Australia 2016 Archive


Double Disillusion – the definitive analysis of the 2016 election out now

I’m very pleased to plug a new book that has just been published from ANU Press about the 2016 federal election entitled Double Disillusion.

The book will be available in hard copy and is also available as a free download.

This is the sixteenth edition in a series of election studies dating back half a century, covering each federal election in depth. It features chapters analysing the overall contest, each of the political parties’ campaigns, the impact of the media and interest groups, and the major policy debates of the election campaign, written by a bunch of excellent academic writers.

I was very happy to be able to contribute the chapter summarising the results in the House of Representatives, including the impact of minor parties, the role of preferences and a run-through of key seats. This is paired with a chapter analysing the Senate results by Antony Green.

Thank you to Anika Gauja and Peter Chen in particular for inviting me to participate in a formal space which I am not normally a part of.

There’s also a bunch of other excellent chapters and I hope readers find it useful and interesting as a definitive take on this bizarre election campaign.


Don’t like senators winning with 77 primary votes? Here’s how to fix it

There’s been a lot of attention over the last few days on the newly elected One Nation senator from Queensland, Malcolm Roberts. His election was a bit of a surprise, whereas we were confident that Pauline Hanson would win her seat, and her two other Senate colleagues were predicted as likely to win since election night.

There’s a lot to pick over about Roberts’ record, but a lot of the focus has been on the fact that he received only 77 primary votes – less than any other successful candidate.

I actually think there is a real problem with our Senate voting system which is exposed by Roberts’ tiny primary vote, but it’s not the one that most media has focused on. This problem isn’t unique to Malcolm Roberts. The reality is most senators, from major and minor parties, receive very few personal votes, and rely almost entirely on voters preferencing according to their party’s ticket to win a seat.

It is remarkable that someone won a seat off only 77 personal primary votes, but I would argue that it isn’t significantly different to the many major party candidates elected on a few thousand votes in large states.

The phenomenon of unpopular candidates winning seats with no public profile is not limited to One Nation – the major parties have a history of gifting Senate seats to party hacks who couldn’t win a seat in the House of Representatives. It doesn’t have to be this way.

In this article, I’ll explain the broader issue, and what I think can solve it.

Read the rest of this entry »


Senate preferences – unpacking the data

One of the exciting things about the new Senate voting system is that we are on the verge of gaining a lot of new data on how people preference. The AEC has released the full dataset of Senate preferences for the last few elections, but the vast majority of preferences followed pre-registered tickets, so the data was less interesting.

This time, everyone chooses their own preferences, and it appears that the vast majority of voters have marked their own preferences. Using this massive dataset we can do a lot of things to learn more about how voters preference, how many follow how-to-vote cards, and how they preference if they don’t. We can also see how these trends vary between seats, states and booths.

I haven’t had time to do any of this at the moment, but some others have done some cool things with the data. I just wanted to shout out to these people, and then address a possible problem with the method of deciding which senators get six-year terms.

David Barry, who does a lot of cool things with election data at his website, has put together a preference explorer. For voters who voted above the line (most of whom numbered six boxes), you can see how many people followed any particular preference flow.

Grahame Bowland has made his own software which takes the AEC preference data and runs his own count, to verify the accuracy of the AEC count. So far the Tasmanian count has passed the test. His page shows how the count progressed in a more user-friendly way than the official AEC PDF file.

Finally, there was some discussion last night around the interpretation of the section of the electoral act which deals with the special count to identify the top-six senators in each state (section 282).

Under section 282, a special count is conducted between the twelve candidates who have been elected from each state. Any votes which give a first preference for anyone other than these twelve candidates is distributed to one of them, and if there is no preference for an elected candidate, the vote is treated as informal. A new quota is struck as 1/7 of the remaining votes, and a new count is conducted. The Senate is then free to use the results of this count, if they choose, to determine who gets a six-year term and who gets a three-year term.

Grahame Bowland and Dean Ashley (who has built his own version of the counting software) both attempted to do a special recount as specified under section 282, but have discovered an ambiguity which makes it unclear who gets elected.

Dean goes into much more detail at his blog, but the short explanation is that it is unclear whether a below-the-line vote which is originally formal, and has a preference for at least one of the elected candidates, but has less than six preferences for elected candidates, is treated as formal in the recount.

It seems likely to me that any vote that was formal stays formal, as long as there is someone to receive the preference – but I can’t say the definitively from my reading of the legislation.


