SANDILE MALINGA: World’s best investors share traits with super-forecasters

When I was growing up, I was inspired and intrigued by the classic Hanna-Barbera cartoon, The Jetsons, set in 2062 — 100 years into the future when it first aired. The show imagined a space-age utopian future: flying cars that folded into briefcases, robot helpers that put C-3PO to shame, universal basic income, and every convenience imaginable at the push of a button. The Jetsons captured the popular vision of the future through a Cold War American lens — technology solving all humanity’s problems and delivering a great quality life for all.

Roll forward to 2025, and I’m a little disappointed my briefcase isn’t flying me to the office and back for my three-day work week. Domestic chores still haunt my family, and food still resembles familiar shapes and isn’t in pill form (though as a lover of a braai, I’m not too upset about that failed prediction). So, while autonomous flying cars remain elusive, the show did get some predictions right — AI, smartwatches, elevators, flat-screen TVs and video calls. The Jetsons highlights the difficulty in making accurate forecasts. As Nobel laureate Niels Bohr once wryly noted, “Prediction is very difficult, especially if it’s about the future!”

A “forecastable” system typically is one that exhibits patterns, causal structures and a degree of stability — enough to make probabilistic inferences that are better than random chance. Regarding the weather, we have very good models that can forecast up to a week ahead, but over longer periods, chaos creeps in and makes our system difficult to forecast.

Still, this doesn’t stop those of us in financial markets from building entire careers around forecasting. Switch on any financial news network and you will find no shortage of experts indulging in the alchemy of forecasting — from where inflation will be in a year (to the second decimal) to where the market will close on December 31, to bond yields — and my favourite indulgence: future asset class returns.

At the beginning of the year, consensus forecasts for the JSE all share implied a 14.75% return for the year. To run a sanity check on these expectations, we looked at annual returns since 1962 (a hat tip to Mr Jetson). As you can see in the figure, the median annual historical return has been 3%, with about half of the years seeing returns of 0%-5.75%. When we compare the actual returns with what the analysts expect, there is quite a disconnect — most analysts’ expectations are far above from the median historical return.

One of the most interesting meta-questions — especially for financial markets practitioners — is why we bother forecasting at all given the overwhelming evidence of how wrong we get it. I believe that classical financial economic theory provides one answer: if we gather all available information, work hard with it and surround ourselves with bright people, we might have a chance at getting these forecasts right.

However, there is an alternative school of thought that suggests that they are inherently unforecastable given that financial markets are populated by humans by nature prone to emotional reactions, rather than coolly evaluating the future like “rational agents”. This creates a system in which chaos tends to dominate — much like with the weather. Now for the good news: it is possible to be better than average at forecasting. On the trading floor, this is an example of what we call an “edge”— something repeatable, testable and consistent over time.

Psychologist Philip E Tetlock and colleagues explored this through The Good Judgement Project, a forecasting tournament in which thousands of volunteers made probabilistic predictions on various events. The most accurate were identified as “superforecasters”.

In a benchmarking exercise, superforecasters were about 30% more accurate than intelligence officers with access to classified information. Like the mid-20th century creators of The Jetsons, they don’t get every forecast right, but they tend to be less wrong than the average, including experts with nonpublic information.

Research into why they were so accurate boils down to several factors:

• Their forecasts exhibit less noise than the average.
• They gather and interpret relevant information effectively.
• They are aware of their own biases and take steps to minimise their impact.
• They update their beliefs as new evidence emerges.

Many of the world’s best investors share these superforecaster traits. They have an admirable ability to make good decisions in a structured way without emotion under uncertainty. They avoid overconfidence and confirmation bias. They are aware of how much they don’t know and seek to reduce that gap by letting data guide their decision-making. And like any good edge, if applied consistently over time, it compounds and produces an outcome that may look like a Jetson-inspired futuristic gadget.

We should all strive to emulate the superforecasters, relying on data to guide our decisions rather than gut feel, especially when the stakes are high and the future uncertain. As we have seen, it can lead us very far from a Jetson utopia.

• Malinga is M&G Investments CIO.