This Small Business Innovation Research (SBIR) Phase I project will develop auction software that allows bidders to specify budget constraints, and will assess the feasibility of this software for conducting large-scale multi-item auctions. Multi-item auction design has been at the frontier of research in economics and computer science over the past fifteen years. Yet no existing mechanism enables effective competition when bidders face serious budget constraints. New, sealed-bid designs that enable bidders to specify a budget encourage bidders to place more and higher bids, better reflecting values. Sellers will therefore receive higher prices and assignments of goods will be more economically efficient.
In multi-item auctions, bidders often cannot risk outcomes where they may be required to pay more than their authorized budgets. Examples are widespread, ranging from online ad placement to auctions of mineral rights and radio spectrum licenses, and rudimentary technologies to account for Internet ad budgets have already been deployed. There is an immediate need for related budget-based technology physical world auctions where the absence of such technology is reducing the number and level of bids. For example, in sales of oil and gas rights by the US DOI, the average number of bids has been only 1.3 per tract, with losing bidders unable to compete for more tracts due to their need to limit budget exposure. Building on economic analyses of multi-item auctions successfully produced under previous efforts, this proposal will take the next step, creating multi-item auction software that allows bidders to work effectively with limited budgets. These auctions will promote more efficient outcomes and higher revenues for sellers. This research will initially assess the feasibility of software for use in oil and gas leasing, which will be a crucial milestone in bringing efficient auctions to other important markets in both the public and private sectors.
Principal Investigator: Paul R. Milgrom Many auctions are conducted for high-stakes goods, such as rights to use the radio spectrum, procurement of electric power by utilities, or sales of rights to natural resources such as oil and gas leases or mining or forestry rights. Sealed-bid auctions have well-known advantages compared to auctions that take place over a number of bidding rounds, but they also encounter an important problem. When there are multiple goods for sale, bidders’ willingness or even ability to pay for each individual item is often closely tied to budget that limit on their overall spending. The need to avoid paying too much in total limits bids and results in lower prices. If an auction system could be created that both allowed bidders to specify budgets and determined allocations in a way that accounted appropriately for budget constraints, bidders would be able to bid more confidently, more goods could be awarded to those who value them the most, and seller revenues would be higher. Our project was to conceive and develop such a system. In particular, when there are multiple copies of each type of item and the seller wants to set a uniform price for each, the challenge is to develop simple user interfaces and associated algorithms and computational strategies to determine which bidders get which goods and at what prices, respecting bids and budgets. Our main research activity has focused on finding robust solutions to this challenge. In the course of this project, we first developed computing methods that should, theoretically, deliver good results in a reasonable amount of time, even in seemingly difficult cases. We then coded software to test our ideas, exercising the software on extreme cases designed to stress the our approach. The software tests have demonstrated the computational feasibility of our approach. More work needs to be done to understand the limitations of this approach, and to build a fully functional commercial system that it easy to use, even when there are a large numbers of buyers and/or sellers and many items being traded.
