Bringing economics into the laboratory

Quantitative Finance, August 2003

Vernon L Smith, professor of economics and law, Interdisciplinary Center for Economic Science, George Mason University; Nobel Laureate (2002), economic sciences

Vernon Smith is justly known as the father of experimental economics. As his citation for the 2002 economics Nobel, which he shared with Daniel Kahneman, says: Smith is responsible "for having established laboratory experiments as a tool in empirical economic analysis, especially in the study of alternative market mechanisms"

Economics has traditionally been regarded as a non-experimental field of inquiry that had to rely on the observation of real-world systems, rather than a science where theory could be tested with controlled laboratory experiments. Thanks to Smith's pioneering work, experimental economics is now established as an essential tool in analysing and designing economic systems. It has been particularly valuable in challenging the idealistic postulates of traditional economics that man exhibits unbounded rationality, pure self-interest, and complete self-control when making economic decisions.

At the age of 76, Smith looks more like a gracefully ageing country and western star than one of the most respected and influential economists of the modern era. His voice retains the easy-going twang of his native Kansas when I call him in Alaska, where he's spending the summer as the first Rasmusen Chair in Economics at the University of Alaska Anchorage. To the European ear, he sounds not unlike the late James Stewart.

Smith's work in experimental economics began in the late 1950s, while Smith was in his first post-doctoral post at Purdue University, Indiana, teaching the principles of economics. He found himself facing what he saw as a great unanswered question: if you have a market represented by patterns of supply and demand, how does it actually work?

"I found economics hard to teach in the sense that we didn't have any way to talk with students about the direct operational relationship between concepts in equilibria and what people actually do in the market with limited information," Smith recalls. "Hayek had kind of brought to the table the idea that the essential thing about markets was not notions about equilibria, but more about information pricing systems and decentralisation. Although there were attempts in economic theory to get some theorems that captured the notion, they were not really very successful."

Smith conceived of a classroom experiment in which the students are divided into buyers and sellers, each of whom was assigned a private maximum buying price or minimum selling price. They were then set around the room to strike the best deals they could with each other, with the results of each trade chalked up on the blackboard.

The experiment was inspired by earlier, neglected work by Edward Chamberlin, Smith's teacher at Harvard, who had conducted a series of classroom experiments in the 1940s. But while Chamberlin regarded his own findings as a falsification of the standard neoclassical market model, Smith found, to his own surprise, that the prices reached in his classroom experiments closely matched the equilibrium prices predicted by theory.[1]

"Information needed to make a decision in the market was not given to any one person," Smith emphasises. "This was Hayek's whole reason why centralisation in any organisation cannot work because information is dispersed in it. He made the neat observation that what markets do is enable all that information to be aggregated into desirable outcomes and there's no one running it. How come things work and no one is in charge of social systems? This was a beginning attempt on my part to understand that."

How things work
Smith ascribes his inquiring drive to his childhood, growing up on a Wichita, Kansas, farm during the Depression of the 1930s. "I always, from the time I was a child, was interested in how things work," he says. "My father was a machinist, a tool and die maker, and he could make anything, he could repair anything. That always kind of fascinated me and that's why I went into physics and engineering."

In 1944, Smith entered Caltech to major in physics, but switched to electrical engineering before taking a first course in economics. "I was fascinated because economics had this structure - I thought this is just physics, I know how to do this. It took me a long time to realise that you didn't have in economics what you have in physics," he says. "In physics, there's this powerful theoretical stuff but also this powerful empirical base. Everything in physics you can directly or indirectly test and demonstrate. In economics, I eventually realised it was top-heavy with theory. It's as though we were doing what Newton did without Kepler. How do we know this stuff works, how do we know it's true to what we can observe?"

Smith applied this approach to a string of economic questions over the following years. One of the most extensively investigated areas was in auction theory, where Smith pioneered the use of controlled laboratory experiments as "wind tunnel" tests of new auction designs. His work now influences the design of auctions for public procurement and privatisation, while the study of more complex auction forms can reveal much about the working of real-world markets.

Research by Smith and his collaborators initiated the systematic study of institutional design for public choice decisions, and of personal exchange; examined stock market bubbles and rational expectations; and investigated the potential application of trading systems to areas as diverse as power generation and airport timeslot allocation.

His work has on occasion proved controversial, particularly his recent work on private markets for trading electricity. An article in the Wall Street Journal last year about the US electric power industry following the California power crisis, which Smith claimed was not a crisis of energy but of bad market design, provoked fury from California Governor Gray Davis - as the WSJ reported it, Gray called the newspaper's editors "f---ing a--h---s" for printing it.

During his research, Smith has also developed many of the tools that underpin the field. Most notably, his induced value theory has become a standard tool in experiment design [2]. This derives from the experimental method used in psychology, by emphasising the importance of providing subjects with sufficient monetary incentives to outweigh the effects of decision costs, and also the importance of designing experiments as repeated trials so that participants can become familiar with the situation.

