Saturday, May 25, 2019

2019 Queensland Senate: Who Will Be Last When The Music Stops?

2019 Queensland Senate

Outgoing Senators: 2 LNP 2 ALP 1 Green, Fraser Anning

Seats won: 2 LNP (Paul Scarr, Susan Macdonald), 1 Labor (Nina Green)

Four-way fight for three seats with one to lose: Gerard Rennick (LNP), Chris Ketter (Labor), Malcolm Roberts (One Nation), Larissa Waters (Green)

Rennick and Roberts should win; last seat between Waters and Ketter is much too close to call,  but Greens seem just a little better placed
Final result probably won't be known until the button is pressed

Warning - Senate races are complex! This article has been rated Wonk Factor 4/5.

Donations welcome!

If you find my coverage useful please consider donating to support the large amount of time I spend working on this site.  Donations can be made by the Paypal button in the sidebar or email me via the address in my profile for my account details.  Please only donate if you are sure you can afford to do so.

In the 2016 Senate count, a double dissolution and lack of knowledge about voter behaviour in the new Senate system meant that races in most states were up in the air until all the ballot papers had been entered and The Button pushed. 

With the increased quota, it is much harder for micro-parties to compete in 2019, and barring upsets not currently on my radar it looks like 2019 has delivered seven fizzers and a cracker.  The cracker is in Queensland, where four parties with most but nowhere near a whole quota each will fight it out on the preferences of over a quota worth of micro-party preferences.    Labor risks the historic loss of its second seat in the state after a disastrous Queensland campaign.

Currently, votes are being gradually verified by data entry, a process that will take weeks, with an end result sometime well into June (I don't know the exact AEC timetable for it.)

As I start this article, the LNP leads on primaries for this race, as it has 2.765 quotas (a quota is one-seventh).  One Nation are next on 0.721, the Greens have 0.681, Labor has 1.604.  However not all votes are yet included in the count, and Ross Leedham, who has been following this closely, is projecting that the primary gap between the Greens and Labor should close a little bit.  The Greens will start the preference distribution with a lead over Labor, but it won't be a big one.

The micro-parties to be excluded are headed by Clive Palmer's United Australia Party (0.241 Q) and the rest include HEMP (.118Q) Katters Australian Party (.107Q), Animal Justice (.091), Fraser Anning Conservative Nationals (.090), Conservatives (.071), Labour DLP (.068), Shooters (.066), Liberal Democrats (.055) and Rise Up Australia (.051).  Various other micros have less than .05 Q.

The total vote for all preference sources is presently 1.229 quotas (17.6%).  However, this 17.6 points includes almost 11% (0.75 Q) for clearly right-wing parties, about 4% (.28 Q) for clearly left-wing parties, and not very much that is even arguably in the middle. Given that One Nation smacked the other three on the preferences of parties like KAP, Shooters, Rise Up and even the left-wing Sex/HEMP ticket last time, it seems a good chance that One Nation will be again a good performer here, even without Pauline Hanson herself on the ballot.  Even though neither FACN nor UAP preferenced One Nation at all, a lot of their supporters will probably preference One Nation anyway.

In terms of the flow between Labor and the Greens, the following are some estimated flows from the parties that ran in 2016.  These are based simply on the breakdown of above-the-line #2 flows that went to one of the four parties fighting for the final seat, and may not be all that representative, but should be better than nothing:

HEMP (as Sex/HEMP): Green 30.7%, Labor 20.6%
KAP: Green 3.8, Labor 23.5
Animal Justice: Green 39, Labor 26.2
Labour DLP: Green 24, Labor 35.6
Shooters: Green 3.9 Labor 16.6
LDP: Green 4.7 Labor 25.7
Rise Up: Green 6.8 Labor 7.0

Someone may want to do a more accurate 4CP-and-exhaust model from the 2016 actual ballot papers (it's a bit beyond my energy levels at the moment.) [EDIT: see update below]

If these flows are roughly accurate and repeated, and a 10% exhaust rate assumed, Labor will gain less than .03 quotas over the Greens from these preference sources, except if the flow is strengthened by one of the right-wing parties crossing quota first (which may not help much as a lot of votes that have been through LNP or PHON will exhaust).  One question is how much help Labor might expect from UAP, FACN and Conservatives (I'd suggest not much at all from at least the latter two).  Labor will also suffer to a small degree from leakage off their surplus and their minor candidates - this might cost them another .005 Q.

A further issue is that given the abysmal primary support of Labor in North Queensland, it could well be that Katter voters (for instance) preference the party more weakly than last time, and are more inclined to preference PHON, LNP and exhaust without bothering to preference Labor.

The UAP vote including their how-to-vote card (if it has any follow rate to speak of) offers potential deliverance for Labor.  The UAP's preferences flow at 2 to the LNP, at 3-5 to parties that will exit before it does, and at 6 to Labor.  If the LNP have been elected before UAP are excluded, and the card has even a say 15% follow rate, then that alone might be enough to catch up.  But I'm not sure the follow rate will be that high given the low follow rate for most HTV cards at the last election.  I also think a lot of UAP voters will preference One Nation.  Even so, if the LNP were over the line, it's quite likely Labor would do a lot better off UAP than the Greens.

This question of whether the LNP (and possibly PHON as well) might cross the line before UAP's exclusion or not is a potential tipping point for Labor's prospects if it's really really close.  On present numbers, the question here is the breakdown of preferences from all the other parties in a six-way split between LNP, One Nation, Greens, Labor, UAP and exhaust.  The LNP currently needs 23.8% of these votes to hit it before UAP is exhausted, and One Nation currently needs 28.2%.  (These numbers will change with more counting.)

Maybe both the right parties will get there by that stage, but maybe not.  If the LNP can cross quota before UAP are excluded then Labor gets a free swing at whatever break on the Greens it can get off the UAP votes, including the UAP-LNP votes, at full value.  But if the LNP need even a few UAP votes to get over, then that will give the LNP a surplus that will be almost all their own votes.  The reason for this is partly that most of Rennick's votes would be LNP votes to begin with, but also that a ridiculous system flaw called Unweighted Inclusive Gregory distortion would slightly expand the value of the LNP votes in the surplus and shrink all the others.  Any vote that has followed the LNP card will at this point go to exhaust.

Something that will help Labor if it appears is a pattern of right-wing voters voting through the card to put the Greens last.  We didn't get to see this last time because Waters was elected too early in the count, and I'm not sure how much of it there would be.

Scrutineering a race like this is almost impossible because of the need to sample minor party preferences accurately across a representative range of divisions.  I suspect the effort involved would dwarf the 15 hours I spent scrutineering to project the Tasmanian Senate race last time.  And even then it might be too close to call.  So my feeling is that we will not know the answer nor have an especially accurate handle on the answer until the button is pressed.  My opening offer is that the current projected Greens lead on primaries might just be enough, but this is complex, and I might be missing something important and obvious, or change my mind later.

If you want to have a go yourself, there's a Not-A-Poll in the Sidebar.  I'll be updating this post now and then with further thoughts, but not for changes in projected or actual primaries that don't significantly alter the overall picture.  You can follow those yourself on the AEC page, and also keep an eye on Ross Leedham's Twitter feed.  I'll be back on button press day (whenever that is) with what I hope will be detailed analysis of what went down in the preference distribution.

Update 26/5: Ross has kindly sent me four-party preferred (plus exhaust) figures for the micro-parties who contested in 2016.  The exhaust rate without reaching any of the competing parties is higher than I guessed above, at 18%.  If the overall preference flow is the same (ignoring changes in party makeup) then nobody will cross quota before the UAP exclusion, and Labor will gain about .057 Q on the Greens in total ignoring any possible flow-on from surpluses.  The actual Labor-vs-Green flows from recontesting parties with over .05 of a quota would actually help Labor by only .024 Q.  (In order the Green/Labor splits were: AJP 31.2/16.2, Sex/HEMP 25.8/17.5, LDLP 17.9/28.5, Shooters 6/14.3, LibDems 7.1/21.3, RiseUp 6.2/9.1, KAP 5.2/17.8).

The 4PP figures are consistent with the gap being difficult for Labor to close, especially given that their preference flows from some parties will worsen.  There are anecdotal suggestions in various directions regarding Chris Ketter possibly getting or missing BTL flows because of his position on same-sex marriage.  My suspicion is that there won't be much impact and what there is could cancel out.

Friday, May 24, 2019

Gladstone Rises Up: An Error In The 2013 Tasmanian Senate Count

There's apparently not all that much going on in the 2019 election postcount, where the only major dramas left at present appear to be which (probably) left party loses in the Queensland Senate and whether anyone can possibly avoid a recount in Macquarie.  When I compare it to 2016, I'm quite surprised at how busy I'm not.

