Predicting presidential election results
Abstract:The 2004 US presidential election proved again how difficult it is to predict vote shares on the basis of polls. Midday media exit polls suggested that Senator Kerry would become the 44th President. Political scientists and econometricians, led by Ray Fair, have promulgated theoretical arguments and empirical results to predict US presidential elections, using macro-economic data and political factors. Respecifying Fair's war variable to include Korea and Vietnam and removing serial correlation improves his election forecasting without public opinion poll variables. This generalized Fair model predicts President Bush's two-party vote share would be 52.3 percent, well below predictions by Fair and prestigious political scientists.
Document Type: Research Article
Affiliations: The McDonough School of Business, Georgetown University, Washington DC 20057, USA
Publication date: 2006-03-20