Classroom Expernomics: Volume 10 (Fall 2001)


A Search-Theoretic Classroom Experiment with Money[1]

Denise Hazlett
Whitman College

This classroom experiment promotes discussion of the social origins and characteristics of money. Students take the roles of traders who face a double coincidence of wants problem. As they recognize the benefits of overcoming trading frictions, students spontaneously begin using a consumption good as a medium of exchange. The setting comes from Duffy and Och’s (1999) experimental version of the Kiyotaki-Wright (1989) search model of money. In the Kiyotaki-Wright (KW) environment, agents specialize in production, but consume a good other than their own product. Specialization combined with decentralized trading introduces the double coincidence of wants problem. In fact, no one could trade if each person held out for his consumption good. For trade to occur, at least some people must be willing to accept a good which they do not intend to consume, but which they hope to trade later for their consumption good. In other words, some people must be willing to accept a medium of exchange. When there exists an item generally accepted as a medium of exchange, then that item is money. Thus the KW setting captures money in its essential role as a medium of exchange. Here, using a medium of exchange reduces the cost of searching for a trading partner who has what you want and wants what you have.

The instructor does not tell students which item they should use as money, or even that they should use money at all. In fact, nothing in the instructions or the title of the experiment ("A Trading Experiment") even mentions money. Instead, students discover for themselves the usefulness of a medium of exchange. Thus the experiment works well for demonstrating what social conditions give rise to money, namely specialization and decentralization. Furthermore, the experiment demonstrates what characteristics of a commodity make it a good candidate for becoming money. Here, the commodity with the lowest storage cost naturally emerges as a generally accepted medium of exchange.

I have used this experiment twice, both times in my upper-level monetary theory course. The first time, with a class of 22 students, we spent the full period (50 minutes) on the experiment, including going over the instructions, running eight trading periods, and briefly discussing our results. The second time, with 16 students, it took 30 minutes to go over the instructions, run eight rounds and have a brief discussion. Eight rounds seemed adequate in both cases. The experiment fit well into our introductory discussion of what circumstances give rise to a role for money, and which items are likely to emerge as money. See the next section for the instructions and record-keeping sheet

Instructions for the Trading Experiment

1. You are about to participate in an experiment which will last several periods. Participants are divided into three types, called Type 1, Type 2 and Type 3. There are also three types of goods in the experiment, called Good 1, Good 2 and Good 3. Type 1 people consume only Good 1. Whenever a Type 1 person consumes Good 1, he or she automatically produces Good 2. Similarly, Type 2 people consume Good 2 and produce Good 3. Also, Type 3 people consume Good 3 and produce Good 1. Your ID tag indicates which type of person you are. There are roughly equal numbers of Types 1, 2 and 3.

2. Because you do not produce the good that you wish to consume, you will have to trade with someone else to get your good. Each period you will be randomly matched with someone else in the experiment. You and the person you are matched with will each be holding one unit of a good. You may trade the good you are holding for the good that person is holding, provided both of you are willing to trade. All trades are one for one, so you may not trade any fractions of a good. There are three possible outcomes of a meeting:

(i) You trade for the good you consume. Then you automatically consume your good, and automatically produce the good your type produces. You store your production good until the next period.

(ii) You trade so that you receive some good which is not your consumption good. Then you store that good until the next period.

(iii) You do not trade. Then you store the good you are currently holding until the next period.

3. At the beginning of the next period, you will again be randomly matched with another participant, and you will decide whether you want to trade with that person.

4. Your objective is to get as many points as possible over the course of the experiment. Points represent the satisfaction you get from consuming your good minus the costs of storing goods. Every time you consume your good, you earn twenty points. Every time you store a good between periods, you pay a storage cost in points. The cost of storing goods between periods is: one point for storing Good 1, four points for storing Good 2, and nine points for storing Good 3.

5. Each player begins the experiment with 40 points, plus one unit of the good which he or she produces.

6. Let’s consider how you earn points. Suppose that you just received in trade your consumption good. Then, you earn the net payoff given in the table below.

Type of Person Points for consuming Storage cost of good produced Net points earned
1 20 Storing Good 2 costs 4 16
2 20 Storing Good 3 costs 9 11
3 20 Storing Good 1 costs 1 19

7. Recall that every period you must pay a storage cost for the good which you hold, whether you consumed that period or not. Please keep track of your points on your record-keeping sheet.



Record-Keeping Sheet for the Trading Experiment


Your Type: _______________ Your name: _________________________



Good you start with

Type of person matched with

Good that person is holding

Did you trade?

Storage cost at end of period

Did you consume? If yes, mark 20 pts

Total points





Details on Matching Traders

The process of matching traders each period requires a bit of time. Using the following matching technique, I ran the experiment and a brief discussion with 16 students in 30 minutes: I put into a hat 16 slips of paper, two of which had the letter A written on them, two with B, two with C, and so forth through the letter H. Each period I had every student draw a slip. The two who drew A’s were matched with each other. They met to trade in a section of the room labeled A. Similarly, the two people drawing B slips had a designated meeting place. As soon as a trader drew a letter, she went to the designated trading area for that letter, and waited for her partner to show up. Each student wore a badge that identified his or her assigned type. As soon as a pair of trading partners met, they told each other what good they were each holding. Then, they decided whether to trade. The actual decisions about whether or not to trade took very little time, compared to the process of matching traders. I suspect that a class larger than about 30 would need to use a computerized random-matching program in place of the hat technique.

