Share This
March 2003 Cover Art

Search Finance & Development

Advanced Search
About F&D


Back Issues

Write Us

Copyright Information

Free Email Notification

Receive emails when we post new items of interest to you.

Subscribe or Modify your profile

Finance & Development
A quarterly magazine of the IMF
March 2003, Volume 40, Number 1


The Lab Man
How experimental economics emerged from the shadows

Jeremy Clift interviews Nobel Prize winner Vernon L. Smith

Vernon Smith isn't one to stand on ceremony. In fact, he had to practice bowing before receiving his Nobel Prize from the King of Sweden. "They were scared to death that I was going to wear my cowboy boots," says Smith, whose fingers are adorned with silver inlaid jewelry from the Hopi Indian tribe. "There was speculation all over Stockholm. I couldn't believe it. Who cares?"

For a man who has built his reputation on controlled experiments to test market theories in the laboratory, Vernon Smith is a bit of a free spirit. His graying flaxen hair tumbles over his shoulder in a ponytail as he gives a quick imitation for his visitors of how he bowed to the king—a stiff formal bow, with just a slight inclination of the head. "For Americans, that takes a little training," he laughs.

Until recently, economics was widely regarded as a nonexperimental science that had to rely on observation of real-world economies rather than the controlled laboratory experiments often used in physical science. "Because of the complexity of human and social behavior," wrote Paul A. Samuelson, author of the classic textbook Economics, "we cannot hope to attain the precision of the physical sciences. We cannot perform the controlled experiments of the chemist or biologist. Like the astronomer, we must be content largely to 'observe'."

Smith changed all that, pioneering the use of experiments conducted in the controlled environment of the laboratory to test economic theories, in particular why markets work the way they do. The lessons from what has become known as experimental economics are valuable for both researchers and policymakers and can be applied widely in such areas as financial market theory and behavior, natural resource economics, and the deregulation of industries such as electric power and water.

Smith, a 76-year-old professor of economics and law at George Mason University in Virginia, shared the 2002 Nobel Prize in economics with Daniel Kahneman, a professor of psychology and public affairs at Princeton University, whose separate research into human decision making helped develop the field of behavioral economics. The Royal Swedish Academy of Sciences, which awards the $1.07 million prize in economics, said both men had "changed the direction of economic science."

The two laureates have not always seen eye to eye, with Smith accusing Kahneman and his longtime coresearcher, the late Amos Tversky, in a 1991 article in the Journal of Political Economy of "ignoring contrary interpretations and evidence over extended periods of time" in work examining the rationality of markets. But whatever differences the two have were not on display when they accepted their prize in Stockholm, when Smith saluted Kahneman "for his ingenuity in the study and understanding of human decision and its associated cognitive processes demonstrating that the logic of choice and the ecology of choice can be divergent."

Passionate advocate

Smith is passionate in his commitment to laboratory testing in economics, a field in which he remains heavily involved, not just in research but also in developing programs and workshops for high school students. "You can't believe how successful that is," he says, sitting in his office at the Interdisciplinary Center for Economic Sciences (ICES), a research center he helped found in 2001 and to which he will contribute his share of the Nobel award.

Initial experiments in the field were aimed at testing what the Swedish Academy said is perhaps the most fundamental result in economic theory: under perfect competition, the market price establishes an equilibrium between supply and demand at the level where a marginal buyer places as high a value on a good as a marginal seller. In Smith's early laboratory experiments, subjects were randomly assigned roles as sellers and buyers; they were individually told the lowest price at which they could sell or the highest price at which they could buy. Given the distribution of such reservation prices, Smith was able to determine the theoretical equilibrium price—the price acceptable to as many sellers as buyers. As early as 1962, when he published the results of his first experiments, Smith found, much to his surprise, that the prices obtained in the laboratory were very close to their theoretical values, even though subjects lacked the information necessary to calculate the equilibrium price. Smith's experiments proved that large numbers of perfectly informed economic agents were not prerequisites for market efficiency—a departure from conventional economic thought.

Smith also initiated the use of wind tunnel–type experiments in the laboratory, where proposed auction mechanisms for privatization and public procurement can be tested in advance. Since these mechanisms are frequently complex and it is difficult to assess their performance solely on the basis of theoretical considerations, the experimental method becomes particularly useful. In similar experiments, Smith has studied different mechanisms for allocating airport landing rights using computer-assisted markets. He has also evaluated various means of organizing energy markets in Australia and New Zealand, where the results have influenced actual market design.

Smith says laboratory testing has its uses for assessing policy choices in developing countries, for example, during crises. "My point is that there are a whole lot of things that we can do in the laboratory to study things like changes in monetary regimes. We can get an understanding in the lab that will give insights in the field. We don't see experimental or laboratory economics as something that motivates only research in the laboratory."

Laboratory experiments do not provide policymakers with iron-clad answers to key economic questions but do offer a quick, cost-effective way to identify market and policy flaws before ideas and theories become major public policy initiatives. When the California energy market was deregulated, for example, the problems were caused not so much by Enron, Smith argues, as by the flawed design of the deregulated market.