This means I have time to post something curious I've been meaning to post for some time.  As is well known, the 2013 Senate count was not the Australian Electoral Commission's finest hour.  In Western Australia, the original count had a tipping point between two candidates, neither of whom could win, but the resolution of which determined the final two seats. The loss of 1370 ballot papers meant that it could not be determined who had won, and as a result the entire 2013 WA Senate election had to be voided and rerun in 2014.  This resulted in the resignations of the Electoral Commissioner and the Electoral Officer for Western Australia and major changes to the way ballot papers are handled.  The farce also contributed to the death of Group Ticket Voting at federal level. Under the system we have now the tipping point would have been irrelevant and the lost ballots may well not have affected the outcome.  Many other issues with the AEC's culture were identified in a review and many positive changes have been made.

I can now reveal that there was also a mistake in the count in Tasmania that year, and it got through every possible keeper and into the final results.  Earlier this year I was compiling some booth swings and I noticed that something weird happened in the Gladstone booth in Bass in the 2013 Senate count.  Gladstone is a small rural mining town in the state's far north-east corner that at least used to have a somewhat wild-west reputation.  The Labor-Green state government and its federally-backed forestry peace deal were not well taken in the area, or in the rest of Bass for that matter, and Andrew Nikolic (remember him?) recorded a 12-point 2PP swing in Gladstone and a primary vote of 57%.

But oddly, in the Senate in the Gladstone booth, the Liberals recorded only 7%.  And the Palmer United Party, who recorded over 8% in the Reps in Gladstone, could only manage two below-the-line votes (rather weird when BTLs were much rarer than ATLs in that election.)  Instead the spoils went to the far-right Rise Up Australia Party, who despite polling only 110 ATL and 19 BTLs in the entire rest of Bass, had somehow managed 82 ATLs and zero BTLs of the 166 formal Senate votes cast in Gladstone.

So perhaps something was driving this, a home-town candidate perhaps?  Well firstly, in that case, why were the Green and Labor votes more or less on a par with their House result and only the Liberal and PUP votes affected?  And also, how to explain that nobody out of these 82 voters voted below the line, a 1 in 472,831 chance based on the behaviour of the remaining Bass Rise Up Australia voters?

As it turns out, there wasn't any home-town effect.  The table below, showing the 2013 Tasmanian Senate groups and also their 2013 Reps and 2010 Senate percentages, indicates what the real problem was (click for larger clearer version)

Clearly, Rise Up Australia has been given what should be the Liberal above the line total.  The Liberals have been given Palmer United's above the line total.  Perhaps Palmer United have been given Rise Up Australia's above the line total, or perhaps there was some other error (two zeros instead of one, or two ones instead of one) in the parties between PUP and the Greens - by the time the list reaches the Greens, it's back on course.

The old Senate system only required the 1-only ATLs to be counted and the numbers of ATLs checked and entered, whereas under the new system every ballot has to be computer-scanned and then confirmed as correct by two human operators.  If the human operators disagree or raise an issue it is raised to a higher level.  A mistake of this kind therefore wouldn't occur.  But this mistake could have happened in 2013 simply if there was some kind of error in data recording and checking processes.  And we know that in 2013 there were a lot of mistakes.

This error then made it all the way to the final count without being caught, and Rise Up Australia are still credited on the AEC website with 82 votes they did not get, costing the Liberal Party about $200 in public funding.

What difference did this error make?

The 2013 Senate count for Tasmania was in fact very close, because of tipping points created by Group Ticket Voting (though it was ultimately not quite as close as the 2016 final seat between the Greens and One Nation).

Jacqui Lambie (then Palmer United Party) won the final seat.  But her victory depended on three tipping points:

1.  The Sex Party's Robbie Swan outlasted Family First's Peter Madden by 821 votes.  Had this not happened, Madden would have won the seat, fuelled by a perfect preference snowball thanks to foolish preferencing decisions by Labor and the Greens especially.

2. Labor's third candidate outlasted the Sex Party's Robbie Swan by 244 votes.  Had this not happened, Swan would have won the seat, even though he doesn't live here.

3. Lambie outlasted the Liberal Democrats' Clinton Mead by 1276 votes (he doesn't live here either).  Had this not happened Lambie would have been excluded, but the LDP would also have lost, and the Liberals' Sally Chandler would have won the seat.

The main error was the Liberals being short 82 votes that were instead credited to Rise Up Australia.  These 82 votes then helped Rise Up undeservedly outlast the Country Alliance, but that made no difference since in either case both would be eliminated.

The 82 misplaced votes then flowed to Democratic Labour.  On their exclusion, they flowed to Australian Independents. On their exclusion, they flowed to Family First at preference 11.  Finally, on Family First's exclusion, they flowed to the Liberals at preference 23, and now everything was back on track.

So ultimately, although these votes shrank the margin at the second-closest exclusion point in the count, they weren't enough to determine the outcome.  And presumably had they been enough, somebody preparing a challenge would have noticed it somewhere (but who knows).

I've devoted a lot of time to arguing the case for the new Senate system in terms of the way it gives voters freedom to effectively direct their own preferences, and ends the election of parties from tiny vote shares through dodgy preference deals.  But here's another very tiny argument in its favour - it forces the checking of every specific ballot, rather than just the adding and recording of totals in which an error of this kind might happen.

PS Nick Casmirri has another such case - 368 Labor votes to Group X in NSW Senate 2010.

Thursday, May 23, 2019

2019 Federal Election Postcount: Mallee

MALLEE (Nat vs ALP - 19.8%)
Anne Webster (Nat) leads, but it is very unclear which of four candidates will make the final two
It is known that Webster easily beats Labor if Labor makes the final two.  I expect her to beat any other opponent too.

In the election leadup I had my eyes on the Victorian seat of Mallee as the most likely to deliver an absolute mess in the postcount.  On election night it seemed to be a bit of a fizzer because none of the independents made 10% in their own right, making it clear that the Coalition was headed for victory.  However the Mallee count has thrown up some interesting complications, and there is a theory doing the rounds that the Liberal candidate Serge Petrovich might be able to defeat Webster if he can make the final two.  I am unconvinced about this theory, firstly because I'm doubting he will make the final two, and secondly because even if he does a rather strong preference flow is needed to get him over the line. I don't think that will happen, Labor HTV card notwithstanding, but in the meantime there's a possibility Mallee will create electoral history.  Never (thanks to Malcolm Baalman for this) has a candidate who finished fourth or worse on primaries in a federal seat reached the final two, and it is possible that this could yet happen.

Mallee is a rural/regional seat in the north-west of Victoria that includes the cities of Mildura and Horsham.  Four MPs, all of them Nationals, have held the seat since its creation in 1949.  The most recent was two-term MP Andrew Broad, who was brought down by a "sugar daddy" scandal in 2018 and did not recontest the seat.  Already reeling from the tabloid misadventures of Barnaby Joyce, the Nationals sensibly preselected Dr Anne Webster, a Mildura-based sociologist.

Following the success of Voices for Indi and the destruction of the Nationals' rural vote in the NSW state election, there was a lot of interest in independent campaigns for Mallee, the two most prominent of which have been by Mildura deputy mayor Jason Modica and former Yarriambiack (I had to copy and paste that) Shire mayor Ray Kingston.  However indies on either side of the Murray did not fare as well as the Shooters had in NSW, despite strong reports of a feel on the ground for change in Farrer.  In Mallee the problem was always that there was no obvious single indie contender who could concentrate the vote.

Mallee attracted the largest field in the nation, 13 candidates.  This is how the primaries presently stand (click for larger slightly clearer version):

It's a long way to 50% but somebody's got to get there!  We also know that the two-party preferred is currently 66.73% to Webster, meaning that she is currently getting 67.7% of preferences ahead of Labor.  Probably she is getting nearly all the Liberal preferences and about 55% of the rest.

Obviously Labor do not win if the final two are Nationals vs Labor.  Depending on the strength of flow between the independents, and to the independents over Labor from the 15% of minor party votes below them, it's possible either Modica or Kingston (I understand more likely Modica) can get over Labor into third.  However the flows between the indies won't be anywhere near 100% so it's hard to be sure that this will happen.  If one of the indies can get into third, Labor preferences should easily push them into second ahead of Petrovich.

If an indie does make the final two, let's be conservative with the Petrovich flow and say only 85% of it goes to Coalition partner Webster ahead of the indie.  In this case, Webster would only need 13% of preferences to beat the indie and would obviously get that easily (especially with UAP and Shooters in the mix).  Even if there is a remarkably weak flow from Petrovich, eg 75%, then the 83% of the rest the indie would need to win is still not realistic.  So I cannot see how the indies could win; their primary vote is simply too low.

