Advantage Labor

Numerous pollsters, some previously unknown, have swung quickly into action to record a very rosy view of Labor’s prospects under Julia Gillard. Nielsen surveyed 993 respondents on Thursday night and found Labor’s primary vote roaring back to 47 per cent, decimating the Greens – down seven points to 8 per cent – and delivering them a thumping 55-45 two-party lead. The Coalition primary vote has nonetheless held up: at 42 per cent, it is only down one point on the famous 53-47 poll of June 6. Julia Gillard leads Tony Abbott as preferred prime minister 55 per cent to 34 per cent, widening the gap achieved by Rudd in his last poll from ten points to 21. Against Kevin Rudd, she scores a not overwhelming lead of 44 per cent to 36 per cent: Rudd himself records slightly improved personal ratings, approval up two to 43 per cent and disapproval down five to 47 per cent. Tony Abbott is for some reason down on both approval (one point to 40 per cent) and disapproval (five points to 46 per cent). UPDATE: Full results courtesy of Possum here. Some have pointed that there are some very curious results in the statewide breakdowns, but this provides no statistical reason to doubt the overall result within the margin-of-error. Self-identified Greens preferences have gone from 68-32 to Labor to 81-19, although this is off a tiny sample of Greens voters.

Galaxy produces a more modest headline figure of 52-48 in a survey of 800 respondents, also conducted yesterday. This was achieved off a 41 per cent primary vote, making it a lot more solid than the 52-48 Rudd achieved his final Newspoll, which was based on 35 per cent plus a hypothetical preference share. No further primary vote figures at this stage, but it’s safe to say that here too Labor has recovered a lot of soft Greens votes. The margin of error on the poll is about 3.5 per cent. Opinion is evenly divided on the leadership coup – 45 per cent support, 48 per cent oppose – but most would prefer a full term to an early election, 36 per cent to 59 per cent. Head-to-head questions on leaders’ personal attributes produce consistently huge leads for Gillard (UPDATE: Possum reports primary votes of 42 per cent for the Coalition and 11 per cent for the Greens).

Channel Nine also had a poll conducted by McCrindle Research, who Possum rates “not cut for politics”. Nonetheless, their figures are in the ballpark of the others: Labor leads 54-46 on two-party, with 42.7 per cent of the primary vote against 38.8 per cent for the Coalition and 12.1 per cent for the Greens. Julia Gillard holds a lead as preferred prime minister of 64.8-35.2, the undecided evidently having been excluded. Sixty-three per cent believed she could “understand the needs of Australian mothers”.

Finally, market research company CoreData have produced a hugely dubious poll of 2500 people conducted “at 11am yesterday”, which has Labor on 29.5 per cent and “Liberal” on 42 per cent. This was primarily because no fewer than 21 per cent of respondents would not vote for Labor “because they did not feel that they had elected Julia Gillard”. Possum is familiar with the company, and says the sample would come “from their online panel, probably not perfectly balanced in the demographics and probably not a great fit for instapolitics”.

We’ve also had today the forlorn spectacle of the final Morgan poll conducted on Rudd’s watch. The face-to-face poll of 887 respondents from last weekend had Labor’s two-party lead widening from 51.5-48.5 to 53-47, with Labor up three points to 41 per cent at the expense of the Greens (down half a point to 12.5 per cent) and others (down 2.5 per cent to 4 per cent).

Morgan has also run one of their small-sample state polls for Victoria, this one culled from various phone polls conducted since the start of the month for a total of 430 respondents. It has the Coalition with a 50.5-49.5 two-party lead, from primary votes of 35 per cent Labor, 38 per cent Liberal, 13.5 per cent Greens, 3 per cent Family First and 7.5 per cent others.

UPDATE: Galaxy offers a full set of results, which puzzlingly offers us separate figures for Thursday and Friday. I’m not clear whether the previously published results were a combination of the two, or if they’re springing a new set of polling on us. In either case, the results for the two days are identical in every respect except that the Greens were a point higher on 12 per cent on Friday, and others a point lower on 5 per cent. Lots of further questions on attitudes to the coup and future government priorities, with 52 per cent believing Labor’s election prospects have now improved against 38 per cent who disagree.

