Farmer0606.Pdf

8.6 econophysics MH 2/6/06 10:41 AM Page 686

NEWS FEATURE NATURE|Vol 441|8 June 2006

CULTURE CRASH Some economists had hoped that physicists might shake up the rigid theories typical of mainstream

economics. But so far, they’re unimpressed by physicists’ handling of the markets. Philip Ball reports. CHINAFOTOPRESS/GETTY

or the past two decades, some physi- their statistics. At face value it is a damning It is tempting to interpret this as a mere acad- cists have been trying to apply their indictment, and raises the question of whether emic turf war. But Ormerod and colleagues are ideas and tools to an area that seems a econophysics will ever make a genuine contri- among the few people in economics who have Flong way from traditional physics. They bution to economic theory, or whether it is taken econophysics seriously. Most economists are exploring the notion that there might be a doomed to remain a fringe interest. don’t know the discipline exists — and if they kind of physics of the economy — an ‘econo- did, they would probably heap derision on it. physics’, as it has been dubbed1. Last year, some Claim to blame The idea that physics might have something of these econophysicists even went as far as to Some econophysicists admit that there are useful to contribute to economics arises suggest that economics might be “the next problems. “Econophysics is a field with very because both fields are concerned with systems physical science”2. uneven quality,” says Doyne Farmer, a physi- of many interacting components that obey spe- But now this unlikely marriage is showing cist at the Santa Fe Institute in New Mexico, cific rules. Statistical physics describes the signs of turning sour. Even those economists who made pioneering contributions to the behaviour of bulk matter based on the forces who at first welcomed econophysics are start- study of chaos before moving into economics. acting between atoms and molecules. Econom- ing to wonder whether it is ever going to deliver Yi-Cheng Zhang of the University of Fribourg ics studies the interactions between economic on its initial promise. Early successes in model- in Switzerland is even more ready with ling financial markets have not led to insights a mea culpa. “My economist friends are elsewhere, some complain. Matters came to a right. The literature is often littered with head at the Econophysics Colloquium, held at garbage. We can find gauge-field theor- the Australian National University in Canberra ies of finance, quantum options and last November. A group of economists attend- so on. In short, anything goes.” ing the meeting were so dismayed with what But others reject the accusations. In “Econophysics they saw many physicists doing that they response to the Canberra critique, Joe penned a forthcoming paper entitled ‘Worry- McCauley, a physicist at the University is a field with ing trends in econophysics’3. of Houston, Texas, who now works very uneven In their critique, economist Paul Ormerod mostly on economic problems, says, quality.” — of the London-based consultancy Volterra and “Would one write an essay called his co-authors accuse econophysicists of a ‘Worrying trends in physics’ simply Doyne Farmer litany of sins: applying inappropriate assump- because a few minor researchers put out tions to economic systems, failing to do their bad papers? Bad papers, even wrong homework properly, getting fixated on a small papers, appear in every issue of every corner of the subject, and being sloppy with scientific journal.”

NATURE|Vol 441|8 June 2006 NEWS FEATURE

Econophysicists have focused on the financial market because it offers a lot of good-quality data.

was revisited in the early 1960s, when mathe- matician Benoit Mandelbrot showed that fluctuations in cotton prices have a statistical distribution that differs from that expected of a typical gaussian process — where each event happens randomly and independently of all others. There were more large fluctuations than a gaussian distribution predicts4. This has sig- nificant implications for economic theories that assume market ‘noise’ to be gaussian — but more importantly, it suggests that big fluctuations, perhaps even market crashes, are not rare anomalies but intrinsic to normal market behaviour. Mandelbrot’s 1963 paper on price fluctuations is now regarded as one of the key precursors to modern econophysics.

