How the AEC is adapting Eden-Monaro for Covid


The AEC has published its plan for how it will adapt the Eden-Monaro by-election to account for the COVID-19 pandemic.

Some of the stuff will be familiar to anyone who has visited any public place in recent months. Social distancing rules will be applied at polling places and the ballot draw, and hand sanitiser will be available when you turn up to vote. There will be limits on how many people can be inside each booth which will likely extend lines for voting.

The AEC will only receive nominations by appointment (no more just rocking up to the office) and voters will be provided with single-use pencils when they vote (although bringing your own writing implement is encouraged).

The AEC will not conduct mobile polling at nursing homes, instead encouraging those inside to use postal voting.

The AEC has clarified that they do not have jurisdiction over campaigning outside six metres of the entrance to the polling place, and thus have not tried to impose any rules limiting the handing out of how-to-vote cards, although they put the onus on candidates and parties to abide by whatever NSW Health rules are in place at the time of the by-election. They have agreed to post a website link for each candidate as part of the candidate information on the AEC website, but they have not gone further as some have suggested by providing how-to-vote information in other ways.

Finally, they anticipate a big increase in postal and pre-poll voting. There was some chatter about possibly allowing pre-poll votes to be counted before 6pm on election night to speed up the publication of results (it’s worth reading Antony Green’s blog post on this topic), but this won’t be happening.

So you should expect a significant delay in results on election night, and if it is close we likely will need to wait a few days for pre-poll votes to be counted to get a better sense. A very close result may take weeks as postal votes can arrive up to 13 days after election day.

Finally, I recommend checking out Michael Maley’s paper for the Electoral Regulation Research Network about how electoral management will be affected by Covid-19. I haven’t had a chance to finish it but I’m looking forward to it.

Eden-Monaro by-election set for July 4


The federal by-election for the seat of Eden-Monaro has been set for July 4, just one day after the Liberal Party announced their candidate.

The Liberal Party has preselected Fiona Kotvojs, the same candidate who ran in 2019. She will be facing off against Labor’s Kristy McBain, who was until recently the mayor of Bega Valley Shire.

Be sure to check out my guide to the Eden-Monaro by-election.

There will undoubtedly be some news in the coming weeks about how the by-election will be affected by the pandemic, and I will return to discuss those measures when we know what they are.

What I’ve been reading: PR in NYC


I’ve been working behind the scenes on my guide to the Queensland election, but I’ve also been taking the opportunity of this quiet election period to do some reading, including on some topics related to electoral analysis. I thought I would put together some blog posts summarising these books and journal articles.

First up, I just finished reading Defining Democracy by Daniel Prosterman. This book covers the history of New York City’s brief use of the Single Transferable Vote proportional representation system to elect its city council in the 1930s and 1940s.

Kelly resignation opens up by-election in Eden-Monaro


The triggering of a federal by-election yesterday for the seat of Eden-Monaro will be a test for the major parties, but also will likely be the first test of how our voting will be affected after the peak of the Covid-19 pandemic in Australia.

Read my guide to the Eden-Monaro by-election.

Australia has been lucky to have the pandemic hit during a relatively quiet period for elections. It did coincide with council elections and state by-elections in Queensland, and it forced the delay of Tasmanian upper house elections originally scheduled for tomorrow. But apart from these there are no more elections scheduled until a rush of elections from August to October, which includes one state election and two territory elections.

It’s not clear how quickly the by-election will be held – the government has some discretion to delay – but when it is eventually held it seems likely that some things will change compared to past elections. We saw a massive surge in postal and pre-poll voting at the Queensland council elections. Eden-Monaro already had a high rate of pre-poll voting, with over 41% of all votes cast via pre-poll in 2019.

I will return to the theme of how Covid-19 may affect how we vote in a future post, but it is also worth pointing out that there isn’t actually much evidence of the disease in the community at the moment, and particularly in the council areas which make up Eden-Monaro.

This handy chart by Nick Casmirri organises NSW local government areas by region and colour-codes them by the recency of Covid-19 cases.