Julia Gillard: day two

Australian politics has entered uncharted waters after yesterday’s brutally efficient leadership coup, but the consensus view is that Julia Gillard is favourite to lead Labor to a victory which might have been beyond Kevin Rudd. One naysayer is Peter Brent of Mumble, a man who has been known to get things right from time to time. Brent’s assessment, published in The Australian yesterday, is that the odds now slightly favour the Coalition, whereas Labor under Rudd would most likely have increased its majority. I think he has it the wrong way around.

Certainly there is a view abroad – Mark Bahnisch of Larvatus Prodeo being one proponent – that changing leaders, particularly when in government, is inherently destabilising and destructive. The New South Wales state government’s game of musical chairs is usually offered as a cautionary tale. However, it is a mistake to compare the federal government with one whose problems are underlying, terminal and, most crucially, age-related. Through Morris Iemma, Nathan Rees and Kristina Keneally, NSW Labor’s primary vote has been super-glued to 30 per cent in the polls, for the simple reason that the leadership hasn’t been the problem.

It was a different story entirely with Kevin Rudd, who led a first-term government with a strong economic record that ought to be well ahead. The main problem lay with a leader whose credibility in the eyes of voters had been irreparably damaged by the celebrated series of policy backdowns followed by the government advertising fiasco. As is now well known, such problems were mirrored within the party. Stunning as events of recent days have been, there has been no mystery about their underlying cause: when Rudd’s poll lead evaporated, so did his authority in the party. All that remained to be answered was whether the party still felt he could struggle through to an election win, allowing the matter to be dealt with less bruisingly after the event.

Key to the decision that he couldn’t was internal polling which reportedly showed Labor headed for a net loss of 18 seats. Purported details of such polling were provided by a party insider to Andrew Bolt, and they tell a believeable story. Included are Labor seats on less than 5 per cent and Coalition seats on less than 1.5 per cent – about 40 all told. The broad picture is of Labor facing swings of 4 per cent in New South Wales and Queensland and as much as 8 per cent in South Australia, but no change in Victoria or Tasmania. In Western Australia, Hasluck would be lost, but no swing can be determined as Brand and Perth weren’t included in the poll. Also said to be a lost cause for Labor was Darwin-based Solomon.

Twenty-one seats in all were identified as Labor losses against three gains, which coming off 88 seats notionally held by Labor would leave them five seats short of a majority. This would involve an overall swing of about 3.5 per cent and a Labor two-party vote of about 49 per cent, slightly below the trend of published polling. Taken together, the evidence pointed to a worrying but by no means irretrievable situation for the government. What proved fatal to Rudd was a lack of confidence, based on recent performance, in his capacity to turn the ship around.

With regard to the likely electoral consequences, Peter van Onselen in The Australian pretty much bangs the nail on the head as far as I’m concerned, as does Niki Savva at The Drum. This from Lenore Taylor the Sydney Morning Herald also caught my eye:

Tony Abbott put a brave face on Labor’s last-ditch leadership change but privately the Coalition was desperately disappointed that it would not face an election against Kevin Rudd.

And it was utterly dismayed the mining industry had – as one source put it – ”succumbed to [Gillard’s] guile” by agreeing to her offer of a negotiating truce in the mining super profits tax war and to take the industry advertisements attacking the government off the air.

The Coalition has gone out on a limb in support of the mining industry and the prospect of a deal between the miners and the government has left it edgy.

Some developments from the upheaval:

• In what would be red-letter news on any other day, Lindsay Tanner made the shock announcement he would quit politics at the next election, making Greens candidate Adam Bandt a short-priced favourite to take his seat of Melbourne. VexNews reports “talk” that Tanner hopes to be succeeded in the seat by academic, commentator and occasional broadcaster Waleed Aly, who would seem just the thing to defuse the threat of the Greens, and Socialist Left warlord Andrew Giles, who wouldn’t.

• Shortly before the spill, VexNews reported that if Rudd went, so might two Queensland marginal seat MPs: Chris Trevor in Flynn and Jon Sullivan in Longman. Trevor said yesterday that Gillard would “always have my full support”, but Emma Chalmers of the Courier-Mail reports from Labor sources that he was contemplating quitting. Chalmers also quotes Sullivan expressing disappointment at the result, but going no further than that.

