Since the early Eighties, national analyses of voting have suggested that unemployment was the electoral equivalent of Sherlock Holmes's 'dog that did not bark'. In 1983, for instance, the Conservatives secured re-election despite post-war record levels of unemployment.
Furthermore, in the most successful national analyses of the effects of economic conditions on voting intentions, unemployment proved unimportant: interest rates and inflation seemed to be the main factors shaping the public mood.
However, our research into the geography of British voting reveals that unemployment has an important effect on support for the Government at the constituency level. Other things being equal, the Government has fared significantly worse at recent elections in constituencies and regions where unemployment was high or rising than where it was low or falling.
In 1992, for instance, we estimate that for each percentage point of the local electorate registered as out of work, the Conservatives lost 0.37 per cent of their 1987 supporters, and won fewer new voters - both from those who voted Labour in 1987 (0.2 per cent fewer) and from those who did not vote in 1987 (0.56 per cent fewer).
In closely contested marginals, the local unemployment rate can make the difference between defeat and victory.
C. J. PATTIE
(University of Sheffield)
R. J. JOHNSTON
(University of Essex)