Newspoll: 53-47 to Coalition

The Coalition finally records an opinion poll lead, as Newspoll breaks the post-election ice.

The ten-week silence of Newspoll – and indeed Australian polling in general, so far as voting intention is concerned – has ended with a result of 53-47 to the Coalition, as reported by The Australian. To this, naturally, must be added the qualification that the pollster never once recorded the newly re-elected government with a lead in the entire three years of the previous parliamentary term. The poll has the Coalition at 44% of the primary vote (41.4% at the election), Labor at 33% (33.3%) and the Greens at 11% (10.4%). The report seems to be saying One Nation is at 3%, which compares with the 3.1% they scored at the election when contesting 59 out of 151 seats.

The leadership ratings have Scott Morrison’s approval at a new high of 51%, up five on the pre-election poll, and down nine on disapproval to 36%. Anthony Albanese’s Newspoll ratings are 39% approval and 36% disapproval, which is a) “the first net positive approval rating for an Opposition leader since 2015”, as noted in the report since Simon Benson, b) the worst Newspoll debut for an Opposition Leader since Andrew Peacock in 1989, as illustrated in this earlier post, and c) the equal lowest uncommitted rating for an Opposition Leader on debut, perhaps mitigating b) a little. Morrison leads 48-31 on preferred prime minister, compared with 47-38 in the pre-election poll, which we can now presume was flattering to Bill Shorten.

No indication at this point as to whether and how Newspoll is doing anything differently. Certainly it looks like business as usual to the extent that the poll was conducted Thursday to Sunday from a sample of 1601, with The Australian’s report trumpeting a 2.4% margin of error that is less than the size of its error at the election.

Call of the board: regional New South Wales

The second in a series that leaves few stones unturned in its exploration of the May 18 election result.

The metropolitan episodes of this series will feature maps and analysis guided by a demographic model to predict seats’ two-party results, so that areas of over- or under-performance might be noted. However, results maps only really work for areas of concentrated population, and it turns out the model works a lot less well when you move away from the cities. In particular, it records historic Labor strongholds in the Hunter and Illawarra as marginal Liberal seats, which I’m guessing results their lack of ethnic diversity, which the model strongly associates with conservatism. This suggests the model needs to be refined with interaction variables to measure the difference in effects between cities and regions, which I’ll hopefully get around to at some point.

Now for the Call of the Board in non-metropolitan New South Wales, broken into four easy pieces.

Hunter region

Newcastle (Labor 13.8%; 0.0% swing to Liberal): The pattern of the capital cities was reflected in Newcastle, the urban core of which swung to Labor while the low density surrounds went the other way. The Newcastle electorate contained exactly as much of each as to cancel each other out, with both major parties down slightly on the primary vote to make way for United Australia and a lift for the Greens.

Shortland (Labor 4.4%; 5.5% swing to Liberal): In neighbouring Shortland, however, Labor emerged with its narrowest margin since the seat’s creation in 1949. There were traces of the inner urban effect at the northern end of the electorate, but the swings elsewhere were severe enough to take 10.0% out of Pat Conroy’s primary vote. Most of that was harvested by new minor party entrants, but the Liberals gained swings of 2.2% on the primary and 5.5% on two-party preferred.

Paterson (Labor 5.0%; 5.7% swing to Liberal): It was a similar story just north of Newcastle in Paterson, where Meryl Swanson, who should have been enjoying at least half a sophomore effect, copped a two-party swing of 5.7%. The primary vote swing of 5.0% was less severe than Shortland because the minor party market was more crowded here in 2016. In particular, this was one of two seats in New South Wales where One Nation ran in 2016, and the only one where they repeated the performance in 2019. Their vote was up from 13.0% to 14.2%, the second strongest of the six New South Wales seats they contested after Hunter.

Hunter (Labor 3.0%; 9.5% swing to Nationals): Labor’s single worst result of the election was Joel Fitzgibbon’s 14.2% primary vote and 9.5% two-party slump in Hunter, reducing his previously formidable margin to 3.0%. The last time Labor was run this close in a seat bearing the name of Hunter was in 1984, and the time before that was 1906. The coal industry effect was unmistakeable: the Newcastle end of the electorate swung about as heavily as the Shortland booths on the other side of Lake Macquarie, whereas the full force landed at Cessnock. The remarkable 21.6% primary vote for One Nation, more than in any seat in Queensland, was fairly uniformly spread geographically. This left them only slightly shy of the 23.5% vote for the Nationals (who, a little oddly in my view, have the right to contest the seat under the coalition agreement), but the gap failed to close on preferences. How close they would have come of overtaking Fitzgibbon at the final count had it been otherwise is a matter for conjecture.