Out of equilibrium But it wasn’t until the early 1980s, when an unusual mix of researchers got together at the Sante Fe Institute, that economists showed much interest in scientific ideas related to complex systems. They were helped by advances in computing. “Once we got desktop computers, we could model systems of many ‘agents’ — market traders, say, or businesses. mist at the University of Western Sydney in agents and allow them rules of behaviour and Arguably, deriving microeconomic princi- Australia. Keen had hoped in particular that see how they evolved,” says economist Brian ples from the behaviour of individual agents econophysics might break his fellow econo- Arthur, who worked at Santa Fe alongside should pose similar problems to deriving mists’ misconceived obsession with equilib- physicists and evolutionary biologists to thermodynamic laws from interatomic forces. rium. “Equilibrium thinking still has them in develop non-traditional approaches to eco- M. OESER/AFP/GETTY M. OESER/AFP/GETTY The rules dictating how interactions play its unshakeable thrall,” he says. nomics5. Such computer simulations of the out between economic agents are admittedly A glance at almost any plot of commodity economy led to models of ‘interacting agents’6 more complex than the forces between atoms, prices over time belies the idea of market equi- that were influenced as much by work on cog- but in conventional economics the rules have librium: the values fluctuate wildly. But in nition and evolutionary biology as by physics. always been grossly simplified to make the neoclassical theory, these fluctuations are In these models, researchers could give the models workable. regarded as background ‘noise’ caused by interacting agents any decision-making strat- For example, the core theory of mainstream unpredictable ‘shocks’ from outside the eco- egy they desired, and therefore study markets economics, the neoclassical model, argues that nomic system, to which the market constantly with different underlying behaviours. “What agents always act with perfect rationality to and quickly adjusts. An attempt to explain we found was quite surprising,” recalls Arthur. maximize their ‘utility’ (for example, profit), such fluctuations using statistical physics was “Under some restrictive conditions, you get based on complete information about the state made as early as 1900 by Louis Bachelier; he market equilibrium, but under other condi- of the market as a whole. In this picture, an eco- proposed an explanation that introduced tions you get much more complicated out- nomic market quickly reaches an equilibrium the theory of random walks, which was later comes.” It seemed there was no good reason to state, in which commodities find the price that developed independently by Einstein to believe that microeconomics always operates perfectly balances supply and demand. explain brownian motion. at equilibrium. “The economy is out of kilter Bachelier’s theory was deemed too strange to most of the time,” says Arthur. That, he says, Model world be taken seriously by economists. But the issue accounts for one of the virtues of econophysics. Economists recognize that real human agents “The core of economic theory is still built do not always act in such a coldly rational way, around equilibrium models, but most models and that they generally have to manage with in econophysics are non-equilibrium ones.” incomplete information. But although Nobel But those economists who have adopted prizes in economics were awarded in 2001 and new approaches such as agent-based model- 2002 for work that recognizes these limitations, ling have become increasingly frus- neoclassical theories — and particularly the “Physicists trated with the intransigence of main idea of equilibrium — remain central to main- suffer from -stream economics. Some have even stream economics. resorted to starting their own publica- Ormerod and his colleagues, and other a belief that tion, the Journal of Economic Interac- physics-friendly economists, had hoped that there must be tion and Coordination, the first issue of econophysics would help them create a universal rules.” which appeared online in May. Zhang new economics that is free from some of the says that three of the four authors of the dogmatic assumptions characterizing the — Paul Ormerod Canberra critique are victims of the mainstream discipline today. “Economics intellectual exclusion imposed by desperately needs econophysics,” claims mainstream economists. “That’s why they had Ormerod’s co-author Steve Keen, an econo- such high hopes when physicists offered what

NEWS FEATURE NATURE|Vol 441|8 June 2006

IS ECONOPHYSICS REALLY NEW? Although the idea of economists the circulation of the planets. frequently alluded to scientific the foundations of statistical borrowing concepts from Pierre-Simon Laplace and ideas and analogies. physics. The early microecono- physics might seem unlikely, it the Belgian astronomer Microeconomic theory, mists Francis Edgeworth and has been a feature of economic Adolphe Quetelet which strives to under- Alfred Mashall drew on some of theory ever since its inception. helped to establish the stand economic phe- the ideas of these physicists, in Adam Smith (pictured) wrote idea that there are nomena by building particular the notion that the his Wealth of Nations, which laid natural laws, akin to them up from the economy achieves an equilib-

the foundations of economic Newton’s laws of behaviour of individ- rium state like that described IMAGE/ALAMY CLASSIC thought in 1776, in an intellec- motion, that govern ual ‘agents’ in the for gases by Maxwell and tual climate that was infused human social systems economy, was estab- Boltzmann. Edgeworth and with newtonian ideas about such as the economy. lished in the late nine- Marshall’s microeconomics led causative forces. Indeed, some And nineteenth-century teenth century — just as to the neoclassical theory of of his contemporaries compared economic thinkers such as James Clerk Maxwell and economics that defines the the circulation of commerce to John Stuart Mill and Karl Marx Ludwig Boltzmann were laying mainstream today. P.B.

seemed to be an alternative.” doesn’t actually conserve. “Conservation of classes, was falsified in the 1970s by M. F. M. So why have some of these physics-friendly money is too far from the truth in a credit Osborne of the Naval Research Laboratory economists become fed up? Although economy like ours,” McCauley agrees. But he in Washington DC, who he regards as the Ormerod and colleagues are highly critical of says that such mistakes are rare. first econophysicist (see ‘Is econophysics mainstream economic theory, they point out More broadly, claim the Canberra authors, really new?’). “Economists are overwhelmingly that “economics is not at all an empty box.” physicists fail to appreciate how diverse and unaware of this contribution. Worse, the The Canberra critique accuses econophysicists changeable the economy is in practice. “Physi- majority have no interest in anything that of ignoring the existing literature — a charge cists suffer from a belief that there must be physicists write.” also levelled at physicists when they began to universal rules,” says Ormerod. “This is not a Arthur counters that economics has a long dabble seriously in biology. handicap in the physical world, but it is in eco- history of assimilating ideas from outside the Zhang agrees this is a fair complaint. “Most nomics, where the behaviour of agents may field. But he explains that while economics tol- econophysicists don’t care about the under- differ across time and place.” But Farmer erates some ‘unorthodoxy’, it also isolates it lying idea, history and cultural background thinks that looking for generalities can still be from the core theory that is taught to students to the problem. All they want is to get a lot of useful: “The typical view among social scien- and practised by academics. There is an atti- data, crunch the numbers better than anybody tists is that one should focus on documenting tude of “fine, but not in my journal,” he says. else, and publish a lot of papers.” But he admits and explaining differences. Physicists have Despite their concerns, the authors of the that doing the job properly is very daunting: jumped in and said, ‘Before we do that, let’s Canberra critique still hold out hope for “Compared to anything I’ve done in physics, spend some energy on first trying to under- econophysics. “It’s just that we don’t want it the requisite reading is enormous. I get a pile stand if there is any way in which everything is going off the rails,” says Ormerod. Yet for of books delivered every month, and I’m just the same’.” McCauley, the goal of econo- overwhelmed.” “Econophysics will physics is not to batter its way into Another criticism raised by Ormerod and Out on a limb displace economics the citadel, but to raze and rebuild colleagues is that econophysics has too narrow Economist John Sutton at the in both the it. “Econophysics will displace a focus. Physicists have made important con- London School of Economics economics in both the universities tributions to finance and industrial econom- agrees that physicists can universities and and boardrooms,” he says, “simply ics, Ormerod acknowledges, but he wants help to identify relationships boardrooms.” — because what is taught in econom- them to bring an open mind to other areas of within economics that are Joe McCauley ics classes doesn’t work.” the economy. “I think they fail to appreciate driven by, say, elementary Still, if econophysics is to survive that financial markets as such are not terribly statistics, irrespective of any assumptions about and prosper, it needs support from somewhere, important in economic theory,” he says. “They agent behaviour. He says he welcomes out- says Farmer. “It is clear that this is not going to are just a special case of a more general theory siders to economics “who may look in different come from economists.” Turning to the physics of markets.” places for interesting relationships”. community for help has not been the answer Why are physicists so finance-centred? Sutton is, however, a rare example of a main- either. Rosario Mantegna, an econophysicist at Ormerod suspects that they feel more com- stream economist who is aware of the econo- the University of Palermo in Sicily, says that in fortable with the generally high-quality and physics literature. Arthur believes that some academia physicists cannot pursue econo- long-term data from financial markets. Most shortcomings of econophysicists may be physics as their main research. “There is no one other economic data exist in small, noisy data caused by their enforced isolation: “They tried in the United States who has gotten tenure sets. Some econophysicists defend the focus to publish in mainstream economics journals, based on work in econophysics,” Farmer on finance precisely for these reasons: but were rebutted. It’s a shame and a scandal agrees. “This is a real pity for physics.” ■ “Finance data are strong enough to falsify spe- that the journals haven’t opened their pages to Philip Ball is a consultant editor for Nature. cific classes of models,” says McCauley. Farmer them.” Consequently, econophysicists set up adds that physics has often been successful by their own journals. This means they don’t get 1. Mantegna, R. N. & Stanley, H. E. An Introduction to tackling problems that could be solved with the contact with economists that would have Econophysics (Cambridge Univ. Press, 2000). 2. Farmer, J. D., Shubik, M. & Smith, E. Phys. Today September the data and the tools to hand. “rounded off their rough corners”, Arthur says. 37–42 (2005). Perhaps more damning, Ormerod and col- Perhaps not surprisingly, some econophysi- 3. Gallegatti, M. et al. Physica A (in the press). leagues worry that some econophysicists have cists reject the Canberra critique outright. “The 4. Mandelbrot, B. J. Business 36, 394–419 (1963). imported concepts from physics that just don’t problem is not with econophysics,” McCauley 5. Anderson, P. W., Arrow, K. J. & Pines, D. (eds) The Economy as an Evolving Complex System (Addison-Wesley, Reading, belong in economics. For one thing, they say, says. “The problem is within the economics Massachusetts, 1988). some have tried to apply conservation laws to profession.” He argues that the neoclassical 6. Tesfatsion, L. & Judd, K. (eds) Handbook of Computational quantities, such as money, that the economy model, which is still taught in all economics Economics Vol. 2 (Elsevier, Amsterdam, 2006).