By Shankar Vedantam
Monday, December 17, 2007
Nearly four decades ago, psychologist Stanley Milgram had a volunteer stand stock still on a busy New York sidewalk and look up at the sky. About one in every 25 passersby stopped to look up, too. When five volunteers were recruited to sky-gaze, nearly one in five passersby stopped to look up.
When Milgram and his colleagues assembled a group of 18 volunteers to simultaneously look up at nothing in particular, nearly one in two passersby looked up to see what was going on, snarling traffic within moments.
In the intervening years a variety of experiments have unearthed similar findings in a number of domains. When people have limited information about something important -- a potential crisis in a building several stories high, a fire alarm that goes off in a crowded theater or a sudden drop in the stock market -- they use other people as guides to their own behavior.
This can be smart if the people on whom you are modeling your behavior know more than you do. But in many situations, no one knows very much. When you follow people who don't know what they are doing, and other people follow you, the resulting feedback allows small events to trigger huge and irrational changes in group behavior.
Such thinking was behind the "Washington's real estate market will never soften" belief between 2000 and 2005, and it's also behind the pervasive pessimism that plagues the same market today, said Robert Shiller, an economist at Yale University and the author of the book "Irrational Exuberance."
Shiller argues that patterns of market behavior have a lot in common with infectious diseases. His book explores the idea of "contagion" in financial markets -- except that instead of the flu, Shiller talks about the spread of dogmas from one place to another.
"I am talking of views that seem intuitively right," Shiller said. "One hears other people saying things and confirming ideas you have. When things are commonly accepted, you file it in your brain as something that is true."
One of those intuitively appealing ideas that was widely passed around during the Washington real estate boom was an example of what Shiller calls the uniqueness bias -- investors' belief that their particular investments are somehow immune to normal economic processes. In the D.C. area, that bias centered on the idea that, in the nation's capital, the real estate market is supposed to be largely immune to the kind of shocks that buffet other places.
But even if Washington real estate investments are marginally safer than those in other places, that has already been factored into the higher price of housing in the area and is therefore of little benefit to new investors, Shiller said.
Another problem Shiller has identified in the feedback loop is the role of hindsight and narratives in shaping market behavior -- a Monday-morning quarterback effect. Every rise and dip in the stock market, for example, is assigned a narrative on a daily basis. A particular set of factors is said to have caused the market to go up or down, and this judgment quickly becomes the general wisdom.
After a sudden stock market dip in 1989 was quickly ascribed to an airline merger, Shiller conducted a survey asking investment managers which they had heard about first, the market dip or the narrative explanation. Although nearly all the investment managers believed the airline merger story, no one had heard it before the market fell.
"Now we are all in agreement the housing market is going to fall," said Shiller, describing the latest version of market groupthink. "That is what happens when you have contagion of an idea. You are going around wearing a tie and you realize that no one else is wearing a tie, and suddenly it seems so wrong to be wearing a tie."
Jian Yang, who teaches finance at the University of Colorado Denver, and David Bessler, a professor of agricultural economics at Texas A&M University, recently showed that contagion played a central role in the 1987 stock market crash, in which several international markets with very different economic fundamentals moved together in lock step.
In a complex analysis that meticulously tracked data from several countries, and that is to be published next year in the European Journal of Operational Research, they showed how market weakness spread from one country to another, in a sort of international version of Milgram's passersby experiment in New York.
Yang said that although many investors recognize the problem of contagion and herding behavior, this does not keep them from following the herd, possibly because they feel less regretful about bad decisions if they know many other people made the same mistake.
As global communication has increased with the advent of the Internet, Shiller and others said, ever-greater feedback systems and contagion effects are likely in real estate and financial markets.
"A lot of Japanese investors admitted that they saw what happened with the U.S. market and had a gut panic" during the 1987 crash, Yang said. Rather than trust their own accurate assessments of Japanese economic fundamentals, they thought, "Maybe something is wrong with my assessment of the Japanese market."