Using time-series models to explain and predict state gubernatorial election outcomes: An application to Oklahoma

Abstract

In Oklahoma gubernatorial elections the candidate of the incumbent governor's party benefits from a healthy state economy, though national economic conditions have little influence. When the President is popular. Oklahoma voters support the gubernatorial candidate of the party not in the White House, evidencing anti-Washington sentiment. A regression model incorporating these influences. Along with a control for party. Successfully predicted the outcome of the 1998 gubernatorial election.