Presidential Elections in Space and Time

Abstract

In this study a state-level model of presidential election outcomes is developed and empirically tested for the years 1960 to 1984, using pooled cross-sectional analysis. The model specifies presidential elections as a function of long-term and short-term, political and economic, state-level and national variables. Several important findings emerge from the analysis. First, state-level presidential election outcomes are not responsive only to state or national factors but are reflections of both. The state of the national economy, presidential incumbency and popularity, state political ideology, and partisanship are found to be the most important determinants of presidential election outcomes. Second, although not as powerful, there is evidence of a fairly strong presidential home state advantage, as well as a somewhat weaker home region advantage. No such advantage is found for vicepresidential candidates. Finally, national factors are found to improve the accuracy of the model in some states (mostly Plains and Mountain West states) much more than in other states.