Part 4: To see the World in a Grain of Sand – Behavioural Economics, the Science of the Future?

Conversations with my Father about the economy are always short. He has only two things to say about it: “Greed and Fear”. Every time it comes up in conversation, as predictable as the tide he’ll tell me this, and nothing else, is what it’s all about.

For years I thought he was an idiot: Everybody knows that Economics is a Mathematical science and when used to explain and predict actual events in the world, economic theories in both Macro and Microeconomics rely, not on such heady emotions as these, but fundamentally on the concept of rationality. Unless all the participants in a market behave ‘as they should’ – that is in the way most likely to maximise their payoff – the model will not hold.

The thing is, though (father forgive me!) on closer inspection, real markets, and certainly the economy as a whole, seldom DO conform to the mathematical ideal. The principle reason, it seems, is that human decisions are seldom rational in the mathematical sense. This shows its face in the form of a number of robust ‘biases’ (when compared to the optimum strategy) in real human behaviour, including the use of heuristics (rules of thumb) rather than logic; and ‘framing’ strategies where anecdotes and stories, rather than objective facts, are used to make sense of the market. These biases give rise to market inefficiencies, that is, situations where the standard relationships between supply, demand and price break down, often to the mutual disadvantage of the participants. (Arbitrage, though, it’s worth noting, is an activity that has grown up specifically to profit from exploiting these inefficiencies.)

Two of the most obvious examples of market inefficiencies are bubbles and crashes. In a bubble, confidence in artificially high. Investors pile money into markets that undergo rapid growth and provide large short term payoffs. Mimicry or herding behaviour, another irrational bias, where investors follow one another into the market blindly, is often a characteristic of this phase. But the growth is unsustainable, the bubble bursts as consumer interest wanes and shares in those companies collapse as they find themselves unable to pay back their investors. A stock-market crash occurs when confidence declines sharply across a number of markets simultaneously and investors all pull out together, further devaluing the remaining stock. That such phenomena are based on nothing more than confidence and collective behaviour is strange – perhaps – in an environment where data and mathematical analysis are plentiful.

For all the talk of irrationality, though, when viewed from a human perspective, many of these behavioural biases make intuitive sense: Loss aversion, for example, describes a reluctance of individuals to sell assets where to do so would incur a nominal loss, even if holding on to them incurs a bigger loss in the future. Similarly, risk aversion describes the phenomenon whereby most of us prefer a smaller, but certain, reward now rather than a larger, but riskier reward in the future, even if, mathematically, we ‘should’ choose the riskier option, or the immediate – but smaller – loss. The ‘bird-in-the-bush’ phenomenon (so widely accepted that it’s a proverb!) describes the (seemingly) arbitrary preference that we have for holding on to what we already have over acquiring something new.

The science of psychology lags some way behind conventional economics in its ability to provide a mathematical description of observed events. Humans are notoriously difficult to test because of the number of variables that have to be controlled to get a meaningful result. We frequently behave differently when we know we’re being tested – in surveys and questionnaires for example. Consider the differences there might be in your strategy in a poker game involving someone else’s money, and one involving your own! One fruitful area of research is in what is sometimes called ‘comparative economics’:

Comparative economists start from the premise that human and animal behaviour exists along the same continuum and look, therefore, at the economic behaviour of animals with the hope of learning something about human behaviour. A lot of work has been done on animals like pigeons and rats using the operant conditioning methodology popularised by Burrhus Skinner in the middle of last century. These studies reliably demonstrate that pigeons and rats respond in a similar way to humans when the reward for their work is diminished (i.e. they work less). This is sometimes taken to illustrate a similar attitude to price elasticity of demand too, whereby the preference for the food reward is diminished as the ‘price’ (the amount of work required to get it) diminishes.

Much more exciting, however, is the recent work by Chen et al (2006) in which they trained a group of captive capuchin monkeys to use a currency system to ‘buy’ food from experimenters. Using this behaviour they were able to demonstrate that most of the biases described above also apply to monkeys, evidence in favour of the view that our own biases and irrationalities might actually be innate, evolutionary and therefore provide us with a survival advantage.

Fundamentally, economic systems ARE determined by human behaviour. These demonstrations of irrationality present significant problems for pure mathematical models. It seems likely that any future economic theory that improves on the current ones MUST recognise that humans are not truly rational actors. The problem for behavioural economics remains, though, that we lack an effective means of using even the little that we know about human psychology and behaviour to accurately predict what a person will actually do in a given situation; to determine if, and when they will act irrationally, and exactly how. And once we consider large numbers of people all ‘behaving’ together, as we do in economic systems, the situation becomes even more complicated.

Behavioural economics is an evolving field. The holy grail is a mathematical description of the human psyche that predicts, at least probabilistically, it’s behaviour given certain environmental conditions. New developments in data mining [link to data mining blog], and the emerging discipline of Social Physics might provide us with the kind of data that we need to accurately model human decision-making, but it seems a long way off.

May 14, 2015