ABSTRACT
News anchors often suggest the outcome of US presidential elections based on gallop poll data. Many researchers have offered their predictions based on economic, political, financial and other factors. Compared to the significant research efforts devoted to forecasting presidential elections, models for midterm elections have been largely ignored. Using 36 years of stock market data, coupled with key economic performance data, midterm-election results, presidential and congressional ratings, the authors suggest a modeling approach for predicting midterm elections, and use it forecast 2018 results.
Keywords
Elections, Midterm Elections, Predicting Election Results