(FORTUNE Magazine) – Have you ever wondered what drives fads? Why, for example, so many people suddenly buy lava lamps, take melatonin, or get their eyebrows pierced? Even if we never understand the appeal of sticking pieces of metal through tender body parts, a new theory on the economics of fads explains how such seemingly absurd waves of conformity unfold.

The theory, hatched in a 1992 paper, may itself become a fad among marketing managers. Among other things, it shows how some products become outlandishly successful. It also sheds light on bank runs, bidding wars, and how certain marvels of mediocrity become corporate stars.

Fads have long bothered economists. Their field's bedrock is that human behavior is basically rational. But that notion is hard to reconcile with the dubious mass thinking that leads to, say, dictator-worship or Internet companies' recent stock prices. So economists seek hidden rationality in such herd behavior. Some maintain that it's induced by the costs society levies on nonconformists, as exemplified by Saddam Hussein's cheering crowds. But then, no one was hurt socially by just saying no to Netscape at $85 a share.

Stabbing again, some economists say bandwagons are propelled by "positive-feedback loops." For example, once a few people adopt the convention of driving on the right side, others must follow or pay big time. But this fails to account for fads' tendency to quickly fade.

Such conundrums, however, can be explained by the theory of informational cascades, proposed by Ivo Welch and Sushil Bikhchandani at UCLA and David Hirshleifer at the University of Michigan. Its main premise is that we often imitate others' actions when making choices with limited information about what's best. That inevitably leads to chain reactions--cascades--of imitation. We may despise conformity, but it's almost a law of nature.

Animal behavior is a nice example. Schooled by evolution's hard knocks, many animals are hardwired to mimic in order to avoid the possibly fatal costs of direct experience. Territorial creatures, like certain birds, tend to cluster because latecomers copy early settlers' choices as an indicator of nearby goodies. Females of various animal species copy other females in choosing mates--males lucky enough to have amorous contact with one or two females get mobbed, while equally winsome males stay single.

We, too, seem dyed-in-the-genes copycats: Teenage girls swooning over the rock star of the moment come to mind, as do politicians, who are well aware of how imitation feeds momentum (which is why early primaries are crucial). Doctors are notorious fashion slaves: Tonsillectomies were once almost routine for children but now are out of style.

The new theory shows how imitative behavior leads to vast but fragile domino effects. Roughly, it posits that we decide what to do by melding our partial information about the true value of the available choices with observations of how others have previously chosen. Even if our first-hand data say nay, a few yeas by others can override our private information and induce us to echo yes. Our decision, in turn, makes others similarly override their own personal data, causing blind imitation to snowball.

Importantly, the process rules out group decisions based on pooled knowledge, so the herd can race off in a silly direction contrary to most of its members' first-hand information. But if later deciders get additional first-hand data conflicting with snowballing yeas, they're likely to say nay and break the pattern. That's because we're all attuned, if only unconsciously, to the fact that the blind imitation driving most fads reveals little about the underlying value of choices--imitators' uninformative choices are trumped by real data. Thus, fads are fragile.

Cascades don't account for all mass conformity--blind imitation probably has little to do with sales of Coke, with which we get lots of personal experience by repeated buys. Other herd drivers may work in tandem with cascades. Many things dynamited communism in East Germany, but the trigger seems to have been an informational cascade that went off after a few high-profile protests in Leipzig. In general, cascades come into play when we lack the time or resources to think through our options, and others' choices are presented to us by indicators such as book-jacket blurbs or a line outside a restaurant.

Besides elucidating fads, the theory offers insights on daily life. Job hunters, beware of gaps in your resumes, for they may signal prospective employers that others have rejected you, setting off a stigmatizing cascade--the longer you're unemployed, the worse it looks.

The workplace is fertile ground for cascades. Recently hired managers, for instance, know predecessors' actions without knowing all the reasons, hence may propagate dubious policies--the Peter Principle's elevation of employees to their level of incompetence may be one result. But stars who get ahead riding cascades may have another thing coming: The imitative waves propelling them can suddenly collapse if decision-makers get first-hand data about their failings. When that happens, the upstart who got the job you deserve may go the way of the Cabbage Patch doll.

--David Stipp