September 10 2007 Mixing the polls in Crikey Latest two party preferreds: 57.5 to 42.5. See here.

ACNielsen says 57 to 43

And here's a PDF of Nielsen over the year. Product warning: do not over-interpret state by state numbers, as the sample sizes are small.

Possum on the Textor-Crosby data

Check out Possum's excellent explanation of the leaked Crosby-Textor stuff Crikey posted last Friday. A useful glimpse into how the party-people do their stuff. 

If I understand correctly, piccies like this indicate that Textor & co reckon that the second most important voter determinant of all is/was "preferred PM". You know my position on that (don't agree), but it's at least debatable. But what about "win expectations" as one of the strongest determinants? That's just silly. 

In general, stuff like this seems to rest on heroic assumptions about cause and effect. But perhaps I don't fully grasp it; will have a closer look in due course.

September 9 Time for a Hawkie?

For those who advocate a last minute Liberal leadership switch, a la Labor in 1983, when Bill Hayden was dumped for Bob Hawke, there's a teensie weensie difference: Labor wasn't a mile behind in the opinion polls at the time. In fact, they weren't behind at all, they were ahead.

Having said that, Andrew Bolt has been making a remarkable lot of sense on Insiders recently, and today his reasoning for giving Peter Costello a go - that people would at least listen to what the PM was saying for a while - was very sound. But I would draw a wider canvas: Tanya on the glossy covers, Peter's jokes becoming funny, both of them on the Rove/Kerri-Anne (if she currently has a show) interview circuit, he loves his footy, and so on.

Lenore Taylor suggested that the problem is deeper: people don't like the Coalition's policies. But (Workchoices a little bit aside) that's a symptom, not a cause. Governments don't start having unpopular policies: they become unpopular. And this. 

Having (again) said all that, Costello would very probably go down as well, and then his chances of ever becoming PM will be close to nil.

(But is he the sort of guy to hang around for 6 plus years in opposition anyway? Been there done that.)

Hanson's chances As you know, Mrs Hanson's hat is in the Queensland Senate ring. Read Graeme Orr recently in the Courier Mail. Earlier discussion had been on her anticipated tilt in NSW, and the extra AEC largesse that would have entailed.

The consensus is that she hasn't a hope in hell. But is that so? She got 4.5 percent - below the line - in 2004  [update: not true, a reader points out. She was ATL but with no party name - see Group K on senate paper], and is above the line this year. True, she'll get close to zero in preferences, which means she'll need almost a quota - about 14 percent.

Is 14 percent an unrealistic expectation in Queensland? One Nation got over 20% at the 1998 Qld election, and Hanson got 36% in the federal seat of Blair in the same year. Ok, she hasn't the pulling power she once had, but ...

She's unlikely to get up, but stranger things have happened, particularly in Australia's strangest state.

September 7 The stories we tell

Where has the conviction politician! gone? The man who understands the Labor party better than anyone in the ALP, who connects like no other with middle Australia? Surely Workchoices and a new opposition leader aren't enough to trnmp all that?

Or could it be Howard was never as fine a politician as they made out? Those who were telling the above story have moved onto a new one without a moment's (self) reflection. They live in the present. 

After the election we'll either be back to the tale contained in the first paragraph, or a new one, largely framed by the victors: who voted for Rudd and why, and how come Howard got an almighty flick. 

Just as we can thank Liberal federal directors for the phoney "Howard's battlers" and "conviction politician!" narratives, Labor smarties will spruik their version of history. It will no doubt contain some nonsense, but at least Australians will no longer be seen as 20 million John Howards.

[Update: relatedly, got this Matt Price piece, written a year and a half ago, from pollbludger, who notes that Matt's words are "not holding up too well". Apparently, "only a mug would bet against the Coalition winning a fifth straight election", in no small way due to the knockabout talent the Libs get into parliament:

Whereas the Liberals routinely attract a healthy mix of hardened political professionals and relatively artless amateurs ..., Labor is left to drown in its nasty, narrow, cloistered, limiting, repulsive, infested, depressing and ultimately suffocating union gene pool.