• According to The Australian’s Jack the Insider, “Liberal Party polling tells (Abbott) that he is starting this contest against Gillard from a long way behind. Kevin Rudd may have had his nose in front but the polling tells Abbott that Gillard would win the next election by the length of the straight.”

And while I’m here, here’s a piece I wrote for Crikey last week on the electoral state-of-play in South Australia. It might be showing its age in some respects.

South Australia was Labor’s forgotten triumph of the 2007 election, replicating on a smaller and less spectacular scale the decisive tectonic shift in Queensland.

The statewide two-party swing to Labor of 6.8% was only slightly below Queensland’s 7.5%, which was borne out in the proportion of seat gains: three out of 11 in South Australia, nine out of 29 in Queensland.

Labor’s resurgence put an end to a slump which dated back to 1987, the last time they had won a majority of the South Australian two-party vote, and 1990, when they last won a majority of seats.

Before that the state had been a source of strength for Labor in the post-war era, notwithstanding that a dubious electoral boundaries regime kept them out of office for much of that time at state level.

This was partly because the state party branch was spared the worst of the 1954-55 split, but also because of the large blue-collar workforce required to service an economy based largely on manufacturing and industry.

The difficulties experienced by these sectors meant the state was hit hard by the economic upheavals of the 1980s, which together with the damage done to Labor by the 1991 State Bank collapse led to a fundamental electoral shift in the Liberals’ favour.

At federal level this was manifested in a series of grim federal election results that reduced Labor to two seats out of 12 in 1996, to which only one seat was added in later terms of the Howard Government.

With one seat having been abolished in 2004, Labor’s doubling of their representation at the 2007 election gave them a bare majority of six seats out of 11, and left the Liberals without a safe seat in Adelaide.

The two Liberal hold-outs in the city were Christopher Pyne’s seat of Sturt and Andrew Southcott’s seat of Boothby, which cover the traditional party strongholds of the east and inner south.

In a tale that will become increasingly familiar as this series proceeds, speculation about the coming election was long focused on the Liberals’ chances of retaining these existing seats, but such talk faded as the new year began and disappeared with Labor’s poll collapse over the past two months.

Labor’s main strength in South Australia lies in the coastal plain north of the city centre, which makes a safe Labor seat of Port Adelaide and marginals of four others which are leavened with more conservative areas beyond.

The electorate of Adelaide covers inner suburbs both north and south of the city, which are respectively strong and weak for Labor, and the growing inner-city apartment population in between, which has proved highly volatile in its electoral habits of late.

In a rare sighting of the “doctors’ wives” effect, Labor’s Kate Ellis bucked the trend of the 2004 election to win Adelaide from Liberal incumbent Trish Worth, and she emerged from the 2007 election with what seemed like a secure 8.5% margin.

However, the Liberals are talking of internal polling showing them “closing the gap”, after staggering swings were recorded in the electorate at the March state election (at which Education Minister Jane Lomax-Smith lost the state seat of Adelaide with a swing of 14.4%).

To the west of Adelaide is coastal Hindmarsh, which combines Labor-voting inner city areas with prosperous and conservative Glenelg in the south. Labor’s Steve Georganas won by the narrowest of margins when popular Liberal member Chris Gallus retired in 2004, before picking up a relatively modest swing in 2007.

North-east of the city centre is Makin, home to newer suburbs in the hills along with the eastern part of Salisbury on the plain. Makin is the only seat in the state which has form as a bellwether, being held by Labor from its creation in 1984 until 1996, Liberal through the Howard years and Labor’s Tony Zappia since 2007.

Further north is Wakefield, which offers even starker contrasts: deep red Elizabeth in the south, rapidly growing Gawler just past the city’s northern limits (where change is favouring Labor, if the state election is anything to go by) and conservative rural and wine-growing areas beyond.

Wakefield was a safe Liberal country seat until it absorbed Elizabeth at the redistribution before the 2004 election. Liberal candidate David Fawcett unexpectedly retained it for the Liberals on that occasion, but his narrow margin was eliminated by Labor’s Nick Champion in 2007 (Fawcett now stands poised to enter the Senate).

