When the Swedish Nobel Committee awarded the 2002 Nobel Memorial Prize in Economic Sciences to Vernon L. Smith, an economist at George Mason University, it simply affirmed what economists have long known: that experimental economics has arrived as a respected and powerful discipline within economics. The committee noted that the award was based on Smith’s “having established laboratory experiments as a tool in empirical economic analysis, especially in the study of alternative market mechanisms.” But what exactly are market experiments, and what can researchers learn from them? Of what importance, outside the academy, is the “study of alternative market mechanisms”?

Economic experiments are not simulations or role-playing exercises. They involve real people who make serious choices. Through their efforts, participants stand to make or lose substantial amounts of money.

The simplest form of economic transaction—and the simplest experiment to conduct—is a two-person exchange. This experiment addresses how a single buyer and seller of an item reach, or fail to reach, mutually agreeable terms of trade for that item. In this experimental setting, the researcher induces value on the item for the buyer and the seller. For example, the person assigned the role of seller might be handed a card that indicates that his cost of production for the item is $10. If he can sell the item to the buyer in the experiment for more than his cost of production, then he will be awarded the difference between his sales price and $10. Likewise, the person assigned the role of buyer might be handed a card indicating that his resale value of the item is $22. This means that if he is able to acquire the item for less than $22, he can then sell it back to the experimenter for $22 and keep the difference.

Although no actual physical object is traded, both the seller and the buyer have an incentive to behave exactly as if one were. The seller will desire a price well above $10 for his item; the buyer will wish to pay as little as possible for his. What will happen? Two outcomes are possible. Either the seller and buyer find a mutually agreeable price between $10 and $22, or they fail to reach agreement. Economics says that both sides have an incentive to make a deal, but it says nothing about how the benefits of that deal will be divided. Economics also has little to say about the frequency of occasions in which the seller and buyer part company without making a trade. Many versions of this simple experiment have been conducted to explore these empirical issues.

The simple experimental design outlined above provides a building block for all subsequent experimental market designs. After all, a market at its core is a place where bilateral trades are facilitated between multiple buyers and sellers. Suppose we want to construct a market with five sellers and five buyers. In this case, we would hand a card to each seller indicating the cost of production. For example, one seller would be given a card indicating a cost of $10. The other four sellers would have costs of $12, $14, $16, and $18. People assigned to be buyers would receive a card indicating their resale value. Continuing the example, suppose these values were $22, $20, $18, $16, and $14. Each seller and each buyer in this design would have the opportunity to make one transaction.

Given the range of values for buyers and the range of costs for sellers, what will occur when they are allowed to trade? Will sellers have the upper hand? Will all trades that might benefit both buyers and sellers occur, or will some beneficial trades fail to take place because of incomplete information or so-called market failure? When trades do take place, will they be across a wide range of prices or a narrow band?

Economic theory in its simplest incarnation of supply and demand makes a strong set of predictions. Consider a graph (see Figure 1) that has price on the vertical axis and quantity on the horizontal axis. The supply schedule answers the question: How many units would voluntarily be brought to the market at various prices? Thus, supply in this experimental structure is an ascending stair-step pattern that starts at $10 and rises $2 per step for each unit in the market. Above $18 the supply curve is vertical, for only five units can ever be purchased in this setting. Likewise, the demand schedule answers the question: How many units will be voluntarily purchased in the market at different prices? Using the same analysis as that for the sellers, we find that the demand schedule is a descending stair-step pattern that starts at $22 and falls $2 per step for each unit demanded in the market. Below a price of $14, the demand schedule also is vertical, for no more than the five units are desired in this setting. For this scenario, textbook economics predicts that equilibrium will be reached where supply equals demand. In this case, that means that four units would be traded at the identical price of $16.

Figure 1

Vernon Smith developed this basic structure for creating an experimental market in the mid-1950s. His first creative insight was motivated by the severity of the textbook prediction for all prices to be exactly $16, as noted above. He asked how everyone in the market could reach a single price. Neither the sellers, wanting high prices, nor the buyers, wanting low prices, would necessarily be happy with this outcome. In other words, how was Adam Smith’s invisible hand to do its work? Vernon Smith found empirically that the market took care of both the buyers’ concerns and the sellers’ concerns simultaneously. This led to his first insight: a defined set of trading rules produces an efficient market price.

Smith’s second creative insight was that exploring these questions could not be done in an institutional vacuum. Half of the experimental structure was missing. The sellers and buyers in this structure cannot trade unless specific rules forming the structure of a trading institution are employed. In his early work, Smith opted to use the rules of a double-oral auction. These auction rules are similar to the rules used for trading at the New York Stock Exchange or the Chicago Board of Trade. It is “double” because both sellers and buyers participate (as opposed, for example, to a silent auction at a fund-raising event, where the seller is passive). It is “oral” because the participants call out their bids and offers publicly. They do this using an important rule called the “bid-asked-price-reduction rule.” What it means is that sellers call out asking prices, which are posted publicly, and all subsequent asking prices must descend from this starting asking price. On the other side of the market, once the first buyer makes his bid, all subsequent bids must ascend from this starting bid price. Trade can occur in two ways. Any buyer can accept a seller’s asking price, or a seller can accept any buyer’s bid.

When Smith ran these first experiments, the mechanics of the invisible hand became visible for the first time! Undergraduate student subjects produced single-price market equilibria, even though none of them desired this outcome. When they repeated the exercise, prices were even tighter around the equilibrium. The number of units being transacted was also “efficient,” exhausting the gains from trade without anyone being in charge of the market.

