Font Size: a A A

Essays on Decision Models

Posted on:2014-10-10Degree:Ph.DType:Dissertation
University:The University of Wisconsin - MadisonCandidate:Onuma, ToshiFull Text:PDF
GTID:1459390005488801Subject:Economics
Abstract/Summary:
First chapter estimates a model of information spillovers in exploratory drilling using data on bids and drilling outcomes on federal oil and gas leases in the Gulf of Mexico. The sample consists of sets of adjacent tracts, an exploratory tract and a development tract. The exploratory tract is a drilled tract where there is no history of drilling in any of its adjacent tracts. The development tract is adjacent to the exploratory tract and active when the exploratory tract is drilled. The importance of information spillovers is identified by estimating the distribution of drilling outcomes on development tracts conditional on the outcomes on exploratory tracts. The value of information spillovers is then obtained by computing the change in the values of the development tract by using exploratory outcome. I find significant value of information spillovers across adjacent tracts, which may lead to an inefficient exploration and production under the decentralization of ownership. In the second chapter I propose a nonparametric test of affiliation. Affiliation is a notion of positive dependence among random variables. If random variables are affiliated, a large realization of one random variable makes other random variables likely to be large. Affiliation is frequently used in economics. In the auction literature, affiliation is a sufficient condition for the symmetric monotone equilibrium.;Previous literature on the nonparametric test of affiliation is unconditional on observable characteristics due to the computational cost. Our test allows testing equations to be fully nonparametric and to be defined as conditional moment inequalities on continuous covariates. Monte Carlo experiments show that the test possesses good power properties. In the third chapter I apply the affiliation test developed in the previous chapter to the data from the oil and gas lease auctions in the Gulf of Mexico. I test affiliation in participations by potential bidders, which are the owners of ongoing leases in the neighborhood of auction sites. Affiliation in participations implies that if one potential bidder participates an auction, it makes other potential bidders more likely to participate. I find that behaviors of potential bidders are affiliated if none of them are participating the auction while affiliation is strongly rejected if some potential bidder is participating the auction. Results suggest a lack of competition among firms which could be a reason why recent auctions do not generate sufficient revenues.
Keywords/Search Tags:Information spillovers, Exploratory, Auction, Affiliation, Drilling, Chapter
Related items