Abstract

In this paper, we run a laboratory experiment to compare two mechanisms in a procurement setting: Right-of-First-Refusal where the incumbent supplier is granted a privileged position versus standard First-Price-Auction. To this end, we have subjects compete against a computerized agent programmed to behave in a risk-neutral way in a FPA, and in a ROFR auction where the “incumbent” bidder is the computer. In contrast with theory, we observe that on average bidders are slightly but significantly more aggressive under the ROFR when their costs are such that they are predicted to behave identically under both auction procedures. For the sub-sample of subjects for whom we can estimate a CRRA parameter, we confirm the theoretical prediction that the buyer’s expected cost is larger in the ROFR if the newcomer is not sufficiently risk-averse.