In this work we present a model, named logistic primary model (LPM), which aims to describe the probability of a Democratic (Republican) victory at the U.S. presidential elections. The proposed model is based on a logistic regression with a unique regressor and exploits the primary results of the candidates to the White House. It follows the idea of the existing primary model (PM) proposed by Helmut Norpoth since 2004, which is a ARIMAX model for the two-party popular votes obtained by the Democratic Party. The LPM, applied to the U.S. election data 1912-2012, shows good performances both in terms of goodness-of-fit and forecasting. In addition, the paper presents an extensive review of the electoral forecasting models proposed in the literature.
Franco Mars, S., Calegari, E., Bagnato, L., Forecasting the United States Election Through Primary Vote Results, <<Forecasting the United States Election Through Primary Vote Results>>, 2021; (150): 1-36 [https://hdl.handle.net/10807/249071]
Forecasting the United States Election Through Primary Vote Results
Calegari, Elena;Bagnato, Luca
2021
Abstract
In this work we present a model, named logistic primary model (LPM), which aims to describe the probability of a Democratic (Republican) victory at the U.S. presidential elections. The proposed model is based on a logistic regression with a unique regressor and exploits the primary results of the candidates to the White House. It follows the idea of the existing primary model (PM) proposed by Helmut Norpoth since 2004, which is a ARIMAX model for the two-party popular votes obtained by the Democratic Party. The LPM, applied to the U.S. election data 1912-2012, shows good performances both in terms of goodness-of-fit and forecasting. In addition, the paper presents an extensive review of the electoral forecasting models proposed in the literature.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
