines of ‘don’t buy food when you’re hungry’.”

This can be done without technology but it is easier to systemise with technology – what Davies calls decision prosthetics. Oxford Risk designs tools to profile financial personalities in order to know how investors act under certain conditions and to send interventions and notifications based on these findings.

“We measure risk tolerance and composure, alongside numerous other dimensions of financial personality, so that we know when someone needs their hand held in a time of crisis or if someone is too laidback and needs to pay more attention at a time when they should be rebalancing their portfolio. The same market event can have different implications for different personalities. We start by understanding your personality. We look at their baseline personality using psychometric testing, study their historical decisions and assess their current state,” says Davies.

The profiling can be very complex and Davies says that the financial services industry lags behind sectors such as elite sport or the military, which both study how people perform under stress. As an example, he says that a fighter pilot would not be allowed in a plane if suffering from anything worse than a cold, but we pay no attention if a portfolio manager has been out till 4am and turns up to work with a hangover.

Davies says much more can be done with behavioural science before the need arises for artificial intelligence. “We are still focused on Iron Man rather than Terminator. When Garry Kasparov was beaten by Deep Blue, a human was no longer the best chess player in the world. However, then we saw the development of Centaur Chess, where computers and humans worked together, beating either working alone. In chess, the computers now dominate, but in the investment world, there are many more variables than chess and we are still a long way from AI alone being better than a combination of a human decision-maker enhanced with AI. The use of AI at present is less about managing behaviour and more about helping to crunch the numbers.”

The use of behavioural science in the funds industry can be seen in the rise of factor investing – funds based on market anomalies arising from behavioural bias and not just compensation for risks that other investors want to transfer – over the past 20 years, says Prasenjeet Bhattacharya, senior investment strategist at Dutch fund manager NN Investment Partners (NNIP) and financial data scientist.

“Contrarian strategies such as the Fidelity Contra fund are a further example of looking for value in companies whose value the manager believes is not fully recognised by the public. The JP Morgan Undiscovered Managers Behavioral Value Fund focuses on companies with significant insider buying and share buybacks. These are more recent in timeline as compared to factor funds. But most of the application of behavioural science is still within stock-picking and I still do not see a trend where it is being more widely used in multi-asset strategies or within ‘quantamental’ type of approaches to investment,” says Bhattacharya.

Like others, he does not believe it is possible to fully eliminate all human bias from decision-making but it is possible to minimise their occurrence. “Biases are nothing but repeatable, predictable mistakes which we do in our daily decision-making.

Unlike many might perceive, higher cognitive sophistication (intelligence) or a greater number of experiences in a domain of work (expertise) do not mitigate our biases and this is because our cognitive processes and emotional processes are inextricably interconnected. The brain is wired to perceive before it thinks – to use emotion before reason.”

For fund managers, the aim is therefore to integrate these behavioural techniques into the investment process. NNIP is currently running a project based on neuro-forecasting. “It is a functional MRI study of risky decision-making to get insights into our analysts’ brains,” says Bhattacharya.

“The objective is to see how we can use the brain scan information to create de-biasing strategies. We have now the technology that can help us deep dive into the human brain and understand how the process of making decisions happens. Using brain-mapping technology, we aim to bridge the behaviour gap in our investment decisions.”

A number of tools aim to decipher behaviour based on people’s digital footprint, but these are mostly reserved for marketing or consumer neuroscience, says Bhattacharya – for example, the use of eye-tracking software to analyse a person’s visual attention span.

Competitive edge
Investment teams could find a competitive edge with their de-biasing techniques through the use of advanced analytics, he adds. “However, this would require the creation of an integrated data set covering the investment decisions and information around the trade like security weighting, selling time, reasoning for the trade action, emotion associated with particular decision.

“Thereafter, along with the strategists or analysts, a hypothesis needs to be developed about the biases that might have negatively affected their investment decisions. The machine learning algorithm then, for example, can unearth repetitive patterns of selling timing and behavioural biases. I can also think of clustering algorithms that group trading decisions with emotional profiles. Such data and analytics-driven feedback loops can prove to be very effective for the investment teams.”

Beating the market with just the ‘information edge’ is a limited strategy, says Bhattacharya. “The complexity of today’s financial system means that investors have more information to consider than ever before and this information, along with sophisticated algorithms, is available to almost everyone. In order to maintain a competitive edge, one has to blend the quantitative insights and information alpha with qualitative judgements and behavioural alpha.

