Author (Your Name)

Kyle Smith, Colby College

Date of Award


Document Type

Honors Thesis (Open Access)


Colby College. Economics Dept.


David W. Findlay

Second Advisor

Clifford E. Reid


This study has two chapters. The first uses Ray Fair 's national economic voting model of U.S. presidential elections to pose and answer specific questions about how voting theory works in practice. The results suggest that economic activity in the year of an election is the primary determinant of voters' perceptions of presidential performance on the economy, while earlier years in the administration's term are not important. Also, voters hold the incumbent party responsible for economic conditions whether or not that party controls Congress. Finally, the results suggest that economic voting generally operates symmetrically -a fall in growth affects the vote as much as an increase in growth does. In the second chapter, a pooled, cross-sectional forecasting model is developed to explain state vote shares in presidential elections from 1972 through 2004. The model takes into account state and national economic conditions and election year political variables. It also controls for states' partisan predispositions in order to produce a forecast using data available prior to the election. The model correctly predicts 88 percent of state results and has an average error of 2.9 percentage points of the state vote.


voting, elections

Multimedia URL

Included in

Economics Commons