In an election year, the term we most hear is “swing”. Labor needs this swing; Debnam achieved that swing. We hear it so often we forget it is an unreality. A mathematical fiction.
“Swing” is just an after-the-event comparison of relative vote tallies. It is neither predictive, nor explanatory.
Most of all it doesn’t measure the size of the mountain to be climbed. The mountain, in our system, is a majority of the two party vote, in a majority of seats. That is how the media should report elections: percentage of the vote for each party, and seats won and lost.
Talk of swing is as meaningful as arguments about whether one’s footy team has a better for and against record this season over last. Interesting arithmetic – for the wonks.
To intuit why swing is a fiction, consider how many voters die, leave the jurisdiction, or enrol for the first time at every election. And the bulk who never budge.
What remains are floating voters. And if we take this lot seriously, they are not “swinging” but, by and large, apolitical. At best they dimly remember who they voted for last time, and it is certainly not a platform from which they swing, pendulously, like a circus act.
This is not to say each election starts from scratch. There is no year zero in electoral politics, not because elections are about “swinging” people away from their last voting habit, but because politics is a set of impressions about past performances and existing personalities.
Our electoral system is stable because economically and existentially, we are a stable country. Elections invariably are referenda on incumbency. Because of that stability and incumbency’s advantages, we don’t have large fluctuations between polls. But to micro-analyse those fluctuations as “swing” is largely an illusion.
It’s most obvious that swing is a fiction when you consider exceptional electoral events. Like the sudden appearance (or disappearance) of significant electoral forces, like popular independents or minor parties. To say there was a “swing of 22% to One Nation in Queensland in 1998” is to admit the term is no more than an arithmetical construct.