I doubt this is the tune Matt and others will sing after the election. They might instead be contrasting Labor's ability to recruit real political operators - no-nonsense folk with nous who take their politics seriously - with the Libs' sad excuse of a partyroom - all those spivs, clowns and used car salesmen. And what's the fascination with people in uniform? That can't be healthy. Etc.

(Matt has a point about the ALP's suspicion of outsiders, but.)]

Old posts

Browsing a page of old posts, found gang of eight interesting and good to be with you prescient. See also the trouble with kim.

September 5 Compare the leaders There are of course lots of differences between these two guys' approaches to the Labor leadership. But a perhaps all-encompassing one is that Latham's strategy was all about Latham. He was out to show he was larrikin, a fair dinkum Aussie, a bit of a tearaway who embodied middle Australia etc.

True, Rudd ran those ads about his up-bringing and values, but his strategy seems to be to attack the incumbent, present himself as something different, but above all be a safe, competent alternative that is easy to take up if people have the inclination. Then let eleven and a half years do the rest.

(Note Mr Kelly's analysis today of the reasons for Howard's troubles barely mentions the Labor leader.) 

Latham, for all that  "outsider" schtick, swallowed whole the media narrative that the incumbent enjoys a special relationship with Australians, he is there because of his extraordinary qualities, and to prevail the opposition leader must become an even better representation of ridgy-didge Aussie values than the everyman PM.

Oh, and show he's a "conviction politician!"

All it did was make Howard look an even steadier hand.

With hindsight

The other day I pointed to a dud-ish call from last year on my part. Here's a better one, six months before the event.

George's data with William's seats

From now 'til the election, I'll construct tables from time to time, some of them using the census etc data that George Megalogenis and the Oz have kindly put online. 

In 2004 I made my own seat pages, attached to pendulums and maps, but don't have time this year so seats in these tables will link to the pollbludger's seats.

Median house price rise

Today's table shows federal electorates by median house price increase since 2004. George dealt with it a few weeks ago. I've also separated them out by state.

You could argue that real estate price rises are a not bad indicator of economic growth and economic "happiness". People whose house-value has markedly increased are likely to feel pleased with themselves (for being such astute investors), the economy, the country, the government and life in general. Those with sinking home values might be rather cranky. 

The tables are large and can be found here.

September 4 The poll mix: 56.5 to 43.5 Me in Crikey today.

Newspoll says 59 to 41. Tables here.

September 3 Brough's battlers?

As you know, commentators can't let go of the "Howard's battlers" narrative, even though there ain't no evidence to suggest more blue-collar workers have voted for Howard over the last 11 and a half years than did for, say, Malcolm Fraser. (Actually, more probably voted for Malcolm.) 

It's a smashing tale, an inescapable ingredient of the news formula, and what can you do? People will just keep employing it.

Even sensible journos buy into it from time to time, and I must gently scold  George Megalogenis for doing so (a little bit) in this otherwise fine and interesting piece on Mal Brough's seat of Longman. 

I've used the census data George has previously posted to make the below table of all Qld seats in order of median household weekly income. As you can see, Longman, (no 23) is indeed down near the bottom. 

The first thing to note is that Longman was created just before the 1996 election, and went into that election notionally Liberal. So if it's a battlers' seat, they were already "Hewson's battlers".

Longman is AEC-classified rural, and rural seats tend to register low incomes, and vote conservative come what may. Certainly the safest National seats in the country have very low on-paper median incomes. (See here.)

But perhaps it's better to describe Longman as a hybrid of outer-suburban and rural; it's not uber-safe. As well, Qld seats tend to be across the board more pro-Coalition. 

From the AEC map, Longman's neigbours are Petrie, Dickon, Blair and Fisher, which are all Liberal-held. Dickson is way up at number 2 in income, while the other three sit down around Longman.

My suspicion is that, in crude terms, outer suburban mortgage-belters will (relatively) stick with Howard at this election, while everyone else (including rural) will move towards Labor. 

Longman, being a hybrid, might throw up a hybrid result.

