August 30 Federal implications of WA election result

The received wisdom  is that a WA Labor loss next Saturday would be bad for Kevin Rudd. But surely anyone who has been been watching Australian politics over the last decade knows it would be the opposite?

Did any of those 20-something thumping Coalition state/territory drubbings auger badly for the Howard government? Did the federal Libs say "oh no, we lost another state election, we're done for federally"?

No - if anything the opposite. A Liberal win next Saturday would, for one thing, make Labor gains in WA in 2010 more likely. But with the number of seats involved small, that's relatively unimportant (and will be possibly more than negated by a mini-bloodbath in NSW.)

Wall-to-wall Labor governments was a negative for federal Labor last year. And if those state Labor governments start falling over now ... it can only be good. 

Under the Howard government, it was always a little sad to see hopeful Labor types trying to draw comfort from the latest massive Labor state win: overlaying state results on federal boundaries; talking about something called the "Labor brand".

If there is a relationship between a party's state and federal fortunes, it seems to be inverse. 

And giving stick to unfriendly state governments is fun - just ask Mr Howard.

If I were federal Labor I would be secretly hoping Colin Barnett becomes WA premier next week. 

Let the boot pass to the other foot.

Beware the big states: gratuitous prime ministerial advice

While we're on the topic: consider Victoria, which will go the polls in November 2010 - about when Rudd is due. If that 11 year old government is wearing out its welcome, it will rub off on federal Labor. Remember the Cain Labor experience in 1990 - eight Victorian seats going to the Coalition, half of them returning to Labor in March 1993, six months after the election of Jeff Kennett.

This on top of the likely albatross of NSW. These two states have between them almost three-fifths of the nation's seats. If I were Kevin Rudd I would consider either: 

(a) a double dissolution in 2009, or 

(b) hanging on until April 2011.

(Either exercise would meet with voter cynicism and so be tricky.)

August 28 New South Wales: the best odds in town

Below are current Centrebet odds for the next NSW election result. For something that is about as close to a certainty as you can get in politics, $1.90 is a fantastic payout for a Coalition win.

It is, for example, more certain than a federal Labor win in 2010 (or whenever), which only pays $1.19. 

A Coalition win at the next federal election is highly unlikely but could conceivably happen if everything right fell into place for them - beginning with the economy hurtling down the toilet. But a NSW Labor win at their next election is all but impossible.

The downside of such a bet is, of course, the two and a half year investment period.

[Update: 28 August 5pm: Somebody's had the credit card out. Coalition win now only pays $1.80 (v $2.00 for Labor).]

August 26 Newspoll says 56 to 44

Dennis here, tables here.

Morgan says 57 to 43.

August 25 Me in the Canberra times

On Mr Rudd, narratives 'n stuff. Here Here

Speaking of the misuse of retrospectively determined narratives ...

August 25 Barack Obama = Jack Kennedy? 

I think everyone would agree that if President John Kennedy hadn't been shot in 1963 the stories told about him would be different, not so centred on glamour, Camelot, charisma etc.

I wasn't alive in 1960 when he was elected, so don't know how true any of those things were at the time. But it was one of the closest elections in American history (sitting between 2000 and 2004) and many (most?) commentators reckon Kennedy (along with Mayor Daley of Chicago) stole the thing anyway. [Update: reader David Walsh corrects me on the detail.]

It's one thing to tell these fantastic stories, about glamour leading to victory, that gee the troops up, but to actually believe them?

Do we live in a perfect world, in which race doesn't matter? Of course not, but an odd collaboration of people is pretending we do: wishful thinkers on the left, ignoring that which under normal circumstances they might shout from the rooftops; and people on the right who would never concede white racism exists.

Let's get real. I've said it before: Barack is going down. Sad but true.

Call his bluff, Malcolm

Brendan has made it through winter. 

Messrs Hartcher and Carney both interesting on Dr Nelson on weekend.

Does anyone really think Peter Costello will take the leadership? His decision not to do so last December indicates he has a reasonably sophisticated view of these things - that the position itself it poison and whoever occupies it will be chewed up and spat out. (Pardon metaphor-mixing.)

Therefore, he knows it would still be silly to take it now.

This brings us to the weird idea that Costello is "watching Nelson's back": if Turnbull challenges, Peter will counter-challenge. 

So Peter ends up as leader even though it's the last thing he wants and was intending to be out of parliament by Christmas. Whoops!