There have been no cases in the last three weeks throughout the Snowy Monaro, Snowy Valleys and Yass Valley council areas which make up the western end of the council. No cases have been reported for two weeks in Bega Valley. There have been more recent cases in the Queanbeyan area but even those are not of an unknown origin.

So unless we see a second wave of cases, it seems likely that there won’t actually be an immediate health crisis coinciding with a by-election, but I would still expect changes in how voters and campaigns behave out of an abundance of caution.

In addition to being a moment in the Covid-19 story, this by-election is a critical moment in federal politics.

You would have seen a lot of people talk about how no government has won a by-election from an opposition since 1920, but I don’t think that statistic really tells us much.

There just haven’t been that many by-elections overall, and they tend to be in relatively safe seats. Most by-elections are caused by the sitting member choosing to retire, and that doesn’t tend to happen in marginal seats.

Let’s look at the by-elections held over the terms of the current government and the last Labor government.

Labor faced five by-elections in Coalition seats in its first term in government. All were held by senior Howard government ministers and all took place in the first year of the government. They tended to be safe seats, with the most marginal being Gippsland, which was held by the Nationals by a margin of 5.9%. Labor contested Gippsland, suffered a negative swing, then sat out the other four. No by-elections were held for the remaining five years of the government, with MPs particularly avoiding resignations during the tight hung parliament after the 2010 election.

Three by-elections were held during the first term of the Coalition government, one of which was in the Labor-held seat of Griffith. There was a swing of 1.25% away from Labor with the departure of longstanding local member Kevin Rudd, leaving Labor with a 1.8% margin. The by-elections held in the last term were mostly held due to section 44 citizenship problems, which makes them quite different.

So in the last thirteen years we only have one case of a marginal opposition-held electorate being contested by the government, in Griffith, and in that case there was a swing to the government, one that would be big enough to win Eden-Monaro. It’s just not enough of a sample to say “this is a thing that doesn’t happen”.

As to Eden-Monaro: it’s obvious that Mike Kelly had a strong personal vote. It doesn’t seem worth the trouble to try and quantify it, but something in the area of 5% seems plausible. Remember that he only held Eden-Monaro by less than 1% in 2019.

So this seat really could go either way, and it would be silly to call this by-election one way or the other.

It’s worth noting that this is quite possibly the single electorate hit hardest by last summer’s bushfires. The south coast was hit hard, but so was the Snowy Mountains and areas to the west including Tumut. The electorate also includes the towns of Queanbeyan and Yass which surround Canberra. They were not directly threatened by fires, but Canberra experienced worse and longer-lasting bushfire smoke than any other major city. We could well see big shifts thanks to voters judging the performance of Scott Morrison during those fires.

This will also be the first political test for the federal government since the onset of the Covid-19 pandemic, so this could also play out in either direction.

Finally I should note that the choice of candidate for the Coalition parties could have a big impact. This electorate covers the state electorates of Monaro and Bega, which are held respectively by the Nationals leader (and deputy premier) and the deputy Liberal leader, arguably the two most important figures in the state government outside of Gladys Berejiklian.

It seems unlikely that Andrew Constance will run, but Monaro MP John Barilaro appears to be seriously considering a run. The Nationals came a distant fourth in Eden-Monaro in 2019, and have never held the electorate at a federal level, but that could change if such a prominent figure were to run.

If Barilaro runs (win or lose) that will trigger a state by-election in his state seat, which looked very marginal before a 9% swing to him at the 2019 state election.

All quiet, but a lot going on


I have been planning to post for a few weeks about the plans for this website going forward following the conclusion of the Queensland council elections.

I’m not going to be actively posting on the website, or publishing any new episodes of the podcast, for the next little while, but a lot of stuff will be going on in the background.

Podcast #35: Queensland local elections wrap-up


Ben is joined by Alexis Pink to wrap up the results of the Brisbane City Council election and the Queensland state by-elections, including how these elections were influenced by the COVID-19 pandemic.

You can subscribe to this podcast using this RSS feed in your podcast app of choice, but should also be able to find this podcast by searching for “the Tally Room”. If you like the show please considering rating and reviewing us on iTunes.

