Economists have always been keen to borrow principles from the hard sciences. In the 19th century Léon Walras and William Stanley Jevons both started their work with a view to importing the insights of physics into the economic sphere. Irving Fisher, the great neoclassical economist whose 1930s work has been rediscovered during this crisis, even wrote his doctoral thesis at the turn of the 20th century under the supervision of a physicist.
This tendency was given renewed impetus in the mid-20th century by Paul Samuelson's application to economics of mathematical principles derived from thermodynamics. The development of computers able rapidly to analyse data made the development of mathematically elegant economic models particularly desirable, driving the acceptance of concepts such as American economist Eugene Fama's efficient market hypothesis.
Most of the “quants” – financial mathematicians – who used such concepts to build financial models always knew that this project had serious flaws. Emanuel Derman, for example, a physicist turned financier who formerly worked at Goldman Sachs, is credited with playing a central role in the development of models in relationship to derivatives. Yet more than a decade ago, he was warning Goldman Sachs clients of the limitations of derivatives models – he compared their relationship to reality to that between a child's toy car and an actual automobile.
Mr Derman remains, to say the least, wary of the idea that efficient markets hypothesis can provide a “complete” guide to finance. “Unfortunately, absolute value theories don't work very well in economics,” he wrote recently. “It's difficult or well-nigh impossible to systematically predict what's going to happen. You may think you know you're in a bubble, but you still can't tell whether things are going up or down the next day.”
Such scepticism has not often been expressed quite so frankly. On the contrary, some quants have furtively revelled in the power that their apparently elite knowledge gave them. “The dirty secret of banking is that lots of bankers have always felt a bit insecure because they did not really understand how this stuff worked – so those who understood it were in a strong position,” observes one banker.
However now that the crisis has exposed their shortcomings, the EMH and the entire model-based approach to finance are facing a radical rethink. A growing chorus of financiers, quants and economists argues that it is wrong to apply simplistic assumptions that underpin the physics-like models to people, since – unlike atoms, say – they can learn from each other and change in response to events. Changes may not happen in a neat, linear fashion.
Donald MacKenzie of Edinburgh university says the real problem with models is that bankers tend to view them as “cameras” that capture how the world works, like the camera that might photograph a physics experiment. Instead, he argues, they should be viewed as “engines” – since the presence of a model tends to change and drive market behaviour in a way that makes it impossible to assume that the past can predict the future.
Nevertheless, no alternative intellectual model – or source of inspiration – has emerged to offer a truly coherent alternative. George Soros, the former hedge fund manager, for example, argues that market participants need to embrace the idea of “reflexivity”, to recognise that markets change in response to participants, and to accept that models are an “engine, not camera”. However, turning this reflexivity theory into any investment manual or strategy has proved difficult.
