An experimental test of the generalized Anglo-Dutch auction

I report an experiment on the generalized Anglo-Dutch auction and the two-stage sealed-bid auction. The generalized Anglo-Dutch auction generates more revenue than the original Anglo-Dutch auction, while the two auction mechanisms are similar in the efficiency and the entry problem. The two-stage sealed-bid auction is experimentally equivalent to the generalized Anglo-Dutch auction with one drop-out, as theoretically predicted. My result suggests that in the design of the generalized Anglo-Dutch auction, the information gained from other bidders bids does not play a great role in raising revenue while the number of bidders in the second stage is crucial