UC San Diego

This dissertation analyzes the role of the president as an agent of the party. It examines both the strategy of presidents as they raise funds for House and Senate candidates as well as the consequences of this strategy on a candidate's total funds raised and share of the vote. Central to this analysis is assessing under what conditions the president helps the party maximize seats in Congress and when president allocates fund-raising events to pursue other objectives. The dissertation focuses on the president's activities in the 2000 congressional election as well as the 2002 and 2006 midterm elections. Doing so allows variation in national partisan tides, the time the president has left in office and other factors that alter the president's incentives. Data for this dissertation comes from observations from journalists as well as data files from the Federal Election Commission. The following analysis uses maximum likelihood models, extends traditional models of vote share to include the president's activities and attempts treatment-effects models. These findings suggest that presidents routinely behave as faithful agents of the party, but changes in national partisan tides can expand or limit their strategy. Moreover, fund-raising events can have a substantial impact on the total funds congressional candidates can raise, while the effect of these events at the polls confirms traditional findings of how money affects the vote