Tasmanian Senate button-push – summary

I’ve got a piece prepared for the Guardian today which discusses the Tasmanian Senate result but I had some thoughts that didn’t fit in to include here.

Firstly, the challenge in winning a seat from below the line isn’t just in gaining enough primary votes to be a contender. On primary votes, Richard Colbeck looked like he was in a strong position to win a seat. He was actually on track to win the tenth seat at the point where there were 24 candidates still racing for the last four seats.

Neither Colbeck or Singh received a single above the line vote. Not one. Colbeck was knocked out before Bushby was elected, so he wasn’t able to receive any. Singh was elected just before Bilyk, on preferences from Colbeck.

Colbeck was successively overtaken by Bushby, then Bilyk, then McKim, then finally McCulloch. Singh only survived because she was close enough to a quota to stay ahead of the pack, and to limp across the line with below-the-line preferences.

As long as we have above-the-line voting, this will remain a significant hurdle for any insurgent candidate. Ideally, they need the candidates ahead of them to drop out of the race early to open up room for them to gain preferences, because it’s hard to see any candidate getting close to the much higher half-Senate quota on their own steam. Singh wouldn’t have won a seat if this was a half-Senate election.

The second thing to note is that preferences are critical. Before this election we relied on tremendous speculation about how an election might work under this new voting system. Some suggested that all votes would exhaust, and it would become a “first past the post” race. That has well-and-truly proven to be wrong. I’m personally surprised by how few votes exhausted in Tasmania, and how close the final seat was to a full quota.

Less than 2.8% of votes exhausted. The last seat went to a candidate on more than 80% of a quota. Over 85% of above-the-line votes were numbered 1-6. It appears that a similar phenomenon is taking place in other states, although you’d expect that more votes will end up in the exhaust pile in states with bigger ballots.

Considering this information we have, we need to assume that a lot of preferences will decide the last seats in every state.

Gee wouldn’t it be nice to have daily interim distributions of preferences, as they do in ACT territory elections!


Close seats – two weeks in

It’s now been two weeks since polls opened, and we are now getting very close to a conclusion of the House of Representatives count.

There is only one conventional seat still in play, which is Herbert. There will also be news today in Melbourne Ports which may either make the seat a serious seat in play or make it a clear Labor seat.


The Liberal National Party, at the time of writing late on Friday night, led by 12 votes in Herbert.

There are 200 absent votes outstanding. Labor has won 51.5% of these votes so far. If the remaining votes break the same way, Labor will gain six votes.

There are 399 prepoll votes outstanding. The LNP has won 53.5% of these votes so far. If the remaining votes break the same way, the LNP will gain 27 votes.

There are at least 44 postal votes remaining, with yesterday being the deadline for postal votes to be received. The LNP won 56.6% of the postal votes counted so far. Assuming there are no postal votes to be processed, 56.6% of the remaining votes would give the LNP an additional five-vote lead.

There are 446 provisional votes outstanding. Many provisional votes turn out to be not valid, and thus are not counted. If there are the same number of valid votes as in 2013 (320 votes), then there would be 154 votes. Labor won 60.8% of those votes cast so far, and if this continues this would give Labor a 30-vote lead.

So that’s 36 votes gained by Labor amongst absent and provisional votes, and 32 votes gained by the LNP amongst prepoll and postal votes. That’s a change of four votes in favour of Labor, which would leave the LNP with a lead of four votes.

“This is too close to call” seems a massive understatement.

Melbourne Ports

Up until now we haven’t had any information about whether the Greens are gaining enough preferences to overtake Labor for second place, apart from vague scrutineer reports. If the Greens overtake Labor, the current Labor-Liberal two-party-preferred count would become redundant, and either Liberal or Green will win the seat off Labor.

I’ve heard that the AEC will today be conducting an indicative three-candidate-preferred count between Labor, Liberal and Greens in Melbourne Ports to identify which candidates are in the top two. If Labor does not reach the top two, presumably we will need a fresh two-party-preferred count between Liberal and Greens to determine who is leading in the race to win the seat.

I expect that we’ll know more before the end of the day.


Close seats – Thursday morning update

We’re getting close to the end! In this update, I am calling two more seats for Labor, leaving only one seat up for grabs.