George's (and the Oz's) electorate census data is great stuff, and I'll perform further manipulations on it from time to time.

August 29 Intestinal fortitude: in the wrong seat?

Saw on telly last night businessman Geoffrey Cousins boasting of his "intestinal fortitude" in running ads against Malcolm Turnbull in Wentworth.

If Mr Cousins really feels that strongly about that Tassie pulp mill, why isn't he causing direct trouble for the man with whom the buck stops? Ie why not a Bennelong campaign? If Howard wins the election the brilliant environmental policy will all be deemed his doing, not Turnbull's.

Perhaps real "intestinal fortitude" would involve actually biting a hand that feeds you.

Memory lane: late 2006

It is the way of our collective memory that it is difficult to recall that, at the time of Rudd's elevation last December, the dreadful Crean-Carr cabal was trying to impose a two-headed leadership monster on the party.

Presumably Rudd quickly and politely told certain folks there was room for only one prima donna in this outfit. And a good thing too.

Did I really say that?

But speaking of that those good old days: check out the first line of this Canberra times piece, and this statement that while I expected Labor to win the election, they were "very unlikely to win the primary vote" 

Both calls seem rather unreal now. I expected Rudd to enjoy a honeymoon and then settle in at around 53 to 47, about where I reckoned Beazley would be now. 

How times change. Probably once the election results are in I'll concede that replacing Bomber with Rudd was a good idea. But let's wait until then.

August 28 Pick the election result and win $1000 

From the Monthly Magazine. You only get one entry, and they close at end September, so if you're wise you'll wait until then.

Still, I've done mine, nominating 91 seats for Labor, 2 Independents - and can't remember how I split the 57 Coalition seats between Libs and Nats.

It might seem excessive, but it's on a par with Hawke's 1983 win. (Specifically, Labor won 75 of 125 seats and then expanded the HoR to 148, with new laws about fair boundaries. I read somewhere once that the notional majority going into the 1984 poll was 34, which would give Labor ... 91 seats. Only just worked it out the exact number now.)

Anyway, I don't want to share my winnings, and not many people will pick a result that big.

Enter from the mag's home page. 

Peter Garrett It was a mistake putting Peter Garrett in environment. The other dodgy appointment, Fitzwhatsis in Defence, was understandable because he'd done so much to bring down Beazley, and anyway he's been largely silent.

As many are commenting, Garrett has been silent too, but in his case it's a shocking waste. Why is the ALP so hopeless with outsiders?

All this was pointed out at the time.

August 27 Mixing the polls

I'm in Crikey today, with latest poll aggregate. My rather convoluted calculations generally put the two party preferred votes closer than the pollsters - and, of course, other aggregates such as Reuters.

With and without Morgan

I'm also doing something else. Morgan almost invariably gives numbers more favourable to Labor than the others, and they are producing a lot of polls at the moment. Morgan may well be “right”, (their primary votes were very "accurate" in the final days of the 2004 campaign) but there are so many of them they crowd out the others.

So I’m providing two graphs, one including Morgan and one without. Click either for larger version.

Excluding Morgan (right), we can see the 2pp hovering around 55 to 45 and 54.5 to 45.5 over the last two months. With Morgan it's around 55 to 45 and 56 to 44.

Go to poll numbers

Galaxy says 57 to 43

In News Ltd tabloids. Here are tables.

August 24 Mr Rudd's hospital take-over

The best thing about it (politically) is that it has a built-in safety-valve: if state governments don't agree to it it goes to a constitutional/other kind of referendum. The big target is therefore deflatable to a small one.

Return of the conviction politician?

Laurie Oakes describes the John Howard as a 'a "whatever it takes" politician', a description no observer of politics today could deny. It's the opposite of the ludicrous "conviction politician!" tag commentators have excitedly bestowed on the PM for years; you can bet no-one will call him a "conviction politician!" between now and the election.

However, if the Government somehow wins, we can, sadly, expect the silly phrase to return. It's one of the few tools many storytellers (I don't include Laurie) have at their disposal, and they'll have to give it a run.

