One of the most interesting aspects of the new-fangled tradition cited by the Nobel Foundation last week is the difficulty — at least on the surface — that experimental economics has had over the years in establishing an institutional identity.
Considering that a single experiment designed to investigate, say, attitudes towards risk, can take $50,000 in ready cash to provide incentives for its subjects, requires a big computer lab and often fails, you’d think that success would have required a university prepared to back a program to the hilt through degrees, appointments and big budgets. And you would be partly right.
Yet the developments that were honored last week actually unfolded some 40 years ago. The award went to a man who has wandered from place to place. And only relatively recently have the foremost universities begun to invest in the kind of computer labs that the Nobel citation described as “wind tunnels” designed to test economic mechanisms.
Understand that the Nobel Award in Economics this year recognizes two new and very different sub-fields, behavioral economics and experimental economics.
Daniel Kahneman, a professor of psychology and public policy at Princeton University, was honored (and his research partnership with the late Amos Tversky was cited) for beginning the work of pinning down certain apparently universal quirks in human judgment that routinely affect economic behavior.
Vernon Smith of George Mason University was recognized for showing how replicable experiments with human participants could be used both in research and teaching to uncover why markets work the way they do — counter to the received wisdom in many quarters that no such experiments were possible.
The two sub-disciplines were compared by Ross Miller, author of Paving Wall Street, a popular recent book on experimental economics, to the differing perspectives of cell physiology and evolutionary biology — life viewed from opposite ends of a telescope. Behavioral economics tends to focus on the tastes, values and cognitive abilities of individuals. Experimental economics investigates the operation of markets as a whole.
In both cases, the idea is to formulate small groups of subjects under carefully-controlled conditions, in order to study everything from the outcomes of different kinds of auctions to the tendency of “free-riders” to enjoy public goods without paying for them. The payoff most often comes when experimenters with different preconceptions criticize and seek to duplicate each others’ work.
Both enterprises are a considerable departure from the time-honored economic tradition of reasoning from the general to the particular — the assuming of can-openers whenever necessary, as if it solved the problem, as in the punch-line of the old joke.
But aside from the inductive tendencies they share, behavioral and experimental economics have little else in common, at least at the present time. There are relatively few cross-citations in their respective literatures. And the feeling is widespread that neither field has yet fully coalesced. This year’s Nobel Award is almost certainly a compromise, designed to get both specialties on the map.
The experimental approach is somewhat harder to understand — after all, everybody’s an armchair psychologist. For a full discussion of its many skeins, consult Alvin Roth’s introduction to the Handbook of Experimental Economics.
“Many contemporary experimental economists carry around with them different and very partial accounts of the history of the field,” Roth wrote. But then the great thing about the Nobel Award is that it settles on one of these stories and personalizes it.
Experimental economics had its modern origins in the Harvard classroom of Edward Chamberlin. A great star of the 1930s and ’40s, Chamberlin was a determined critic of doctrines of perfect competition.
Elements of monopolistic behavior, he argued, pervaded the market for nearly every good whose characteristics were differentiated, even slightly, from all others. Even gasoline sold on one street corner was different from gasoline sold on another, he said. Most firms were able to manipulate the kinds of products they sold, as well as their prices and quantities.
At one point, Chamberlin devised a classroom experiment to “prove” the impossibility of perfect competition. He passed out cards to his graduate students, marked with price caps for the buyers, floors (or reservation prices) for the sellers. He instructed the possessors to circulate, pair up and make deals.
The strike prices, reported to Chamberlin, were posted on a blackboard. And the results, usually at odds with the implicit schedules of supply and demand that the professor had drawn up beforehand (based on the values he assigned), were said to show how little competition mattered.
Enter Vernon Smith, an electrical engineer from California Institute of Technology with a masters in economics from the University of Kansas. Smith came to Harvard to get his economics PhD. He sat in on Chamberlin’s experiment, and like the others in the class, disparaged it.