If I were Malcolm Turnbull I would force his hand. Let Costello (if his threat is serious) take the hits for 18 months, then chop him down - just in time for the next election.

If Costello is not serious, Malcolm becomes opposition leader. (It would kill him of course.)

August 19 More on Saturday's WA Newspoll

Further to August 16 post, Antony Green has all the data from the 2005 election and estimates Saturday's Newspoll at around 52 to 48. The 2005 Green preference flow was about 75 to 25.

Antony writes that "a 1% difference is within the margin of error, so it's not worth getting too  hung up about preference estimates", which is of course true as far as it goes, but in our imperfect world journalists over-interpret opinion poll "movements", especially when the results are hovering around 50-50.

More on Northern Territory election

Malcolm Mackerras finalises his two party preferred numbers:

August 18 Nielsen says 55 to 45

In Age and SMH.

Here is the table. 

(Note (as I just did) that Nielsen is now giving two sets of 2pps: their usual numbers, ie which party respondents state will get their final preference; and estimating from primary support using 2007 election minor party flows. 

The latter, in today's poll, comes to 54 to 46, and in general over the year has been the same or slightly better for the Coalition than the headline numbers.)

August 16 Malcolm Mackerras ranks our PMs

In the Weekend Australian. Here is the accompanying table. The most recent contestants (current occupant aside) are at numbers 5 and 10 to 13.

WA opinion polls

1. Newspoll has Labor on 51, but it should be 52 or 53.

Newspoll in the Oz, in the West says 51 to 49, from primaries of 42, 42, Greens on 10 and others on 6. 

The last election, in 2005, saw 41.9, Libs 35.6, Nats 3.7, Greens on 7.6, Family First on 2, Christian Dems 2.9, and One Nation 1.6. Two party preferred was 52.3 to 47.7. 

In other words, in 2005 a 2.6 percent Labor lead on primaries grew to 4.6 percent after preferences, a meagre, by modern standards, preference flow to the ALP of 55 to 45. That was due to all those right of centre minor parties getting, between them, 6.5 percent.

Newspoll, as you know, estimates two party preferred from total preference flows of non-major parties at the last election. So they're going with 55 to 45. But not only have they registered a higher Green support in the latest survey, it's also a big call to assume all those right-wing minor parties will again do (in aggregate) well.

If you split the Green vote 8 to Labor and 2 to the Coalition (as they went federally last year), and the six "others" 2 to 4, you get 52 to 48. Split the "others" down the middle and you get 53 to 47.

The WA government is comfortably ahead on this Newspoll.

2. Westpoll 

Pollbludger reports a Westpoll from our old friend Pattersons. They say 50 50.

August 15 My bets update

Recently I have thrown money at Centrebet for: 

John McCain. Phenomenal odds - $1 got me $3.50 - for something I see as very likely. (Also hedged on "any other candidate" in case McCain is no longer the candidate on election day. He is, putting it delicately, no spring chicken.) [Update: mind subsequently changed.]

The WA Liberals. I got $4.25. While not actually favouring them to win at this stage, I reckon they have a better chance than that, and if opinion polls begin reflecting this, the odds will too and I can then hedge a little. 

That's the plan, anyway. As I've written before: don't try this at home.

And a while ago I got $1.47 for an almost certain change of government in New Zealand. (You can now only get $1.26)

August 14 Malcolm goes North and says 50.2 to 49.8

Malcolm Mackerras drops us a line:

"Dear Peter

I have just been visiting Mumble and I note that you have the Labor-CLP vote shares in 2005 as 58-42.

The actual 2005 figure for Labor was 59.1 per cent of the two-party preferred vote. [Me: yes, sorry, 58 to 42 was a guess from memory.]

I have now done a tally of the current two-party preferred vote count and it is 50.2 per cent for Labor, on my estimate.

I say that because the 2005 figures for Arnhem and Macdonnell have been plugged into the 23 contested seat actual votes in 2008.

It may be said that the Labor figure should be lower because I am allowing for no swing at all in the two seats where Labor members were returned unopposed.  

My point is that the two members have just completed their first terms. Therefore, they would, in normal circumstances, benefit from "sophomore surge" in 2008. If we allow for that, and also allow for some general swing against Labor, then I think it entirely reasonable simply to add in the actual votes cast in 2005. The two seats are majority Aboriginal. In the comparable seat of Nhulunbuy the swing against Labor was only 1.3 per cent.