This leaves a total of:

  • 76 – Coalition
  • 68 – Labor
  • 5 – Others
  • 1 – undecided
SeatAbsentProvisionalPre-pollPostalCurrent Labor leadProjected Labor lead


There are still about 5000 absent and prepoll votes in Cowan. Labor has a substantial 1030-vote lead. Labor has only polled 45% of the prepoll vote, which is the reason why I’m expecting a drop in the Labor lead, but Labor should still win. This seat is called for Labor.


There are very few votes left in Herbert, but the race is still extremely close. Most of the remaining votes are prepoll and postal votes. Labor has won 46% of the prepoll vote and 43% of the postal vote, which is why the model suggests the LNP will increase their lead. If Labor could win 50% in both these vote categories they would likely win a slim victory, but that seems unlikely.


Most remaining votes in Hindmarsh are absent votes, and Labor has won 56% of these votes up until now. Because of this, Labor is on track to increase their lead. Called for Labor.


Senate preferences – how might they flow? Part 2

Following yesterday’s post analysing the NSW, VIC and QLD Senate counts, here’s my analysis of Western Australia, South Australia and an update of the Tasmanian race.

Read the rest of this entry »


Senate preferences – how might they flow? Part 1

Up until now, I’ve largely ignored preferences when working out who are the leading contenders for Senate seats. While the new system will likely make it impossible for parties on a tiny vote to jump ahead of a number of contenders, preferences are still likely to matter.

I visited the NSW Senate count on Saturday in Moorebank. Apart from seeing the new scanning technology in action (a story for another day), I got to observe ballots being data entered for a few hours, and got a sense of some general trends.

The main takeaway is that nearly everyone is preferencing, but very few people have preferenced more than six parties above the line. In my time, I only saw three ballots which had a ‘1’ and no other numbers above the line, and a similar number who had numbered more than six boxes. This observation has been confirmed by Kevin Bonham at the Tasmanian count and appears likely to be a national trend.

This tells us that preferences will definitely be flowing, particularly between parties with some similar policies, but that preferences won’t flow all the way. Plenty of votes will exhaust after electing one person and the last seat could still be decided on substantially less than a quota.

In this post I’m going to run through the latest figures in the Senate count and the general political make-up of the preference available to flow. We don’t really have detailed information on how minor party preferences will flow. Even if you were to scrutineer, there are so few of them in such a large bundle of votes that you don’t get a decent sample.

In this post I cover the races in NSW, Victoria and Queensland. I’ll come back tomorrow with a similar analysis for Western Australia and South Australia. I’ll do a separate update on Tasmania, as the below-the-line voting there makes that race unique and worth following more closely.

The other thing to note is that these primary vote totals are based on an incomplete sample: as more declaration votes are counted the numbers should shift. I also ignore the effects of below-the-line voting, as we don’t yet have any data on this for the mainland states.

For each state, I group together all the votes for minor parties (and small leftover quotas for larger parties) into four groups: left, right, centre and random. “Random” covers unclassifiable parties like the Citizens Electoral Council, the Socialist Equality Party, Online Direct Democracy, etc. “Centre” covers Derryn Hinch, Nick Xenophon Team, the Veterans Party and the Motoring Enthusiasts. The rest should be clearly left or right.

Read the rest of this entry »


Close seats – Tuesday morning update

Sorry there was no update yesterday – I only have limited time each day and I devoted it to my takedown of electronic voting in the Guardian.

In today’s update I’m calling three more seats for the Coalition giving them 76 seats. Labor is leading and look likely to win two more, with Herbert extremely close.

These three seats that are newly called for the LNP are the Queensland seats of Capricornia, Flynn and Forde. This gives us the numbers of:

  • 76 Coalition
  • 66 Labor
  • 5 others
  • 2 leaning ALP (Cowan and Hindmarsh)
  • 1 tight (Herbert)

I’ll also try to address the topic of Melbourne Ports at the end.

SeatAbsentProvisionalPre-pollPostalCurrent Labor leadProjected Labor lead


We have less than 1000 postal votes remaining and my model estimates around 1000 absent votes remaining, plus between 1400 and 1800 prepoll votes. The current LNP lead of 691 is expected to grow, and considering the relatively small number of votes left to be counted it’s not possible to see Labor recovering. Called for LNP.