Brisbane 2020 – election-day turnout crashes


The election results from Brisbane have been pretty messy so I haven’t tried to stay on top of all of it. At the moment it looks like Labor, the Greens and the independent have retained the 7 wards they currently hold and the LNP is leading in all 19 of their wards. There are three wards the ABC has listed as “in doubt” while we wait for some more concrete results and those are the three I’d be watching.

I wanted to look at a different angle, so I downloaded all of the voting data by booth (not as easy as I would have liked) for Brisbane and compared it to the 2016 results to get a sense of how turnout has changed.

When you look by category, it looks like only about one quarter of eligible voters would have cast a vote on election day, while the pre-poll rate tripled to more than 36% of eligible voters. With more postal and telephone votes yet to be counted, it seems possible that early votes will make up more than half of the eligible electorate, and about two thirds of total votes.

Queensland election – state of play one day later


I was planning a deeper analysis looking at the proportion of votes which have reported, and what it tells us about how low the turnout was on election day but unfortunately I’ve had some issues with the data coming out of the ECQ, so I’m going to keep this post a bit simpler as a summary of what we know.

It looks likely that Adrian Schrinner has been elected to his own term as lord mayor, a fifth term in a row for the Liberal National Party. His primary vote has been climbing as more votes have reported, and he now sits on 47.2% of the primary vote, with Labor’s Pat Condren trailing on 31.1%. This translates into a swing of 6.3% against the LNP and 0.9% against Labor.

The Greens appear to have performed well all across this election, which was somewhat obscured by the technical difficulties in election results last night and a presumption that some results were skewed early on. The Greens lord mayoral candidate Kath Angus is currently on 15.6%, a swing of 5.4% compared to 2016.

About 60% of turnout has been counted so far in the mayoral race, and the sample is fairly even across the city (although Calamvale is very underrepresented in the sample).

We don’t have a preference count yet but it looks certain that Schrinner will win with a reduced majority.

The Greens also did well in a number of key wards. Sitting Greens councillor Jonathan Sri is sitting on 48% of the primary vote: a swing of 15%. We don’t have a preference count but he should have no trouble clearing the 50% barrier.

The Greens also gained a large swing in Paddington ward, which overlaps with the Greens state seat of Maiwar. Donna Burns gained a 10.9% swing. We will need the preference count to know who wins here – the ABC’s preference estimates predict the race is too close to call. The Greens have also moved into second place in the Central ward but will need to do well on preferences to win.

At the moment the LNP has won twelve wards, Labor has won five and others have won two. This leaves seven in play.

Apart from Paddington, there are six other contests still not decided.

The ABC has the LNP leading in Bracken Ridge, Central, Holland Park and Northgate, while Labor is leading in Calamvale and Enoggera, but it’s important to emphasise that these are all based on preference estimates: we still don’t have preference flows in any of these seats. We are also still missing most pre-poll figures and the postal votes are yet to be counted.

These early votes make up a massive part of the electorate, and we don’t know how different they will look in the current environment. It’s conceivable they could shift some of these races back to the LNP.

We also don’t know what impact the lack of how-to-votes will have on preference estimates, but there have been reports of Greens preference flows to Labor in Currumbin being weaker than you would expect in such a race.

If those seven undecided wards all go to the leading party that will result in a council with two more Labor councillors than the current council, and will leave Labor slightly closer to taking control of the council, but won’t materially impact on the LNP’s grip on power.

I will return to this topic once we have more information, later in the week.

Queensland elections – election night live


7:07am – Overall we are seeing a trend of swings against the LNP and to the Greens, with Labor gaining in some places and losing in others. The Greens in particular appear to have strengthened their position in the inner-city wards of Coorparoo, Paddington and The Gabba, where it looks likely that Jonathan Sri will be re-elected with a much bigger vote. I should warn, however, that all these figures are very preliminary and do not include preference counts. We don’t know how the remaining postal and pre-poll votes might vary from the election day vote.

7:05am – We have some primary votes from each council ward but without preference counts and without comparing where the booths have come from it’s hard to compare. I’ll do some work on this later today. For the eight wards I’m most interested in, this is the current situation:

  • Coorparoo – With 14.53% counted, LNP is down 4.3%, Labor down 3.4%, Greens up 10.2%. Very small sample so I would want to see more.
  • Doboy – With 13.16% counted, the LNP is on 53.5% in a two-horse race.
  • Enoggera – With 23.18% counted, the LNP is down 7.2%, Labor is up 1.9%, Greens are up 0.7%.
  • Holland Park – With 21.26% counted, the LNP is down 4.5%, Labor is down 4.6%, the Greens are up 9.6%.
  • Northgate – With 24.32% counted, the LNP is down 3.3%, Labor is up 0.3%, Greens are up 2.9%.
  • Paddington – With 27.89% counted, the LNP is down 5.4%, the Greens are up 13.7%, Labor is down 8.4%.
  • The Gabba – With 38.62% counted, the Greens are up 16.2% to almost 49%, Labor is down 4.1%, the LNP is down 9.6%.
  • The Gap – With 22.09% counted, the LNP is down 2.2%, Labor is down 2.4%, the Greens are up 3.8%.

6:48am – Over 40% has been counted for lord mayor of Brisbane and there appears to have been a swing of about 8% away from the LNP, but none of that has boosted the Labor vote. The Greens are still doing quite well, polling over 15%. It’s not clear if the data is skewed to particular areas. I’ll look into that for the post later today.

6:45am – We have 56% of the vote counted in Bundamba and it looks similar to where we were last night. Labor is about 15 points clear of One Nation so should win on Greens preferences.

6:41am – We have close to 40% of the results in Currumbin now, and the LNP leads Labor by about 3%. Antony Green reports that party scrutineers have received a paper version of two-candidate-preferred results with the LNP “a handful” of votes ahead. There are a lot of postals and pre-polls to come and it seems likely they will favour the LNP, so they are favourites here, but those postals and pre-polls may not behave the same as always due to the pandemic so I’d want to wait and see.

6:35am – I’m just coming back here with a few updates on what we’ve learnt overnight. I won’t be liveblogging today but will return with a post this evening.

9:02pm – I’m going to be stopping the liveblog here. I’ll return tomorrow afternoon with an update. We now have almost 11% of the primary vote for the Brisbane lord mayor’s race and it still appears to have a very high Greens vote – I’m still assuming this is due to the selection of booths and doesn’t reflect a final result. It does look likely that Labor will retain Bundamba but we have not gained any extra data in quite some time. There appears to be a problem with the Currumbin count. Elsewhere it looks very clear that Tom Tate will be re-elected as mayor of the Gold Coast.

8:31pm – We’re really not getting very much. I’ll keep checking in, but expect more tomorrow.

8:20pm – There appears to be a problem with the Currumbin voting results so I wouldn’t take the small number of votes recorded there too seriously.

8:16pm – With almost 35% counted in Bundamba, Labor’s Lance McCallum has slightly increased his primary vote lead over One Nation: 42.3% to 27.4%. He’s looking pretty good here.

8:14pm – Almost 20,000 votes have been counted for mayor of the Gold Coast and Tom Tate is still above 50%.

8:13pm – So we now have a lot more data for the Brisbane lord mayoral race. Adrian Schrinner is currently leading on 43.4%, with Labor’s Pat Condren on 30.4%, and the Greens’ Kath Angus on 18.7%. This is off a sample of almost 6% which appears to lean towards the best Greens areas in the western suburbs, so it’s too early to say. At the moment this is a swing of about 10% away from the LNP and a small swing against Labor.

7:46pm – We won’t be getting any preference counts tonight for lord mayoral and council races in Brisbane (I assume the same is true in other councils) so we may be able to call races where the leading candidate is far out in front but close races won’t be decided. We should expect preference counts in Bundamba and Currumbin.

7:45pm – It will be quite some time before we have results that are sufficiently complete and details to know exactly how many people voted using each method in this election, but I put up a post this afternoon analysing how we should expect a massive crash in the number of people voting on election day. Since posting this I have heard from a number of people on the ground that a two-thirds drop in attendances sounds plausible.

7:42pm – Incumbent Gold Coast mayor Tom Tate is well out in front with 53.6% of the primary vote, with his main rival Mona Hecke a distant second on 24.4%, off almost 10,000 votes, or 2.8% of the total. This count is still too early but he is in a good position.

7:41pm – We do have over 1% counted for three other big city mayoralties. Teresa Harding is well out in front with 43% in Ipswich (off a sample of 1671 votes). 2345 votes have been reported in Moreton Bay and Peter Flannery is narrowly in the lead with 26%, with Adrian Raedel and Chris Thompson both over 20%.

7:39pm – Meanwhile just 460 votes have been reported for the Brisbane lord mayoralty.

7:38pm – The vote count is very early in the City of Brisbane. I’m following eight wards tonight (Doboy, Northgate, Coorparoo, Holland Park, The Gap, Enoggera, Paddington and The Gabba) and we only have vote counts from one of them. We only have 62 votes in Paddington so I won’t bother describing that count further. This looks likely to be a slow count.

7:30pm – The number of votes in Currumbin is much smaller – only about 3.5%. At the moment LNP candidate Laura Gerber is leading with 47.9% ahead of Labor’s Kaylee Campradt on 37.6% but the numbers are very small.

7:28pm – We have small numbers of votes reporting from the two state by-elections. In Bundamba, Labor’s Lance McCallum is well out in front with just over 25% counted, with 41.9%. One Nation’s Sharon Bell is in second place on 28.2% while the LNP and the Greens are both in the mid-teens. We don’t have any preferences, but on those numbers it’s very likely Labor would win.

6:00pm – Polls have just closed in the Queensland council elections. I’ll be live-blogging tonight, although it’s worth noting that a very high rate of pre-poll and postal voting means we will likely need to wait to know who is winning close races.

Queensland council elections – how low can election day go?


We are still waiting for results from today’s Queensland council elections, but this election is on track to look very different to past elections, with only a small proportion of Queenslanders voting on the day.

The Electoral Commission of Queensland has estimated that about 1.2 million Queenslanders have cast pre-poll votes over the two weeks of voting. 570,000 Queenslanders also registered for a postal vote, and 40,000 people registered for telephone voting (although this does not mean all these people will cast a vote).

When you look at the vote counts from the 2016 election, you can see that these figures suggest a total transformation in how Queenslanders vote, and suggests that the proportion of the electorate turning out today would be dramatically reduced, mostly driven by the COVID-19 pandemic.

This table shows the number of voters by method of voting at the 2016 election. I have just looked at mayoral votes, since there were less voters in uncontested electorates for their mayoral election. There were 18 small councils which only used postal voting in 2016, and 8 other councils where the mayor was elected unopposed. I have excluded them from the count.

There was turnout of 83.2% in the councils analysed in 2016.

Vote categoryVoters% of votes% of enrolled
Postal & unenrolled221,6789.17.6
Did not vote491,35516.8

Ordinary votes made up 63% of the total vote, and over half of all enrolled.

If you group into votes cast early and votes cast on election day, election day made up 70% of the vote, and 58% of total enrolment. Early votes made up 30% of the vote and 25% of total enrolment.

Now let’s compare these figures to what we know about 2020. We know that about 3.2 million voters are enrolled.

There has been 1.2 million pre-poll votes and 570,000 postal vote applications. That translates to 37.5% and 17.8% respectively (although that’s an upper ceiling for postal votes). Plus telephone voting has jumped from less than 500 votes to as many as 40,000, or 1.25% of the total electorate.

This adds up to possibly as many as 56% of the total electorate having already voted before today.

I think you also have to assume a drop in overall turnout, with some voters simply failing to turn out due to concern about the risk to their health. So if turnout was to drop to, say, 75% of the electorate (down from 83% in 2016), you could imagine as little as one in five eligible voters may have cast their vote today. This would be barely a third as many as voted on election day in 2016.

There has been an ongoing trend of increasing numbers of voters choosing to cast their votes early (although mostly through pre-poll voting – postal voting has been largely stable). But I don’t think we would ever expect to see these kinds of trends outside of a global pandemic.