Theoretical Predictions

For the utility function and storage cost given in the instructions, theory predicts a unique pure strategy Nash equilibrium. In this equilibrium, called the fundamental equilibrium, everyone’s best response is always to trade for a good which is less costly to store. Thus, Good 1, as the cheapest to store, becomes a generally accepted medium of exchange. The experiment therefore demonstrates how money arises endogenously in an economy with specialization and decentralization. Furthermore, the experiment shows in a straight-forward manner the desirability of low storage costs for the medium of exchange.

However, a different storage costs and utility functions could yield a different unique pure strategy equilibrium, called the speculative equilibrium. Specifically, given a low enough cost of storing Good 3, Type 1’s always trade for Good 3. They hope to use it in trade with a Type 3 person who has produced and stored Good 1. In this equilibrium, Types 2 and 3 always trade for a lower storage cost good. So, Types 2 and 3 use a fundamental strategy, whereas Type 1’s use a speculative strategy.


Running this experiment using the parameters described in the instructions, I found that within three periods everyone had settled on using fundamental strategies. Everyone was always willing to trade the good they were holding for a lower storage cost good. In the subsequent discussion, students reported that Good 1 spontaneously became a universally accepted medium of exchange because of its low storage cost. After the initial periods, no one was willing to accept a good for which they would have to pay a higher storage cost than that of the good they currently held. (Of course, if the good offered was their consumption good, students accepted it regardless of the storage cost. They would be consuming it immediately and paying a storage cost only on the good they subsequently produced.) Type 2 people therefore found it impossible to unload the costly-to-store Good 3’s which they produced, unless they met a Type 3 person. Type 2’s without the good luck to be matched with a Type 3 ended up with very small or even slightly negative profits after several periods of storing good 3. In comparison, a lucky Type 1 or 2 had 60 or 70 points by the end of the eight trading periods.

I also ran a version of this experiment using parameters that generate a speculative equilibrium. Here, the storage costs on Good 3 were low enough that Type 1’s would accept Good 3 in hopes of getting matched with a Type 3 holding Good 1. The resulting speculative strategies generated a more complicated follow-up discussion. As theory predicts, the Type 1’s were always willing to trade for good 3. After the initial periods, everyone else held out for a lower storage cost good, provided they couldn’t get their own consumption good. So, students did note in the discussion that Good 1 served as a universally accepted medium of exchange. However, they wrestled with the role of Good 3, because Type 1’s were also willing to accept it as a medium of exchange. In some ways this discussion was productive. It allowed us to emphasize that by definition, money is a generally accepted medium of exchange, not merely a medium exchange accepted by a particular subgroup of society. However, I do not recommend using this version. The more complicated nature of the equilibrium seemed to make it harder for students to form an overall impression of what other traders were doing. For instance, in the follow-up discussion, many of them reported thinking that people other than Type 1’s were playing speculative strategies, when they actually were not.

In fact, in computerized research experiments using versions of this set-up, Duffy and Ochs (1999) found that subjects had a pronounced tendency to play fundamental strategies. Subjects would play the fundamental strategies even given the speculative equilibrium parameters and under informational treatments meant to promote speculative play.

Questions for Discussion

The following questions could serve to start a follow-up discussion, or as the core for a laboratory report. Note that to answer Question 2, students must have access to information about all the trades in the experiment.

1. What trades were you willing to make and why? Did you have a particular trading strategy, and if so, what was it? Was your strategy effective at maximizing your total points?

2. Did any item serve as a generally accepted medium of exchange in the experiment? If so, what item was it, why were people willing to accept it, and how was the pattern of trades affected by the existence of a medium of exchange? What were the advantages having a generally accepted medium of exchange in this economy? If not, why was there no generally accepted medium of exchange?

3. What would the effect on trading strategies have been if the storage costs of all the goods had been equal?

4. Can you think of any markets where some item other than currency serves as a generally accepted medium of exchange? If so, what are the advantages and disadvantages of using this item instead of currency?

Request for Beta-Testers

This experiment is one of six designed for macroeconomics courses as part of a National Science Foundation curriculum development grant. Feedback from instructors willing to beta-test it would be most welcome. Please contact Denise Hazlett at for more details.


This experiment demonstrates how specialization and decentralization endogenously give rise to money. Furthermore, the experiment promotes discussion of the characteristics of an item which make it a good candidate for becoming money. Here, the commodity with the lowest storage cost spontaneously emerges as a generally accepted medium of exchange.


Duffy, John and Jack Ochs. "Emergence of Money as a Medium of Exchange: An Experimental Study," American Economic Review, 1999, vol. 89, number 4, pp. 847-77.

Kiyotaki, Nobuhiro and Wright, Randall. "On Money as a Medium of Exchange." Journal of Political Economy, August 1989, vol. 97, number 4, pp. 927-54.

1. Support for this work was provided by the National Science Foundation's Course, Curriculum and Laboratory Improvement Program under grant DUE-9950688.

Back to Expernomics Main Page