The most interesting final two would be a within-Coalition contest between Webster and Petrovich.  A within-Coalition final two happened in 2013 but in that case the Coalition partners had 66% of the primary vote, rather than 47.3% ! Labor has preferenced the Liberals ahead of the Nationals in Mallee, but Labor voters are not very obedient when it comes to being told to preference the Liberals.  In 1993 according to the ABC over 40% of Labor voters ignored such a card and preferenced the Nationals, and I can find no three-cornered contest recently where Labor preferences have broken to the Liberals even to the extent of 60-40.  If Petrovich doesn't get the preferences from Labor I greatly doubt he'll get them from the rest, especially as Webster is a local while Petrovich isn't.  Thus I am sceptical that Petrovich will gain more than a few percent (if even that) off the Labor votes, but I'm open to being convinced otherwise if anyone has very rigorous scrutineering data.

All I've heard regarding scrutineering so far is comments that preferences are going all over the place irrespective of cards, and that is not a surprising thing in a contest where various candidates have their own home town and their own team of people they know, and where voter understanding of the region's politics would be high.  I'm grateful to Michael DiFabrizio (Sunraysia Daily) for the remarkable information that the top six candidates in the count have all topped at least one of the division's 100 polling booths on primaries!  (See graphical representation in his article.  Is this an Australian record?) Scrutineering this one would be very challenging because of the question about exclusion order and what I would expect to be wild variations in flow across different booths.

As for the future direction, I'm not sure we'll get anything significant before the throw of preferences which I expect could happen the week after next.  In theory the AEC could realign the count, but they couldn't be sure that the final two actually won't be National vs Labor, and worse still, they run the risk that if they do realign it, they might get the wrong opponent again.

Mildura looks more like a curiosity than a contest, but I thought I should cover it in detail anyway.  Barring further major developments, I won't update this post much until the preference distribution is finished.

Tuesday, May 21, 2019

The Miracle Is Over: The 2019 Australian Federal Election Poll Fail

Nice 2PP.  Shame it's for the other side ...

"I have always believed in miracles" said re-elected Prime Minister Scott Morrison very late on Saturday night.  But many (not all) of us who study national Australian polls and use them to try to forecast elections have believed in a miracle for one election too many.  The reason we believed in this miracle was that it kept delivering.  While polls failed to forecast Brexit, Trump and two UK elections in a row (among other high profile failures) Australian national polls continued to churn out highly accurate final results.  The two-party preferred results in final Newspolls from 2007 to 2016 are an example of this: 52 (result 52.7), 50.2 (result 50.1), 54 (result 53.5), 50.5 (result 50.4).  

Predicting federal elections pretty accurately has long been as simple as aggregating the polls, adjusting for obvious house effects and personal votes, applying probability models (not just the simple pendulum) and off you go; you generally won't be more than 5-6 seats wrong on the totals.  While overseas observers like Nate Silver pour scorn on our little polling failure as a modest example of the genre and blast our media for failing to anticipate it, they do so apparently unfamiliar with just how good our national polling has been since the mid-1980s compared to polling overseas.  As a predictor of final results, the aggregation of at least the final polls has survived the decline of landlines, volatile campaigns following leadership changes or major events, suspected preferencing shifts that frequently barely appeared, herding with the finish line in sight, and come up trumps many elections in a row.  This has been put down to many things, not least that compulsory voting makes polling easier by removing the problem of trying to work out who will actually vote (another possibility is the quality of our public demographic data).  But perhaps it was just lucky.

For sure, the same success has not been seen in individual seat polling (extremely unreliable for many possible reasons), and state polling hasn't been great lately, with a 3.3 point 2PP blowout in the Victorian state government's landslide re-election.  And there have been misadventures in the past, like Nielsen's 57-43 to Labor in 2007 (caused by doing their final poll too early) and Morgan's 54.5 to 45.5 to Labor in 2001 (which lost them their contract with the Bulletin).  But even so, these tended to cancel out errors by other pollsters, meaning that aggregation of the final polls hasn't been much more than a point wrong since 1993.  Even in 1993, contrary to popular myth, the final polls probably should have been taken as pointing to a slightly more likely than not Keating victory.  We have to go back to 1980 for the last "polling disaster" case (some background in here) where polls clearly predicted the wrong winner.  And at the last few elections, national polls had been excellent, if a little on the high side for Labor in 2010.  It seemed some lessons of some earlier failures had been learned.

And so, when warning signs appeared, in the form of both the Coalition moving within historical striking distance, and then a ridiculous level of herding first called out in an excellent post by Mark the Ballot, I went through the motions of warning that there was a realistic if slim chance that the polls were all baloney. Based on historic projections, there was maybe a 25% chance that the Coalition would win anyway (a la Trump vs Clinton).  But after so often flogging the poll failure risk horse and having my fears proven groundless, my heart wasn't totally in it.  After seeing blowouts compared with final polling in five of the last six state elections around the country, and in seat polling where clusters of 50-50s and 51-49s were often followed by lopsided margins, it looked more likely that Labor would win by more against a government that nine months ago had set itself on fire.  To the extent that this was a widespread view (especially after the death of Bob Hawke) it was another incidence of Nate's First Rule: polls can be wrong, but trying to second-guess which way they will be wrong if so nearly always backfires.  

A Failure In Two Parts

To outline the nature of the failure briefly, especially for unfamiliar audiences overseas,  this is how the final polled primary votes and two-party preferred (2PP) compare with the current counting towards the final result. 

(* Average excludes Essential, which did not provide a breakdown for UAP.)  

The primary for the Liberal-National government was underestimated by about three points and the opposition Labor Party was overestimated by two points.  Errors on the other parties were minor.  The two-party preferred (2PP) figure on the far right hand side, however, is the key figure in Australian polling, because which of the Liberal-National and Labor parties wins the nationwide two-party-preferred vote will usually form government (with some exceptions when it's very close).  The 51.6% is my estimate based on swings in the ABC's election night projections; currently the live count is at 50.9% but this live count excludes several seats where the Government and Opposition are not the final two candidates, or where they were wrongly expected not to be.   [Update 23 May: It is looking like my estimate was conservative; Antony Green now estimates 52.1.]

Depending on exactly where the 2PP ends up, a small part of the circa 3-3.5 point miss on the 2PP may be caused by modelling error regarding how the preferences of the minor parties would flow.  Nearly all of it, however, has been caused by getting the primary vote for the major parties wrong.  (The exception is the Ipsos poll which has a long-unfixed habit of getting its Green primaries a few points too high, and also tends to get Labor low compared to others - which cancels out as 82% of Green preferences flow to Labor.)

So that's the first part of the error - that, on average, parties had the 2PP wrong by about three points.  This by itself might be dismissed as part of the "house effect" of polling at this election.  But there is more: the final seventeen published polls (and the Galaxy exit poll as well) all had a rounded 2PP for the government of 48, 48.5 or 49.  The sample sizes of these polls varied from several hundred to around 3000, so if the average had really been around 48.5 the smaller polls would have had about a one in three chance of landing in this band randomly, while the larger polls would have done so about 60% of the time.  The chance of all seventeen doing so - if they are independent and purely random samples - is a little under 1 in 200,000.  But in fact polls aren't purely random and the use of weighting in the polls should increase their margin of error and make the streak even less likely.  As Professor Brian Schmidt pointed out in the Guardian, the mathematics do not lie.  Some of the polls were not pure samples, some of the polls were not independent of the same pollster's other polls, or some of the polls were not independent of each other. 

The failure is amplified because not a single poll in the entire term of government showed the government leading, with the exception of five obviously dodgy respondent-allocated 2PPs in the short-lived YouGov-Fifty Acres series.  The government hadn't even tied a poll, except a single Ipsos on respondent preferences, since 2016.  The old saw about the only poll counting being the one on election day held true, except that in Australia now, election day goes for three weeks, robbing the pollsters of the usual excuse for this sort of thing (that people simply changed their minds.)

Is this like Trump or Brexit?

Recent major failings of polling like Trump and Brexit have created an impression that polling is getting much worse, when actually, worldwide, it isn't - it's as mediocre as it always was.  It just happens that there have been failures on some particularly momentous and close contests that have cast polling in a bad light.

But the failure that has happened here is actually worse than the Trump failure.  In the Trump case, national polls were actually quite accurate - they projected that Donald Trump would lose the popular vote, which he did, though not by quite as much as they projected.  It's the equivalent of a 1 point 2PP failure in Australia, which in this instance would probably have seen Bill Shorten in the Lodge but without a floor majority. The main failure in the USA was in the polling in a few particular states crucial to Trump's win in the Electoral College.

It is also slightly worse than Brexit.  The average error in the Brexit case was very similar, but the polls were not herded - two pollsters' final polls had No winning, albeit by less than it did.

Why were the polls, on average, wrong?  

One of the problems in saying why the polls were, on average, so wrong is that Australian pollsters, especially YouGov-Galaxy which also administers Newspoll, don't tell us very much about how they are doing their polling.  This frustrating opacity is a big contrast to many UK pollsters whose released results come with lengthy and detailed reports (example).  For instance, we know that when Galaxy took over Newspoll it commenced augmenting phone polling with online samples, but the breakdown of the two methods isn't published.  Another example is that in late 2017 the pollster changed the way it dealt with One Nation preferences in constructing its 2PP estimate.  The change was well justified, but the pollster did not make it publicly known until psephologists had gradually detected the issue over the coming five months, during which time the Australian had continued to claim that the poll was using last-election preferences.

Getting accurate samples in polling is increasingly difficult.  No major Australian pollster still uses purely landline-based sampling, nor has since 2015, but one still often sees claims that they do.  Nowhere near everyone has a landline, answers it, or takes a recorded-voice call.  Live phone calls are expensive and prone to social-desirability bias (as is face-to-face polling), a possible source of Ipsos' constant inflation of the Greens vote.  Pollsters lack access to a complete database of mobile phone numbers, so some people can be reached by mobile phone and some can't.  Online polling is another solution, but not everybody likes spending their time filling out surveys on a computer for trivial returns.  This particular failure has cut across all of these polling modes.

Here is one explanation that has been offered that is definitely false:

* Margin of error: People casually familiar with margin of error are claiming that the failure was within the margin of error of the polls.  But margin of error applies to the results of a single poll, not a string of them.  One poll might be wrong at the outer edges of its margin of error, but if even two polls in a row by the same company do this in the same direction then there is already a problem.  (This is a variant of error 3 in my list of margin of error myths).  Also the failure was outside the margin of error of the largest polls taken.  The 17 clustered polls collectively had a sample of about 23,600 on which the margin of error would be 0.6%.  Four or five times that margin puts you nine standard deviations from the expected sample if you were correct - in other words, extremely improbable.

Here are some explanations that have been advanced that in my view are non-starters:

* Late swing: the idea here is that those making up their minds on the day swung to the Coalition but made that decision too late to be included in the sample.  The problem with this is that prepoll voting was going on more or less throughout the period of these wrong polls, and prepolls have actually showed a greater swing to the Coalition (2 points 2PP) than election-day booth voting (0.8 points) - figures as of Sunday morning.  Of course, the prepoll voting mix has changed a lot with the massive increase in prepolling but even so to expect that to make 4 points of difference the other way seems a bit much.

* Rolling late swing: the idea here is that some voters were intending to vote Labor but chickened out because of scare campaigns on the way to the ballot box (or post box) whether they voted before polling day or on it and voted Coalition instead.  This theory avoids the problem with the prepoll and booth swings mentioned above.  However if this was the case the polls should have shifted to the Coalition through the campaign by about a point as these voters reported back their actual vote. This didn't happen.  All else being equal this should have also led to a larger gap than last time between the votes of those who had already voted and those voting on the day in the few polls that provided a breakdown of this.  In the one I saw, Ipsos, it didn't.

* "Shy Tory effect": the idea here is that conservative voters are afraid of telling the pollster they vote Liberal because they think the interviewer will think they are a bigot.  But unless a respondent is very paranoid, they're hardly likely to care about admitting they vote Coalition to a robopoll or an online survey.  Also, there is no systematic skew to Labor in recent Australian polling - the Victorian election with its 3.3 point skew to Coalition in final polls being an example.  If ever Tories had a campaign to be shy about, that one, with its "African Gangs" beatup, would surely be it.

Here are some explanations that in my view are plausible (at least as partial explanations):

* As with overseas polling failures, pollsters may have been oversampling voters who are politically engaged or highly educated (often the same thing).

* Connected to the above, some pollsters may have been underweighting (or failing to set quotas for) other important information.  There is too little information available about what Australian pollsters actually adjust their samples for, but what is available often refers just to age, gender and location.  Pollsters should not be expected to declare their exact formulae, but a list of everything a pollster weights for would be useful in picking cases where the polls might be missing something.

* There could well have been an unusual skew to the Coalition among politically disengaged voters who are basically unreachable by any method by pollsters, but are still required to vote (a way in which compulsory voting can make polling more difficult not less).  Pollsters just have to trust that the unreachables behave like those they could reach with the same demographics

I may add others as I see them mentioned. 

Herding, smoothing etc

Even if some explanation can be found for the average skew of the polls at this election, it doesn't explain the run of seventeen polls in a row with the same wrong result.  This sort of thing is often referred to in polling studies as herding.  Nobody wants to be the lone pollster with the completely wrong result while all the others are right (a la Morgan in 2001) so the myriad subjective choices that can be made in polling may result in struggling pollsters being more likely to get results similar to better pollsters.  But if the normally good pollsters then make mistakes, the whole herd is dragged off course.  In reducing the risk of being the outlying pollster, herding pollsters increase the risk that everyone is wrong.  An appearance of herding has been common at recent Australian elections including the 2016 federal election, and sometimes this takes the form of one or more pollsters who have had different results to Newspoll early in a campaign saying the same thing as Newspoll at the end.  

It's not clear that there was actually any herding this election.  An alternative possibility is that pollsters could be self-herding, and just coincidentally happened to do so around the same range of values as each other.  This could happen if pollsters were using some form of unpublished smoothing method to stop their poll from bouncing around and producing rogue results.  Galaxy has always had uncanny stability, and when it took over Newspoll we started seeing such things as, at one stage, the same Newspoll 2PP six polls in a row.  Essential has also in the past been prone to get "stuck", but seems to have behaved more naturally recently.  Galaxy has also displayed strange behaviour in seat polling at both the last two elections - when it repeat-polls a seat, the difference between the two polls is on average little over a point on 2PP, about half what it should be randomly even if nothing has happened in the campaign in that seat.  Also in 2016, Galaxy's seat polls had strangely underdispersed swings.

An article by former Nielsen pollster John Stirton raised concerns about the strange lack of volatility in Newspoll since Galaxy took over but the claim has been neither confirmed nor denied to my knowledge.  (There's not necessarily anything wrong with it, but if a poll isn't a pure poll, this should be declared.)

It's worth noting that the primary votes do not display the same level of herding as the 2PP, for reasons including Ipsos' trademark inflation of the Greens vote, and also Essential having One Nation too high.  But it's largely irrelevant because the 2PP is the figure that is used to forecast results and that would be the obvious concern for any pollster worried about their reputation.  In theory, herding the 2PP could also be a factor in decisions about how to allocate preferences from minor parties.  The four different pollsters (counting Galaxy and Newspoll as one) applied four different preferencing methods.  Galaxy is known to have changed theirs at least twice during the term, while what Ipsos did is unclear based on published information.

Were the seat polls better?

David Briggs of YouGov-Galaxy has referred to the many seat polls that Galaxy issued in the final week as pointing to the correct picture and has complained that when these results were released they were scoffed at because they didn't fit the narrative.

I'm aware of 16 such results from the final week.  In fact, on current figures (which won't change much) they were on average 3.2 points better for Labor on 2PP than what actually happened, so they were just as bad as the national polls in that regard.  Across the whole campaign, the average 2PP/2CP error per seat for the 21 Coalition-vs-Labor/Greens seats polled by Galaxy/Newspoll (some seats polled more than once) was about 2.7 points.   It's true that all these polls showed only two wins for Labor in Coalition seats, both of which they actually won.  It's also true that these polls showed Labor behind on average in two of their own seats, both of which they lost.  And of the 18 such seats where these polls showed one side ahead on average, they picked the right winner in 16, with two in doubt where if the wrong side gets up, it will be by a very slim margin.  So that's impressive.

But these seat polls did not reveal anything to cast doubt on the picture in the national polls, because they had the same swing as the national polls.  In the final week Galaxy had Labor at 50-50 in four Coalition seats and 49-51 in five Coalition seats.  If these were broadly accurate samples Labor would have won three or four of these seats by chance, but Labor has actually won none, has lost one of its own seats (Herbert) where the final Galaxy had it at 50-50, and is a tossup in another (Macquarie) that Galaxy had it leading in 53-47.  Moreover, only one of the nine 50-50 or 49-51 seats has ended up even remotely close.  There were massive errors in two Queensland seats (Labor held Herbert and LNP held Forde) that Galaxy had as 50-50; both are now showing the LNP over 57%.

The final week polls were also, yet again, somewhat under-dispersed for their small sample size, with a standard deviation of around two points both on the released 2PPs and the predicted swing from the actual election, compared to three points in the actual results and four points in the actual swings.  (About three points is normal.)  This repeats an issue that was seen in the last election, and raises the question of how on earth Galaxy keep getting results at both national and seat poll level that are less variable than they should be by chance.

Were the internal polls better?

When there is an unexpected result one usually hears that the losing party had this result in its internal polling all along, released only to a select list of senior people.  This is even though this is invariably news compared to what little could be gleaned about internal polls from pre-election spoonfeeding of the media and, in rare cases, leaks.  As a general rule, any claim about internal polling that is politically convenient for the source making it should be ignored unless they publish the full details.

In this case, we have got a rare glimpse under the hood of Labor's internal tracking polling, which reveals that Labor chose to contract YouGov-Galaxy for their internals even while the same company was also providing Newspoll and Galaxy polls to the Murdoch press (which was editorially hostile to Labor during the campaign).  This is quite a surprise in itself, especially following the concerns about another pollster, uComms, having union ties that were not well known to clients.  It raises the question of whether Australia has enough good pollsters to be able to avoid hiring someone who might be conflicted.  (There's an illuminating history of other such issues by Murray Goot. It is known to me that ReachTEL often declines contracts to avoid having a conflict in the market.) YouGov-Galaxy have pointed out in the article that the two contracts were "siloed", meaning that the polling teams used for Newspoll and for the Labor research were entirely separate.

The tracking poll covered 20 seats with an average pre-election 2PP of 51.4% to Coalition.  As an average through the campaign, the tracking poll showed Labor getting a 0.8% swing, enough to probably govern in minority but probably not with a majority.  At the end it kicked up to a 1.4% swing, probably enough for a slim majority government and completely consistent with the late reports from an unnamed Labor source that they expected to win about 77-78 seats (but hoped for more).

The interesting thing is that the tracking poll dipped below easy-win territory (51 and above) after only five days of the campaign, and at some points even had Labor losing based on the swing in these marginals.  If Labor were taking this polling seriously they should have noted that it was telling a different story to the national polls early in the campaign, and should have realised that they needed to cut back on risks.  Labor's tracking poll was as wrong as everything else at the end, but it wasn't always so.

There are, however, suggestions that the Liberal Crosby-Textor internal polling was rather accurate.  Again, we need to see detailed figures on this.

Where to from here?

I mentioned before that Australia's last fully-fledged polling failure in a national election was in 1980.  Commenting on that case in 1983, David Butler wrote "Nowhere in the world has a debacle for the polls diminished their use in subsequent elections".  But I wonder.  The frequency and diversity of Australian polling (especially at state level) had already declined sharply in recent years compared to the period 2013-5, though Australian pollsters had done little to deserve that.  This downturn was sometimes attributed to the Donald Trump upset casting pollsters in an unflattering light, though I think the increased dripfeeding of journalists with stories culled from internal/commissioned polling also has a bit to do with it.  Where does an already weakened industry go from here?

Obviously pollsters will have to review to try to determine what went wrong.  They need to make this review broader than just in-house and bring in independent experts (retired former pollsters, perhaps overseas pollsters, professionals from related industries) to provide a detached view of the fiasco.  The reviews must seek to explain not only the error on the averages but also why the polls clustered in the last weeks of the campaign, resulting in no pollsters releasing Coalition win results that, even if apparently stray, would have at least increased awareness that the result wasn't completely in the bag.  There is a stark contrast between the last four elections and the elections before them at which final results from different pollsters varied widely.

But I also think the Australian polling industry, and also the media who use polls, need to try to win back the trust of the public by being far more upfront about what polls are doing and how they are doing it.  This extends both to the public polls, which need to be far more comprehensively reported than they are now, and also to commissioned seat polls.   Seat poll results spoonfed to media by left-wing groups in the leadup to this election were often extremely left-skewed (much worse than the national polls), with the exception of GetUp!'s polling of former Prime Minister Tony Abbott's impending defeat in Warringah (which was quite accurate). 

The unhealthy synergy of activists and parties spoonfeeding poll-shaped objects to reporters, thus giving them free content in return for what is nearly always uncritical coverage of deeply dubious issues polling questions, must end.  Pollsters that wish to be reputable need to commit to standards alongside those of the British Polling Council, so that any poll reported in public (whoever commissions it) is available for public discussion of its methods and full results.  Ideally, media should refuse to report polls unless the source is willing to release their full results, but that's probably a bit much to hope for.  (It's encouraging that at this election some journalists were refusing to accept seat polls with very small sample sizes, but there's a long way to go.)

As serious as the issues with the public polling are, they pale besides the rottenness and shoddiness of so much of the commissioned polling being reported on, which affects the whole industry's reputation.  When I have been able to get hold of results for commissioned polls, I have often seen errors I did not expect - naming some candidates but not others, absurd breakdowns by gender, previous-vote reporting that demonstrates a biased sample, spreadsheet errors and so on.

Polling in Australia attracts an incredible volume of daft conspiracy theories.  Any Newspoll showing a benign result for the Coalition is dismissed as a Murdoch plot, even if independent polls are getting the same result or worse.  Discredited tropes created by the famous author and clueless polling crank Bob Ellis still continue to spread on Twitter years after his death.  The stupidity and evidence-aversion level of a lot of this stuff is quite staggering and pollsters could hardly be blamed if it was giving them a dose of Stockholm Syndrome.  But at the same time, the pollsters are partly responsible for the ignorance that breeds in a vacuum, because they've given the public the mushroom treatment for so long, and then they seem surprised when opinionistas and "Twitter experts" make outdated comments about how their polls work.   How they cannot see that this is self-inflicted is beyond me.

I hope this failure will lead to a better and more open polling industry.  

Sunday, May 19, 2019

2019 Senate Postcount: Main Thread

Carry-Over from 2016 Senate: Coalition 16 Labor 13 Green 3 CA 2 AC 1 PHON 1
Expected 2019: Coalition 19 Labor 13-14 Green 5-6 PHON 1 Lambie 1

Currently Coalition is likely to hold 35 seats and need two of Centre Alliance, One Nation and (Bernardi+Lambie) to pass bills.

Donations welcome!

If you find my coverage useful please consider donating to support the large amount of time I spend working on this site.  Donations can be made by the Paypal button in the sidebar or email me via the address in my profile for my account details.  Please only donate if you are sure you can afford to do so.

Welcome to my main 2019 Senate postcount thread.  This will contain outlooks for each state which I will update.  I may move any state that I do any very complex modelling on to a different thread.  In the case of Tasmania, this is only likely to happen if Lisa Singh's below the line vote starts projecting to such a level as to create a serious contest between her and Catryna Bilyk.

Some states will receive much higher detail level than others on account of the competitiveness of races.  Where races appear uncompetitive I won't be posting frequent updates.

On this page, a quota is c. 14.28%.  A candidate will be elected if they reach quota, but in the case of the last seat or two may not need to get quota to be elected.  Votes are initially counted by party (whether above the line or below) and are then gradually sorted from "unapportioned" into ticket votes and candidate votes (BTLs).  This is a long and messy process.  Initially candidate below-the-lines will be much lower than where they eventually get to so please don't say "oh so-and-so is on only 6 votes" until the unroll is finished.  As of Tuesday, the unroll for Queensland has just started.

DISCLAIMER: All assessments are provisional unless clearly stated as definite calls.  While these counts are not as difficult to project as Group Ticket counts in Victoria and WA, unexpected outcomes may nonetheless occur and I have had limited time to closely study all possible issues in the counts.

IMPORTANT: Please be aware that how-to-vote cards are weakly followed in the Senate and as a result preference flows are generally not very strong.  Do not use how-to-vote cards for modelling Senate counts unless you are familiar with the follow rates for each party's card from the 2016 election.  In most cases, a primary vote lead of 2% will not be overturned by preferences in this system.  A possible exception is in ACT.  


New South Wales

(Seeking re-election: 2 Coalition 1 Labor 1 Green 1 UAP 1 Lib Dem)
Expected result: 3 Coalition 2 Labor 1 Green

Intro (Sunday):

NSW is currently 53% counted to primaries.  At a quick look the counts tend to include normal booths but often not prepoll voting centres.  These are the leaders:

Liberal/National 2.65 quotas
Labor 2.14 quotas
Greens 0.66 quotas
One Nation 0.33 quotas
Shooters 0.17 quotas
HEMP 0.15 quotas

At this stage the NSW result looks straightforward with a 4.5% gap between the leading contenders for the last two seats and anybody else, which should be much too large to bridge on preferences even if the numbers settle slightly with more counting.  

Assuming it stays like this the only interest is in the Jim Molan below the line votes as he seeks to dislodge the Nationals' Perin Davey from #3 and win the third Coalition seat.  As a BTL-only candidate Molan cannot receive above the line preferences unless Davey is excluded.  NSW doesn't typically have a lot of BTLs for other candidates and so Molan would need a large share of the Coalition's leftovers in order to exclude Davey before she could gain too much on him from other parties.  He might need something like half a quota, but I'll model this exactly if he even looks like getting half the Coalition's excess.  Across a state as large as NSW and with such a low BTL rate this seems as difficult as it always has, but I don't have any data on it.  There will be some leakage when Molan is (presumably) excluded but a lot would have to change for it to place the seat at risk.