“I foresee a cultural shift within the active investing community where we move towards more ‘quantamental’ kind of strategies (a combination of fundamental and quantitative) because neither quantitative strategies nor human judgements alone can fully capture the complexity of the financial markets. Within the quantamental or man-plus-machine approach of investing, insights from behavioural science will play a crucial role in strengthening the overall investment process.”

For Chris Woodcock, head of product at behavioural data analytics service Essentia Analytics, the most exciting development in behavioural science has been the number of tools designed to record and analyse investor decision-making behaviour so that patterns and biases can be identified and addressed. “We can now tell an investor if his or her sunk-cost or loss-aversion biases are leading to missed opportunities and help them take corrective action – that didn’t used to be the case,” he says.

The goal, he adds, is to make investors aware of their biases rather than to remove them altogether. “Biases are fundamental to human nature, and we aren’t presuming to change that. We don’t believe they can be removed, in fact – mitigated, perhaps, or circumvented, but not removed.”

It is essentially about self-awareness, says Woodcock. “We offer investors a data-driven feedback loop that keeps them tuned in to their behaviours and biases – and the consequent effects they have on their portfolios. We are like a coach: we monitor our clients’ performance and ‘nudge’ them when we see potential issues emerging.

“This allows them to make course corrections such that they spend more time and energy doing what they’re good at and less doing what they’re not so good at. We are very fond of the Socratic aphorism ‘know thyself’ – if there’s one piece of advice all investors would do well to adhere to, that would be it.”

Technology has played a fundamental role in the development of behavioural science, says Woodcock. “Without cloud computing, machine learning and the algorithms we have developed to track and analyse investor decision-making behaviour, Essentia would not be able to do what it does. Our machine learning algorithms are what uncover the patterns and anomalies and behaviours that may represent biases or poor decision-making habits, and provide data-driven illustrations of where and how investors can add additional value to the portfolio.”

While the technology is at the core of the offering, the human element is still an essential component, says Woodcock. “It’s like a high-performance engine in a race car – it’s very impressive, but it needs a driver to put all that raw power to productive use. The data analytics produced by our algorithms must be interpreted and adapted to each investor’s decision-making process to add any value and we employ human coaches, all ex-portfolio managers, who work one-on-one with our clients to put all that data to good use.”

Growing fast
There are still areas where the available tools are scarce, says Woodcock. “While data analytics products have been around for decades, tools that are designed to not only measure behaviour but also help users make better decisions are few and far between. However, it is a nascent area that is growing fast.”

Woodcock believes that these behaviour science tools can become essential for investors trying to add incremental returns to their portfolios and that it is just a case of raising awareness. “We know there is latent performance in portfolios that is unrealised due to behavioural habits and biases. The use of behavioural finance has documented it and our research has proven it.”

He says that greater use of behavioural tools can uncover an average annual excess return in the region of 100 basis points that had otherwise been lost to behavioural bias or other common decision-making deficiencies. “We call it ‘behavioural alpha’ – the added value that a manager can bring to a portfolio by mitigating their own behavioural bias and inhibiting the wrong behaviour. The alpha is there just waiting to be released. What active manager wouldn’t want to release it?”

As well as improving performance, the use of behavioural science could also prove important in the battle of passive versus active management, says Brown Advisory’s Dillon. “I think it can really differentiate active managers as it allows our investment processes to be continually refined, creating increased value for our investors.”

Oxford Risk’s Davies agrees that behavioural science could also be a vital tool in the contest between passive and active management. “It will be increasingly important to try and isolate the value of human decision-making and the ‘active’ element in investing. If humans are worth anything, this element has to be enhanced and be seen as essential.

“Passive investing now has a more sophisticated message because of the use of rules, and active management needs this as well. It is not just about improving decisions but about having a better story to tell.”

©2019 funds europe

HAVE YOU READ?

THOUGHT LEADERSHIP

The tension between urgency and inaction will continue to influence sustainability discussions in 2024, as reflected in the trends report from S&P Global.
FIND OUT MORE
This white paper outlines key challenges impeding the growth of private markets and explores how technological innovation can provide solutions to unlock access to private market funds for a growing…
DOWNLOAD NOW

CLOUD DATA PLATFORMS

Luxembourg is one of the world’s premiere centres for cross-border distribution of investment funds. Read our special regional coverage, coinciding with the annual ALFI European Asset Management Conference.
READ MORE

PRIVATE MARKETS FUND ADMIN REPORT

Private_Markets_Fund_Admin_Report

LATEST PODCAST