Australia 2013 Archive


14 votes, 13 votes, 12 votes

As a follow-up to Saturday’s post I wanted to revisit the question of how much the WA Senate recount result was effected by the missing votes from the four booths in Forrest and Pearce.

On my post the other day, I estimated that the missing votes improved the position of the Christians relative to the Shooters by fifteen votes. At the time I warned that this margin may not be exact because it was possible that some other votes cast at these booths that weren’t lost may have changed in the recount.

Commenters at the Truth Seeker blog have analysed the booth results from the four problematic booths individually, and it turns out that it is possible to identify precisely which votes were missing, as it appears that the missing votes were entire bundles of votes for a particular group at a particular booth, and you can identify which parties are recorded as receiving no above-the-line votes at any of those four booths.

William Bowe at Poll Bludger has neatly summarised these missing votes by party and booth, which makes it possible to confirm that the net change was actually thirteen votes. It appears that one Australian Fishing and Lifestyle Party vote was removed and one Australian Christians vote was added through the legitimate recount process, which explains the difference between my original figure of fifteen votes and Truth Seeker’s total.

With these adjustments, it seems our best estimate suggests that the missing votes would have produced a one-vote victory for the Shooters at the key count, leading to the ALP and PUP winning.

At the point where one vote is decisive, it seems very hard to avoid a by-election.


Numbers point to WA Senate by-election

Update: The result today saw Scott Ludlam (GRN) and Wayne Dropulich (Sports Party) elected instead of the PUP and ALP candidates who had won in the first count. The margin at the key point is 12 votes, a net turn-around of 26 votes. The case seems set to head to the Court of Disputed Returns.

Original post: After the AEC announced on Thursday that 1375 votes were missing in the WA Senate recount, many people quickly jumped to the conclusion that a WA Senate by-election was needed to resolve the situation.

This was slightly premature, as it was still possible that the recount would be decisive enough that those votes wouldn’t make a difference.

However after examining the latest numbers on the AEC’s Virtual Tally Room, I believe that there are two alternative methods of coming to a result that produce different winners, and this probably means that a by-election will be needed.

There are four booths where votes are missing, these are:

  • Bunbury East (Forrest)
  • Henley Brook (Pearce)
  • Mount Helena (Pearce)
  • Wundowie (Pearce)

However, it is not the case that all of the votes at these booths are missing.

According to the latest figures on the AEC’s Virtual Tally Room, 3445 votes have been counted at these booths (including informal votes). This compares to 4799 before the recount started. Confusingly, this adds up to 1354 votes missing. I don’t know why this diverges from the public figure of 1375, but I’m going to set that aside for now.

There are five parties that we need to track for the purposes of determining who will win.

The critical count that is decisive was the point where the Shooters and Fishers defeated the Australian Christians by 14 votes. This produced a victory for the second Labor candidate and the Palmer United Party. If this 14-vote margin was reversed, then the last two seats would have gone to the Greens and the Australian Sports Party.

Prior to this count, three others parties had been excluded and had passed on their above-the-line preferences to either the Shooters or the Christians. The Australian Independents and the Fishing and Lifestyle Party preferenced the Shooters, and the Climate Sceptics preferenced the Christians.

For these purposes I am ignoring all other parties and only looking at the net change in votes between the Shooters/AFLP/Aus Independents grouping and the Christians/Climate Sceptics grouping.

The disappearance of the 1354 votes at those four booths produced the following net change at those booths from pre-recount to post-recount.

Shooters and Fishers6854-14
Australian Independents84-4
Fishing and Lifestyle Party2423-1
Shooters Total10081-19
Australian Christians6866-2
Climate Sceptics31-2
Christians Total7167-4

The missing votes massively disadvantage the Shooters grouping – by a net 15 votes, which is enough to flip the result.

It should be noted that this is based on the assumption that there were no changes in the 3445 votes from those four booths that were counted, but this is unlikely. So the composition of the missing votes could be slightly different to what is listed above.

In the rest of the state, despite quite a lot of votes being challenged and anecdotal reports suggesting many votes had flipped, overall the Shooters grouping has lost only one seat relative to the Christians grouping.

This table lists the vote before and after the recount for the remainder of the state, with all votes at the four key booths excluded.

