Evaluating rival forecasting models of the 2005 general election in Britain – An encompassing experiment

Abstract

This paper examines two rival forecasting models of election outcomes for Britain. The first is a model based on a revised version of the ‘cube rule’ called the seats–votes model which is designed to predict the number of seats won by parties in the House of Commons. The second is based on the idea that incumbent parties are rewarded by the electorate for a good performance in government, particularly in relation to the economy, and punished by a poor performance. The seats–votes model appears to have an edge over its rival, when it comes to forecasting elections. However, tests show that neither model encompasses the other, so that both models can learn from each other when it comes to improving the accuracy of forecasts. Insights from a revised model which incorporates variables from both versions are then used to simulate the outcome of the next general election in Britain in 2009 or 2010. A wide range of possible outcomes suggested by these simulations produce a hung Parliament, where no party has an overall majority in the House of Commons.