Investment – in common with so much of what human beings do every day – involves gathering and processing information in order to reach decisions. But just how might investors do this in the speediest and most efficient way? Are they better off working alone or in groups and, if in groups, how tightly-knit should these be? Welcome to the world of organisational theory and the concept of ‘clustering’.
For some years now, received business wisdom has held that people who work in teams come up with better solutions – hence the time, effort and money big organisations tend to spend on open-plan offices, ‘hot-desking’ and other such ideas as they seek to encourage their employees to communicate and co-operate more. But is this time, effort and money well spent?
Boston-based academics Jesse Shore, Ethan Bernstein and David Lazer wanted to test this idea and the result was a paper published earlier this year, the short title of which is Facts and figuring. Using data from a novel laboratory experiment on complex problem-solving, they set out to investigate “how an organisation’s network structure shapes the performance of problem-solving tasks”.
In their experiment, 352 college students competed in various rounds of a Cluedo-esque game. Based as it was, however, on a US Defense Department exercise where participants have to predict four aspects of a fictional terrorist attack – the planned time, location and target as well as the identity of the perpetrators – it was not so much a ‘whodunit’ as a ‘who-will-do-it’.
The students had to ask questions of a Google-like database and then use the information they gathered to form their conclusions. Shore, Bernstein and Lazer wanted to analyse how people approach “both exploration of information space (for facts that may be important pieces of the puzzle) and exploration of solution space (for theories, or interpretations of facts, that combine puzzle pieces into an answer)”.
They also, as we flagged at the start, wanted to see if groups were better than individuals at searching for and interpreting information and if tightly-knit groups were better at this than more loosely structured ones. One plus for the tightly-knit variety, the paper found, was that “clustering promotes exploration through information space”.
In plain language this means the closer people work together, the more efficient they are at searching for new information. They know the sorts of questions their colleagues have already asked and so they do not ask them again. Clustering therefore leads to people making fewer searches while the searches they do make are more effective.
The trouble is, another of the paper’s findings was that clustering “inhibits exploration through solution space” because the element of mutuality that helped people search for information now becomes a disadvantage. Or as Shore, Bernstein and Lazer put it: “The mutual awareness of each other’s theories results in a convergence in interpreting that information, reducing the exploration of theory space.”
Simply put, clustering can lead to negative behaviour, such as copying and settling for less – for example, if somebody in a team comes up with a promising but intermediate solution, colleagues tend to focus all their attentions on that and merely look to tweak it. Confirmation bias also raises its head in that one person’s choice of a particular answer may “seemingly provide social proof of its value”.
Operating by yourself, on the other hand, may lead to time being wasted on some redundant searches but you may also be in a better position eventually to find superior or more effective solutions. In turn, this raises the possibility of whether individuals operating in looser groups – so-called “inefficient networks” – might help produce better solutions and indeed the paper suggests it might.
“It is in the less connected – and thus less efficient – networks that we will find individuals who are not yet exploiting the current best solution conducting more exploration of the solution landscape and bringing more potential solutions into the network,” it says “Indeed, agents in inefficient networks eventually converged on better solutions, collectively, than agents in efficient networks.”
Does this have any relevance to the world of investment? Did you ever doubt it? Here on The Value Perspective, we would certainly regard sell-side analysts, for example, as extremely well-connected groups – as efficient networks. Everyone knows what everyone else is doing, which makes them very good at finding and publishing information – but are they perhaps processing it in a sub-optimal way?
Similarly, some giant US fund houses, say, are never shy to proclaim the virtues of their vast global networks but does that necessarily mean they are coming up with the best solutions? We would argue investors are better off aiming to combine the best of both worlds – working as loose teams of individuals and then perhaps, towards the end, comparing their conclusions with those of others.
“By theoretically and experimentally disentangling exploration for information from exploration for solutions – two core but separate domains of problem-solving – we interpret the results of this study to indicate that clustering has opposite effects in those two domains,” is the conclusion reached by Shore, Bernstein and Lazer.
“It promotes exploration through information space but depresses exploration through solution space. Whether increased clustering improves or impairs performance will therefore depend on whether the immediate task or problem-solving stage benefits more from exploration of facts or from the figuring that comes through the exploration of theories that interpret those facts.”
In investment, where everybody can to a greater or lesser extent dig out the same information, it tends to be how you interpret what you find that can make all the difference.