Cheers

Malcolm"

[Me: I wonder whether perhaps one should factor in a small one or two percent swing in each of those seats. I think allowing for a one percent swing takes the overall 2pp down to 50.1, at two percent you're down to 50.0. Five percent gives you 49.8 percent. You're right, a deeply trivial quibble.]  

[Update here.]

August 12 Newspoll says 57 to 43.

Tables here, Dennis here, Brendan's net satisfaction graph here.

While previously, headlines were all about Dr Nelson's low preferred PM, now it's switched to satisfaction.

Can't take a trick, poor Brendan, but at least he's enjoying, in a last hurrah, a northern summer somewhere.

August 11 Silly season: NT election stories

The days after an election, particularly with an unexpected result, are very silly ones, as people who are paid to do so come up with explanations.

The Australian's wild colonial boy, Paul Toohey, at least admits he doesn't really know what happened. (But proceeds to explain what happened: an arrogant government - the catchall explanation for every election loss/shock.)

Let's face it, we don't know why Labor won in 2005 with 58 [?] percent of the vote and nor do we know why last weekend's affair was about 50-50. Maybe 50-50 is less deserving of explanation than 58-42.

To just make reasons up, as many people must (because it's their job), is worse than useless, because some of these things will become accepted as gospel truth. Then we all believe all sorts of rubbish.

On the other hand, it is of course desirable to try to get to the bottom of these things.

Anyway, being a pattern-guy, I'm now looking forward to the next WA opinion poll. Perhaps another "arrogant"/out of touch etc government is about to get the heave-ho.

In other NT news, a man from the ABC named Antony emails about recent post below.

August 10 NT election result very close

It's  very close, with a possible hung parliament. That $100k punter must be nervous.

This was the first state/territory election since last year's federal change of government, and I've suggested before that a change in one would see the other turning.

It does make you wonder about Western Australia  next month. Had already put a little money (at wonderful odds of $4.25) on the Liberals, and have just now thrown a little more. Which is not to say I expect them to win at this stage, just that these are good odds. (Don't try this at home: my betting record overall is not great.)

• For results, see Antony Green, and NT EC.

August 9 Northern Territory election today

Haven't paid it much attention, but no-one seems to expect anything other than a comfortable Labor win. A reasonable expectation.

However, note that Claire Martin won power with just 48 percent two party preferred in 2001 (against 52 percent for the then government). Do tiny electorates make the vote-seat equation less predictable? Possibly the opposite; have to think about that. [Read Antony Green.]

• Charles Richardson on how to get rich quick in yesterday's Crikey.
  
• The recent latest from Pollbludger.

August 7 My other hat Been researching online the first Australian returning officers - who they were and what they did. More here.

August 6 Costello - me in Crikey today

Here.

August 4 Messiah complex alive and well Yesterday on Insiders, Andrew Bolt could not contain his excitement at the prospect of Peter Costello taking the Liberal leadership.

Liberal fans in general, and others keen to see a contest, are salivating at the thought of their chap showing this Kevin Rudd upstart what for.

Fools! The position largely maketh the man, not the other way around, and if Peter does become opposition leader he probably won't last until the next election. It would ruin him.

Irrespective of who replaces Nelson this year, it is likely to be the one after that who faces the next federal poll. (Unless there's a double dissolution.)

August 1 Bridge-burning exercise

Did you read Christian Kerr in Crikey the other day? Ouch.

July 29 Newspoll 57 to 43

As I think Jim Middleton noted on Insiders recently, the government's primary vote is "stuck" around the mid to late 40s. This was supposed to be a bad thing for the government.

New governments get a good run, with most people supporting Rudd's ETS. Tables here, Dennis here.

Morgan's most recent: 55 to 45.

Mr Costello ... ... loves being talked about.  He deserves the experience of being opposition leader in 2008; it would change his world.

July 21 Nielsen says 54 to 46

Primary votes 43 to 40

Mr Milne 's baton remains in the knapsack.

July 19 Ticket voting in WA lower house

Antony Green on ticket voting, including a WA attempt in the late 1980s to introduce it for the Legislative Assembly.

Two topical definitions at Wikipedia

Green paper and white paper.

Rudd campaign book

You may recall Pamela Williams' 1997 book The Victory, about the 1996 election campaign. It was a ripper read (even if the dialogue was a tad matey).

Back then, Paul Keating's office more or less boycotted the exercise, so they copped the blame for everything. Those who did cooperate - eg Gary Gray and John Della Bosca - came up smelling like roses: if only "Captain Whacky" had heeded their wise words, etc.

