The triumph of generalists over specialists
Our team is often asked why we choose to organise our investment process along generalist rather than sector specialist lines. The answer is partially empirical, in that we have noticed the cross pollination of ideas across sectors is helpful in generating investment ideas, but crucially, the answer is one of risk management; it helps protect our investors against potential loss inflicted by subconscious behavioural biases that may exist within the team.
Taking a generalist approach helps offset some of the cognitive dissonance we can possess as investors, such as the bias of ‘investing in the familiar’. This heuristic famously manifests itself in ‘home bias’, or the tendency to overweight investment in one’s home country. However, we believe ‘investing in the familiar’ bias also emerges when individuals work as sector specialists where minutia is often triumphant over relevance, and decisions are made without comparison to a wider, potentially richer, pool of assets.
Interestingly, we note the idea of approaching investing as generalists is supported by research dedicated to the topics of broadcast search and superforecasting:
There is a growing body of work indicating that many of the best ideas emerge from combining insights across areas of expertise that don’t seem intuitively connected. Problem solving across disciplines is often achieved through a ‘broadcast search’ process where external solvers outside a knowledge discipline are invited to submit solutions to complex problems.
An early example of this process took place in the 18th century when a broadcast search was instigated to solve the “longitude problem”; finding a practical method for determining longitude at sea. Advising the British Parliament on a £20,000 prize to be awarded for a solution to this puzzle, Sir Isaac Newton made an ultimately misplaced assertion that only an individual taking an astronomical approach could find the solution. Proving Newton wrong, surprisingly, was an unknown carpenter and clockmaker, John Harrison. Harrison’s winning solution was innovative on two counts; it did not rely on astronomy, and the design of the chronograph differed markedly from those developed by leading clockmakers, evidencing a novel understanding of materials, science and mechanics (Randall 1996).
A more recent example centres around the development of the world’s first practical quantum computer. The computer was developed in 2006 by an individual who describes himself as interested in various areas of theoretical and applied physics and mathematics without being an expert in any.
Geordie Rose, founder of D-Wave Systems asserted that all necessary cutting-edge theories for approaching building quantum computers already existed, but they resided in various distant knowledge areas:
“There is a perception over here that the problem lays there and a perception over here that the problem lays there or there. But when you build this patchwork quilt, what you discover, interestingly, is that in none of the particular patches do the experts [of each patch] think there is a problem... They always think it is in one of these other patches. So a premise of our company was that all the individual components were good enough, but nobody was sitting on top of it synthesizing the whole project” (Nagle & Teodoridis 2017).
In his book, Superforecasting – the Art and Science of Prediction, Philip Tetlock describes superforecasters as ordinary, everyday individuals that are logical thinkers who know a little about everything, rather than a lot about a little. He shows that generalist superforecasters are, by a wide margin, statistically better at predicting events than experts covering a topic in depth. Tetlock also found that when these individuals were grouped into smaller teams, they became even more accurate. For example, when assessing events of geopolitical importance, a group of generalist superforecasters from a diverse range of backgrounds, with access to nothing more than Google, meaningfully outperformed a group of experienced intelligence analysts with access to classified information – reportedly by 30%.
Implications for our investment team
Tetlock found that superforecasters remain in a perpetual beta mode, realising that nothing is certain and that reality is infinitely complex. For us, the lessons from the broadcast search method and superforecasting lie in taking a generalist approach, remaining intellectually curious and building an inclusive analytical culture that will provide the best environment for our team to deliver excess returns to investors.