The only seat in Adelaide which conforms neatly with the mortgage belt marginal seat stereotype is Kingston, covering the city’s outer southern coastal suburbs. Labor’s Amanda Rishworth recovered this seat for Labor in 2007 after it was lost in 2004, interest rates having had a lot to do with it on each occasion.

The diversity that characterises the other marginals is significant, as it leaves their members as susceptible to rebellions in party heartlands as to the normally more decisive ebb and flow of the mortgage-payer vote.

This is where the mining tax could cause problems for Labor, as many blue-collar workers perceive a connection between the mining boom and the industrial and manufacturing sectors which employ them.

While South Australia is rarely given a guernsey as a “mining state”, BHP Billiton’s massive Olympic Dam project single-handedly allows the industry to punch above its weight, as it is associated in the public mind with the state shaking off its “rust belt” reputation from the 1990s.

Uncomfortably for Labor, BHP Billiton says the tax will jeopardise a $20 billion expansion to the project which is currently under consideration, a process that will certainly not be completed before the election.
Premier Mike Rann captured attention last week when he claimed any decision to stall the project would cost Labor four or even five seats.

For all that, the Liberals have big hurdles to clear if South Australia is to produce any of the seats it needs to overhaul Labor’s majority.

The problem is a lack of low-hanging fruit — even the most marginal of Labor’s six seats, Kingston, sits on an imposing margin of 4.4%.

Furthermore, the March state election suggests Labor has a trump card in the form of a ruthlessly efficient marginal seat campaign machine, which helped Mike Rann hang on to office with just 37.5% of the primary and 48.4% of the two-party vote.

The only seats in the state which swung to Labor were the two most marginal, Light and Mawson (respectively in Wakefield and Kingston federally), and the critical eastern suburbs seats of Hartley and Newland likewise held firm against a torrid tide. Elsewhere, Labor suffered double-digit swings nearly everywhere they could afford to.

Federal Labor will be hoping to achieve similar successes in working-class areas with a campaign to focus minds on industrial relations, thereby shoring up valuable support in Makin and Wakefield in particular.
Beyond Adelaide, the state’s three non-metropolitan seats are of limited electoral interest, notwithstanding the vague threat the Democrats and now the Greens have posed in Mayo, where Jamie Briggs struggled over the line in the September 2008 by-election that followed Alexander Downer’s resignation.

That leaves Barker in the state’s east, which covers rural territory which has never been of interest to Labor, and the outback electorate of Grey, which has transformed over the past two decades from safe Labor to safe Liberal — testament to the decline of the “iron triangle” cities of Whyalla, Port August and Port Pirie, and reflecting the experience of Kalgoorlie west of the border.

Rudd vs Gillard: 9am tomorrow

And you thought a week was a long time in politics. Two hours after the first intimations of action, Kevin Rudd has announced he will face a leadership challenge from Julia Gillard at 9am tomorrow. Speaking at his press conference, Rudd invoked indigenous issues, the mining tax, pensions and climate change in a clear pitch to the party’s left, whom he called on to stand tough against the machinations of the Right faction heavies who have brought the situation to a head.

For my money, if the party room’s electoral prospects are what matters to it, there is little choice for it but to back Gillard. The warlords have moved against Rudd because they are brutally aware that it is he who is dragging them down in the polls and threatening their re-election prospects. In his absence, the government will be able to modify damaging policies as Rudd could not afford to, for fear of being called out over another “backflip”. Labor would also enter the election with a credible and certain story to tell about the next three years, the lack of which defeated Howard more than any single factor with only the possible exception of WorkChoices. Then there’s the feel-good factor of our first woman prime minister, which most voters recognise as overdue. Finally, I suggest that Glenn Milne’s thoughts last week on the dynamics of a Gillard-Abbott election battle would end up looking highly prescient after the event.

Over to you.

UPDATE: And with what great timing, we get the long-awaited quarterly cumulative Newspoll. This combines polls five from April to June, which successively had Labor’s party vote at 54, 49, 50, 51 and 52, allowing state and demographic results to be provided from a reasonable sample size. The state breakdowns show a surprisingly mild move against Labor in Western Australia, from 51-49 behind in January-March to 53-47 in April-June. While Labor has crashed seven points on the primary vote to 31 per cent, the dividend has gone entirely to Greens and “others”. If the result was uniform, Labor would hold its own on those numbers. The only other state with Labor trailing is Queensland, where Labor fell from 51-49 ahead to 52-48 behind. New South Wales and Victoria recorded little change with Labor leading 52-48 and 56-44, while their lead in South Australia dived from 55-45 to 51-49.

There was little sign of recent turmoil among voters over 50, among whom the Labor vote held steady on 37 per cent. It was a case of other age groups falling to that level: Labor fell five points to 39 per cent among 18-34s, and seven to 36 per cent and 35-49s. The Coalition primary vote was up three points among men to 43 per cent but steady on 39 per cent among women, who have instead sent votes lost to Labor to the Greens and others.

UPDATE 2 (Thursday morning): Not sure how much it’s worth now, but The Advertiser ran a poll this morning of 530 voters from the seat of Adelaide, where Labor holds a margin of 8.5 per cent but has been said to be in trouble. The poll doesn’t entirely bear this out: Labor’s primary vote was down 7 per cent from the election to 41 per cent, but the Liberals are also down from 37 per cent to 35 per cent – suggesting the undecided had not been distributed – and most of the dividend went to the Greens, up 6 per cent to 16 per cent. In two-party terms, Labor retained a handsome 57-43 lead.

UPDATE 3 (Thursday afternoon): Comments thread talk tells us Galaxy are in the field, suggesting we can expect the first poll of the new era either in the Sunday News Limited tabloids.

Newspoll marginal seats survey

The Australian has published marginal seat Newspoll results from two key seats in New South Wales, Lindsay and Page, and three in Queensland, Flynn, Dawson and Longman. Whereas yesterday’s national poll gave Kevin Rudd cause for relief, today’s results are grim for Labor in every case but Page – and particularly so in Lindsay. According to Dennis Shanahan’s report, the poll shows:

• A 12 per cent swing in Lindsay, which David Bradbury holds for Labor on a margin of 6.8 per cent. Kevin Rudd’s disapproval rating in the electorate is 61 per cent, and both Tony Abbott and Julia Gillard are preferred as prime minister (Abbott by 44 per cent to 40 per cent). UPDATE: Psephos in comments questions the utility of a poll conducted simultaneously with the Penrith by-election, and there may well be something in this.

• There is a collective 6 per cent swing against Labor in Flynn, Dawson and Longman, held by Labor with margins of 2.2 per cent, 2.6 per cent and 1.9 per cent, resulting in a Liberal two-party lead of 54-46. However, Kevin Rudd apparently leads as preferred prime minister.

• It’s a very different story in the north coast New South Wales seat of Page, where the poll shows Labor picking up a swing of 2 per cent.

Full tables presumably forthcoming.

UPDATE: It turns out that the Lindsay and Page polls, and the combined poll of the three Queensland seats, each had samples of 600 and hence margins of error of about 4 per cent. We also have today courtesy of Essential Research data from the weekend’s survey on firmness of voting intention, which – believably enough – shows the Greens’ surge has been muted by an increasing proportion of “soft” support. The “very firm” share of the Greens vote has fallen from 41 per cent to 31 per cent, “might change” has gone from 39 per cent to 45 per cent and “soft” has gone from 18 per cent to 22 per cent. While Labor’s vote has gone up and the Liberals down, both parties have a higher proportion of firm and a lower proportion of soft votes.

UPDATE (23/6): Today The Australian reports that Galaxy has polled 1600 respondents in Brisbane, Petrie, Bowman and Ryan on behalf of WWF Australia, which shows the Liberal National Party with a 51.5-48.5 lead – reversing the result from the last election. The 3 per cent swing this points to, if uniform, would cut Labor very fine in Brisbane and Petrie without actually costing it either of them. Although the Michael Johnson ruckus in Ryan might have dragged down their result. Labor is said to be on 34 per ent of the primary vote, down 9 per cent on the election, but the Coalition are only up one to 46 per cent. More evidence then of Labor bleeding votes to the Greens and getting most back as preferences – depending on how you do the calculation, of course.

Newspoll: 52-48 to Labor

Full results available from Peter Brent at Mumble. Labor’s 52-48 lead is a slight improvement on 51-49 from three weeks ago, and under the circumstances will come as an enormous relief for the Prime Minister. One sting in the tail is that Labor’s primary vote remains steady on a parlous 35 per cent. The Coalition is down one point to 40 per cent and the Greens are on 15 per cent, one point off their record-breaking effort from three weeks ago. The two-point slack has been taken up by “others” on 10 per cent.

Another sting in the tail is that the preferred prime minister rating has swung to Abbott: Rudd is down three points to 46 per cent and Abbott is up four to 37 per cent, which is respectively a personal worst and the best result achieved by a Liberal leader on Rudd’s watch. This is despite the fact that the leaders’ approval ratings are basically unchanged. Kevin Rudd’s approval is steady on 36 per cent and his disapproval is up a point to 55 per cent, while Tony Abbott is respectively up a point to 38 per cent and steady on 49 per cent.

A further question on prospective standard of living produces a neutral result: “improve” and “get worse” are both on 17 per cent, with 65 per cent nominating “stay the same”.

Next cab off the rank: Essential Research, which should be through at about 1pm EST.

UPDATE: Hats off to Dennis Shanahan, who shows he’s not scared of a renewed round of opprobrium from the Laborsphere.

UPDATE 2: Essential Research joins the party by also showing Labor’s lead up from 51-49 to 52-48, although it gets there by showing a primary vote recovery for Labor (up three to 38 per cent) at the expense of the Greens (down three to 11 per cent), with the Coalition down one to 40 per cent. Again, there’s a sting in the tail for Kevin Rudd – 40 per cent say Labor would have a better chance of winning if they changed leaders, against only 37 per cent who say he is the best person to lead the party to the election. However, the results on this measure are substantially worse for Tony Abbott – 29 per cent and 47 per cent. Kevin Rudd remains preferred prime minister over Abbott by 47 per cent to 30 per cent, and also over Julia Gillard by 36 per cent to 33 per cent. There’s also a very interesting finding on troops in Afghanistan, with 61 per cent saying out troops should withdraw.

Penrith by-election live

UPDATE (FRIDAY): At a safe distance, here is my Crikey piece on the by-election which was originally subscriber only.

By-elections have been a feature of Australian political life since the first democratically elected colonial parliaments were established in the mid-nineteenth century.

Compiling a comprehensive set of historic by-election results at state as well as federal level would involve painstaking research through the archives of various electoral authorities and newspapers-of-record.

It is possible that, buried in some such dark and dusty place, there might be found details of a by-election defeat as bad for a major political party as the one suffered by New South Wales Labor in Penrith on Saturday.

However, the lack of any such precedent in recent memory, together with the trend towards greater electoral volatility over time, suggests it isn’t very likely.

We do know that Labor has suffered the worst by-election swing recorded in the long history of New South Wales, thanks to exhaustive historical results compiled for the New South Wales Parliamentary Library by Antony Green.

The closest historical parallels that spring to mind are indeed from the Premier State: Bass Hill in 1986, when Neville Wran’s previously safe seat fell to the Liberals upon his retirement, and — distressingly close at hand as far as the government is concerned — Ryde and Cabramatta just after Morris Iemma’s departure in September 2008.

With respective two-party swings of 22.2%, 22.9% and 22.0%, Penrith looks to have surpassed all three, recording an election night swing of 25.5%.

One result that has its measure is the 1991 by-election for Geraldton in Western Australia, held as Carmen Lawrence’s Labor government groaned under the weight of revelations from the WA Inc Royal Commission.

On that occasion Labor’s primary vote fell from 47.6% to 16.6%, while the combined Liberal and Nationals vote went from 43.5% to 66.5%. Labor finished third behind the Nationals and thus did not even make the final two-party cut, making it impossible to determine a two-party swing.

It might also be worth mentioning Labor’s forfeit last year in the Tasmanian upper house seat of Pembroke, which the party was too scared to contest after the resignation of its sitting member Allison Ritchie.

(It should be stressed that this is limited to two-party contests, and thus excludes the fairly common occurrence of major party support being gouged by the emergence of popular independents. And while I’m making asides, it’s interesting to note that the by-election hall of shame is dominated by Labor).

In Penrith, Labor finished the night on 24.4% of the primary vote, losing almost exactly half of their 48.7% from 2007.

Three-quarters of the dividend was collected by the Liberals, up from 32.6% to 50.9%, while the Greens vote more than doubled from 5.5% to a still quite modest (for a by-election) 12.8%.

While the result has never been in doubt since former member Karyn Paluzzano announced her resignation in May, Labor might initially have hoped for something a little less bruising.

At the time of the 2008 by-elections the then Premier, Nathan Rees, had a Newspoll approval rating of 39%, which — mediocre as it was — proved to be a honeymoon peak. It clearly says something very alarming about the state of the Labor brand that Kristina Keneally has been able to do even worse with an approval rating of 47%.

Several factors suggest themselves as explanations for Labor’s ability to plumb new depths, of which the most newsworthy is the decline in Labor’s federal fortunes since Ryde and Cabramatta. Evidence of Rudd Government policy failures feeding into a general questioning of Labor’s competence in the electorate might be anecdotal, but it’s substantial in volume.

Of course, those rushing to judgement on the Prime Minister need to recall that the by-election follows a horror stretch for the State Government, even by its own abysmal standards.

There is also reason to think Penrith was especially unfavourable terrain for Labor to face a by-election.

As the experience of the corresponding federal electoral of Lindsay has shown, this is an area of fickle political loyalties which never fails to jump on the bandwagon when a swing is on.

The electorate’s outer suburban location also places a premium on transport issues, with local voters having suffered the worst of freeway gridlocks and a creaking public transport system.

A somewhat more nuanced picture of the result can be gained by comparing the two distinct parts of the electorate: Penrith and its surrounds, which account for about 80% of the voters, and the very different electoral terrain of the Blue Mountains, which the electorate touches upon at Blaxland, Glenbrook and Lapstone.

Labor was hit hardest in the former, down by about 26% compared with 21% in the Blue Mountains, and the dividend there more heavily favoured the Liberals.

There have been suggestions the Greens could have hoped for more from a Labor collapse on this scale, but the distinction between the two areas suggests their failure to do so says as much about Penrith as the state of the parties at large.

Penrith boasts slightly above average incomes but below average educational attainment, making it weak territory for the Greens. By contrast, the Blue Mountains has almost double Penrith’s proportion of professionals, and is thus a lot more representative of the kinds of seats where the party will be hoping to defeat Labor.

It is thus highly significant that the Greens outpolled Labor in every one of the four Blue Mountains booths, polling a collective primary vote of over 23% compared with about 11% in and around Penrith.

Accordingly, the result holds little comfort for Labor in the Greens targets of Balmain and Marrickille, and suggests they will struggle to stay ahead of the Greens in the neighbouring seat of Blue Mountains — academic though that may be, given the near certainty of it falling to the Liberals.

And while federal implications of state by-elections should generally be treated with caution, the scale of the overall result gives Labor ample reason to be nervous not only Lindsay, but also the other nearby marginals of Macquarie and Macarthur.

UPDATE (MONDAY): I have a subscriber-only article on the by-election in Crikey today.

My earlier preference projections proved too optimistic for Labor: with all the two-party booth counts conducted, the swing at the close of play is 25.5 per cent. This is a result without any precedent I’m aware of, at least as far as two-party contests are concerned. The swings in Ryde (22.9 per cent) and Cabramatta (22.0 per cent) in September 2008 were presumed at the time to show Labor as low as they could go, but tonight’s result is measurably worse.

The result comes despite an improvement in Labor’s leadership position since September 2008 – at state level, at least. According to Newspoll, Nathan Rees achieved the earlier swings on a honeymoon approval rating of 39 per cent, while Kristina Keneally has suffered Penrith on 47 per cent. However, Kevin Rudd’s approval rating in that time has dropped from 54 per cent to 36 per cent. Talk of federal connections to state election results can usually be taken with a grain of salt, but in this case it’s hard to believe Labor’s federal collapse has played no role in giving state Labor new depths to plumb. It certainly doesn’t bode well for them in Lindsay.

Swing 2PP/Swing
Thain (ALP) 8200 24.4% -24.4% 33.7%
Wright (GRN) 4221 12.6% 7.3% -25.5%
Saunders 724 2.2%
Ayres (LIB) 17067 50.9% 18.3% 66.3%
Leyonhjelm (ORP) 638 1.9% 25.5%
Sanz (DEM) 297 0.9% 0.0%
Green (CDP) 1516 4.5% -1.8%
Selby (IND) 884 2.6%
ALP LIB ALP 2PP Swing
Blaxland 20.4% 45.7% 37.3% -22.0%
Blaxland East 14.0% 51.8% 29.1% -21.9%
Braddock 29.4% 45.9% 41.2% -25.3%
Cranebrook 24.2% 52.2% 34.3% -28.4%
Emu Heights 23.0% 52.2% 33.4% -23.9%
Emu Plains 23.5% 55.6% 32.5% -20.8%
Glenbrook 15.9% 48.4% 33.1% -19.1%
Jamisontown 22.8% 52.7% 33.2% -27.5%
Kingswood 27.5% 43.7% 41.0% -26.1%
Kingswood Park 30.8% 46.7% 41.6% -25.3%
Kingswood South 29.2% 45.6% 40.7% -24.2%
Lapstone 16.0% 50.9% 32.4% -17.2%
Leonay 17.0% 65.2% 24.3% -20.4%
Nepean High 22.3% 55.5% 30.2% -25.4%
Penrith 27.9% 46.2% 40.0% -24.4%
Penrith North 28.2% 51.7% 37.9% -25.1%
Penrith South 27.8% 49.9% 37.8% -21.3%
Stuart Street 25.8% 51.2% 35.5% -19.1%
York 25.7% 53.0% 34.8% -25.5%

8.44pm. Penrith High is the final booth to report primary votes.

8.31pm. Kingswood Public School pushes swing up from 23.4 per cent to 23.6 per cent.

8.18pm. I suppose federal Labor can argue this result is no different from the Ryde by-election of September 2008, at which time it held an enormous lead in the polls – though I’m not sure how many will listen (it would be useful for them if late counting pulled the swing below 22.9 per cent). The riposte would be that Labor had a popular state leader this time and should have done better.

8.15pm. Labor haven’t come within cooee of winning a booth. The Greens have more than doubled their vote, but to still modest levels – Penrith not exactly being the latte belt. There’s a six point gap between the pro-Liberal and anti-Labor swings, which is reasonably good work for the Liberals.

8.12pm Jamistown Uniting Church Hall added. Two to go.

8.04pm. Cranebrook booth added – one of Labor’s worst results, but no real difference made.

8.02pm. So at this stage Labor seems to have suffered a slightly worse result than Ryde (22.9 per cent) and Cabramatta (22 per cent), and a good deal worse than Lakemba (13.5 per cent).

7.52pm. Another two booths added, five more go, and the swing continuing to settle at 23 per cent. I’m using real world 2PP figures now.

7.43pm. Jamistown Public and Penrith South results added. NSWEC link working now, BTW.

7.40pm. Three more booths including very large Jamison High School booth see the swing settle at about 23 per cent. Antony Green has 2PP counts and I don’t.

7.35pm. Stoopid NSW Electoral Commission PDF now points to wrong link.

7.33pm. Lapstone, Leonay and Emu Heights very slightly reduce the size of the swing.

7.31pm. St Dominic’s College booth added.
7.28pm. Kingswood Park booth keeps the picture consistent.

7.24pm. Blaxland High and Penrith PCYC booths very slightly improve the picture for Labor, the two-party swing now looking at 25 per cent. The Greens outpolled Labor at Blaxland High.

7.08pm. Preference distributions are guesses at this stage.

7.06pm. Swing of about 30 per cent against Labor at small Nepean Hospital – 267 votes all up. Not a good start for them, you would have to say.

6.15pm. Welcome to live coverage of tonight’s bloodbath in Penrith. I guess we should get first figures in about 20 minutes. In the meantime, enjoy this illuminating booth results map for the 2007 election.