UC San Diego

Chapter 1 studies the properties of the two most commonly used legal institutions, the inquisitorial system and the adversarial system. In the former, the decision maker takes a decision based on her own acquired information, whereas under the latter the decision maker requires the advocates to present their acquired information prior to making a decision. When information is both verifiable and costly to collect, I show that the decision maker expects to make fewer decision making errors in the adversarial system than in the inquisitorial system. The main factor behind this result is that the advocate with the burden of proof values information more and, consequently, works harder to collect information under the adversarial system than the impartial decision maker under the inquisitorial system. This larger effort exerted by the advocates leads to more informed decision making under the adversarial system in spite of the advocates' incentives to distort information. Chapter 2 studies the problem of an uninformed decision maker who acquires expert advice prior to making a decision. I show that it is less costly to hire partisan agents than impartial agents, especially under advocacy, and that the decision maker prefers partisan advocacy to other forms of institutions. I also extend the literature, originating with Dewatripont and Tirole (1999), to a setting with contracts that condition on information provided and not just the decision made. Chapter 3 studies the robustness of fully revealing equilibria (FRE) in multidimension-multisender cheap talk games. A FRE is outcome-robust (strategy-robust) if there is an equilibrium whose outcome (strategy) is close to the FRE outcome (strategy) when the noise in senders' observations is small. I show that there is no outcome- robust FRE in the model of Levy and Razin (2007), and discuss the connections between these new notions of robustness and the existing stability concepts studied in the literature