Shooters and Fishers13,56513,573+8
Australian Independents4,0344,039+5
Fishing and Lifestyle Party5,7035,706+3
Shooters Total23,30223,318+16
Australian Christians21,41021,438+28
Climate Sceptics1,4911,480-11
Christians Total22,90122,918+17

There appears to be two possible ways to produce a result using these votes:

  1. Only count those votes that have been managed to be recounted, with the missing votes excluded from the count, which will likely result in a very slim victory for the Christians, and thus for the Greens and the Sports Party.
  2. Substitute votes cast at the four booths where votes are missing for the count from prior to the recount, which will likely result in a slim victory for the Shooters, and thus for the ALP and the Palmer United Party.

It is also possible that changes to below-the-line votes that were previously counted as informal could shift the count, but it is clear that the result remains extremely close and likely to not produce a clear outcome. In such a scenario, the case for a statewide Senate by-election becomes quite strong.


Post-election gender balance update

After an election where the vast majority of candidates running were men, the male-dominated Coalition won seats, and today a new cabinet will be sworn in with only one woman, it may be surprising to know that the number of women in the new Parliament will be increased.

In the House of Representatives, the number of women has increased from 37 in 2010 to 39 in 2013, out of a total of 150.

In the Senate, the number of female Senators will fall from thirty to 28 when the new Senate takes office on 1 July 2014.

Overall, this results in a net increase of one woman in the new Parliament, although the number of Senators could vary from 27 to 31.

Both Labor and the Coalition have increased their proportions of women in their House delegations. The ALP lost three men and three women in the Senate, while the Coalition has the exact same number of men and women in the Senate as before the election.

The Greens Senate delegation has only changed slightly, with the addition of a seventh woman to their team of ten.

The main backwards move is the non-Greens crossbench in the Senate. Nick Xenophon and John Madigan are currently on track to be joined by five more men, with only a small chance that one woman could be elected for the Palmer United Party in Tasmania.


Correction: due to a coding error I had one LNP member from Queensland listed as female, when he is actually male. The attached spreadsheet and the table above have been adjusted.

I’ve identified six seats where I think it’s conceivable there could be a change to effect these numbers:

  • Fairfax – Clive Palmer is currently leading by an extremely slim margin over the LNP’s Ted O’Brien. His election wouldn’t effect the overall gender balance but would have increased the proportion of men in the Coalition party room.
  • ACT Senate – The Liberal Party’s Zed Seselja is likely but not certain in winning over the Greens’ Simon Sheikh. Again no change in overall balance but would reduce male proportion of Coalition and increase male proportion of Greens.
  • NSW Senate – The Liberal Party’s Arthur Sinodinos is very likely to win, but slim chance for Greens’ Cate Faehrmann, which would change gender balance.
  • Tas Senate – The favourite for the final seat is the Liberal candidate, a woman, but there is a possibility either a female Palmer United Party candidate or a male Sex Party candidate could win.
  • Vic Senate – The male Motoring Enthusiasts Party could lose to the Liberal Party’s Helen Kroger.
  • WA Senate – The male Greens Senator Scott Ludlam could lose to female Labor Senator Louise Pratt.

You can also download an updated list of all candidates who ran in the election, including their gender.


Seat in focus: Fairfax

Clive Palmer’s push to take the Sunshine Coast seat of Fairfax off the Liberal National Party looked like it had been successful on election night, but a very poor performance on special votes has Palmer’s lead falling so fast that it is likely to reach zero before the counting finishes.

Click here to read my pre-election profile of Fairfax.

As of last night, Palmer was leading by 507 votes in Fairfax.

Vote typePUP %LNP %To be countedProjected diff.
Absent57.5942.412,266PUP +344
Pre-Poll43.2856.722,268LNP +304
Postal37.6762.332,227LNP +549

Clive Palmer is performing extremely badly on postal votes. Over 2000 more postal votes are expected, and if they break as badly as those postal votes already counted, the LNP will gain enough votes to take the lead.

Palmer should gain some votes from the 2266 absentee votes, but he is also losing in pre-poll votes, which should cancel that out.

Overall, these figures ad up to a net gain of 509 votes for the LNP, which turns Fairfax into almost a dead heat.

These figures project an LNP win by seven votes. If the result is that close, expect court challenges and a complete recount. A win by a few hundred votes by either Ted O’Brien or Clive Palmer is also still a possibility.

