Abstract

We study multidimensional franchise auctions under uncertainty about variable and fixed costs. The main result is that in a wide class of cost functions the best scoring function for first- and second-score auctions is the auctioneer's objective itself, regardless of her beliefs and even when she is completely ignorant about participants' costs and market demand. Therefore, since social welfare is usually not computable under complete ignorance, in such cases one has to choose a different function which is necessarily suboptimal. Our results, however, allow to exactly identify what one can obtain by employing a generic scoring function. The paper moreover discusses how one can make a reasonable choice of the latter.