The Past as Prologue: How to Forecast Presidential Elections
23 Pages Posted: 17 Aug 2020
Date Written: August 10, 2020
The authors apply a novel forecasting technique called Partial Sample Regression to predict the outcomes of U.S. presidential elections. This technique first measures the statistical relevance of past elections. It then employs an obscure mathematical equivalence – that the prediction from a linear regression equation equals a relevance-weighted average of the values for the dependent variable – to forecast election outcomes from a subsample of prior relevant elections. This technique has been applied successfully in finance to predict factor returns and the correlation of stock and bond returns. The authors apply Partial Sample Regression to predict the outcomes of the past five presential elections as well as the 2020 election. They also report which past elections were identified as being statistically most relevant for each of the elections they predict.
Keywords: Informativeness, logit, logit transformation, mahalanobis distance, partial sample regression, relevance, statistical similarity
JEL Classification: C00, C01, C02, C10, C13, C50
Suggested Citation: Suggested Citation