Mission control: Principal agent theory as a model for national security policy decisions

This dissertation argues that the president is able to effectively control the national security bureaucracy to achieve his goals. It uses the principal agent model to demonstrate that presidents can serve as principals and assert control over bureaucratic agents. They do this by using structural control—controlling the number of missions, operating rules and communications of bureaucratic agents in order to increase their effectiveness at accomplishing the stated mission and eliminate interference from Congress.; I focus on three general case areas of weapons acquisition during the 1950s and early 1960s: ballistic missile development, aerial reconnaissance, and satellite reconnaissance. In each case the president and his executive team structured—or failed to structure—the bureaucratic agents that undertook efforts to develop weapons Programs. These cases demonstrate that presidents are aware that the structure of an agent can affect its responsiveness to him and its ability to effectively achieve his goals. They also demonstrate that presidents understand that the agent's ties to Congress affect the agent's responsiveness and he therefore seeks to control these ties. Furthermore, he realizes that controlling the information that an agent transmits to Congress—either through the creation of information channels or through more extreme measures, such as security classification—can be vital to limiting congressional interference. Finally, they demonstrate that the more a president can control structure, the more likely he is to achieve his desired goals. Highly structured programs are the most successful. Programs that are relatively unstructured are more likely to fail.; The executive powers of the president are a significant counterpoint to the legislative powers of the Congress in the national security field. I argue that the president acts as the preeminent principal when it comes to establishing such policy and that the divided nature of the principal has a significant effect on both presidential strategies and outcomes. Finally, I argue that there is a range of variation in how much the president can control—in some instances he can be remarkably effective, so effective, in fact, that his opponents never even know what he has done. In other instances, practical and constitutional restraints on his actions can frustrate his ability to achieve his goals.