Implementation by Iterative Dominance and Backward Induction : An Experimental Comparison

We report experimental results on the relative performance of simultaneous and sequential versions of the Abreu-Matsushima mechanism. Under the simultaneous version subjects typically use undominated strategies, but apply only a limited number of iterations of dominance. Consequently the unique strategy surviving iterative elimination of strictly dominated strategies is rarely observed. Under the sequential version, subjects also typically use undominated strategies, but apply only a limited number of steps of backward induction. Thus the backward induction outcome is also rarely observed. The sequential version results in fewer observed outcomes corresponding to the predicted outcome than the simultaneous version