"The thing I finally evolved in my thinking is whenever I go into a new area to do experiments, if there's a framework of theory that goes with that, either it works or it doesn't work," Smith says. "In either case you have to go in much further. If it works you've got to ask can I come up with a design where it doesn't work - can I find very pathological conditions of supply and demand where this thing doesn't converge? The idea there is to better understand the limits of this thing you've discovered. And if it doesn't work, there are things you can do to make this thing work.

"These projects often become whole research programmes because we're not satisfied with just moving from one thing to another. I would like to see ultimately more people with skills in mathematical economics and economic theory who would take on the challenge of finding ways of modelling some of these results. We need new tools. Our conventional way of thinking about modelling I don't think is adequate. I don't know what the answer is - I've thought a lot about it and tried to articulate the problems, but I don't have a way to capture these things in a formal theoretical framework."

Planes, brains and early man
In 2001, after 26 years at the University of Arizona, Smith moved to George Mason University, Virginia, as a founder member of the Interdisciplinary Center for Economic Sciences. The move was based on the Center's proximity to Washington DC and the desire to make ripples in public policy circles. Research here is focused on the laboratory method of experimental economics, and also draws on fields as diverse as computer science, philosophy and psychology.

Two fields of inquiry at the Center are particularly interesting, Smith says. One is the idea of using the laboratory as a testbed for developing markets and market-like structures to improve decision-making and replace hierarchical managed systems. An example is an area where Smith has been doing research since the early 1980s - the allocation and charging structure of landing and take-off slots at airports.

"We don't have good airport access pricing for the airlines - the landing and take-off fees tend to be fixed independent of the time of day," Smith says. "Conditions will frequently be changing in airports, but that's not reflected in the price of airports and utilisation of resources. One of the problems with that is there's often huge congestion problems. What we're working on is some pilot experiments that will explicate that problem."

Smith and colleagues are testing the idea that airlines should be able to trade timeslots when schedules are disrupted. Rather than being delayed along with every other scheduled flight, the operator of a plane filled with high-paying business passengers with connecting flights to meet would be willing to pay a higher price to take an earlier slot.

"Our idea looks to us as a good place to use prices in a kind of spot market - not in buying and selling the assets, but simply in leasing them," Smith says. "The FAA has some interest in this, and we think of this as a good facility for airport management during ground-delay programmes. Our idea is if people get used to using this sort of market during ground-delay programmes, they might transfer the systems to other times.

"Like all markets, what drives this is the fact that information is dispersed - it's the individual airlines that know their manifests on each flight," he adds. "Pooling that information in a managed system would be a problem - but if there's prices out there and it's in my interest to trade a slot, I will do it and still maintain my private information. That's basically what markets are all about. Unfortunately people don't understand that well but experiments can get them to see the importance of these principles."

The other key area is in linking up economics, cognitive psychology and neural science to study individual decision making. "Brain imaging is going to revolutionise how economists think about individual decision making as we learn more about how people process information and make decisions," Smith says. "Right now the exercise is mostly to do with what circuitry are we using and what is that circuitry for. We're still at the beginning of that, but I think that's a very exciting field."

Smith is working with ICES colleauge Kevin McCabe to use MRI scanning on people playing simple economic games. "There's a difference in the activation of the brain between those who cooperate and those who don't - we're getting very very distinctive differences," he says. "These difference are very much related to the parts of our brains we use for inferring intentions in other people - children when they're around three and a half years of age start to acquire the skill of being able to infer what other people are thinking based on what they say and what actions they're taking. In one of these extensive two-person games, if I see you make one move my brain automatically sees the intentionality in that."

Smith also has a long-standing interest in anthropology and evolution, and has published several papers on the economics of the earliest societies. "I've been long interested in the economics of hunter-gatherer societies and the evolution of humans - our background and the split off from chimpanzee line five million years ago," he says. "The reason I'm very much interested in that is because you see reciprocity in other animals as well as humans, but reciprocity is trade. These are trading systems and in humans that's become a way of supporting specialisation. There's forms of evidence of socialisation that occurred long before there was anything like the formal markets you find today, but they were still exchange systems. Linking up our understanding of laboratory behaviour with the cultural and biological evolution of humans is all part of understanding what it means to be human and for us to have societies."

References:
[1] Smith VL 1962 "An experimental study of competitive market behavior" Journal of Political Economy 70, 111-137
[2] Smith VL 1975 "Experimental econonics: induced value theory" American Economic Review, Papers and Proceedings, 274-279
For a collection of Smith's recent work: Smith VL "Bargaining and Market Behaviour: Essays in Experimental Economics" Cambridge University Press 2000
For more information on the Interdisciplinary Centre for Economic Sciences, see http://www.ices-gmu.net/