Yesterday there was the big news that a bundle of 200 Greens votes (presumably mostly flowing to Labor as preferences) accidentally in the Liberal pile, and this shifted the trajectory significantly towards Labor. There are only 183 postal votes waiting to be counted, along with about 3000 absent votes and 2756 pre-poll votes. It looks likely Labor will win but I’d want some more of those votes to be counted before calling the seat.


We are down to less than 2000 postal votes. The AEC claims to only have 1500 remaining absent votes but this is about 2000 less than were counted last time – it seems likely some more votes are yet to be reported. The LNP looks to have won this seat. Called for LNP.


Very few postal votes left over and about 3000 votes of other types. While the margin isn’t massive there’s not much room to move. Called for LNP.


This is the closest seat in the country now. We’ve still got a few thousand postal votes remaining, along with just under 4000 absent and prepoll votes. Labor is winning a decent majority of the absent votes but narrowly lost the first batch of pre-poll votes. My model currently gives the LNP a majority of 20 votes, which is well within the margin of error. If Labor picks up ground in the pre-poll that will put them in front.


Labor look to be strengthening here after coming close to falling behind. There are very few postal votes left over, with around 4600 absent and prepoll votes yet to come in. Labor is winning the absent votes comfortably. We don’t have any pre-poll data but the model assumes a narrow Liberal majority. The model expects the Labor margin to grow from 583 votes to 850. Leaning Labor, but not ready to call.

Melbourne Ports

There was a story in Fairfax yesterday about Melbourne Ports still being in place. Unfortunately we don’t have much public data to go on. The seat will be won based on preferences which aren’t currently being recorded. All I will say is that it is possible for the Greens to overtake Labor, but is a mighty task. Beyond that we can’t do any more than note that Greens and Liberal scrutineers believe that Danby is likely to lose. Kevin Bonham has done his own analysis.


Close seats – Sunday morning update

I didn’t get a chance to do an update yesterday morning as I headed over to Moorebank to observe the NSW Senate count – I’ll come back to that topic soon with a post on the Senate race.

Seat summary

I’m calling Gilmore for the Coalition, as I flagged on Friday morning.

I’m also pulling Cowan back into the “very close” category from being “leaning ALP” thanks to a very poor batch of absent votes which has cut back on the Labor margin and put the Liberal Party in front in the projection.

Since I last updated, the LNP took the lead in Capricornia and Flynn, as expected. The projected lead in Capricornia has widened, but it has actually narrowed in Flynn.

My current numbers are:

  • 73 Coalition
  • 66 Labor
  • 5 others
  • 6 tight

If you take the projections at their current word, Labor would win Hindmarsh and lose the five other tight seats, although I suspect the model is overemphasising the Liberal vote in Cowan. Unless we see a major shift back to Labor on absent and pre-poll votes, it seems like we are heading for a Liberal majority government.

With Bob Katter, Andrew Wilkie and Cathy McGowan agreeing to give confidence and supply to the government, they are guaranteed of minority government. Probabilistically they are very likely to win their three remaining seats from the five outstanding seats.

Just a reminder that the first four columns of data in this spreadsheet represent the AEC’s official data on how many votes are left to be processed in each category (bearing in mind that pre-poll and absent numbers are expected to grow).

SeatAbsentProvisionalPre-pollPostalCurrent Labor leadProjected Labor lead


Labor has slipped behind the LNP in Capricornia in the last two days, and the projected LNP lead has grown. We’ve also started to see absent votes counted in Capricornia, and like in 2013 these votes are favouring the LNP slightly.


The first batch of absent votes in Cowan was very poor for Labor, which cut down on the Labor lead, and put the Liberal Party on track to win. It seems unlikely that the remaining absent votes will break in the same way, and if we see more you’d expect Labor to recover ground in the projection.


After a long decline, Labor has fallen behind the LNP in Flynn, and I expect the LNP to probably retain the seat. The projected LNP margin has actually dropped a bit since Friday morning.


The LNP lead in Forde has grown from 687 to 915, and the projected final lead has also grown slightly.


I’m calling Gilmore as the 73rd Coalition seat. The Liberal lead has only grown by 144 votes, but the projected lead has also grown, and they look clearly on track to win.


The Labor lead in Herbert has dropped from 449 to 302. I am projecting a Liberal win, but their projected lead has also dropped from 459 to 300. I can still envision a Labor win if they pick up in the remaining postal and pre-poll votes.


Labor has substantially recovered ground here since Friday – the Labor lead has grown from 68 to 247, and the projected Labor lead has substantially grown.