More: Can science reinvent the economy?
Borrowing techniques from seismology to calculate the likely frequency of market fluctuations is all very well, but it has the same fundamental limitation as earthquake prediction. It does not tell us what we really want to know: exactly when and where the next cataclysmic event will be. That kind of predictive capability would demand what seems an unobtainable wish 鈥� a comprehensive, bottom-up theory of why markets move as they do.
At the moment, that ideal is confounded by the sheer number of individual interactions in any real market. The result is a gaping divide between microeconomics (the study of how individual people and companies in a market make decisions to buy and sell) and macroeconomics (the study of movements in economy-wide indicators such as GDP, inflation and unemployment). Trends in macroeconomics are the sum of microeconomic decisions, but attempts to extrapolate from the one to the other are by necessity grossly oversimplified, says Eric Weinstein, a physicist who works for the Natron Group, a hedge fund in New York City.
Advertisement
They often assume, for example, that agents in a market all have an unchanging list of all the things they want, ranked in the order they want them. There鈥檚 no room for them to change those preferences, by learning, for example, or becoming interested in new products. That puts all of economic theory out of kilter at the first step.
Might physics help in bridging the divide? Physics also deals separately with the microscopic 鈥� the individual movements of particles in a gas, say 鈥� and the macroscopic, for example when the sum of those movements creates a pressure that enables a gas to push a piston. But it also has mathematical frameworks, such as statistical mechanics, capable of bridging the gap between them.
Weinstein has been involved in one of the most audacious attempts to meld physics and economics: showing how gauge theory, the mathematical underpinning of the quantum field theories of the standard model of particle physics, might be the key to a rational theory of economics based only on physically observable quantities, rather than hypotheticals akin to the economists鈥� list of ordered preferences. Impressed by this idea, physicist Lee Smolin of the Perimeter Institute in Waterloo, Ontario, Canada, has .
It is part of a wider movement towards 鈥渁gent-based鈥� models that are better equipped to deal with the non-equilibrium behaviour of markets. In these models agents are treated as particles, albeit ones imbued with adaptive behaviour. They make mistakes, try to learn from them, and change their beliefs and expectations about the market on the fly. Because most humans tick in broadly similar ways, all those behaviours can be dealt with statistically.
The catch, says economist Blake LeBaron of Brandeis University in Waltham, Massachusetts, is that the models cannot be wrapped up tidily in directly solvable formulae like those of traditional equilibrium theories. Instead, they have to be explored with computer simulations. Still, they can show how great rallies and crashes emerge as avalanches of buying or selling behaviour among uncertain investors.
Some physicists suggest that these models can be used to identify particular moments when markets are likely to make big moves. of the University of Miami in Florida has used agent models to show how 鈥渨indows of predictability鈥� open up when many agents come to hold similar beliefs, meaning that they are likely all to act in a similar way in the near future. In one demonstration, he and his colleagues were able to use this technique to predict the movements of a foreign exchange market with more than 90 per cent accuracy (Physica A, vol 299, p 222).
The new ideas pour cold water on some choice morsels of economic orthodoxy 鈥� notably, that the complex financial instruments known as derivatives stabilise markets. Derivatives are contracts to buy or sell an asset at a set future time, and conventional economic theory says the additional flexibility of balancing any current losses with the prospect of future gains smooths out market wrinkles.
Recent work, notably by economist William Brock of the University of Wisconsin, Madison, and colleagues, incorporates adaptive models of agent behaviour to suggest that classical economics might have got things backwards. People participating in markets will quickly copy investment strategies that seem to be working, creating swings in the future price of assets that are actually fundamentally destabilising.
Despite the insights it could give, such non-equilibrium modelling remains at the edges of accepted economic theory. 鈥淭he old ideas get drilled into economists as they get started,鈥� says LeBaron, 鈥渁nd I think people have a hard time detaching.鈥� Smolin also cautions against assuming that ideas imported from physics can contribute anything substantial to an established field. 鈥淏ut there are good reasons to think that the present is a propitious moment for seeking to improve economic theory,鈥� he says.