Modeling the Number of Bids Received for Outer Continental Shelf Leases by Poisson-Type Models (includes associated papers 12296, 12297. 12338, 12369, 12570 and 12710 )

Abstract

Summary Since 1954, the U.S. federal government has held hydrocarbon lease sales for areas on the Outer Continental Shelf (OCS). The U.S. DOE is charged with developing lease sale policies designed to increase competition on the offered tracts, Increased competition has been assumed synonymous with increased number of bids (NOB). To study the influence of alternative bidding systems on the number of bids received, a mixed Poisson-type model has previously been employed. Here we show why this model is not statistically supportable. A truncated model is proposed and is shown to be statistically justified for the number of solo bids over all sales and marginally supported for the number of joint bids on sales before the joint-bidding ban. Introduction Since 1954 the federal government has held hydrocarbon lease sales for areas in the OCS region of the U.S. In a typical sale, from 75 to 300 tracts were offered and 30 to 90% of these tracts received bids. The number of sealed bids received by tracts varied from 0 to 18 with approximately exponentially decaying frequency. The U.S. DOE is charged with developing bidding systems that will increase competition on the offered tracts. At present, if one sale generates a larger average number of bids per tract than another sale, the first sale is considered to have been more competitive. Dougherty and Lohrenz rightly question the efficacy of judging different bidding alternatives "by comparing the average number of bids per lease" alone. Their paper attempts to address this inadequacy by developing a model for the number of bids received considering classes of differing" bidder interest." (A related model has been developed independently by Engelbrecht- Wiggans.) Unfortunately, the Dougherty-Lohrenz model was not properly tested for goodness of fit. In the next section of this paper we statistically evaluate and reject the mixed Poisson model as proposed. Then, in the ensuing sections, a number of departures from the previous approach are considered. Solo and joint bidding patterns are examined separately. Then, certain ambiguities regarding the interpretation of the number of tracts that were offered but received no bids are discussed. The removal of such tracts from consideration led to the formation of a new, truncated Poisson-type model that provided some significant improvements in goodness of fit. Later we examine the effects on the distribution of NOB of the 1975 ban on joint bidding among major oil companies. The Dougherty-Lohrenz Model Dougherty and Lohrenz proposed a Poisson-type model of the following form to fit the number of bids (NOBS) on offered tracts. Pk=F1 k+F2P(N=k/ 2)+F3P(N=k/ 3), ........................ (1) JPT P. 1263^