Endogenous Information Acquisition: Essays in Applied Game TheoryPublic Deposited
This dissertation endogenizes information acquisition in two-player games across three different settings. The first chapter explores when moral hazard in a principal-agent contract can lead to pareto improvements when it is preceeded by information gathering. The second chapter studies how product differentiation affects the amount of market research done by firms that compete on price. The third chapter examines the role of costly risk in a production process when there is a debt contract between an investor and an entrepreneur. In Chapter 1, I consider an agent who designs an experiment that reveals information about a state to a principal. The principal subsequently decides whether or not to implement a project. If she does, then she offers a limited-liability contract to motivate the agent to exert effort, which together with the state stochastically determines the project's output. If effort is contractible, the contract ``holds up'' the agent so that conditional on implementation his payoff is independent of the principal's beliefs. In equilibrium, he provides only enough information to maximize the probability that the principal implements the project. In contrast, if effort is non-contractible then the principal must promise the agent rent to motivate effort. Since the promised rent varies across beliefs, the agent may provide more precise information. Thus, although the non-contractibility of effort lowers the principal's payoff at a given belief, it can improve welfare by mitigating hold-up and encouraging information provision. ", 'In Chapter 2, I apply the main result in Persico (2000), that decision-makers acquire more information when their payoffs are more risk-sensitive, to a duopoly model of Bertrand competition with uncertain demand following Vives (1984) in order to show how the amount of private market research firms undertake depends on competition, measured as the level of product differentiation. I decompose the relative marginal return of research across competition levels into two effects, a competitive profit effect and a coordination effect, and show how each of these depends on competition. When the cost of market research is sufficiently high, the amount firms invest in market research is decreasing in the level of competition. In contrast, when the cost of market research is sufficiently low, firms perform the most market research at an intermediate level of competition. I partially extend this result to a public market research setting. In Chapter 3, I extend a simple model of debt between a liquidity-constrained entrepreneur and an investor to allow one of the players, according to the governance structure, to choose either risky or safe production at time 1. Risky production causes capital to depreciate, lowering the value of collateral and production at time 2. When the entrepreneur is tempted to choose risky production in order to foreclose more often in the low state and less often in the high state, he must offer more collateral to the investor. In this way, his inability to commit to safe production can lower his expected value from the project in equilibrium. I provide necessary and sufficient conditions such that the entrepreneur strictly prefers for the investor to have governance over time 1 production in order to overcome this commitment problem.