The Flegg effect is far worse than recent Courier Mail reports led me to believe. Today’s release of Newspoll figures allows for a comprehensive analysis of just when Bruce Flegg started to affect federal voting intentions and just how far the rot has spread.
Flegg became Liberal Party leader in August 2006. In April-June 2006 the two-party preferred federal Coalition vote was 55% in Queensland. Then in Jul-Sep 2006 it dived to 51%. So, let’s say that the initial Flegg effect was 4% (although it would be nice to know what the figure was in July, as I might be unjustly giving Flegg the credit for an effect caused by his predecessor Bob Quinn). There was certainly no other negative movement like this in any of the other states. In fact the Liberal vote in South Australia went up at the same time.
The Flegg effect seems to have cut in again in Jan-Mar 2007 with a further dip of 4% in the Coalition’s two-party preferred vote. Except that this dip is mirrored in every other state except South Australia, where the decline was 11%. Obviously word about Bruce Flegg was permeating around the country and voters were starting to mark the Liberal Party down as a result.
But familiarity sometimes breeds comfortableness. Or what else can you make of the effect that the vote in Flegg’s home state of Queensland has stayed constant since then, but the Liberal vote has deteriorated even further in NSW?
Yeah, well, if you believe any of this you’re obviously running strategy for the Queensland Liberal Party. The washout of the polls is that between last election and this there has been a 12.2% swing against the Liberal Party in NSW, 9% in Victoria, 11.1% in Queensland, 9.6% in SA and only 5.4% in WA. All the east coast states are pretty consistent. If voters in Queensland mention Flegg it is likely to be because they’re looking for a rationalisation of their voting intention.
When we poll voters more of them mention water and climate change as reasons for how they will vote. This probably provides a better explanation, although rainfall doesn’t seem, on a cursory analysis, to exactly match voting trends either.