Bubble vision

We all know that the rational consumer is a fiction. So what does drive our economic behaviour? Dr Benedetto De Martino says the answer lies in our theory of mind.

CAM 75 Easter Term 2015

What drives a person to pay three years’ worth of wages for a tulip bulb, £1.2m for a shoebox in Chelsea, or their life savings for a 64-digit computer-generated number called a Bitcoin? Greed? Mass hysteria? Plain stupidity? Or something more complex? Bubbles – in the price of anything from bulbs to Bitcoins – have long fascinated and infuriated economists. They appear with depressing regularity, and when they pop, they leave a trail of economic and human misery in their wake. Yet we still have no explanation as to why they happen. Which is why cognitive neuroscientist and neuroeconomist, Dr Benedetto De Martino, principal investigator at the Brain Decision Modelling Laboratory, is taking a rather different approach.

De Martino’s standard fourth-floor office seems an unlikely location for cutting-edge behavioural science, but there are a few clues: the incomprehensible scrawls on the whiteboard; a china Victorian phrenology head; Peter Bernstein’s study of risk, Against the Gods, jostling with William J. Baumol’s solid Economics: Principles and Policy. On his desk sits a huge computer, dizzying amounts of data arrayed across its screen. It is these numbers – taken from thousands of brain scans – that are at the heart of De Martino’s investigations.

Why we act as we do in a bubble isn’t about logic, or rational and irrational behaviour, as traditional economics would have us believe – humans aren’t fully rational. Rather, it’s about what cognitive scientists call ‘theory of mind’: the ability to understand other people’s intentions. Most of us have it: not having it is a significant marker for autistic spectrum disorder. In many contexts, having theory of mind is a huge advantage – it enables you to put yourself in someone else’s shoes. But in the context of an emerging bubble, theory of mind could well be the reason why we fail to realise what’s really going on – until it’s too late.

Have you heard the phrase ‘it’s like comparing apples and oranges’? Well, this is probably one of the most stupid sentences you can say

“People have a very strong tendency to attach intention to anything,” De Martino points out. “Not just about humans, either. We anthropomorphise. We say the gods were angry and that is why the weather is bad. We talk about the market ‘panicking’. So we end up seeing intention where there is no intention. We see a price going up, and we start to think: maybe the other people know more than I do, I should buy.” So everyone is following each other because they think others know more – but actually, as the screenwriter William Goldman wrote about predicting box office hits, nobody knows anything.

“We are trying to get inference from other minds, even if there aren’t other minds there – just noise,” continues De Martino. “And that is a good example of how difficult it is to define what exactly is rational and what is not rational. For example, we showed that in the context of financial bubbles, having theory of mind makes you worse. It makes you take bad decisions and lose money as a consequence. But in another financial context, having theory of mind improves your performance.”

Indeed, a 2010 study on insider trading by Antoine Bruguier, Steven Quartz and Peter Bossaerts found that when people were trading in markets with insiders (individuals with access to non-public information about a company) those with theory of mind were very able to pick up on the signs that something untoward was happening in the markets. Their ability to read the intentions of others was a plus.

Representing value

Understanding how value is represented in the brain is also pivotal to understanding this kind of economic decision-making and that, De Martino says, is one of neuroeconomics’ biggest leaps forward.

“Have you heard the phrase ‘it’s like comparing apples and oranges’?” he asks. “Well, this is probably one of the most stupid sentences you can say. We compare oranges and apples all the time. You go to the supermarket and you choose between meat or fish.” People actually conflate two kinds of decisions, he points out. One is perceptual: is this piece of fish larger than that piece? In this case, you need to compare the same types. But the most interesting decisions are value-based decisions, in which your brain has its own common currency that allows you to compare things.

“[This currency is] very flexible and subjective, and it depends on your own internal state. There is a region in the brain, in the prefrontal cortex, just between your eyebrows, that allows you to compute those abstract values,” said De Martino. “Say I love ’80s pop and I hate heavy metal: that area will be more active when I choose to listen to ’80s pop. But if I like heavy metal, that area will be active when I choose to listen to metal.”