Thursday 1:50 There are some early BTLs in which Molan is recording many times the BTL rate for both major party ticket leaders, by ratios of 8-10.  These are probably from very unrepresentative booths (haven't checked where, as it isn't easy to do) but it is interesting that he is doing this anywhere.  In 2016 both major party ticket leaders got nearly 1% on primaries but in this case they are non-incumbents and may very well get a lot less.

7:40 Some more BTLs in and the impression so far is that these are still from unrepresentative urban booths.  Molan may pick up a few percent in the end but there is no reason off these early figures to believe that he can win.

Friday 10:30: Ross Leedham who has a far better handle on this one than me is projecting Molan to land between 1 and 4% (nowhere near enough).



(Seeking re-election: 2 Coalition 2 Labor 1 Green and Derryn Hinch)
Expected result: 3 Coalition 2 Labor 1 Green 

Intro (Sunday): 

The Victorian count is relatively slow at 43% counted.  The following are leaders:

Coalition 2.38 quotas
Labor 2.27
Greens 0.82
One Nation 0.192
Hinch Justice 0.186
Labour DLP 0.175
UAP 0.162
Shooters 0.123
HEMP 0.112
Lib Dems 0.101

The first five seats are obvious but the last seat is not yet clear, as nobody fighting for it has very much.  The Coalition may move more clear of the field with the counting of prepolls and postals, or Labor's position might improve (I haven't looked at where these votes are from).  The problem for Labor on the current numbers is that the left-wing micro-party preferences will be splitting between them and the Greens (to the extent they follow ideology at all rather than spraying and exhausting) and the other problem is that there aren't all that many left-wing preferences,  though Labour DLP will help a bit.  However, if Derryn Hinch can get ahead of Labor #3 and One Nation he may be able to draw preferences from everywhere and mount some kind of challenge if the Coalition isn't too far ahead.  Hinch's vote has really collapsed since the last election and it seems absurd to even be considering his prospects when he hasn't polled 3% in a half-Senate race, but it does highlight the possibilities of the system in Victoria.  Still most likely a Coalition win at this stage.

Tuesday 6 pm: The Coalition's position has improved to 2.43 quotas.

Wednesday 6 pm: The Coalition's position has improved to 2.47 quotas.  The Coalition is now 4% ahead of Hinch which seems like it should be too much to be bridged.  Hinch could pick up an extra boost if the Labor exclusion tips the Greens over quota with a modest surplus but I am only talking hundredths of a quota here.  At the moment I think the Coalition's lead should be too much but I am yet to look in detail. Certainly the flow of Coalition preferences to Hinch (mentioned by Glenn Druery in the Age) is irrelevant.



(Seeking re-election: 2 LNP 2 ALP 1 Green + Fraser Anning)
Expected result: 2 Coalition 1 Labor + 4 way race for 3 spots
Three-way race between Greens (leading), One Nation, Labor and LNP - one to lose (probably Labor or Greens)

Queensland Senate has been moved to its own thread.

Intro (Sunday):

The Queensland count is not very advanced at 43%, but we know that the LNP thumped Labor in the House of Reps by a bruising 57-43 (compared to polls around 51-49 or 50-50).  These are the primaries:

LNP 2.55 quotas
Labor 1.65
Greens 0.82
One Nation 0.69
UAP 0.23
HEMP 0.14
KAP 0.116
AJP 0.095
FACN 0.094

At this early stage we have the Greens 0.13 quotas ahead of One Nation and Labor 0.10 quotas ahead of the LNP.  Labor are also not far behind One Nation.  Whoever is last out of these four after preferences loses, but the LNP (and also One Nation) should do well out of UAP who are the next biggest source of goodies, Clive Palmer's tilt having clearly failed.  The 2PP from the Reps implies that these primaries might be understating the conservatives at this early stage, and bear in mind that certain North Queensland seats (especially Flynn) tend to have big turnarounds in the LNP's favour on postals.  It is also especially notable that at this stage the urban seats are slightly over-represented in the count compared to the far-northern regionals, so the count appears skewed to the left (so Labor's miserable 24% primary might be an overestimate!)

So it may well be that Labor is in trouble here, but we'll have to see where the parties stand when the primary count is more complete, as maybe the urban skew issue will affect the Greens more than them.  One Nation's Malcolm Roberts is going relatively well so far and based on the skew issue might do even better.

Tuesday 5:30 pm: Ross Leedham has been unskewing the primary vote count and has recently posted this projection for the final seat (my quota numbers in brackets afterwards)

Greens - 10.10% (.707)
ON - 9.97% (.698)
ALP - 9.11% (.638)
LNP - 8.87% (.621)

It's possible we won't know the loser of this race for a long time but given the right-wing skew of the preference base it could be that One Nation and the LNP might both win leaving Labor vs Greens for the final seat (and it's not impossible Labor could catch up there.)  The projection does not account for declaration votes.  In total right wing micros have about 11% vs about 4% for left micros, so it is hard to see the left parties both getting up at this stage.

Friday 8:00 -  In the live count the LNP now has 2.76 quotas, One Nation 0.71, Greens 0.69, Labor 1.61.  The Labor vs Greens race is impossible to call and likely to remain that way until the button.  I am hoping to look at the patterns from last time to get some sort of handle on it, but this involves complex modelling, and my feeling is the preference flow to Labor from KAP etc will weaken.  One Nation are not necessarily safe either as there is no love lost between them and FACN and it's possible the LNP will get more conservative preferences than they will.  But given the overall right-wing skew of the preference pool it's rather difficult to see them missing out.  Most likely Anning's voters will preference One Nation anyway, and really the only risk is that other parties preference One Nation much more weakly on account of Pauline Hanson herself not being on the ballot.  I think both conservative parties should win easily.

Saturday 5:00: Ross Leedham has been tweeting some unflattering figures for the Greens concerning their performance in PPVCs, of which significant numbers remain uncounted.

Sunday: Moved to own thread.

Western Australia

(Seeking re-election: 2 Coalition 2 Labor 1 Green 1 One Nation)
Expected result: 3 Coalition 2 Labor 1 Green

Western Australia is at about 52% counted and these are the leaders:

Liberal 2.82 quotas
Labor 1.98
Green 0.89
One Nation 0.37

There's no point mentioning the rest because the first three are so far ahead that I can't see further counting bringing One Nation into the mix.  I don't expect to be commenting further on WA.


South Australia

(Seeking re-election: 3 Liberal, 1 Labor, 1 Green, Retired at election: Tim Storer)
Expected result: 3 Coalition 2 Labor 1 Green

South Australia is about 57% counted and here are the leaders:

Liberal 2.54 quotas
Labor 2.20
Green 0.83
One Nation 0.32
UAP 0.20
Centre Alliance 0.18
HEMP 0.15
Animal Justice 0.12
Conservatives 0.11

At the moment the Liberals are over 3% ahead of One Nation, a large lead that is unlikely to be caught based on either remaining primary votes or preferences.  The Centre Alliance has bombed out completely, suggesting its current Senators will struggle to get elected in 2022.  Labor also doesn't seem to be a threat as the usual issue occurs here of the left votes splitting two ways between them and the Greens.  I will keep an eye on this one but don't expect the outcome to change.



(Seeking re-election: 3 Labor, 1 Liberal, 1 National, 1 Green)
Expected result: 2 Liberal, 2 Labor, 1 Green, Jacqui Lambie
Identity of second Labor winner to be confirmed, probably Catryna Bilyk
Lisa Singh has polled a large below the line vote but does not appear to be winning

Tasmania is at 63% counted and here are the leaders:

Labor 2.17 quotas
Liberal 2.17
Green 0.89
Jacqui Lambie Network 0.60
One Nation 0.25
United Australia 0.18
Shooters 0.12

There are also two indies with about 1% (0.07 Q) apiece - Craig Garland running as a grouped independent and Steve Mav running as an ungrouped BTL-only indie.  (Although Mav has lost his deposit, his 1.04% is still about 25 times better than any other ungrouped-column candidate in Australia.)  Steve Martin's Nationals attempt has also bombed out getting just more than these two. Lambie's vote has held up remarkably well despite a lack of resources and so much competition from similar parties, and indeed she seems to have taken a few votes from the majors.

The seat breakdown by party here is not going to change.  There is, however, the issue of Lisa Singh's below-the-line vote.  Singh was demoted to what is wrongly called an "unwinnable" position in 2016 and won from it, getting 6.1% of the vote.

