General Election Forecasts in the United Kingdom: A Political Economy Model

Michael S. Lewis-Beck, University of Iowa
Richard Nadeau, University of Montréal
Eric Bélanger, University of Montréal

Abstract

With the notable exception of Mughan (1987), forecasting attempts in the United Kingdom have been solely concerned with developing popularity functions. This paper formulates a vote function model to forecast general elections in the UK. The model relies on three independent variables: approval of the government’s record, inflation, and a variable controlling for the “cost of ruling.” The model is estimated over the period 1955–1997 using a substantial six-month lead time. It performs reasonably well, and offers a plausible and parsimonious explanation of government vote support. The vote function is used to generate a forecast of the next British general election. It clearly predicts a majority government for Labour, assuming the election would be held in Spring 2001.