The Ups and Downs of Elections, or Look Before You Swing

Abstract

Why not use swing as a measure of electoral change in the United States? ask David Butler and Stephen D. Van Beek in PS (1990). This article shows why one should look before one swings, for the conditions of meaningful use are so limiting that swing is unsuited for the great majority of countries, including many elections in the United States and Britain. Instead of swing, it is better to employ a more precise and more robust measure of the ups and downs of parties.DefinitionSwing is the average of the change in the share of the vote won by two parties contesting an election. If party A gains 5 percent of the vote between election x and election y, and party B loses 5 percent, then there is a swing of 5 percentage points from B to A. By definition, swing is a measure of movement between two, and only two parties. In an attempt to control for the effect of other parties, the two-party measure of swing treats the votes for parties A and B as equal to 100 percent of the total vote, whatever the share of the vote taken by one or more “third” parties.