As with every other seat, I’ll now focus on the geographic breakdown of the vote. I have broken the electorate into the same four subdivisions as before the election: Nambour, North-East, South-East and West.

Voter GroupGRN %ALP %PUP 2CPTotal votes% of ordinary votes
Other votes7.2417.8445.8932,845

Clive Palmer won a majority in three out of four areas, ranging from 54.5% in the north-east (covering Coolum Beach) to 57.5% in the Nambour area. The LNP won a narrow majority in the south-east, which covers Buderim and Maroochydore, and is the largest part of the ordinary vote.

Clive Palmer is losing the ‘other votes’ badly, mainly due to a poor performance on postal and pre-poll votes. If election day votes were the only ones counted, Palmer would win easily.

After the fold is a map of the electorate, and a break-out map of the south-east of the electorate. Read the rest of this entry »


Sunday night counting update

There are four seats I am now tracking where I don’t think it’s entirely bedded down: McEwen, Fairfax, Indi and Parramatta.

UPDATE: Now covered all four key seats.


Labor MP Rob Mitchell currently leads by 97 votes, over the Liberal Party’s Donna Petrovich.

Vote typeALP %LIB %To be countedProjected diff.
Absent55.6444.362,264ALP +256
Pre-Poll4,845LIB +115
Postal47.7452.26442LIB +20

There are almost 4900 pre-poll votes that have yet to be counted. This is by far the largest chunk of votes to be counted. No pre-poll votes have been counted yet, so we can’t calculate a projection. However I have made an estimate that they will help the Liberal Party slightly, as the Liberal Party did 1.32% better in 2010 on pre-poll votes in McEwen than they did on ordinary votes. If you add those to the ordinary vote total, you give the Liberal Party 51.2% of the pre-poll vote, or an advantage of 115 votes.

There are still quite a lot of absentee votes, that go to Labor, with very few postals to help the Liberal Party.

While there are up to 1700 provisional votes, only 250 were counted last time. They favoured Labor, but I’m guessing the swing to the Liberals mean the provisional votes work out roughly evenly.

On my estimate, I expect the Labor vote lead to grow by 121 votes, to a total of 218 votes.


Labor MP Julie Owens currently leads by 482 votes, over the Liberal Party’s Martin Zaiter.

Vote typeALP %LIB %To be countedProjected diff.
Absent50.9449.063,911ALP +73
Pre-Poll51.3248.681,458ALP +38
Postal45.6954.31250LIB +22

Almost all postals have been counted, withrelatively few prepoll votes to be counted. Almost 4000 absentee votes are to be counted, and these favour the ALP slightly.

The absent votes are decisive, and they are breaking slightly for Labor. At the moment this means the ALP should increase their margin. To win the seat, the Liberal would need to win 56.5% of the remaining absentee votes, compared to a current rate of 49.1%

On my estimate, I expect the Labor vote lead to grow by 89 votes, to a total of 571 votes.


Independent candidate Cathy McGowan currently leads by 515 votes, over Liberal MP Sophie Mirabella.

Vote typeIND %LIB %To be countedProjected diff.
Absent54.5245.481,127IND +101
Pre-Poll41.9158.091,991LIB +323
Postal42.9257.0854LIB +8

The postal votes helped Mirabella, but they are done. The 2000 remaining prepolls, if they break at 58% like the first 2000, will give an additional 300 to Mirabella, but that won’t be enough.

On my estimate, I expect the Liberal vote deficit to reduce by 230 votes, but McGowan to win by 285 votes.


Clive Palmer of the Palmer United Party currently leads by 502 votes, ahead of Liberal National candidate Ted O’Brien.

Vote typePUP %LNP %To be countedProjected diff.
Absent57.5942.412,266PUP +344
Pre-Poll43.2856.722,268LNP +304
Postal37.6762.332,227LNP +549

Clive Palmer is performing extremely badly on postal votes. Over 2000 more postal votes are expected, and if they break as badly as those postal votes already counted, the LNP will gain enough votes to take the lead.

Palmer should gain some votes from the 2266 absentee votes, but he is also losing in pre-poll votes, which should cancel that out.

Overall, these figures ad up to a net gain of 509 votes for the LNP, which turns Fairfax into almost a dead heat.

