"And the Winner is..." Predicting Presidential Elections

Publication Year:
Usage 654
Downloads 426
Abstract Views 228
Repository URL:
Lin, Amanda
Presidential elections; 2016; Prediction; Nate Silver; Econometrics
thesis / dissertation description
This paper establishes a model to forecast the Presidential election outcomes, particularly the 2016 United States Presidential Election by analyzing two distinctive approaches: predicting election wins through voting function, and using approval ratings as proxy for votes. I examine and replicate previous models for vote share of the Democratic Party and of the government for elections from 1948 to 2012. Then, I construct a model for approval ratings based on economic and non-economic variables. My findings have direct implications for forecasting elections and the political business cycle.