Is Rudd responsible for Turnbull’s unpopularity?

Something that might be worth having a bit of a squiz at is how the polling metrics for the Coalition have changed since Turnbull replaced Nelson as leader – if for no other reason that the results are rather interesting.

To start with, we’ll take a look at the swing – the change in both the primary and two party preferred vote since the election – that each leader has/had experienced. For Nelson we’ll use the 3rd quarter 2008 data as his final polling and for Turnbull we’ll use the latest.

You’ll notice that it’s a fairly mixed bag, with Turnbull improving the Coalition’s position in some places and making it significantly worse in others compared to where Nelson was at when he was replaced. If we measure the change that has occurred in the Coalition vote under Turnbull rather than just use the raw figures, the Turnbull effect starts to become a little more obvious.

Just keep in mind that the social demographics don’t have any two party figures attached to them. Turnbull has improved the Coalition position in NSW and Qld, killed it in SA and WA, while having improved the vote substantially in the non-capitals at the expense of the capital cities. On the social demographics, only the older age group – their historically strongest demographic – has shown an absolute move away from the Coalition under Truffles. All other age groups and both men and women have had a move toward the Coalition under Turnbull, at least as far as the primary vote goes.

So we’ve got some fairly enormous variation in the results at play here. Yet, how much of this is explained by Turnbull’s actions and public perceptions of his leadership?

Let’s now drag the net satisfaction ratings into play (the satisfaction rating for Turnbull minus his dissatisfaction rating) to see if the change in the Coalition vote under Turnbull has tracked or has some relationship with the change in his ultimate satisfaction level.

This is where it gets a little bizarre. Across every cohort and sub-sample measured, Turnbull has far more people dissatisfied with his performance than satisfied compared to Nelson. Yet that dissatisfaction change doesn’t appear to relate at all to the vote change, which we can see when we directly compare the change in net satisfaction that occurred under Turnbull with the change in primary voting intentions that changed under his leadership.

Qld is the real standout, being the state that is not only more dissatisfied with Turnbull compared to Nelson, but is also the state where Turnbull has increased the Coalition primary vote the most!

In SA and WA, we at least have the net satisfaction changes and the voting intention changes walking in the same direction to some meaningful degree, but with very differently sized satisfaction-to-vote-change effects, the relationship a little questionable – especially considering the relationship (or lack thereof) in the other cohorts. Whatever is happening with the voting intentions, it seems to have very little to do with Turnbulls behaviour in most places.

However,  there is something that explains a good chunk of the movement in the LNP vote. If we run a scatter plot where we compare the change in the Coalitions primary vote under Turnbull against the change in Rudd’s net satisfaction rating over the same period, we get:

Even though Rudd’s net satisfaction rating us up across the board over the period, this suggests that the larger the improvement in Rudd’s net satisfaction rating, the more likely the vote was to move away from the Coalition, while the smaller the improvement in Rudd’s net satisfaction rating over the period, the more likely it was that voters moved toward the Coalition.

It’s food for thought – the primary vote levels of the Coalition under Turnbull may have very little to do with the actual actions of Turnbull or the Coalition itself, but may in fact be simply a product of whatever Rudd happens to be doing at the time. At best, the Coalition may be having an impact around the margins, but one which gets washed out by the government’s behaviour.

It would certainly feed into the whole “oppositions don’t win election, governments lose them” theory.

Perhaps the only real “Turnbull Effect” we can see is one of impotence when it comes to actually changing the vote. He wouldn’t be the first Opposition leader to experience such a thing – and as is generally the case, it’s usually the Party behind him that is as much responsible as the leader himself.