On my estimate, the LNP would win by seven votes, which would likely result in stretched out legal fights over who wins the seat. It’s also possible that late counting could result in either side winning by over 100 votes.


Seat in focus: Indi

Indi has been the seat that has been the most interesting seat during the first week of counting. Cathy McGowan has led for most of the week, but was on track to fall behind before an additional 1003 votes that had not been correctly entered were found, and expanded her lead.

Click here to read my pre-election profile of Indi.

Cathy McGowan’s margin is now 837 votes. Most of the postal votes have now been counted, and these substantially helped Mirabella. Only small batches of absentee and prepoll votes have been counted. At the moment, Mirabella has a 56.5% lead in prepoll votes and McGowan has a 54.52% lead in absentee votes. There is, however, a lot more prepoll votes to come in, which should mean that Mirabella will narrow the gap further as they are counted.

Type of voteIND %LIB %To be counted

I estimate that if the current trends continue in the remaining votes in the three major categories (and the 377 provisional votes split evenly), then Sophie Mirabella will gain an additional 507 votes, narrowing the gap to 330 votes.

At this point, it’s a bit too close to comfortably say McGowan will win. Of course, it’s possible McGowan will pick up ground in the next batch of prepoll votes. Less than 700 of these prepoll votes have been counted, which is too little to be sure that they are representative of the entire sample. It’s also possible Mirabella will pick up ground on both absentee and prepoll votes and win, but she would need to do a lot better.

Now let’s turn our attention to the geographical breakdown of results. Again, I have used the same geographic divisions as I did in my pre-election profile. The biggest city in the electorate, Wodonga, has its own subdivision, as does the rural council area of Indigo immediately south of Wodonga. The next biggest city, Wangaratta, makes up 4465 out of the 11,190 votes in the ‘west’ subdivision.

Voter GroupGRN %ALP %IND %LIB %IND 2PPTotal votes% of ordinary votes
Other votes2.7911.8629.8346.5848.6136,907

Cathy McGowan won majorities in the west, Wodonga and Indigo. Sophie Mirabella won majorities in the south and east, as well as a majority of other votes.

The ‘other votes’ category doesn’t just include the categories discussed above. Since 2010, prepoll votes cast at regular prepoll centres within the electorate are counted on election night as ordinary votes, and are counted in the list of booths. These prepoll votes varied between booths. Over 16,000 votes were cast at the Wangaratta and Wodonga prepoll centres, and McGowan won slim majorities at both of those booths. Overall, McGowan won a majority in all of the ‘other votes’ before you include the 7720 postal votes counted so far – these went to Mirabella so strongly that she wins that category.

McGowan did particularly well in the big cities of Wangaratta and Wodonga. While McGowan won a slim 51.5% majority in the west of the electorate, that actually included a 56.7% majority in the four Wangaratta booths and only 48.1% in the other 26 booths in the area.

After the fold are maps of the results across the electorate, including special maps for Wodonga and Wangaratta.

Read the rest of this entry »


Seat in focus: Parramatta

Parramatta in Western Sydney currently is sitting as the second-most-marginal seat in the country. Labor MP Julie Owens is leading by 482 votes: a 50.32% margin.

Click here to read my pre-election profile of Parramatta.

Owens started out leading by over 1000 votes, but has not yet fallen behind despite losing ground.

Most postal votes have now been counted, while quite a lot of prepoll and absentee votes are yet to be counted. The Liberal Party’s Martin Zaiter polled very strongly in postal votes, while Owens is leading in those other categories. This race will be very close.

The analysis is based on the breakdown of the electorate into four sub-areas, the same used in the pre-election profile.

Voter GroupGRN %ALP 2PP %LIB swingTotal votes% of ordinary votes
Other votes4.9948.1516,759

The ALP won a majority in three out of four areas on ordinary votes, with the Liberal Party winning 54.9% in the north-east, closest to Bennelong and the Liberal Party’s northern heartland. While the Liberal Party gained swings in all four areas, none of them were enough to change the majority party in any area.

The Liberal Party is currently leading with 51.9% in the ‘other votes’ category.

The Liberal Party gained modest swings in most booths, apart from four booths. The ALP gained significant 5% swings in two booths near Granville.

Most booths saw small swings of less than 5% to the Liberal Party, although booths in the centre of Parramatta rose to 7%, 8% and 11%.