Mickey goes to Oxford Street National polls today generally show a swing of about seven to nine percent to the ALP from the last election. This means pollsters can take their box of tricks to any place in the country and, on average, measure a swing something like that. It generates lots of stories in various seats and states.

Today it's a battle of the party pollsters in Wentworth, with 400-size Labor survey v Liberal data revealing no-one's heard of the Labor guy. (Can 90% of electors anywhere name the challenger in their seat?)

If Labor gets a national swing of 7% or more, all sorts of seats will fall, probably including Bennelong and Wentworth. It would be Labor's biggest win since WWII, which doesn't mean it won't happen, but that it probably won't. 

And as Malcolm Turnbull points out, leaked party polling should always be treated as dodgy.

August 22 "Leadership."

This was Labor's 1996 election slogan; its point being that irrespective of what you thought of Prime Minister Paul Keating, you had to agree he was a strong leader who stood for certain things and wasn't afraid to take unpopular decisions. The other fellow was evasive, claiming to agree with the government on just about everything under the sun, and you really couldn't trust him.

Yesterday on telly John Howard pronounced that "love me or loathe me, people know I stand for things, while Kevin Rudd doesn't" (or something like that). Like "working families have never been better off" it was eerily familiar.

Policy deja vu

And from memory the only two 1996 opposition policies of note dealt with Telstra (partial privatisation) and IR (unfair dismissal). The rest was vague, that the country would continue to be run the same way, with just a tweaking of "the vibe" (less hand-wringing about past treatment of Aboriginals, for example).

After the election the story about the election changed - as they do - with Paul Kelly, for example, writing that Howard "out-muscled" Keating. Now the PM even patiently explains to his state colleagues that oppositions really must present a comprehensive case for a change if they hope to prevail. 

Keating seemed to believe his own press in the end too.

And another thing: industrial relations

A reminder: during the campaign, any IR emphasis will be not on Workchoices, but the opposition's policy and big-bellied unionists. It's Labor's Achilles heel. (I know I'm not telling you anything.)

August 21 Newspoll says 55 to 45

Here's Mr Shanahan; here's the data.

Note the questions were asked from Friday evening to Sunday evening, so only a fraction of respondents would have been aware of the Rudd strip club story.

August 20 New York strip club

Assuming the 'inappropriate behaviour' stuff is rubbish ....

Would a yarn like this have led the news 20 years ago? Or even ten? 

This is the problem with importing Yankee campaigning techniques to this country: they are based on certain assumptions about the audience. Australians have never been as prudish/hypocritical about such things as Americans. And, contrary to beatups you've heard based on a religious party getting under 2% of the vote in 2004, they still aren't.

Future Liberal leaders?

Jason Koutsoukis in the Sunday Age on who should lead the Liberals if they lose the election. (Self-promotion disclosure: he quotes me.)

ABC's 'Insiders' yesterday touched on the topic, with the worst suggestion - Tony Abbott - coming from Annabel Crabb. If the Libs lose, people like Abbott will be seen as part of the problem. For the last 11 years the narrative has been that all Australians are all Liberal-loving and common sense conservative; for the next few the template will be that we are progressive - in an unideological, results-oriented, pragmatic sense.

That is, something like the general view circa 1983-96.

And the accepted yarn will be that the Libs must move towards the centre/be more like Labor/lose their ideological baggage.

Tony Abbott wouldn't last five minutes as opposition leader.

August 17 Reuters aggregate of polls This is Reuters poll trend over the last approximately 12 months. Click for larger version.

They don't include Galaxy, which I think is a mistake,  not least because it has turned out to be very accurate.

I've been meaning to do my own poll aggregate. Will probably begin next week.

Punters don't know shit, pt xxiiv

Belatedly on my part, Simon Jackman in Crikey the other day noted that the number of seats favouring Labor in the betting markets falls just under a majority of the HoR, while the overall betting favours Labor. Simon's also been fiddling with such things on his site.

