Competitive Market Institutions: Double Auctions vs. Sealed Bid-Offer Auctions

Forty-eight experiments were conducted to discover whether the double auction (DA) institution's robustness with respect to numbers of agents, design parameters, shifts in parameters, and variations on the trading rules in yielding convergence to competitive equilibrium (CE) outcomes are properties common to all institutions of contract. All the experiments used the PLATO computer system to display experimental instructions, administer rules, compute and display information generated by the market, and compute individual subject payoffs. The results generated several conclusions: 1. Prices are nearer to their theoretical CE values in the DA than in the sealed bid-offer (PQ) mechanism. 2. Overall efficiencies are significantly higher under DA exchange than under the PQ mechanism. 3. The unanimity tatonnement version of the PQ mechanism does not improve performance over the PQ with short-term contracting. 4. The P(Q) mechanism, which allows buyers to submit demand schedules, yields better results in terms of price and efficiency than does PQ. 5. Adding the unanimity tatonnement to this last mechanism yields performance at least as good as DA. 6. The percentage of strategic responses is approximately constant whether the trading group is small or large, experienced or inexperienced.