Bubble physics

IF THE US HOUSING market crashes--next month, next year, or whenever--someone will have to take the blame. Accusations will fly that banks were irresponsible in their lending, or that people got just a little too greedy. Meanwhile, economists and social scientists are increasingly coming around to a rather different view: that market ''bubbles" in houses, stocks, or anything else are just about as inevitable as earthquakes or floods, rising up out of the stream of humanity by virtue of the most elementary of human behavioral traits.

In a strange scientific twist, this idea is now gathering support from an unlikely source: physicists. According to a recent study by two French scientists, our inclination to imitate one another--in everything from buying houses to having children to clapping at the end of concerts--makes mass behavior prone to dramatic and seemingly inexplicable swings that echo the patterns found among the building blocks of matter. We've all heard that people are like sheep. But maybe we're more like atoms.

. . .

It's an old idea in economics that people make decisions ''rationally"; when faced with alternatives, we take all factors into account and act to ''maximize" our own well-being, in financial terms or otherwise. Yet acting rationally requires adequate information on which to base a decision, and finding that information isn't always so easy. In the absence of complete information, we often do what we see others doing in the implicit hope that they might know something we don't.

In an influential paper published 13 years ago, David Hirshleifer of Ohio State University, working with fellow economists Sushil Bikhchandani and Ivo Welch, argued that imitation can make human behavior take on a somewhat ''mechanical" aspect, and probably underlies a host of sudden social and economic transformations, ranging from the revival of New York's Times Square in the early 1990s to the Internet stock craze of the late 1990s. Since then, economists and social scientists have begun to accept and explore this idea in earnest.

But now it is physicists, of all people, who are helping to illustrate just how widespread a phenomenon imitation may be. Earlier this year, Quentin Michard of the School of Industrial Physics and Chemistry in Paris and Jean-Philippe Bouchaud of the Atomic Energy Commission in Saclay, France, developed a surprisingly general theory which suggests that social imitation should make runs in financial or housing markets, say, or trends in birth rates, work in a remarkably similar way.

Imitation means that one person's decision influences that of another; if many people buy cell phones, I will be more likely to buy one too. As Michard and Bouchaud point out, this basic effect has a direct analogy with the physics of atoms. Within a piece of iron, for example, the atoms are like microscopic magnets, with north and south poles; you can think of them as tiny arrows pointing in various directions. Put the chunk of iron in a strong magnetic field, and all the tiny arrows will tend to line up with it, like soldiers falling into formation. But even in the absence of such an overwhelming external force, these atoms also influence each other; if many point in one direction, they tend to coerce others nearby to line up in the same way also.

If you think of the direction of an atom as an ''opinion" or ''behavior," then you could think of what's happening at the atomic level as imitation--what one atom does influences what others might do. On the face of it, this seems like just another analogy. Yet physicists have spent a century developing the mathematics that lets them understand the collective patterns that tend to emerge when hundreds of millions of such atoms interact with one another, or get pushed around by external influences. What's more, they have found that the way collections of atoms behave often depends only very weakly on the precise details of how the individual atoms interact with one another. So physicists have good reason to believe that crude models--even those that leave out many details of human thinking--might still give accurate results.

Using this mathematics, but directing it toward the social world, Michard and Bouchaud conclude that the general influence of imitation should be to exaggerate our collective social response to real world trends; even small external changes can lead to disproportionately large consequences. As they put it, ''Imitation leads to distortion, and to a decoupling of cause and effect."

. . .

The theory predicts that the same should be true with people--and real data seems to back this up. Studying plummeting European birth rates in the late 20th century, and the rapid adoption of cell phones in Europe in the 1990s, the two physicists found that these changes took place far too fast to be explained by independent decision-making based on external factors alone. Many people had to be making decisions by imitation rather than by independent judgment. ''These natural trends," Bouchaud explained in an email, ''were substantially amplified and exaggerated by peer pressure." On a smaller scale, they found that their model also predicts accurately how episodes of clapping begin and then abruptly end following a concert.

''This work is provocative and intriguing," says Hirshleifer, who suggests that the physicists' unusual approach could become increasingly important in helping to understand the social world. As he points out, many traditional economists would be critical of the approach because it lacks firm rooting in the decision-making process of individuals based on analysis of costs and benefits. But with their less detailed approach, the physicists derive predictions that go well beyond those possible through standard economic reasoning.

''The assumptions they make seem reasonable," Hirshleifer adds, ''and their empirical findings certainly make this work seem promising."

It is surprising that by treating people as if they respond with the simplicity of atoms, you can still explain real world social trends. We're accustomed to thinking that the social world is so hard to understand because people, as individuals, are so hard to fathom. But maybe, paradoxically, our own behavior is often simpler than we think. It's how we influence one another that makes our world so complex.

Mark Buchanan is a science writer based in Cambridge, England.