In this case it might seem that half Labor's excess over one quota (so that would be 8.4%) would help Singh beat #2 Catryna Bilyk or perhaps Lambie.  But this is not the case, because from that point Singh could only attract below the line preferences until such time as Bilyk was eliminated, while Bilyk would keep getting above the line preferences from excluded micro-parties.  We saw this in the 2016 count where Bilyk had gained 1.6% more than Singh at the time when Singh was elected.  A further disadvantage for Singh is that the available preference pool for parties with an ATL box is slightly larger this time than in 2018 (I get 18% compared to 14.4%).  The target for Singh to win seems to be a primary of approaching 9.5%, about 1.5 times her 2016 vote.  I have seen samples from four booths that showed that she was getting something at least comparable to last time, and much higher in one small booth, but as the primary vote unrolls we will get a better idea.  In the seemingly unlikely case that Singh is competitive I'll transfer this to a new thread.  My suspicion at the moment is she'll finish seventh, as she would have in 2016 had that been a half-Senate race.

Wednesday 5:15 Unrolls of some early booths have started.  These are too small to draw reliable conclusions from as two are hospital booths that may have changed their list of hospitals.

Thursday 2:40: Still only a very minimal unroll of unrepresentative booths, with under 1000 votes unrolled.  I also have the unroll from a single booth from scrutineering.  At this stage I am tracking both average swing by polling place on the Singh vote and total swing from votes counted, and waiting until the sample becomes significant.

Thursday 7:40: Two significant booths in with a zero swing on the Kingston PPVC booth (Franklin) and 1 point against Singh in Goodwood (Clark).  Not an encouraging start at reaching a very difficult target.

Friday 7:50:  I have 4.6% of the count unrolled at matching booths so far.  Some large prepolls were added in Braddon today.  I have Singh tracking for the 6-7% range at the moment but this is still very rubbery.

Monday 2:30: Off 7.3% of the count unrolled at matching booths, I have Singh tracking for around 6.4%.

Monday 10:40: Off 13.7% I still have Singh tracking for around 6.4%.  The count is, however, dominated by Bass votes with no Lyons votes included at all.  I'm using the average of two projections to project at the moment (swing by booth and swing by total votes) and the latter gives a very slightly higher figure than the former.  I'll switch to using wholly the latter once the range of booths is more representative. 



(Seeking re-election: 1 Labor 1 Liberal)
Expected result: 1 Labor 1 Liberal (CALLED)

A quota in the ACT is a third (33.33%).  The ACT primary count is at 51% complete with the following scores:

Labor 1.190 quotas
Liberal 0.896
Greens 0.588
Pesec (IND) 0.158
UAP 0.062
SUS-A 0.048
FACN 0.027
CDP (ungrouped) 0.014
ungrouped inds 0.017

The out-there scenario that has been considered for the ACT Senate from time to time is the Greens taking the Liberals' second spot.  At present they are .308 Q behind with .348 Q in spare Labor and Pesec preferences.  But for the Greens to pass Zed Seselja off these preferences, the Greens would need to gain at .88 votes/vote, when in 2016 they only gained at .54 votes/vote from Labor.  The 2019 ballot paper will have slightly lower exhaust, and Labor votes might flow more strongly anyway, but even so a .88 flow off Labor is currently impossible. Even if it were possible the UAP, FACN and CDP preferences would help Seselja.

I also don't have any data on flows from Pesec but I would doubt his preferences would flow more strongly to the Greens than Labor's, or at least not much more.  So I think the Greens are at least 0.1Q (3.3%) behind where they need to be relative to the Liberals at the moment and need to make massive improvements on remaining primaries before this could be taken seriously.  It may be that the Liberals will improve instead on postals.

Thursday: The Liberals have now improved to very close to a quota with their lead out to nearly .44 quotas, though this may come back slightly.  I have had scrutineering reports that while the Labor flow to the Greens has increased slightly, the flow from Pesec is only slightly advantageous.  On either of these bases I would be happy to call it, let alone both.



(Seeking re-election: 1 Country Liberal 1 Labor)
Expected result: 1 CLP 1 Labor (CALLED)

A quota in the NT is one third (33.33%).  The NT has a slow count with 31.5% counted and the CLP are on 1.12 quotas, Labor 1.08, Greens 0.36 and the rest no more than 0.14 apiece.  Even if one major falls below a quota the others cannot catch them.  I won't be updating NT.

2019 House of Reps Postcount

Coalition has won the election, almost certainly with a small majority

Apparently won Coalition 77 Labor 67 Green 1 CA 1 KAP 1 IND 3

Seats Assumed Won By Coalition From Labor: Longman, Herbert, Bass, Braddon, Lindsay
Seats Assumed Won By Labor From Coalition: Gilmore, Corangamite, Dunkley
Seat Assumed Won By IND From Coalition: Warringah
Seat Assumed Won By Coalition From IND: Wentworth 

Seats that were being covered but now assumed won: Boothby and Chisholm (Coalition retains), Lilley, Cowan  (Labor retains), Bass (Liberal gain). 

Seats currently in significant doubt:

Cowan (ALP) - Labor leading and expected to win
Macquarie (ALP) -  Liberal slightly in front, too close to call, may require recount

If all current leads hold Coalition will win 78 seats, Labor 67, KAP 1, CA 1, Greens 1, IND 3.

This is the main thread for the 2019 House of Reps postcount.  A few days ago I expected I would be unrolling separate threads to unravel three-cornered contests in Melbourne, a complete mess in Mallee and so on, but at this stage none of that has happened.  I have done a quick thread for Mallee though as there's some interest in it.  There is a weird situation in Hunter, where One Nation are currently two points short of beating the Nationals into second, but I don't see any reason to think they can get into second, let alone win if they do.  Indeed the Nationals' margin in second should increase.

I have been very busy with media commitments regarding the Great Australian Poll Fail and will be unrolling comments on the seats in doubt later this afternoon.  For each seat there will be an intro post and then comments will follow until the seat is no longer of interest.  When the seat is no longer of interest it will be moved to the bottom.  Other seats may be added.  Any seat that goes to a recount will get its own thread for the recount, as may any unusually complex seat that is competitive.

At first the posts will be brief, and more detail will be added when I've got other articles up and running.

Macquarie (NSW) (ALP 2.2%)

Macquarie in outer Sydney often looked like a bit of a trouble spot for Labor and attracted some seat polling.  This looks like being the closest seat with a 625-vote lead to incumbent Susan Templeman (50.37) off ordinary votes but half of that was wiped out in the first 2000 postals.  In 2016 the Coalition gained 0.39% 2PP in the postcount so if that repeats this one will go down to the wire.

Monday 5 pm: Sarah Richards ahead by 23 votes after more postals.

Tuesday 5 pm: An extremely close count with Richards now ahead by 68.  Under 2000 postals likely to be added compared to twice as many absents so Templeman can come back.  It's hard to rely on postcount projections staying the same from election to election so we need to see if either candidate builds a serious lead.

Tuesday 7:30: Richards' lead out to 133, I believe on rechecking.

Wednesday 7:27: Richards has won a batch of absents, but that doesn't necessarily mean much because we don't know where they're from (one cannot project off absents for this reason).  Nonetheless lead now out to 196.

Thursday 11:40: Richards' lead now down to 131.  Another parcel of absents favoured Templeman.  An important thing to watch is whether either candidate can win by more than 100.  A win by 100 or less after the distribution of preferences triggers an automatic recount.  In Herbert in 2016 the recount was fast-tracked.   If the margin exceeds 100 then the petitioner for a recount would need very good reasons.

Friday 8:10: Richards' lead is now down to 71, not sure what on.

Seats previously covered but now assumed won:

Boothby (SA) (Lib, 2.8%)

Boothby was the Coalition's most endangered marginal in South Australia, but it helped there that Nicolle Flint had a personal vote bonus as a first-term MP.  Late on the night this seat was projecting to her by a narrow margin.  Currently her lead is 1442 votes.  I assume this is way too much but am waiting to see if there is any reason why the seat is still given as in doubt by the AEC.  In 2016 there was a slight postcount move to Labor in Boothby.

Chisholm (Vic) (Lib, 2.9%)

Chisholm is Julia Banks' former seat, the Coalition's only gain at the 2016 election.  It was a Labor retirement at that election so there is no personal vote effect.  Liberal candidate Gladys Liu (an accomplished former chess administrator, by the way) was widely criticised during the campaign and generally expected to have lost the seat as a result, but it has actually ended up on the wire.  This is a seat in which the Liberal Party controversially deployed mock-AEC coloured signage that Labor has complained bitterly about (restrictions on this sort of thing need tightening.)

Presently Gladys Liu (Lib) trails Jennifer Yang (ALP) by 134 votes, and will gain another 25 or so from a very small prepolling booth not yet added to 2PP.  In 2016 the Liberals increased their vote by 1.11 points after the postcount in Chisholm despite not holding the seat.  Anything like that again will see Liu retain the seat.  Nonetheless the seat was not projecting that strongly before the projections were switched off so let's see what happens here.

