Do nine-year-olds converge to competitive equilibrium outcomes in a continuous double auction (CDA)? To find out, we conducted a CDA with 37 nine-year-olds from the New York metro area who were visiting Princeton in the Community of Scholars program. Specifically, we divided the children into two groups, one group of 20 kids (Group 1), and one with 17 kids and one counselor (Group 2) and held simultaneous CDAs.
Using standard induced supply and demand curves with each child either demanding or supplying one unit of the commodity, we used a modification of the CDA design contained in Davis and Holt (1993) designed (by an elementary school teacher) to be understandable by nine year olds. We then ran a practice game to make clear the rules of the game.
Following this, we ran three real games with the last game shifting the supply and demand curves up by equal amounts. At the end of the session, children were paid in cash according to their trading profits (in the event that they made zero or losses, we paid $1.)
We now present a summary of the trading results:
Groups 1 & 2 Pooled
| Game | Average Price | Predicted Price | Quantity Traded | Predicted Quantity |
| 1 | $6.89 | $7-$8 | 11 | 10-12 |
| 2 | $6.79 | $7-$8 | 13 | 10-12 |
| 3 | $11.88 | $12-$13 | 10 | 10-12 |
As is apparent, prices tended to be lower than competitive equilibrium predictions; however quantities were quite close. Examining the play of the groups separately is revealing:
Group 1
| Game | Average Price | Predicted Price | Quantity Traded | Predicted Quantity |
| 1 | $7.00 | $7-$8 | 4 | 5-6 |
| 2 | $7.25 | $7-$8 | 5 | 5-6 |
| 3 | $12.20 | $12-$13 | 5 | 5-6 |
Group 2
| Game | Average Price | Predicted Price | Quantity Traded | Predicted Quantity |
| 1 | $6.82 | $7-$8 | 7 | 5-6 |
| 2 | $6.39 | $7-$8 | 7 | 5-6 |
| 3 | $11.55 | $12-$13 | 5 | 5-6 |
Group 2 behaved significantly differently (according to a Wilcoxon Rank-Sum test) than Group 1. Group 2 tended to trade higher quantities at lower prices than Group 1. Relative to theoretical predictions, Group 1 is quite close. This is perhaps not too surprising in light of Gode and Sunder's (1989) work on convergence of boundedly rational traders in CDAs. Nonetheless, the convergence of Group 1 to equilibrium predictions in so few rounds and despite demand and supply shifts is still interesting and perhaps not entirely expected.
If you have any comments, have run any other experiments with children or are interested in additional details including the protocol for this experiment, send me an email: