In yesterday’s post I explored some of the intra-party dynamics you get under Hare-Clark when multiple candidates from the same party are competing.
In this post I wanted to map out the relative strength of different candidates from within a single party across an electorate.
There’s some similarities in style to this post I did for the last ACT election in 2020.
Competition between candidates within a party isn’t just a matter of self-interested candidates trying to overtake each other. There are strategic benefits for the party.
Let’s take a scenario where a party has polled 2.4 quotas, split evenly between three candidates. A minor party has polled 0.7 quotas, entirely concentrated behind one candidate.
If the votes were cast under a Senate-style above-the-line system, the third major party candidate would have little chance of closing that 0.3 quota gap. But under Hare-Clark, all three of them are on 0.8, each able to gain preferences. It’s quite possible the major party would win three seats, and the minor party would miss out. To a certain point, evenly distributing votes amongst your candidates improves your chances of winning seats.
I have heard reports of parties deliberately spreading out their campaign efforts between each candidate by getting them to campaign in different parts of a seat. That is a common practice in Ireland, but it’s apparently used more quietly in Hare-Clark systems in Australia.
On the other hand, it’s possible different candidates have different support just because that’s where their supporters are based. I won’t try and judge why there is a geographic spread, but there is one.
This table shows what share of each party’s vote in each electorate went to the top-polling candidate, second-polling candidate etc:
There are some electorates where one candidate dominates the race. The most extreme case in 2021 was the Liberal vote in Bass, where then-premier Peter Gutwein polled over 80% of the party’s votes. Rebecca White in Lyons wasn’t far behind. Generlaly the most-popular Liberal incumbent in the other electorates had a big lead on the secondmost-popular. But the gap between first and second is close for the Labor groups in Bass and Franklin.
So for the rest of this post, I’ll run through some maps for the more interesting electorates and parties. I won’t bother with a map showing Peter Gutwein or Rebecca White topping their party’s votes everywhere.
Labor in Bass
Michelle O’Byrne polled almost 45% of the total Labor vote in Bass. Janie Finlay came second, with a third of the vote. This left not many votes for the other three Labor candidates.
O’Byrne won most booths understandably, but Finlay’s support base has a clear geographic basis, doing well in the western suburbs of Launceston, as well as the West Tamar region.
Labor in Braddon
Shane Broad polled a smaller percentage of his party’s vote than any other top-polling candidate. He only slightly outpolled Anita Dow, and Justine Keay came third with more than a fifth of her party’s vote.
There’s a strong geographic divide. Shane Broad won most booths in the eastern end of the electorate, in the areas around Devonport and Ulverstone. Dow topped the Labor vote in Burnie and areas further west.
Interestingly Justine Keay doesn’t really win many booths, but she came second in a lot of booths.
Labor in Franklin
Dean Winter and David O’Byrne polled very similar figures in Franklin, and had a clear geographic pattern. O’Byrne dominated in the Clarence area, while Dean Winter dominated in Kingborough. Toby Thorpe also managed to top a few booths around Huonville.
Liberals in Braddon
The current premier, Jeremy Rockliff, received almost half of the Liberal vote in Braddon, but the remaining vote was split fairly evenly, with the second, third and fourth-placing candidates all getting between 12% and 16% of the total Liberal vote. For this map I have excluded Rockliff to see the pattern for everyone else.
Roger Jaensch was the fourth-placed Liberal, and has a very clear support base in the Waratah-Wynyard council area, and spilling over into Burnie.
Adam Brooks dominated Devonport, while Ellis dominated Ulverstone and surrounding areas. The westernmost areas were split between Brooks and Ellis.
Liberals in Clark
Like in Braddon, the Liberal vote in Clark was dominated by Elise Archer, but there was a reasonably similar vote for the next two candidates: Madeleine Ogilvie and Simon Behrakis. Ogilvie ended up narrowly defeating Behrakis, and he had to wait for Archer’s retirement last year to enter parliament.
The seat of Clark is divided down the middle by the council boundary between Glenorchy (north) and Hobart (south), with a small part of the Kingborough council at the very southern end.
Ogilvie dominated most booths in northern Clark, while the City of Hobart was split between Ogilvie and Behrakis areas.
Independents in Clark
Okay so this isn’t technically competitors within a party, but the race in Clark was interesting as there were two strong independents. Until this election, strong independents have been uncommon.
These two independents, Kristie Johnston and Sue Hickey, also had different geographic centres. Johnston was Mayor of Glenorchy and Hickey was a former Lord Mayor of Hobart.
And what do you know, there is a clear divide. Johnston dominated all of Glenorchy, but her territory also spread down into northern and western Hobart. The larger Johnston area is reflected in her slightly higher vote, but both candidates polled votes everywhere. I clicked around the map and was unable to find a booth where the leading candidate doubled the other’s vote.
Of course, since 2021 there has been quite a bit of change. A number of MPs have retired or left their parties, and the 7-seat electorates will change the race. But these may give some insights to some candidates’ strong areas.
Have there been quite clear instances where the parties vote being spread amongst multiple candidates has caused them to win an additional seat? Or is it mostly theoretical?
I suspect Michelle O’Byrne not winning the ALP intra-party primary vote in Bass is related to the fact that West Tamar was NOT in Bass when she was a Federal MP. Seems like it was in Lyons at the time, but moved to Bass in the 2017 redistribution.