Northern coast

Lyne (Nationals 15.2%; 3.2% swing to Nationals): David Gillespie held almost steady on the primary vote while Labor fell 2.5% and the Greens fell 2.9%, reducing the flow of preferences to Labor. Fact I hadn’t noticed before: the Liberal Democrats can score pretty well in Nationals seats with no Liberal running, in this case 5.8%.

Cowper (Nationals 11.9%; 0.7% swing to Labor): After all the hype about Rob Oakeshott’s prospects, the result was remarkably similar to his failed bid in 2016. Pat Conaghan, who replaces Luke Hartsuyker as the Nationals member, added 1.1% to the party’s vote, scoring 47.1%, while Oakeshott was down 1.8% to 24.5%. That still left him well clear of Labor, up 0.2% to 13.8%, and he landed 6.8% short after preferences, which was 2.2% more than in 2016. The Coalition-versus-Labor two-party count produced a 0.7% swing to Labor, perhaps reflecting Hartsuyker’s retirement.

Page (Nationals 9.4%; 7.1% swing to Nationals): Kevin Hogan, a Nationals member who vaguely kept his distance from the Coalition after the putsch against Turnbull, achieved the biggest margin ever recorded in a seat that has been an arm wrestle since its creation in 1984, the margin never previously exceeding 5%. Hogan was up 5.3% on the primary vote and 7.1% on two-party preferred, the latter being the biggest swing against Labor in New South Wales after Hunter. There were two areas where Labor held its ground: just outside Coffs Harbour at the electorate’s southern extremity, and behind the hemp curtain at Nimbin in the north.

Richmond (Labor 4.1%; 0.1% swing to Labor): As just noted, the area around Nimbin bucked the trend of a heavy swing against Labor in Page. This regional effect was even more pronounced at the Byron Bay end of Richmond, where a number of booths recorded double-digit swings to Labor. Many of these booths are in fact won by the Greens, who only succeeded in treading water overall in the face of competition from Sustainable Australia and Involuntary Medication Objectors (though the latter, critics of this region take note, only polled 1.2%). The Tweed Heads end of the electorate was and is a different kettle of fish, recording low support for the Greens and a two-party swing to the Nationals. With the two ends pulling in different directions, the distinctiveness of the Byron Bay region is further enhanced, as illustrated by the image below (which would naturally tell a similar story for the Greens primary vote).

South-eastern

Cunningham (Labor 13.4%; 0.1% swing to Labor): The size of the two-party swing typified a dull result, in which Labor fell slightly on the primary vote, the Liberals were up slightly and the Greens vote hardly changed. The primary vote difference presumably failed to translate into a Liberal two-party swing because the Christian Democrats vacated the field after recording 4.1% in 2016.

Whitlam (Labor 10.9%; 2.8% swing to Nationals): The Liberals made life hell for some of us by declining to field a candidate here and leaving the seat to the Nationals, so that a two-party swing could only be calculated by comparing Labor-Liberal to Labor-Nationals. This the AEC, for one, declined to do. By that measure, Labor’s Stephen Jones suffered a swing of 2.8%. In the Liberals’ absence, the combined Coalition primary vote was down from 32.7% to 25.5% as Liberals unwilling to plump for the Nationals opted for the United Australia Party, whose 8.8% was their second best result in the country after Riverina.

Hume (Liberal 13.0%; 2.8% swing to Liberal): Labor dropped 5.3% on the primary vote here, though it went to independent Huw Kingston and the United Australia Party rather than Liberal member Angus Taylor, who was down slightly.

Gilmore (LABOR GAIN 2.6%; 3.3% swing to Labor): One of the few seats that went to Labor’s this was Labor’s eighteenth biggest swing nationally, and the fifth biggest in a seat that can’t be described as inner urban. The primary vote for Labor’s Fiona Phillips was actually down 3.0%, as seven candidates took the field compared with four in 2016 – among whom was spurned Liberal independent Grant Schultz, who came in fifth with 7.0%. Katrina Hodgkinson failed to light up the scoreboard as Nationals candidate, scoring 12.5%. The drop in the Liberal vote exceeded this, so that the combined Coalition primary vote was down 3.6%, similar to Labor. That Labor nonetheless enjoyed a solid and decisive two-party swing suggests a reasonable share of Nationals votes leaked to them as preferences.

Eden-Monaro (Labor 0.8%; 2.1% swing to Liberal): Eden-Monaro’s fame as the bellwether seat was further buried as Mike Kelly held off a swing of 2.1% to hold on by 0.8%. The Nationals might have done better to have stayed out, polling only 7.0% and contributing to a 4.3% drop in the Liberal primary vote. Labor was down 2.7%, the Greens up 1.2%. There was maybe a slight tendency for Labor to hold up better in urbanised areas, but no clear geographic pattern overall.

Interior

New England (Nationals 17.6%; 1.2% swing to Nationals): Barnaby Joyce’s remarkably strong result at the November 2017 by-election was proved to be no fluke, as he gained in 2.5% on the primary and 1.2% on Coalition-versus-Labor two-party in the face of even greater adversity this time. The former accomplishment was no doubt assisted by the absence of Tony Windsor, who polled 29.2% in 2016, although another independent, Adam Blakester, polled 14.2% this time to take second place over Labor, landing 14.4% short after preferences.

Calare (Nationals 13.3%; 1.5% swing to Nationals): The only seat Shooters Fishers and Farmers contested in New South Wales after their state election triumph in March, they managed third place with 17.4% of the primary vote. Nationals member Andrew Gee, a sophomore, was down 2.9% to 44.7%, and Labor was down 4.9% to 22.1%.

Riverina (Nationals 19.5%; 3.0% swing to Nationals): The only seat in the country where the United Australia Party broke double figures, to which it owes a small field of four candidates that didn’t include a Liberal, leaving Palmer’s outfit as the only non-left alternative to the Nationals. Nationals leader Michael McCormack gained 2.7% on the primary and 3.0% on two-party.

Farrer (Liberal 10.9% versus Independent): Kevin Mack was one of a number of highly regarded independents who struck out on the night, managing 20.5% of the primary vote – not nearly enough to disturb Liberal incumbent Sussan Ley, who despite shedding 7.2% on the primary vote still ended up with a straight majority of 50.7%. Ley won by 10.9% after preferences, and suffered a 0.7% two-party swing against Labor.

Parkes (Nationals 16.9%; 1.8% swing to Nationals): Both major parties were well down on the primary vote, incumbent Mark Coulton shedding 7.9%, in the face of solid performances by the Liberal Democrats (8.1%, another example of the no-Liberal-candidate effect), independent Will Landers (7.2%) and the United Australia Party (an above-average 6.3%).

Western Australia redistributed (state)

A draft new set of state boundaries in Western Australia produces little to frighten the horses.

Update

Over the fold is a table showing an almost-complete set of Labor-versus-Liberal/Nationals two-party margins, excluding a few seats where the 2017 result was Liberal-versus-Nationals (Moore and Roe) or Labor-versus-independent (Baldivis). This treats Kingsway as the successor to Girrawheen, and Girrawheen as the successor to Mirrabooka. I am now calculating the Labor margin in Kingsway at 9.1%, which is modest enough that Labor would lose the seat at a bad election, like 2013. This amounts to a 7.6% cut in the old margin from Girrawheen – so if, as I suggested, Labor runs Margaret Quirk in Kingsway and gives Girrawheen to Janine Freeman, who is technically homeless with the abolition of Mirrabooka, Quirk would consider that regrettable.

As noted in the original version of this post (also over the fold), Labor has been short-changed by the redistribution’s determination to preserve the existing number of country seats, but finds ample consolation in a number of helpful revisions to marginal seats:

• Labor’s margin in Balcatta, which the party lost for the one and only time in 2013, goes from 5.8% to 8.0%, as it loses marginal territory (at least on 2017 results) in the north to Kingsley and gains Labor territory in the east from Mirrabooka.

• The change just noted to Kingsley also nudges the dial there very slightly in Labor’s favour, from 0.7% to 1.2%.

• In Burns Beach, the loss of territory in the south to Joondalup and gain in the north from Butler boosts Labor from 2.5% to 4.9%.

