Lessons in misplaced trust and dollars lost

A US researcher with a passion for British beer is challenging assumptions that underlie many economic models - and finding that losing a few cents can be an unforgettable aid to classroom learning. Huw Richards reports.

A doubtless apocryphal anecdote tells of a Talmudic scholar inducted into the Austro-Hungarian army in the first world war. Against all expectation, he turned out to be a crack shot and was sent to the front.

In his first battle, the Russian enemy charged and the scholar was ordered to fire. Nothing happened and his enraged sergeant demanded to know why. "Those are men out there," protested the horrified scholar.

The observer of economics watching the construction of ever-more subtle and sophisticated models, each with something of the beauty that attaches to an elegant structure, whether material or not, may be inclined to essentially the same reaction when real people come into the equation.

Such questions are the preoccupation of Jim Engle-Warnick, a postdoctoral junior research fellow at Nuffield College, Oxford. A 37-year-old Ohioan, he came comparatively late to academic economics, gaining his first degree in electrical engineering then working for several years for Goodyear before doing an MBA and spending two and a half years working in Koysice, Slovakia. He returned from Europe, with, not unnaturally, "a strong interest in the problems of transitional economies and the role played in them by issues such as trust".

Starting graduate work at the University of Pittsburgh, he planned to focus on those issues and based his subsequent work on "experimental economics", as practised by his supervisor, Alvin Roth.

He remembers: "I arrived at Pittsburgh in something of a golden age for experimental economics." The excitement generated there was noticed not only at Nuffield, which awarded Engle-Warnick one of its three-year prize fellowships after he completed his doctorate, but at Harvard University, which has appointed Roth to a prestigious chair and equipped a laboratory for his experiments.

Engle-Warnick admits that his slightly unconventional trajectory has probably underpinned his attraction to methods that question the cherished fundamentals of classical economic theory. But it is a matter of questioning and refining, rather than a full-frontal assault on the entire structure.

He points to the continuing importance of classical economic theory as a means of analysis. And he does not suggest downgrading the importance of mathematics or modelling. "The issue is one of finding out where traditional assumptions do or do not work and building better theories based on those findings," he says.

Among the key elements here is the classic assumption that economic actors - whether individuals or institutions - will invariably attempt to maximise their income from any set of circumstances. Some economists have argued for "satisfisation" - reaching a level at which the actor is content with the outcome - as an alternative model.

Engle-Warnick believes in maximisation, but with a distinctive twist:

"People do maximise. The question is what do they maximise? It may be something other than straightforward economic product."

He is living proof of his own contention. Were he simply concerned that his academic career be as lucrative as possible, the logical next step from Nuffield would be to return to North America. But Engle-Warnick, a quietly spoken man with a passion for British beer, would be happy to stay on this side of the Atlantic and has already refused an offer from the United States.

His research is based on examining human reaction to economic circumstances and contexts. This is done by "creating a market in a laboratory". A group of about 20 people is placed in a room. They are divided into groups of buyers and sellers, told the actions available to them and the consequences of those actions, and they then act out their roles on a network of linked computers.

The likeliest participants are students. This might seem to court the risk of mickey-taking skewing results, but Engle-Warnick reports: "Students here seem to be admirably serious-minded when they participate in the experiments." It helps that they play the games involved in experiments with real money, albeit in small amounts, and that they get to keep it at the end.

Much of Engle-Warnick's work has been on trust. In a recent paper, he cites the argument that "trust may lead to better functioning institutions and facilitate economic transactions" and Francis Fukuyama's view that "the level of trust influences the degree of economic success in markets".

In one experiment, he divided participants into pairs - one a "trustee" and the other a "trustor" and gave each 40 cents. The trustor was invited to hand his 40 cents to the trustee. If he did so, the trustee was given another 40 cents, so that he had $1.20 in total. He was then asked to choose between returning half the money to the trustor or keeping it all himself.

A theory based on maximisation predicts that the trustor will hang on to his 40 cents rather than risk losing everything. He found instead that a surprisingly large number of participants were prepared to trust the trustee and handed over their 40 cents - many were then disappointed when the trustee chose to keep all of the $1.20, leaving the trustors with an average of less than 40 cents each.

The key point in this experiment was that participants were told the game would be played only once, so they would neither come into contact with the other person again nor develop a reputation for being either trustworthy or untrustworthy.

Engle-Warnick then factored in the assumption that there was an 80 per cent chance that they would be paired with the same person again and that the game might run through any number of rounds. By varying the length of games, he found that the crucial variable was how long the game immediately before had been. But such effects were also comparatively short-term. This suggests that although trust fluctuates in the short term, it is a fairly robust quality in the longer term. Still, "it is pretty tough to get conclusive results", Engle-Warnick admits. "If you are lucky enough to find an answer, it typically raises another question."

Such research can be frustrating, but Engle-Warnick finds experimental methods highly effective in teaching students. He draws on one, from Experiments with Economic Principles : Microeconomics by Theodore Bergstrom and John Miller, to illustrate the economic impact of prohibition on the market for drugs.

This operates by dividing students into ten drug users - five addicts who will pay a high price for the drug and five who are casual users and will pay less - and ten suppliers. The game is played twice, first with no prohibitions and then with the assumption that the police will confiscate 50 per cent of suppliers' stocks and fine them. This has the effect of raising prices by 250 per cent, driving the casual users out of the market, but leading to higher total expenditure as the addicts pay the increased price.

Engle-Warnick has played the game on numerous occasions, but with remarkably uniform results. He says: "This makes it an excellent teaching tool. The markets in which the students participate look to them like chaos, but the same result invariably comes out of them. You explain how that happens and how you can plot and predict the results on supply-and-demand curves. Because it is based on something in which they have participated, rather than just graphs in a book, students are much more likely to understand those curves, what they mean and the way they can change, and so find it easier to take in the theory underlying them."

The Talmudic scholar would doubtless be intrigued.