chapter  10
How best to auction natural resources
ByPETER CRAMTON
Pages 28

The first step is defining the product: the term of the license, the lot size, royalties, and tax obligations. An important part of the product definition is the identification of what terms are biddable and what terms are fixed. Next a number of basic design issues must be resolved: sequential vs. simultaneous sale (with lots sold either one after another or all at once), dynamic vs. static auction (using either an ascending auction process or a single sealed-bid), the information policy (what bidders know when they place their bids), and reserve prices (the minimum selling prices). Collusion and corruption also must be addressed. The structure of bidder preferences is an important input in the design choice. The items for sale – the right to explore and develop natural resources on a particular geographic lot – are sometimes substitutes and sometimes complements. Bidders’ values are interdependent, since each bidder has private information, such as from surveys and seismic tests, that is relevant in determining the largely common value of the lot, based on the net value of the extracted resource. This preference structure suggests, it will be argued below, that some version of a simultaneous ascending auction is best, since this will promote efficient pricing and packaging of the lots. In this chapter I consider a number of alternative auction formats. At one extreme is the first-price sealed-bid auction used in the US for offshore leases. The bidders simultaneously submit bids for each desired lot. Each lot is awarded to the highest bidder at the winning bid price. This simple format is suitable for marginal lots with nearly additive value structures (that is, the value of a package is equal to the sum of the values of the individual lots) and small value interdependencies across bidders. It also may mitigate collusion. At the other extreme is the package clock auction (Ausubel et al. 2006, Cramton 2009). As explained below, this is a version of the simultaneous ascending auction often used in the auction of radio spectrum. The package clock auction is a method of auctioning many related items over multiple bidding rounds, allowing bids on packages of items. The auction begins with a clock stage. The auctioneer names a price for each lot and the bidders respond with the set of lots they desire at the specified prices. Prices increase on lots with more than one bid. This process continues until there are no lots with multiple bids. At this point there is a supplementary round in which bidders express values for any desired packages of lots. An efficient assignment of lots is found based on the supplementary bids and all the bids in the clock stage. Prices are determined from the competition among the submitted bids. The package clock auction encourages effective price discovery in the clock stage and the supplementary round promotes an efficient assignment and competitive revenues. Although this approach may appear complex, it is actually simpler for bidders than common alternatives. The price discovery (the development of prices over many bidding rounds) reduces guesswork and focuses the bidders’ attention on the relevant part of the price space. Then the supplementary round gives the bidders a means to further express package preferences and finetune the assignment of lots. The approach is well suited for high quality pros-

pects, with complex value structures depending on the particular package of lots won as well as the private information of other bidders. Still other designs between these two extremes are appropriate when the bidder preferences are not so complex that package bidding is essential and not so simple as additive values. Just as a fisherman tailors his equipment to the desired catch, an auction designer must tailor the auction format to the structure of bidder preferences and other aspects of the setting. I begin with some motivating insights from auction theory and practice (Section 2). Then, in Section 3, I consider bidder preferences and some of the basic design issues in natural resource auctions. Section 4 addresses problems specific to developing countries. Sections 5 and 6 examine the experience with oil rights auctions and auctions in other sectors. Section 7 presents the package clock auction. Section 8 considers a number of alternative auction formats and makes recommendations based on the particular setting.