These results came as a big surprise. Textbooks say that for the market to equilibrate, there must be perfect information. But the subjects produced market equilibria having no knowledge about others and with little experience, if any, trading in the double-oral auction. Finally, when Smith manipulated the number of sellers and buyers, he found that astoundingly small numbers of sellers and buyers—for instance, four of each—could produce competitive equilibrium. Prior to this research, the textbooks said that infinite or “numerous” numbers of each were required. Smith’s early research challenged this convention and opened up the possibility that many apparently “thin” markets (having few sellers and buyers) in the real world produce competitive outcomes.

By the late 1970s Smith was examining all types of market institutions: English and Dutch single-object auctions, sealed-bid auctions, posted-offer markets (like a grocery market, where stores place take-it-or-leave-it prices), treasury bill markets, and others. Smith found a computer system at the University of Arizona that was progressive for its time, offering both real-time networking and touch-screen communication, and soon began computerizing all of his experimental markets.

I have conducted experiments of the simple-oral-double auction with groups ranging from eight-year-olds to Communists to professional traders. In every group, the auction has always produced the competitive result. It is the best economics education a student can absorb if a teacher has only one hour.

Experimental economics not only has allowed us to see how command and control regulations in the market affect behavior and produce unintended consequences, but also has helped address how public goods might be provided using market principles.

Another economist whose work on experimental economics has led to substantial insights about markets is Charles Holt of the University of Virginia. Holt, together with Anne Villamil and Loren Langan, conducted an experiment that showed that even when sellers had more than one unit of a good for sale and even when withholding output from the market could drive up the price substantially on the units they did sell, competitive pressures caused them to price at a level close to the competitive level. Work by Holt and others has also shown that when sellers can offer secret discounts from posted prices, collusive agreements tend to break down. These experiments, along with many others, buttress the late George Stigler’s contention that competition is a hardy weed, not a delicate flower.

Experimental methods have been used to understand not only markets, but also politics. One leader in this area has been Charles Plott of the California Institute of Technology. Plott and Michael Levine showed that someone who rigs the agenda can push those who vote on the agenda to the outcome preferred by the rigger. The bottom line is that he who controls the agenda has a large say over the outcome.

Experimental economics is also used to solve some knotty problems in U.S. public policy. Consider two examples.

At certain times of the day in large American cities, more jets want to land and depart than can be handled by the airport. One obvious economic solution to this problem is to auction off the right to land and take off during the congested periods so that the fixed number of scarce “slots” is sold to the highest bidders. Although this logic is correct, Vernon Smith realized that it is incomplete. The problem is not just the fact that you want to land at O’Hare airport in Chicago on Friday in the 4:00 through 5:00 p.m. time slot. Typically you want other conditions to be met as well. You might also want a slot out of O’Hare between 6:00 and 7:00 p.m. Additionally, you may be flying from Chicago to another congested airport like Atlanta, and so you will want a landing slot there as well, and so on.

At its essence, slot allocation is a problem of balancing supply and demand. But the constraints associated with so many crowded, interconnected airports with so many airlines and aircraft competing for space make the simple problem seem impossibly complex. Smith did not think so. He was able to develop a system of combined auctions that solved this problem. These auctions were exhaustively tested in his laboratory and are now used as allocation tools in national airport management.

Experimental economics has also been applied to electricity regulation. Electricity has three properties that make it different from other economic commodities. First, it is the only product for which supply and demand have to be equal at all moments in time. Electricity suppliers promise to meet the use by demanders—when you switch on your lights, you expect them to come on. Second, electricity is hard to store. Third, electricity does not really move directly from its source to its ultimate user. Rather, when electricity is provided to a power grid, the grid is like a great pond whose water level has just increased. These three factors have inhibited the trading of electricity across regions in the United States. Smith, always on the lookout for market solutions that could improve efficiency, cracked the complex technical problems associated with how to trade something so seemingly amorphous as electricity. His work in this area provided the basis for a radical new system of electricity and energy trading that swept the country during the 1990s. Smith advocated an open trading system for electricity in both the wholesale and retail markets. States that have adopted his system fully—mainly western states other than California—have benefited greatly. Other states that still use old regulatory regimes, mainly eastern states, or, like California, that applied only a partial market framework, have struggled.

The tools of experimental economics allow us not only to understand known issues with new precision, but also to discover whole new classes of otherwise unknown phenomena.

About the Author

Don Coursey is the Ameritech Professor of Public Policy Studies at the University of Chicago’s Harris School of Public Policy.

Further Reading

Holt, C. A., L. Langan, and A. P. Villamil. “Market Power in Oral Double Auctions.” Economic Inquiry 24 (1986): 107–123.
Plott, C. R. “Industrial Organization Theory and Experimental Economics.” Journal of Economic Literature 20 (1982): 1485–1527.
Plott, C. R., and M. Levine. “A Model of Agenda Influence on Committee Decisions.” American Economic Review 68 (1978): 146–160.
Rassenti, S. J., V. L. Smith, and R. L. Bulfin. “A Combinatorial Auction Mechanism for Airport Time Slot Allocation.” Bell Journal of Economics 13, no. 2 (1982): 402–417.
Smith, V. L. “Economics in the Laboratory.” Journal of Economic Perspectives 8, no. 1 (1994): 113–131.
Smith, V. L. “Electric Power Deregulation: Background and Prospects.” Contemporary Economic Policy 6, no. 3 (1988) 14–24.
Smith, V. L. “Microeconomic Systems as an Experimental Science.” American Economic Review 71, no. 3 (1982): 467–474.
Smith, V. L. “Theory, Experiment and Economics.” Journal of Economic Perspectives 3, no. 1 (1989): 151–169.