(In West Tamar. Not the whole seat.)
There’s some examples in this post https://kevinbonham.blogspot.com/2012/10/getting-gininderraed-another-for-hare.html
Not mentioned in there – in 2021 Labor might have lost their second seat in Bass to the Greens if they didn’t preselect high profile Janie Finlay.
@Leon
Janie Finlay had recently (August 2020) ran in the upper house seat of Rosevears which covers West Tamar. She did well – only lost to the Liberal candidate by 260 votes. I think that’s why she did better there than O’Byrne in the 2021 election.
In Tasmania do party candidates ever issue their own HTV cards separate from the party HTV? It would say Vote 1 for Mr X and then all the other candidates from that group – without of course numbering because of the Robson Rotation.
August T, Ben’s account of the Gininderra effect is called that because it happened in the electorate of Gininderra in the ACT. I forget exactly which year. Two ALP candidates both with votes close to a quota were just ahead of the Green candidate and won both seats.
Doug, that situation occurred in the 2012 ACT election. Labor candidates Chris Bourke and Yvette Berry both had 0.7 of a quota, which was more than the 0.6 quotas held by the Greens candidate Meredith Hunter.
Therefore, even though the partial Labor quota of 0.4 was less than the partial Greens quota (0.6) at the final exclusion count, the individual Labor candidates had a higher vote and were elected first. This confirms that Hare Clark is more of a battle between candidates rather than a straight party contest seen in the Senate.
This shows that in some cases under Hare Clark, it is better for the major parties to run a few strong candidates who can split the vote fairly evenly rather than concentrating it all under one single ‘lead’ candidate.
In the 2021 election, after the recounts, some of the vote filtering/washing went at a higher percentage to some of the parties depending on who’s votes were distributed first and where their preferences tended to flow.
Have you done any voting analysis on the preferential flow patterns from last election and what that might look like if some of the minor candidates preferences fill up quotas for the others sooner than the actual split of 1st preference votes?
The reason I ask this is that in 2021 the following happened to preferences:
In Bass, Liberal 1st preferences was equivalent to 3.6 quotas, after distribution it went down to 3 quotas.
In Braddon, Lib 1st preferences was equivalent to 3.43 quotas, after distribution it went up to 3.77 yet they secured only 3 seats,
In Clark Libs had 1st preferences equivalent to 1.91 quotas, went up to 1.95 after distribution, and were given 2 seats.
In Franklin Libs 1st pref was 2.54 quotas, went up to 2.64 after distribution, they were awarded 2 seats.
In Lyons Libs 1st pref was 3.07 quotas, went up to 3.21 after distribution, they were awarded 3 seats.
In total across the state they actually secured 14.55 quotas worth of total 1st preference votes, this was adjusted down to 13 seats.
For the ALP it was 7.91 quotas worth of 1st preference votes, yet this ended up becoming 9 seats.
For the Greens it was 3.75 quotas worth of 1st preferences, this ended up at 2 seats in the parliament.
The last seat went to Kristie Johnston in Clark.
The reason I’m asking is it seems that the preference flows a lot quicker to ALP than either LP or the Greens, this effective means the ALP quotas fill up sooner in the count than the other 2 parties. Are you able to come up with modelling that might apply to this analysis to give a bit of an indication of what might happen on Saturday?
Ucomms poll analyst had the following assessment made on 1st preference % distribution:
Analyst Kevin Bonham said the Liberals would be expected to win 14 of the 35 seats if this poll is accurate, Labor ten, the Greens four, the JLN 2–3 and independents 4–5.
This could well be different depending on who’s votes are “washed” first in each recount don’t you think?
Interesting, I haven’t looked into that Cameron.
I’d be interested to see your take on the flows and their trends.
In Bass in 2021, Gutwein was 1st, followed by Courtney (both Libs and all because of Gutwein’s personal vote. Then LP got the 3rd seat on the 20th count. ALP got their 1st on count 44, and their 2nd seat on count 52.
In Braddon, Libs got the 1st seat with an massive personal vote again, but the second seat in the electorate wasn’t declared until count 32 by which stage over 2300 votes had been exhausted in the rotations. This was an ALP seat. ALP got the 3rd seat at the 33rd count, followed by LP who got the 4th and 5th seats on count 34 but with 93.44% & 93.12% of a quota respectively. (The reason they didn’t do a further count was because the next person to exclude was another LP candidate who had at that stage received 90.44% of another quota therefore it wasn’t possible for it to have gone to another party)
In 2021 in Clark, Cassy O’Connor (greens) was elected 1st but not until the 9th count, then Libs were second on count 18, followed by ALP on count 23, then finally ALP and IND declared following count 33.
In Franklin, LP got the 1st seat on the back of an over quota of personal vote to their lead candidate. Then Greens secured the 2nd seat on count 20. It then took until count 40 for the final 3 seats to be declared. (2 ALP, 1 LP)
In Lyons, ALP and LP both got a seat on the 1st count due to high personal votes. LP got seat 3 on count 32, followed by ALP seat 4 on count 41, then finally LP seat 5 on count 44.
with so many extra candidates in each electorate this time around, and with the quotas so much lower it will be very interesting to see how it plays out across the counts.