General election forecasting in Britain: A comparison of three simple models

Abstract

The advent of three-party politics in Britain with the February 1974 general election has introduced an uncertainty into electoral and parliamentary politics unprecedented in the post-war period. In these circumstances, election forecasting has assumed a special interest and significance for academics, politicians, political commentators, and the like. This article presents and assesses the performance of three forecasting instruments, the ‘incremental’, ‘opinion polling’ and ‘economic’ models. They are estimated over the period 1951–1983 and are then used to predict the share of the vote won by the governing, opposition and Alliance parties in the 1987 general election. All are successful in the sense that they forecast the continuation of the Conservative party's electoral dominance. with Labour and the Alliance a poor second and third. Only the economic model, however, generates a reasonable forecast of the gap separating the major parties and it is used to predict the distribution of parliamentary seats between them. It is seen to be substantially more accurate for the government than for the opposition, which is itself a reflection of the uncertainty introduced into British politics by the emergence of a significant third party in recent elections.