If we have a look back in the polling rear view mirror, one of the things that stand out is the consistency of the gap between each of the major party primary vote estimates and their respective two party preferred estimates. To highlight, we’ll run a couple of charts using all of the polls taken by every pollster since January 2008.
What also stands out is that the Labor “vote gap” is much larger than the Coalition’s, reflecting the beneficial preference flow that Labor receives from the non-major party vote (read Greens).
If we compare the primary/two party “vote gap” for each of the majors since January 2008, we get something that looks like this (I’ve used a locally weighted polynomial regression as a trend line to knock out the polling noise)
Since October last year, the 10 point differential between Labor’s primary and two party preferred estimates has been rock solid, as has the Coalitions 6 point gap. In fact, during the course of the whole Rudd Government, the gap has only moved within a 2 point range for each party on trend estimates. This suggests to us that the Liberal Party is experiencing a structural two party preferred deficit to Labor of around 4 points.
This has long been a problem for the Coalition – Labor has hoovered up a larger share of minor party preferences ever since One Nation left the scene. As long as the Coalition ran a primary vote 4 points higher than Labor – as they tended to do under Howard post-1998 – they were pretty much safely ensconced in government.
But the Australian political landscape has significantly shifted since then. As George Megalogenis has pointed out a number of times, the combined primary vote of the broad centre-left in Australia has formed a clear public majority. We can see this by summing the Newspoll ALP and Green primary votes and running a trend line through them (here we’ll use just Newspoll figures).
This brings us back to the Coalition primary vote deficit problem. Under the last few elections of the Howard government, they needed to be on a primary vote around 4 points clear of Labor’s, but if we look at the primary vote differential between Labor and the Coalition as an All Pollster trend line over the term of the Rudd government, the enormity of the Coalition’s problem here becomes stark:
For the Coalition to be returned to government, that Primary Vote Gap needs to be a negative 4 rather than the double digit positive numbers it’s been in for most of the last 16 months.
Even though the Coalition has pulled a couple of points from the ALP primary and added it to their own since April, that is simply small change compared to what is required. A 54/46 or even a 53/47 poll result isnt the manna from heaven that Coalition backbenchers are being lead to believe – sure it’s not another 58/42, but in the broader scheme of the Coalition pursuit of the Treasury benches, it’s barely even a start.
12 Comments
It’s going to be a fairly big problem for the LNP. Assuming the average Primary/TPP gap holds up, the ALP could squeak home with a primary vote of around 40 – and that’s assuming it’s losing primary votes to the right, rather than the left.I remember some discussions in 2001 about whether the ALP could win a general election with a primary with a 3 in front of it.
Given the rock-solidness of the ALP polling to date, ‘the narrowing’ is going to have to be pretty sharp to give the LNP even a fighting chance.
Just out of interest, how many seats would have fallen to the ALP if an election were held last weekend (with a uniform swing, using whichever poll average you feel would be best)? I think that’s the question that shows how far behind the LNP really are – they’re struggling just to get back the people they’ve lost since November 2007.
What was with the vote gap movements between May and October of last year? Effects from the ETS? (The Green Paper was out in mid July)
(Also the emissions targets, which were recieved with some disappointment, came out in September – which seems to correspond with the bump in both the Coalition and ALP vote gaps.) Some of each major party’s supporters (temporarily) changing their first preference to Green?
…or is a lot of this just an artifact of the way pollsters allocate preferences?
TD,
The best estimate for simulating an election recently was using the January-March Newspoll quarterly breakdowns, which would have delivered a Parliament consisting of 104 Labor members – a 21 seat gain.
Caf, there was a fair bit of polling up and down there as the final spasms of the Nelson leadership combined with Truffles taking the helm washed through the system.
It’s hard to tell if it was issue based or leadership based volatility there – probably a good slab of both.
Poss
Way, way, way off topic but have you heard of the Policy Analysis Market and it’s variants? It uses an Intrade system to measure the likelihood of an outcome with national security implications in it’s more popular form.
Agents and analysts from the CIA, FBI, NSA, Pentagon etc etc are given credits which they can bet on a market. The market has lots of possible bets. As an example, imagine it has one “The House of Saud to fall in the next 12 months”. Say it trades at 5%, but then analysts from various agencies see a few patterns and it goes up to 10%. Then the agencies know to look really hard at what’s going on there and get info from other agencies which may all tie together and put the price up to 50%, meaning operatives are sent in to stabilise Saudi Arabia.
Assassinations, terror plots and coups are also predicted. The version where the public could bet as well got snotted on for obvious reasons.
http://en.wikipedia.org/wiki/Policy_Analysis_Market
My gripe with this kind of analysis is that the Liberal/National Coalition is two parties, not one.
What happens if you treat them as independent parties?
Is there an interactive pendulum anywhere on the net? I think there were a couple around prior to the 2007 election but am unaware of anyone doing a pendulum for the present Parliament.
zoomster
Will this do?
http://www.abc.net.au/elections/federal/2007/calculator/
Dio – I’ve only ever heard of it in passing. Afraid I know little about it.
Zoomster – They might be faux independent parties, but their fortunes are still pretty much tied together. With a swing against the Coalition, a couple of coastal Nats seats will be in danger of falling in the same way that urban Lib seats would be in danger (and pretty much in the same rough proportions to the overall number of seats each party holds).
On the interactive pendulum, Antony probably has one somehwere on his ABC site. If not, I can build one – I’m just a bit hesitant to do it at the moment with the redistributions in NSW and QLD being imminent – the boundaries of quite a large number of seats could change making the exercise a bit premature and redundant.
Zoomster – you beat me to it!
Nup, that’s an oldie.
OK, Poss, take you up on that once the redistributions have happened.
What sort of poll would see the primary vs TPP diverge significantly? I understand that the TPP is synthesised from the Nov 2007 preference flows. As long as each pollster keeps a consistent method, I suspect you would have to see a big shift from Greens to FF (or equivalent) to move the gap.