Having been raised by a socialist mother, Smith started out as a skeptic about the efficiency of markets, "but my experimental subjects revealed to me the error of my thinking," he wrote in a 1994 article in the Journal of Economic Perspectives. Born in Wichita, Kansas, he earned his bachelor's degree in electrical engineering at the California Institute of Technology in 1949 but quickly became fascinated by economics. His scientific training made him dissatisfied with the way economists collected data. "The sciences are much closer to rigorous observation than economics," he says. "The data used by economists have not usually been gathered for scientific purposes, whereas in science, most of the observations are gathered by the scientists themselves. I didn't appreciate all of those things when I started, so that was ultimately what attracted me to experiments—the dissatisfaction with the observational state of economics."

What is experimental economics?

Experimental economics allows for the controlled study of markets, trading rules, and the behavior of participants. It enables policymakers to "bench test" competing policy options by comparing the likely outcomes of alternative sets of rules. Most experiments are conducted in a laboratory setting where researchers observe groups of people buying, selling, and bidding in games designed to replicate the environment of the real economy. Participants trade using real money and can often keep their profits. Researchers learn how and why markets react to changes in regulations and trading rules by observing how subjects respond during different stages of the experiment.

Smith, who received his Ph.D. from Harvard in 1955, has taught at a number of U.S. universities, including Purdue and Arizona, where he set up his experimental laboratory. Many other universities have since set up their own laboratories, and his work has led to the rapid growth, over the past two decades, of experimental methods in economics. At ICES, he and coresearchers led by Smith's collaborator, Professor Kevin McCabe, are branching out into behavioral and neuroeconomics, which bridges the disciplines of economics, psychology, biology, and philosophy. A natural meeting ground for economics and neuroscience is the study of personal exchange, where cooperation yields gains from exchange, but these gains can be undermined by opportunistic self-interest.

Economists are the exception

Participants in Smith's experiments come from many fields and include corporate executives, market traders, graduate students, and trained economists. "Among the more amusing results that have come out of experimental economics," says Joseph Stiglitz, who shared the 2001 Nobel award, "are those concerning altruism and selfishness. It appears (at least in the experimental situations) that experimental subjects are not as selfish as economists have hypothesized, except for one group—the economists themselves."

"We find that people are more trusting and achieve more cooperative outcomes more often than game theory predicts," says Smith, who now considers himself a libertarian, "and actually they all make more money doing that."

Vernon Smith on. . .


Countries have really got to be able to manage their economies much better or they are going to be hurt and won't get the benefits of globalization. Getting the benefits is a discipline if they will just learn that. And being bailed out by the IMF, even with rules and everything, is not the way to solve their problems.

Foreign aid

The history of foreign aid has been pretty disastrous in the sense that it's country-to-country transfers, and in a lot of undeveloped countries with corrupt and dictatorial governments, it just increases their power. The way you get development is for investors to invest in those countries and see how their money is going to be used. Public policy, I think, should be directed at removing artificial barriers to that.

The next crisis

We think the next crisis in California is going to be in water. California has cycles of wet and dry. We need to get a better way of allocating water in the whole West, and even here on the East Coast (of the United States), the quality of water is starting to be a problem.

The stock market

I don't know when the low of this market will be, but I know that it's a lot lower than it was three years ago so I'm optimistic. I think it's a buying opportunity, but I may be wrong, so don't follow my advice.

On receiving the Nobel Prize

It was a lot of fun. It is really a celebration of the dignity of humankind, with the intellectual achievements of it being an important part of that dignity.

If Smith's methods are so useful to policymakers, what was the reason for the resistance in the past from some economists to his techniques? "I think it's basically that economists are not trained in these methods," Smith says. "There's a big investment; there was for me and for the people who helped because we weren't trained in laboratory experiments either. Now, of course, it's easier to get the training, but there are all kinds of practical reasons why you don't want to be the first. It's just like econometrics. It took a long time before econometrics was accepted, and I think experimental economics, which is even further from the traditional methodologies, is going to take even longer."

Critics of experimental economics worry that subjects bring to experimental situations modes of thought determined outside the experiment and that therefore the experiments are not as clean and the inferences not as clear-cut as in the physical sciences. Frank Shostak of the Mises Institute argues that experimental economics, far from being the wave of the future, actually stifles understanding by treating humans as laboratory animals.

But many in economics welcome the challenge to the old rationalist paradigm of mainstream economics, which assumes that all participants have the same information and act perfectly rationally, markets are perfectly efficient, and unemployment does not exist. "What we don't really know and haven't really applied ourselves to understanding," Smith says, "is how it is that people who don't know what we know as economists can get to the equilibria that we describe by methods that they wouldn't have the foggiest idea how to use. That's the big mystery in economics, and there's hardly anybody working on it because they believe in these constructivist models. Experimental economics," he adds, "won't have achieved anything near its potential until it changes the way economists and theorists think about their problems."

  Vernon L. Smith, 1991, "Rational Choice: The Contrast Between Economics and Psychology,"
Journal of Political Economy, Vol. 99 (August), pp 877–97.
  ———, 1994, "Economics in the Laboratory,"
Journal of Economic Perspectives, Vol. 8 (Winter), pp 113–31.

Jeremy Clift is on the staff of Finance & Development.