• No doubt the 2017 election is as bad as it will ever get for the Liberals in Hillarys, but I am calculating that Labor would have won it in 2017 by the barest of margins, after falling 4.1% short at the election. Marginal territory has been gained in the north from Joondalup, and Liberal territory in the south has gone to Carine.

• The transfer of part of Liberal-voting Leeming to Riverton in the north boosts Labor from 1.0% to 2.0% in Jandakot.

• Tweaking of the boundary with Fremantle improves Labor’s margin in Bicton from 2.9% to 3.6%.

• A territory swap with West Swan boosts Labor from 7.3% to 9.2% in Wanneroo.

• An exchange of rural territory in the south for Mandurah’s fringes in the north boosts Labor from 1.4% to 2.3% in Murray-Wellington.

Conversely:

• In Joondalup, which gains in the north from Burns Beach and loses in the south to Hillarys, Labor’s margin is reduced from 0.6% to 0.1%.

• In Swan Hills, a Labor margin of 14.5%, which belies its history as a tight marginal seat, reduces to 12.0%, as Ellenbrook suburbia is exchanged for parts of the Swan Valley.

Continue reading “Western Australia redistributed (state)”

Essential Research leadership polling

The second set of leadership ratings since the election is featured in the latest release from Essential Research, which may also offer a hint of how it plans to respond to the great pollster failure.

The fortnightly Essential Research release is the second since the election to encompass the monthly leadership ratings. These offer positive signs for Anthony Albanese, who is up four from his debut on approval to 39% and down one on disapproval to 24%, while Scott Morrison is slightly improved in net terms, with approval steady on 48% and disapproval down two to 34%. Morrison’s lead as preferred prime minister is effectively unchanged, shifting from 43-25 to 44-26. The poll also features a series of questions on the ban on tourists climbing Uluru, which 44% support and 30% oppose, and 69% professing awareness of the issue.

Of particular interest in this release is the revelation that Essential is inquiring about respondents’ income, which appears to be a new development. The only detail provided in the polling results is that Morrison has 59% approval among higher income earners, but the appendices go to the trouble of telling us that Essential has set three income cohorts for its surveys: low (below $52,000), high (above $104,000) and medium (in between).

I suspect this means Essential’s response to the pollster failure will be to start using income to weight its results. This is a departure from the Australian industry norm of weighting only by geography, gender and age, and would also seem to be a bit unusual internationally. An American pollster noted last year the practice had fallen out of favour there due to the high non-response rate to questions on personal income. The preference is to instead weight to other factors which themselves correlate with income, notably education and, particularly in Britain, social class.

The poll was conducted Wednesday to Sunday from a sample of 1091. In the Guardian report accompanying the poll, the elephant in the room was addressed thus:

There has been controversy post-election about the reliability of opinion polling because none of the major surveys – Newspoll, Ipsos, Galaxy or Essential – correctly predicted a Coalition win on 18 May, projecting Labor in front on a two-party preferred vote of 51-49 and 52-48. The lack of precision in the polling has prompted public reflection at Essential, as has been flagged by its executive director, Peter Lewis. Guardian Australia is not currently publishing measurements of primary votes or a two-party preferred calculation, but is continuing to publish survey results of responses to questions about the leaders and policy issues.

Also in The Guardian today are results from a separate Essential Research poll, this one for Digital Rights Watch concerning recent police raids on journalists. In response to a question noting raids on “the offices and homes of News Corp and ABC journalists who reported on national security issues”, 40% said they were very concerned, 34% slightly concerned and 26% not concerned. Similar results were produced on questions relating to metadata and police powers to break into online communications systems. The poll was conducted Wednesday to Sunday from a sample of 1089.

UK Conservative leadership: Johnson 66.4%, Hunt 33.6%

Boris Johnson to become British PM after a crushing victory in a Conservative members’ vote, but suffers a parliamentary defeat even before becoming PM. Guest post by Adrian Beaumont.

Guest post by Adrian Beaumont, who joins us from time to time to provide commentary on elections internationally. Adrian’s work on electoral matters for The Conversation can be found here, and his own website is here.