Christine Jackman's Inside Kevin07 will be launched next week, and I think is being extracted in the Weekend Oz. There's this online, which must surely put a smile on a certain man's face. But he's lucky they didn't install him.

[Update: Today's Magazine does indeed contain an extract, a large part spruiking some Pommy Labour guy called Alan Milburn, who is a - you guessed it - political guru. (His qualification seems to be winning a seat in 1992 and being on the backbench when Tony Blair won in 1997. He was subsequently a minister.) The man of action flies Down Under in 2006 and immediately realises what's wrong and what it will take to win. This Beazley guy must go, for starters ... 

All reads well, but as is the the nature of these exercises is highly post hoc: that which has happened was caused by the things that happened before it (and which we happen to discover and find interesting enough to take note of). Beazley hopeless, Rudd just the ticket, from December 2006 everything had to fall into place for Labor to win (I'm anticipating now), and fall into place everything did ... That day in the campaign when Rudd called a halt to the bidding contest, the earth shifted. Back at Liberal HQ, they knew that day that they were in trouble. "Mate", somebody said to somebody else, "we're in trouble".

Apologies for the cynicism, but most political analysis is inventing reasons for things we don't really understand, and allocating to individuals responsibility for developments that would have happened anyway.

For example (digressing further), John Howard won four elections, therefore only he as Liberal leader could have won four elections. 

If your job is to tell stories, then stories you must come up with.

This is a criticism of the genre, not the author. As I said, nicely written; am looking forward to the whole thing.]

July 17 All stand for the joint committee 

Me in Austalian Policy Online on the JSCEM.

July 16 Climate change leak

Petrol excise to be cut simultaneously. No great surprise.

July 15 Newspoll says 55 to 45

Here, no table apparently.

July 14 Electoral fallacies: the don'ts of punditry

Here, for no particular reason, are a couple of common fallacies folks fall into when assessing electoral behaviour. I've written on them before.

Thinking only in the moment

During the last term of the Keating government (1993-6), a paper (I can't recall who it was written by) did the rounds that "showed" why the Coalition faced huge hurdles ever getting back to power. Put simply, a majority of voters were now either in receipt of sort of federal government payment, or were employed by the federal government. In this way the society was structurally biased against the party of free enterprise.

This, of course, turned out not to be true, and probably the proportion of such electors has since grown.

More recently, during the Howard years (1996-2007), lots of people happily offered reasons why Labor would have great difficulty winning again: they stood for nothing, tradesmen didn't vote for them, membership too small, Howard too clever, ALP too left-wing, dead-wood, no talent.

Since November 24, of course, the boot has shifted to the other foot.

(Over in the USA, articles and whole books were written over recent years explaining how structural changes in American society meant the Democrats were done for in the medium term. Then along came the 2006 mid-term elections - Whoops.)

Comparing seat results

An empirical version of the above compares particular categories of seat historical voting behaviour - but only back to the previous government. I saw a study recently that showed regional seats had, since 1996, moved significantly to Labor, culminating in many of them finally going Labor in 2007. This, the paper said, showed important long-term shifting alliances. But it ignored the fact that many of the seats were held by the last Labor government at some time over 1983 to 1993.

The "Howard's battlers" myth rested on the same mistake, only looking at voting data back to the Hawke-Keating years. If people had looked back further they would have seen Malcolm Fraser got at least as many "battler" votes as John Howard.

The Parramatta fallacy

Until it went Labor in 2004, Parramatta was a great symbol of Howard's battlers, of real, unpretentious Aussies voting for commonsense conservatism. It worked because the suburb Parramatta was well-known, in Sydney certainly and to an extent elsewhere in the country, as Westiesville: ugh boots, flannos, hoons in Holden Monaros etc. (The perception, not the reality!) If they now voted for Howard, well, that meant something didn't it?

For example, ABC radio current affairs would regularly travel out to interview people in the suburb Parramatta and, sure enough, they'd find a few people to say how much they adored Johnny Howard.

But the problem was that while the seat of Parramatta had (like the country) moved to the Coalition side of the ledger, the suburb Parramatta, although having swung (like the country) to the Coalition, was still solidly Labor.

Here are Labor two party preferred votes at all booths in the suburb for Liberal Ross Cameron's three wins.

Booth votes in the suburb of Parramatta 1996 - 2001

As well, since its creation in 1901 Parramatta (the seat) has been won by conservative parties about 75% of the time.

