Author: Vivek BhattacharyaPublisher:
ISBN:
Category :
Languages : en
Pages : 184
Book Description
This thesis consists of three chapters broadly in industrial organization, with a focus on contests and auctions, and game theory. Chapter 1 develops a new model of multistage R&D procurement contests, in which firms conduct research over a number of stages to develop an innovative product and then supply it to a procurer. I show that the primitives of this model-the cost of research, the distributions of project values and delivery costs, and the share of the profits captured by the firms-are non parametrically identified given data on R&D expenditures and procurement contract amounts. I then develop a tractable estimation procedure and apply it to data from the Small Business Innovation Research program in the Department of Defense. I find that within a particular contests, there is low variation in the values of the proposed projects, which are drawn early in the process, but considerably larger variation in the delivery costs, which are drawn later. The DOD provides high-powered incentives, sharing about 75% of the surplus with the firms. I then suggest simple design changes to improve social surplus but find that many of these socially beneficial design changes would in fact reduce DOD profits. Chapter 2, which is joint with James Roberts and Andrew Sweeting, studies the benefits of regulating entry into procurement auctions, relative to standard auctions in which bidders are allowed to enter and bid freely. Specifically, we study the relationship between auction outcomes and the precision of information bidders have about their costs before entering the bidding stage of the contest. We show that the relative performance of a standard auction with free entry and an "entry rights auction," which restricts participation in the bidding phase, depends non monotonically on the information precision. We finally estimate the model on a dataset of auctions for bridge-building contracts let by the Oklahoma and Texas Departments of Transportation. Entry is estimated to be moderately selective, and the counterfactual implication is that an entry rights auction would significantly increase social efficiency and reduce procurement costs. Chapter 3, which is joint with Lucas Manuelli and Ludwig Straub, proposes a model of "signal distortion" in a game with imperfect public monitoring. We construct a framework in which each player has the chance to distort the true public signal, and each player is uncertain about the distortion technologies available to his opponent. Continuation payoffs are dependent on the distorted signal. Our main result is that when players evaluate strategies according to their worst case guarantees-i.e., are ambiguity-averse over certain distributions in the environment-players behave as if the continuation payoffs that incentivize them in the stage game are perfectly aligned with their opponents'. We then provide two examples showing counterintuitive implications of this result: (i) signal structures that allow players to identify deviators can be harmful in enforcing a strategy profile, and (ii) the presence of signal distortion can help sustain cooperation when it is impossible in standard settings. We then extend our equilibrium concept to a repeated game, show that it is a natural generalization of strongly symmetric equilibria, and then prove an anti-folk theorem that payoffs are in general bounded away from efficiency.