Two-party-preferred votes in Parramatta at the 2013 federal election.

Two-party-preferred votes in Parramatta at the 2013 federal election.

Two-party-preferred swings in Parramatta at the 2013 federal election.

Two-party-preferred swings in Parramatta at the 2013 federal election.


Friday night counting update

As it stands, there are still a bunch of seats left outstanding.

Seven seats are on margins of less than 50.6%, and I’ll run through a couple of key races here:

Apart from these seven seats, the Coalition holds 87 seats, the ALP holds 53 seats, and there are three crossbench seats (Bandt, Katter, Wilkie).

  • Barton – The Liberal lead has widened to 895 votes in Barton. Labor was leading by 42 on Tuesday. Most absentee and postal votes have been counted, which both favour Liberal, but no prepoll votes. It’s likely the Liberals will win.
  • Capricornia – The LNP has turned a 152-vote deficit into an 872-vote lead since Tuesday. Another roughly 2000 postal votes (favouring LNP strongly) and 2000 absentees (favouring ALP narrowly) as well as all of the prepoll votes are yet to come in. LNP likely to win.
  • Eden-Monaro – Most declaration votes, apart from 1000 provisional votes, 500 absentee, 450 postal and 1400 prepoll votes, have been counted. The Liberal Party’s lead has widened from 569 votes on Tuesday to 864, and Labor MP Mike Kelly conceded today. Liberal to win.
  • Fairfax – Since Tuesday, Clive Palmer’s lead has narrowed from 1,391 votes to 718. Approximately 8000 out of 15,000 declaration votes remain uncounted, and the largest share of these are prepoll votes from outside of the area, along with 2300-2400 each of absentee and postal votes. Clive Palmer performed very strongly (57.6%) in the absentee votes counted so far, and very poorly (62.3%) in the postal votes. While Palmer is narrowly trailing in prepoll votes counted so far, this gap won’t be enough and he should be elected.
  • Indi – The seat everyone has been watching: Cathy McGowan currently leads by 895 votes. McGowan is trailing badly in postal votes (57.61% for Mirabella) and leading strongly in absentee votes (55.27%). There are 1000 postal votes and 1600 absentee votes to be counted, but the bulk of the remaining votes are over 4000 prepoll votes and 1000 provisional votes to come. If you assume the remaining postals and absentee break the same way, Mirabella would need between 58% and 59% of the other votes to win. Very likely McGowan will win.
  • McEwen – Like in 2007 (and despite a very favourable pro-Labor redistribution), McEwen is the closest seat. Liberal candidate Donna Petrovich leads by 153 votes, smaller than the 331-vote lead at the end of Tuesday. There are a lot more absentee votes than postals yet to be counted (4300 vs 1400), and Labor did better in absentees than the Liberals did in postals. However there are almost 5000 prepolls yet to be counted. The ALP should be able to regain the lead on absentee votes, but the race will be decided by the prepolls. In 2010, the ALP polled 1.3% worse on prepolls than in ordinary votes – which would give the seat to the Liberals. This seat is definitely too close to call.
  • Parramatta – The only close seat where Labor leads, with Julie Owens ahead by 482 votes. Most postal votes have been counted, which heavily favoured Liberal candidate Martin Zaiter. The ALP is leading in prepoll votes (1491 to be counted) and narrowly leads amongst absentee votes, which are the largest portion of the outstanding votes. If the remaining votes break the same way, the race will stay about as close as it is now. Too close to call.

This adds up to 90 Coalition, 53 Labor, 5 crossbench and two seats (McEwen and Parramatta) genuinely so close you can’t pick someone as a favourite.

I have already posted seat profiles, with updated maps and vote breakdowns for McEwen, Barton and Capricornia, and Parramatta will be going up tomorrow. I plan to also post profiles of Indi, Fairfax and Eden-Monaro over the course of this week before wrapping up the post-election coverage.

If there’s another particularly interesting seat you’d like to see given the same coverage please let me know.


ALP and NZ Labour experiment with leadership in parallel

On Thursday, we saw former Labor Senate leader Stephen Conroy come out attacking the new Labor leadership rules pushed through by Kevin Rudd during his very brief second stint as Prime Minister.

He said the Coalition are “getting away with murder” while Labor’s leadership remains in a vacuum.