So ABC Radio was visiting the wrong suburb; if they'd wanted Howard voters they should have gone to Northmead.

July 11 That nice Mr Morgan .... Says 59  to 41, although feelings about the country's direction continue to dive.

Perhaps there are a couple of ways to look at this. One is that pessimism about the future must surely translate into lower voting intentions for Labor some time in the future. 

The other way is to say that despite this pessimism, Labor is still a mile ahead: imagine how well they'd be doing if voters believed the outlook was rosy!

Opposition leader approval ratings

Along similar-ish lines, I have a counter-to-most view of the relationship, if any, between opposition leaders' approval ratings and election outcomes.

Since this piece  a couple of years ago, the Rudd opposition experience - mile high approval and voting intention leads, culminating in a relatively modest performance at the ballot box - has supported my thesis.

July 10 Too many Divisional Returning Officers at the AEC

This week I attended the APSA conference in Brisbane, and gave a paper on the Electoral Commission. It's actually a larger version of a previous Dem audit discussion paper.

APSA paper here.

Gippsland by-election turnout

[Update: reader points out  I hadn't realised the counting hadn't finished. As of 16 July total turnout is 85,717.]

By-elections invariably attract lower turnouts than general elections, for rather obvious reasons: who governs the country is not at stake. In a safe seat, the motivation might be even less. This has been written about by Brian Costar and others.

In Gippsland last month, 84,567 folks turned out to vote out of a total electoral roll of 95,580, which is 88.5 percent. Last November, 90, 851 out of 94, 967 voted, which was 95.7 percent.

However, there were 3,157 absent votes counted at the November election, and absent voting is not available in by-elections. As well, pre-poll votes could be cast in any of the 149 other electorates in the country for three weeks prior to, but not on, last month's by-election day. (It's actually more like 120 physical offices, see APSA paper.) There are also no overseas voting facilities at by-elections, but the number involved would not be great. (Overseas people can and do postal vote, however.)

Of course, some people, knowing they could not absent vote, would have prepolled instead, but if we quite unreasonably assume them to number zero (ie ignore them), adding three and a bit thousand to the turnout still leaves us way short.

Working the other way, the Gippsland electoral roll would have been in better shape last month than last November, largely because federal elections are great cleansing events: non-voting letters are sent out and come back 'return to sender'. 

So you would expect the June Gippsland electoral roll to have less "fat" than it had last November, which should in fact mean, all other things being equal, a higher turnout (as a percentage).

But on yet another hand, people who were put on the roll as a result of successfully casting a provisional vote on election day are counted in the November 24 turnout number, but not in total enrolment. They only numbered 93 in Gippsland, but there is another thing. People who unsuccessfully attempted to vote last year may have (I'm not sure) been chased up by the AEC and put on the roll if appropriate.  These things would add to total enrolment and lower the by-election percentage turnout number last month. 

(Some 907 people unsuccessfully attempted to cast a provisional vote in Gippsland on November 24. If many of them subsequently re-enrolled, that would account for some of the growth in the roll over seven months from last November. But if they all voted that would result in a tiny increase in percentage turnout.)

So many variables. Will return to the topic at some stage.

Of "famous" recent by-elections, Ryan and Aston in 2001 saw turnouts of 88.7 and 92.5 percent. Cunningham in 2002 got 89.5. Their turnouts at the 2001 federal election were 94.5, 96.9 and 95.4 percent respectively.

The end.

July 5 2008 Australia's most boring interviewee? Penny Wong, despite in November 2006 being one of those rarest of creatures, a lemming for Beazley, has done well under Kevin Rudd, eventually getting climate change after Peter Garrett put his foot in it once too often.

She was also the first member of the new government to front Insiders, the day after the election, where she played a straight bat: no euphoria, no glimpses of personality, nothing vaguely interesting ... a lot like her boss's speech the previous night.

In the seven months since then, Wong's modus operandi hasn't changed, as could be seen last night on Lateline, where she was to allegedly discuss the Garnaut Report. To say she was "on message" would be to imply a message existed. Every question was deflected.

Some politicians can make platitudes and non-answers sound interesting. Others, like say Bob McMullan (an inexplicable frontbench no-show), are able to keep their cards close but still appear to let the audience into their confidence.

Not Wong. Wayne Swan is a dangerous extemporiser next to her.

Yes, the government still has a green paper to get through, but a glimpse into its thinking would not go astray.

We know Senator Wong can send an audience to sleep - and maybe that's the idea. But will that be enough when the government has explaining to do?