The conflict between transparency and secrecy is a particularly stark instantiation of the principal-agent problem in public law. It is the province of institutional design to come up with effective means to ensure that government actors act in accordance with voters' desires. Much of the time, elections and other disciplining mechanisms (such as impeachment for judges and indirect political control for unelected members of the executive branch) deter official behavior that diverges too widely from voters' interests, at least over the long term. However, because these traditional incentive-alignment methods require political involvement by the public, they cannot prevent self-interested behavior by political actors if the voters have no way of learning about the malfeasance. Transparency seems essential from this perspective, since it allows voters to monitor the actions of their agents. Transparency and accountability are seen as inherently linked, and secrecy is considered by many to produce large agency costs. The main point of disagreement is where to determine the point at which government operations go from being "open" to being "closed" in order to strike a "balance" between secrecy and accountability. This Note explores alternatives to this dichotomous conception of secrecy and accountability. It considers the problem of secrecy and transparency from the perspective of the principal-agent relationship and advocates creative approaches that focus on reducing the total agency costs in the relationship between voters and their representatives. Toward that end, this Note explores three mechanisms that could, if used appropriately, minimize agency costs while still allowing the government the freedom to make important decisions or conduct sensitive operations in secret. This Note's goal is not to critique existing secrecy law systematically, nor to propose a feasible replacement system. Rather, it aims to explore new ways to think about the secrecy/transparency dilemma and to suggest that it is worth investigating mechanisms (including, but not limited to, the three suggested here) that capture as many of the benefits of secrecy as possible while minimizing agency costs. Part I evaluates secrecy and transparency in terms of the potential agency costs that each creates. Parts II, III, and IV each explore a different, potentially agency cost-minimizing mechanism of secrecy. Each Part describes the mechanism and its ideal implementation. Each Part also investigates two examples of the strategy in American law: one that is well designed in light of the preceding discussion of optimal implementation, and one that is more poorly designed. Part II explores proxy monitoring, in which one governmental actor (or other third party) polices another governmental actor's use of secrecy. Part III examines bottom-line disclosure, in which the public is allowed to monitor the performance of the government, but only along a specified "bottom line" metric. Part IV discusses delayed disclosure, in which the government is required to reveal its decisions or actions after a specified period of time