This is interesting, but expecting the betting market to be always internally logical is unrealistic. The punters are just monkeys who reflect current received wisdom, and it's the received wisdom that contradicts itself. Why wouldn't it? - it's flaky, based on a mixture of facts and myths.

(On the clinical question of why people who believe Labor will win overall aren't making a quid from the seat-betting, or vice versa, and so evening the odds out, my guess is that the bookmaker's take, which is reflected in even-Stephens odds of about $1.85 to $1.85 (rather than $2 to $2 as it would be without a middleman), makes this unprofitable.) [Update: or as a reader succinctly emails: "the vigorish would prevent arbitrage opportunities to bring the two into line."]

Queensland

Me in Crikey today.

August 15 George data update

The URLs of the first two instalments of the Oz's George Megalogenis's Census '06 electoral data (ie all he's released) have been prettied up and moved to here and here. Don't forget to read conditions of use at bottom here.

August 14 Mr Keating on APEC: another plug

Am yet to devise a policy on plugs. In the meantime, Paul Keating is speaking about APEC at the Seymour Centre (Sydney) next week. Flyer here.

A table revisited: Two party preferred cut-off points

I've done this table before, but before the 2004 election. As you know, five elections since 1949 have seen the winner of the two party preferred vote lose the election. We could say that in these cases the electoral terms favoured the winning side, in that they only needed less than 50% to win. 

But the terms always favour one side or another, in that the on-paper cut-off between winning and losing is never exactly 50-50.

This rather numbery post, with table, can be found here. 

August 13 ACNielsen says 55 to 45

Here and here.

State pendulums

See Simon Jackman's state by state pendulums.

August 12 Galaxy in Bennelong

In the Sunday Tele, says 53 to 47 - a roughly seven percent swing on the last election, a point or two under what national polls are currently indicating. 

Taking three points from Jason

From Jason Koutsoukis 's interesting piece in the Sunday Age we can springboard to three handy points. The first two deal with recognising reality over post hocery.

1. Two conditions for a change of government

Young Jason had coffee with "British Labour MP John Spellar, a long-serving minister under Tony Blair who is now enjoying life as a backbencher". Spellar told Koutsoukis that "two overriding conditions" led to Blair's 1997 win. "The first was that the electorate were ready for a change, and the second was that Blair was non-threatening."

"What about Blair's sweeping narrative and the grand Labour platform [asked Jason]?"

"That didn't really start to take shape until we were in office. Not before," replied Spellar.

Hear hear.

2. The Hewson rallies

Second is a reference to John Hewson's "dreadful rallies" in the 1993 campaign. They were indeed dreadful, but at the time went down a treat with commentators, who reckoned they were creating/consolidating something called "momentum". The only dissenter I can recall was Rod Cameron.

Only after the election did the silliness of these exercises become obvious, but then, as is the nature of these things, everything Hewson did was instantly declared wrong.

3. How can Howard win?

And third, Jason says it's difficult to see the government winning. He's right, and I reckon it can only happen under one of a couple of scenarios.

(a) Something catastrophic happens to Rudd's leadership and he has to step down. Highly unlikely, of course.

(b) Labor does something really big and silly on the policy front, like promising huge income tax cuts. No doubt there are some in the camp advocating such a thing as we speak - "Mate, our support is soft; Mate, the punters want to see some gumption; Mate, we need to consolidate the momentum  ..."

One would imagine these people are in a small minority, but the scenario can't be ruled out.

click for the tune

Aug 11 2007 Westpoll: Labor in front in WA Regular readers will know that Pattersons, the market researcher who does the Westpoll for the West Australian, earns a Mickey on two fronts: small sample sizes (approx 400); and misrepresenting their accuracy.

With this in mind: the latest Westpoll, published in today's West Australian, has Federal Labor ahead 54 to 46 in WA. Previous Westpolls have had the Coalition ahead. Tables below.

Note that a third Mickey-determinant (presumably on the newspaper's part), calculating two party preferreds to two decimal places from rounded primary numbers, has thankfully been abandoned.

• Latest Morgan poll, somewhat on the small side (589), says 58.5 to 41.5

In un-electoral news, we warn the Czar. Again.