But a couple years later, in the autumn of 1955, Smith found himself awake at 3 a.m., thinking of Chamberlin’s “silly” experiment. He was an assistant professor at Purdue University, teaching four sections of elementary economics to undergraduates What if he changed the rules to make the classroom experiment a little more like a real market?
Specifically, he wondered, what if each trader’s quotations were called out to the entire trading group, buyers and sellers alike, instead of being the subject of myriad two-way conversations? And what if the trading sessions were strung out over a series of “days,” instead of confined to a single period, in order that buyers and sellers could learn from experience? Maybe then, he could avoid repeating “the textbook supply-and-demand song-and-dance.”
Smith ran the revised experiment — a “double auction,” because it featured both bid and asked prices — for the first time in January of 1956. The outcome was just the opposite of what Edward Chamberlin had expected — and exactly what competitive price theory said the result should be. Prices rapidly converged to an equilibrium, close to the values predicted by theory, even though the students lacked the information necessary the calculate the overall result. Naturally, Smith didn’t believe his results. So he ran another version, and another.
By 1960, he was ready to publish his results. He wrote them up and sent them to the Journal of Political Economy — the same place Chamberlin had published news of his experiment a few years before. The Chicago economists who published the journal would love the results, he reasoned. After all, he had shown that the market worked under much weaker conditions than previously had been thought. A few traders, operating with strict privacy, with little learning and no experience, could make it work.
“A great discovery, right? Not quite, as it turned out,” Smith later wrote. The paper was dismissed, and dismissed again. “At Chicago they already knew that markets work. Who needs evidence?”
Only when editor Harry Johnson overruled three referees (presumably Milton Friedman was among them) was the paper finally published. Smith’s memoir can be found in “Essays in Contemporary Fields of Economics,” edited by George Horwich and James Quirk, and published “In honor of Emanuel T. Weiler.”
Purdue in the late 1950s was one of the most exciting economic centers in the world, and Em Weiler was the reason why. An economist by training and an entrepreneur by nature, Weiler came to Purdue from the University of Illinois in 1953, after the Illini economics department had gone super-nova. Weiler managed to turn an undergraduate service department into a sophisticated theoretical PhD program and merge the resulting economics department into a business school without creating a fracas — though his professors were known to the old-line engineering faculty as “the Strange Ones.”
It was a banner time. The raft of new techniques being developed in mathematical economics was viewed with suspicion by the nation’s more established universities. The potential of computers to transform analysis was understood by few. Through clever recruiting, Weiler built a research center that, for a time, rivaled the other great start-up in the years after World War II — that at the Massachusetts Institute of Technology.
Thus the department that Vernon Smith joined in the autumn of 1955 included Stanley Reiter, Lance Davis and Jonathan Hughes, who were applying quantitative methods to economic history in the project that came to be known as cliometrics. George Horwich, Jim Quirk and Ed Ames were moving aggressively beyond the frontiers established by John Hicks and Paul Samuelson into the methods of dynamic programming and general equilibrium that in the 1970s would take economics by storm.
Cowles Commission nabobs Jacob Marschak and Leonid Hurwicz served as unofficial godfathers to the Purdue department. Sooner or later virtually every math econ scholar in the world came to West Lafayette to talk. A stream of talented graduate students included Morton Kamien, John Ledyard, Nathan Rosenberg, Nancy Schwartz and John Wood. With his wavy blonde hair, his motorcycle, and his enthusiasm for all problems economic, Vernon Smith was an exotic addition to the pack.
Alas, it didn’t last. Smith left Purdue for Brown University in 1967. The next year he moved to the University of Massachusetts at Amherst, which was starting an ambitious program in mathematical economics. Many of the Purdue economists moved to Northwestern University. Quirk went to Caltech. Em Weiler gravitated towards business, became ill, and untimely died. West Lafayette regressed toward the mean of Big Ten universities.
One of the last to arrive in Purdue in its heyday was Charles Plott, a freshly-minted PhD from the University of Virginia who wanted to learn some mathematical economics. He and Smith taught a course together, and Plott acquired a lively interest in laboratory methods. Smith went back East and, in due course, Plott followed Quirk to Caltech.
