Procuring Innovation Through Bid Preference: An Experiment

Procuring innovation involves motivating research efforts to design an innovative project and then implementing that project in a cost effective manner. In this paper, we study how to combine these two goals in a single procurement mechanism. To this end, we run an experiment in which subjects make costly investments in innovation, and then the most innovative project is selected and awarded through a first-price procurement auction. In one of our two treatments, the bidder who proposes the most innovative project in the first stage receives a bid preference in the auction. The theory predicts that the bid preference scheme should indeed boost innovation, but at the cost of a higher awarding price in the auction. Our experimental results confirm the former prediction; however, we find that the bid preference does not result in an extra-payment for the buyer. To explain this puzzle, we search for the presence of a sunk cost fallacy in individuals’ behavior. We provide robust evidence of this bias from two sources. First, in a subsequent Becker-DeGroot-Marschak task played by subjects after the main experimental task, we show that incurring a sunk entry cost causally increases subsequent offers. Second, a Quantal Response Equilibrium model of our procurement of innovation game confirms the presence of a significant reverse sunk cost fallacy. However, the introduction of the bid preference appears to weaken the strength of this bias, which ultimately explains our puzzling result regarding the awarding price.