“We’ve got no leader, no frontbench, no shadow spokespersons who are able to lead the debate for us, and this will descend into complete and utter farce,” he said.

“We have a situation where the US might bomb Syria [and] we have no official party spokesman, we have no leader.

“These new rules were a farce when they were put in place – rules that have left us helpless.”

It’s not surprising that Conroy isn’t a fan of a process that takes power out of the hands of the Labor leadership and gives it to members, but there are real questions raised about the downsides of lengthy leadership contests.

The reality is that, in many countries, parties go an extended period of time without a permanent leader while choosing a successor. In this period, an interim leader is usually appointed.

A lot of the details about how Labor will conduct their contest remain vague, unsurprisingly considering that the idea only emerged a few months ago: it’s unclear if the ‘rank and file’ vote will be announced before the caucus votes, and it seems likely that the most senior leader not running will take over as acting leader, which would be Penny Wong if Anthony Albanese decides to run.

In 2010, when the UK Labour Party chose a new leader following Gordon Brown’s defeat, Brown quickly resigned and his deputy Harriet Harman served as acting leader throughout the contest, and there was a full panel of shadow ministers during this period.

There is no reason to think the same won’t apply in the case of the ALP’s leadership election.

Right now, the New Zealand Labour party is going through the same process as Australian Labor. The party decided after the 2011 election to implement a direct election model, where 40% of the vote is cast by members, 40% by MPs and 20% by unions. Apart from the 20% union vote, this reflects the model proposed for the ALP.

At the moment three candidates are running: Grant Robertson, David Cunliffe and Shane Jones have nominated, and are out campaigning. All three have been given the opportunity to ask questions of the Prime Minister during Question Time, and have participated in at least one televised debate.

Of course it’s possible that a direct leadership election can reveal divisions and bring up conflict, but the last three years in the federal ALP have demonstrated that divisions and conflict can get out anyway.

On the other hand, a direct election has the potential to engage new supporters, give members more of an investment in supporting the party leadership and providing the leadership candidates a platform to present their messages.

It’s also true that, in most circumstances, candidates who engage in negative or nasty campaigning won’t be rewarded by party members, even if swinging voters may do so in a general election.

Of course, a direct election of a leader creates a different incentive when it comes to Australia’s culture of cuttign down leaders. I’ve previously blogged about how Australian leaders tend to be much more likely to be replaced withuout facing an election than leaders in similar countries that directly elect their leaders.

If a leader is directly elected, they have a greater mandate from the members and have moral authority which will limit the power of the MPs to remove them (no wonder Conroy isn’t a fan). The 60% threshold (75% in government) practically limits MPs’ power, but the requirement to have a direct ballot also makes a sudden coup impossible.

These factors then makes the idea of cutting down a leader less attractive, and motivate MPs and members to support them in situations where they might otherwise look for an alternative.

Because of this, it’s possible that, for once, the leader who wins this contest may be capable of sticking around in the long term.

Read my previous posts about direct election.

Read David Farrar’s analysis of the NZ Labour leadership contest.


Seat in focus: McEwen

McEwen was the closest seat at the 2007 election, and this time it’s one of the closest again. Currently the Liberal Party’s Donna Petrovich leads the ALP’s Rob Mitchell by 331 votes, which amounts to a margin of 50.19%.

McEwen covers parts of northern Victoria to the north of Melbourne, including Sunbury, Craigieburn, Seymour, Gisborne and Puckapunyal.

Click here to read my pre-election profile of McEwen.

Mitchell’s 66-vote lead was turned into a lead for Petrovich on Tuesday, with nothing more happening since then.

Over 8000 postal votes have been counted, with another 4000 to come along with 6000 other special votes.

The analysis is based on the breakdown of the electorate into four sub-areas, the same used in the pre-election profile.

Voter GroupGRN %PUP %LIB 2PP %LIB swingTotal votes% of ordinary votes
Macedon Ranges9.916.1154.699.0410,78016.95
Other votes5.546.4254.0620,076

The ALP polled most strongly, with 55.4%, in the Hume area at the southern end of the electorate, despite a 9.4% swing to the Liberal Party.

In 2010, Labor won a majority in all four areas, but large swings across the electorate has created Liberal majorities in Macedon Ranges and Mitchell.

Below the fold, view two-party and swing maps for the electorate.

Read the rest of this entry »