The result of the Conservative members’ ballot for leader was declared on July 23. Boris Johnson defeated Jeremy Hunt by 66.4% to 33.6%, out of the almost 139,000 valid votes cast. Johnson will become British PM on July 24, after Theresa May formally resigns. He has repudiated the Northern Ireland backstop arrangement in Theresa May’s deal. It is very unlikely that an agreement with the European Union can be reached without such a backstop, making a no-deal Brexit more likely.

On July 18, the Commons voted by 315 to 274 to accept an amendment that would make it more difficult for Johnson to prorogue (suspend) Parliament to force a no-deal Brexit. Just one Labour MP voted with the Conservatives while 17 Conservative MPs voted with Labour, and there were notable Conservative abstentions, including current Chancellor Philip Hammond.

While this was a decisive defeat for Johnson’s position, commentator Stephen Bush says that Johnson’s big gamble is not that the Commons will oppose no-deal, it is that the Commons will be unable to agree on something (revoking Brexit, a second referendum, a general election or making Jeremy Corbyn PM) that will actually prevent no-deal occurring. The Commons will rise on July 25 for its summer recess, and is not scheduled to return until September 3. There is a three-week recess for party conferences from mid-September until early October, so there will be less time for Parliament to deal with Brexit. It is unlikely there will be a no-confidence vote before the summer recess.

In five of the seven most recent national polls, Labour had two to six point leads over the Conservatives, but trailed the Conservatives by four points in two YouGov polls. YouGov is the most Conservative-leaning pollster, and it is no coincidence that Labour’s 18%, which I mentioned in my last UK politics article, was from a YouGov poll. There has been some drop in the Brexit party and Liberal Democrat vote, so it has become more of a standard two-party contest.

The most recent ComRes poll had a clear warning for Johnson.  Current voting intentions in that poll were 28% Labour, 25% Conservative, 19% Brexit party and 17% Lib Dem. When asked how they would vote if a new election were held before October 31 with Brexit undelivered, vote shares became 28% Labour, 22% Brexit, just 18% Conservative and 18% Lib Dem. If Brexit is delivered before the next election, vote shares are 32% Conservative, 29% Labour, 18% Lib Dem and 10% Brexit.

Hypothetical polling should be taken with a grain of salt, as people are not good at predicting how they will react to an event.  If Parliament is seen as obstructing Brexit, the Conservatives could do better in the first scenario, although the number of Conservative rebels in such a scenario could itself be damaging. In the second scenario, serious negative economic consequences of a no-deal Brexit would be likely to hurt the Conservatives.

The Brecon & Radnorshire by-election, which I covered in more detail last time, will be held on August 1. A poll has the Lib Dems easily winning with 43%, followed by the Conservatives on 28%, Brexit party 20% and Labour 8%. This result would be a Lib Dem gain from the Conservatives. I will be writing about this by-election result and post-Johnson polls on August 2.

US Democratic presidential primaries and Trump’s re-election prospects

I wrote for The Conversation on July 18 that Joe Biden is leading the Democratic presidential primaries, and is followed by Bernie Sanders, Elizabeth Warren and Kamala Harris. Donald Trump’s ratings are well below where they should be given the strong US economy. Even with the current economy, he will probably lose in November 2020. If the economy tanks, e.g., due to a no-deal Brexit, he is far more likely to lose.

Call of the board: Sydney (part two)

A second, even closer look at the electoral lay of the land in the Sydney region at the May 18 federal election.

On reflection, my previous post, intended as the first in a series of “Call of the Board” posts reviewing in detail the result of the May 18 election, was deficient in two aspects. The first is that patterns in the results estimated by my demographic model were said to be “difficult to discern”, which can only have been because I didn’t look hard enough. In fact, the results provide evidence for remarkably strong incumbency effects. Of the 12 Liberals defending their seats in the Sydney area, all but Tony Abbott outperformed the modelled estimate of the Liberal two-party vote, by an average of 4.0%. Of the 15 Labor members, all but two (Julie Owens in Parramatta and Anne Stanley in Werriwa) outperformed the model, the average being 3.4%.

The other shortcoming of the post was that it did not, indeed, call the board – a now-abandoned ritual of election night broadcasting in which the results for each electorate were quickly reviewed in alphabetical order at the end of the night, so that nobody at home would feel left out. You can find this done for the Sydney seats over the fold, and it will be a feature of the Call of the Board series going forward.

Continue reading “Call of the board: Sydney (part two)”