When exploring acquisition targets, firms typically begin with the possibilities offering greatest option value and work their way down, as prescribed by optimal search theory. Yet the market designs economists have often prescribed, involving simultaneous or ascending prices, stymie this process. As a result they may be arbitrarily inefficient when one accounts for the costs bidders must invest to learn their value for acquiring different items. We present a model that incorporates such costs, and a simple descending price procedure that we prove robustly approximates the fully optimal sequential search process quite generally. Our results exploit a novel characterization of Weitzman's "Pandora's Box" problem in terms of option pricing theory that connects seamlessly with recent techniques from algorithmic mechanism design.
read more...↧