When the experiment in department-building at UMass unexpectedly blew up, Plott and Quirk brought Smith to Caltech. He had been an undergraduate there twenty-five years before, and for a time social science at the Pasadena campus enjoyed a golden age. By some accounts, Smith had stopped working on experimentation before he arrived. By other accounts, his interest remained avid all along. In any event, his California sojourn in 1973-75 was highly stimulating, and work with Plott and their student Ross Miller quickly produced seminal new results.
Efforts to keep Smith in Southern California failed. He accepted an offer to build a new center for experimental economics at the University of Arizona in Tucson. Plott continue to build the experimental programs at Caltech. Had the Swedes chosen not to combine the honors for experimental and behavioral economics in a single award, presumably Plott would have shared the prize with Smith.
But experimental economics in the ’70s was becoming a very different place. Game theory had opened vast new possibilities for investigation. Economic theory and experimentation had become a two-way street. The National Science Foundation funded a number of laboratories around the country. A new generation of graduate students, unburdened by the doctrinal battles of the ’50s and ’60s, walked on and began using new techniques with great success.
And when Alvin Roth addressed a World Congress of the Econometric Society in 1985, he was able to divide the exploding literature of experimental economics into three broad categories, depending on investigators’ aims. There were “speaking to theorists,” “searching for facts” and “whispering in the ears of princes.” He still had material left over for the category, “searching for meaning.”
If there’s a moral here, perhaps it has to do with the difficulty of sustaining first-rate social science in universities overseen by state legislatures. The flowering at Purdue got its start only when legislators launched a communist witch-hunt at the University of Illinois, 75 miles away across the state-line. Purdue’s bloom faded from more natural causes, but the University of Massachusetts flirtation with math-econ ended in failure. Arizona built its university swiftly and imaginatively, especially under governor Bruce Babbitt. By early ’90s, however, the legislature had turned on it and aggressively cut Tucson’s budget — pruning, but not eliminating, its world-class economics lab.
Last year Vernon Smith decamped for the former junior college turned star-packed multiversity that is Virginia’s George Mason University — yet another state school, albeit a highly successful one.
Leadership in experimental economics has been migrating to the nation’s private universities. Harvard hired Alvin Roth from the University of Pittsburgh in 1995. Stanford is building an extensive new capability. So is New York University. Centers around the world — Barcelona, Bonn, Jerusalem, Osaka, Hong Kong — routinely contribute top talent and compete successfully for hires. “We’ve won the battle for the journals,” says Roth, “but not yet the battle of departments.” Caltech, despite the tiny coterie of barely fifteen social scientists on its faculty remains the discipline’s spiritual home.
Important policy results now routinely emanate from experimental laboratories. The possibility of speculative bubbles has been all but nailed down. Those who insist the market “always gets it right” no longer get a hearing, and much work is directed at policies (most of them having to do with supplying or adducing information) that would diminish the likelihood of various kinds of speculative frenzy. How about auctions for initial public offerings? How about futures markets for individual stocks?
The hot topic of the moment is reputation. Those little buyers’ ratings that are beginning to appear in e-commerce are as old in principle as the Better Business Bureau. But interpreted digitally and in terms of public policy, they herald a better way of dealing of markets in which the existence of “lemons” is a threat. See, for example, Eliciting Honest Feedback in Electronic Markets.
Meanwhile, at 75, Vernon Smith is economics’ newest celebrity. He sports a pony tail and indulges in tastes acquired in Tucson for cowboy boots, noisy bars and the Texas two-step. He quotes Kahil Gibran. Sometimes he writes in blank verse himself. Smith has become a thoroughly romantic figure, the kind that his old friend the economic historian Jonathan Hughes lionized in The Vital Few, a classic paean to entrepreneurs.
But experimental economics itself, once lonely and easily ridiculed, has become a lively and thickly-populated field. And Smith’s primitive experiment, the double oral auction, increasingly performed in classrooms around the world, has become as powerful a demonstration that markets work as is life experience itself.