6 pm Sunday: Now 169 ahead and a huge postal count to come (a few already in which broke 60-40 to her) so I don't see why Liu will not win.

Tuesday: Liu now over 900 ahead.

Lilley (Qld) (ALP 5.8%)

Labor were at least vaguely nervous about Lilley on Wayne Swan's retirement; I know internal polling was done on it.  The personal vote loss plus the generally poor result in Queensland has fed into a big swing to the government. Anika Wells (ALP) leads with 50.75%.  In 2016 there was a 0.35% swing back to the government in the post-count, so that's close enough to make it worth keeping an eye on.

Monday 7:55: Wells is down to a 901 vote lead now (50.55%) after some unfriendly postals - but not super-unfriendly.

Tuesday 5:00 Wells has benefited from a counting correction and is now up by 1288 (50.79).  There are still maybe 8,000 postals but even at the current break rate those wouldn't be a threat even ignoring other classes of vote.  I'm moving this one to assumed won but will keep an eye on it.

Bass (Tas) (ALP, 5.4%)

Bass is the infamous revolving door seat around Launceston that has evicted the sitting party at 7 of the last 9 elections (easily the Australian record holder) and appears to have done it again, but in spite of Ross Hart's concession on election night it still isn't totally over.  However the Liberals' Bridget Archer is very likely to win.

After all booth primaries were included, Ross Hart (ALP) trails by 437 votes giving Bridget Archer 50.36% 2PP.  In 2016 the postcount in this seat increased the Coalition's margin by 0.78%, but in 2013 when the seat last had a Labor incumbent, the benefit was only 0.04%.  In general Labor doesn't often gain in postcounts, especially not by 0.36 points, but we need to keep an eye on it just in case something happens.  The other possibility early on in any race this close is counting error.

Monday 2:30 - Very, very interestingly the first batch of postals in Bass has broken only weakly to Archer (50.75%) extending her lead by only 22 votes to 459.  Normally the first batch are the worst for Labor!

Monday 4:00 In 2016 postals broke to the Liberals 6.4% stronger than the booth vote, absents the same, provisionals 7 points worse out-of-electorate prepolls 10% better.  But in 2013 when Labor last held Bass. postals were only 2.4 points better, provisionals 6 points worse, absents 1.4 points worse and out-of-electorate prepolls 1.4 points worse.

Even if the 2013 patterns on non-postals hold, Hart would need remaining postals to break to him to win.  Labor did make some effort on postals in the seat, though less than the Coalition.

However a wild card here is the possibility of swing through the campaign in this seat.

Monday 8:00 A further lot of postals has shrunk Archer's lead to 392.  A Liberal candidate has gone backwards on postals.  You read that correctly.  I'm not going to attempt to project this seat for a while because of the weird postal pattern and the history of varying postcount behaviour in this seat. Just sit back and see what the numbers do.  If pretty much anything increases Archer's lead by a couple of hundred or so she should win - Hart can't really afford any type of votes to go against him significantly (which could happen on out-of-electorate prepolls.)

Tuesday 2:40 A new lot of postals has moderated yesterday's weirdness a bit, so all postals so far are now breaking to Archer very slightly and her lead is up to 463.  That makes it a lot harder for Labor.

Wednesday 11:10 Archer lead out to 476, apparently on rechecking.

Wednesday 12:30: A further small increase on postals, Archer now leads by 510.

Wednesday 2:30:  A small decrease presumably on checking, Archer leads by 504.  Assuming postals do nothing from here Hart would need a roughly 57-43 break on absents and out of electorate prepolls even assuming nothing on provisionals.  This difference from the booth count was achieved by Labor in only two electorates in 2016 (Murray and Cowper, both of which Labor was uncompetitive in.)  Moreover, Bass was one of the worst electorates for Labor on this comparison in 2016, and not spectacularly good in 2013 either.  I have moved this seat to "assumed win" status and it will probably be definitively called soon.

Thursday 3:40: With a further slight increase and disappearance of most of the remaining absents I have called Bass for the Liberals.

Cowan (WA) (ALP, 0.7%)

Cowan was one of Labor's big wins in 2016, with Anne Aly taking the seat.  Aly is currently showing with a 0.5% swing to her so a 1.2% lead.  There was a 0.46% move back on post-count in 2016 and just maybe if that is bigger and there is a shift in out of electorate prepolls there might be something here but I doubt it.  Keeping an eye on it for a bit.

Tuesday 5:00  Aly's lead has been cut back on postals to 786 votes (50.54).   Most of the postals are gone now so this points more to her probably having a narrow margin, but it seems doubtful the gap would close completely.

Thursday 4:00: Aly gains on absents to lead by 836 (50.55).  There are still a lot of prepolls to come in this division.

Friday 8:00 Aly's lead down slightly to 786.  Still many thousands of votes to go.

Monday 5:30: Aly's lead is now out to 810.  There are still several thousand votes left but not that many postals and there is no evidence of a meaningful break to the Liberals on any other category.  So I have moved this seat to assumed-won status.  I'll have another look if the gap drops below, say, 500.

Saturday, May 18, 2019

2019 Federal: Late Night Live Commentary

Coalition has won the election - small majority likely but perhaps a minority

Apparently won Coalition 74 Labor 64 Green 1 CA 1 KAP 1 IND 3

Seats Assumed Won By Coalition From Labor: Longman, Herbert, Bass, Braddon, Lindsay
Seats Assumed Won By Labor From Coalition: Gilmore, Corangamite, Dunkley
Seat Assumed Won By IND From Coalition: Warringah

Seats currently in doubt (projection = AEC projection)

Eden-Monaro (ALP) - Labor leading on projection
Boothby (Lib) - Liberal leading narrowly on projection
Chisholm (Lib) - Liberal leading narrowly on projection
Cowan (ALP) - Labor leading on projection 
Lilley (ALP) - Labor leading narrowly on projection
Macquarie (ALP) - way too close to call on current projection
Wentworth (IND) - Sharma (Lib) currently well ahead and expected to win

If all current leads hold Coalition will win 77 seats, Labor 67, KAP 1, CA 1, Greens 1, IND 3, with one unclear.


Election night arrangements and election watching tips (2019)

Donations welcome!

If you find my coverage useful please consider donating to support the large amount of time I spend working on this site.  Donations can be made by the Paypal button in the sidebar or email me via the address in my profile for my account details.  Please only donate if you are sure you can afford to do so.


My coverage tonight and to come

I will be doing live blogging for the Mercury from 6 pm.  The link to the live coverage is here.  The link to the tweet that links to the live coverage is here.  I am not sure yet how long the coverage will go or whether I will need to take any breaks to write articles.  There is an email address on the live coverage for people to ask me questions during the coverage.

I will probably not be checking emails, tweets or for comments on this site regularly during this time, and I ask journalists not from the Mercury not to call me until the live blog has finished.  However if you have interesting scrutineering samples from Tasmania - especially of the rate of below-the-line voting for Lisa Singh in a Senate booth (please say which booth!) - you're extremely welcome to SMS them to me (0421428775) or email them to me (  I probably won't be able to reply immediately.

The live blog is paywalled and I don't know if there will be a paywall free link or not. Subscription options are available starting from $12.  (In a kind of reverse Antony Green situation, a public media source did try to poach me for this election but negotiations collapsed from the point where I mentioned the concept of being paid.)

Election Eve: Final Newspoll and Seat Betting Roundup

2PP Aggregate (Final): 51.6 to Labor (Last-election preferences)
(51.3 with One Nation adjustment)
Seat projection assuming polls are accurate 79 Labor 66 Coalition 6 Others

The final Newspoll marks the end of my poll aggregation for this term (barring any late offers) though in the last few weeks, I really needn't have bothered.  Aggregation is most useful when polls are bouncing around, when they have marked house effect differences relative to each other or when there is something happening that is causing voter intention to move.  In the last few weeks, almost nothing has happened and seventeen consecutive polls have landed their 2PPs somewhere in the range 51-52.  There's really nothing for an aggregate to do!

Friday, May 17, 2019

A few election notes (especially re Tasmania)

Just a few quick notes on a number of things, mainly to bring links to various articles to the top.

Election night coverage

On election night I am very pleased to say I will again be live blogging for The Mercury, starting from 6 pm and probably going til about 11 or maybe later, except for any time when I may or may not need to stop blogging and write a quick article for them.  The link will be posted here when known.  It is possible the coverage will be paywalled; if so there are various subscription offers starting from $12, or it may be there will be a paywall-free link.  Once that has been finished, after a brief break I will relocate to home and then continue on here for a while.


Some links to recent items on my site that may be useful:

How To Make Best Use Of Your 2019 Senate Vote

Tasmanian House of Reps Seats

Tasmanian Senate Guide