During his bubble research, De Martino found that this same brain region became highly active in traders who were the most keen to ride on the bubble. They were keen to buy above an asset’s fundamental price. It made sense to them, as they valued it more.

As De Martino readily admits, you can’t test people on their theory of mind or their understanding of value before you allow them to trade – and, indeed, many bubbles are fuelled by those outside the original, specialist market, such as the recent Bitcoin craze. But what you can do, he says, is start to understand what mechanisms are driving that behaviour. In other words, if you know why someone is doing something, you can think about what action you might take to change what he or she is doing. People take part in bubbles for different reasons. Some are cynical, knowing that the bubble will eventually pop and they carefully plan their exit. Others genuinely believe that the bubble is not a bubble and it will never pop. Very different kinds of interventions might be required for the former than for the latter. This kind of thinking is already being translated into action in the real world: it feeds into the ‘nudge theory’ pioneered by Richard Thaler and Cass Sunstein in their 2008 book, Nudge: Improving Decisions About Health, Wealth, and Happiness, which in turn was a direct influence on the setting up of the Behavioural Insights team, first launched as part of the Cabinet Office.

Thinking fast, slow and different

De Martino arrived in Cambridge last year as the Sir Henry Dale Fellow (Royal Society and Wellcome Trust). His work is underpinned, he says, by a deep desire to get away from “black and white thinking” and dig deeper into the complexities of the human mind. It’s why he chose to specialise in neuroeconomics, pioneered by his Wellcome mentor Professor Daniel Kahneman, author of the bestseller Thinking Fast and Slow, and the winner of the Nobel Prize for Economic Sciences in 2002. (“He told me that he never took an economics class in his life,” says De Martino. “It’s pretty cool that you can get the Nobel Memorial Prize for a subject you’ve never studied.”). Kahneman, together with his colleague the late Amos Tversky, were among the first to use empirical evidence to challenge the assumption of rationality dominant in classical economic theory. Their work was pivotal to the birth of neuroeconomics, a new discipline that combines psychology, new imaging technology and economic concepts in order to find new ways of understanding human behaviour. This new, young discipline has not been without its detractors.

“People think it’s all about bits of the brain lighting up,” De Martino says. “But that is a profound misunderstanding of our goals. Finding different areas that light up – that was early neuroimaging, almost like phrenology. But how do the different parts of the brain work together to control complex behaviour? What are the neural dynamics of these computations? That’s what we are interested in. For example, economic theory will tell you that it doesn’t really matter how I frame the question. You should make your decision based on the benefit that you get. But we have studied the neurobiology of nudging – the fact people are incredibly susceptible to frames.” He gives the example of a dieter picking up a 70% fat-free yogurt. “Great, you think, and you buy it. Now, imagine the same snack was advertised as 30% full fat. It’s exactly the same percentage of fat. You understand that. But you’re probably not going to buy it. Sometimes what you’re saying doesn’t matter – it’s the context around it. We are now starting to understand how this contextual information is represented by the brain and how it shapes our daily choice.”

The strangeness of confidence

De Martino’s team are now investigating another less than rational human trait: confidence. It is a strange concept, he says, and one that everyone believes they know how to define but is a lot more complex when you look below the surface. De Martino is investigating its link with uncertainty, combining complex brain imaging technology with Bayesian probability – the idea, derived from the work of Thomas Bayes, the 18th-century Presbyterian minister with a passion for mathematics, that when you make a decision, you always have a prior belief. This, in turn, is shaped by sensory evidence, creating a ‘posterior’ belief.

What really matters in that process, he says, is the uncertainty. “If your prior belief is very uncertain, any information I get from you is going to drive that belief. But if you have a very strong prejudice, you need to give me way more information to change my prior belief. For example, I am Italian. I am very precise and I am always punctual. But people have a prior belief that Italians are always late. So people will want way more evidence to change their prior belief because they have much more certainty about that prior belief.”

Research is still in its early stages, but De Martino is confident – in its most basic sense – that exciting ideas will emerge. “What I want is to build a model of the human brain that shows how we really are,” he says. “Not an ideal standard of rationality.”

CAM is published three times a year, delivered direct to 210,000 readers around the world.