Mapping the candidate vote in the ACT


In ACT elections, it’s not just the total vote for your party that matters. It’s also how that vote is distributed between your candidates, particularly if you are one of the major parties.

Kevin Bonham coined the term “Ginninderra effect” to refer to what happened in the electorate of Ginninderra in 2012, where Labor polled 2.39 quotas compared to 0.61 quotas for then-Greens leader Meredith Hunter, but won three seats. A slightly less dramatic but similar result took place in 2016, with Labor polling 2.48 quotas and the Greens polling 0.58, but Labor winning three seats.

Labor can win three seats in these scenarios because their vote is distributed relatively evenly amongst their candidates, in particular amongst their top three candidates. Vote exhaustion means the last few seats are usually decided with less than a full quota, so three evenly-positioned Labor candidates can outpoll a Greens candidate with a vote higher than Labor’s third quota share.

In this post I’m going to explore how the vote for different candidates for each major party varies across an electorate in a variety of ways.

This table shows the vote for each of the major party candidates in the five electorates as a proportion of the party’s total vote, with the candidates ranked by their total vote.


There’s great variety in how much the vote is concentrated. More than half of the vote for the party was concentrated in one candidate for Labor in Kurrajong and for the Liberals in Murrumbidgee, where party leaders Andrew Barr and Jeremy Hanson were running.

The leading candidates also tended to perform more strongly in Yerrabi, but in the other seats no candidate gained more than 30% of the party’s total vote.

The vote was particularly even in Brindabella and Ginninderra, with a gap of less than 8% between the first candidate and the third for both major parties in both these electorates.

I’m going to zoom in on those two electorates, mapping out the distribution of the vote and looking a how it varies by sub-area. It’s worth noting that all of this data is pre-redistribution. No candidate can run in more than one seat so it’s not possible to calculate redistribution estimates at the candidate level.


Firstly in Ginninderra, let’s start with Labor. Yvette Berry topped the primary vote with 10.1%, while Kim Fischer polled the lowest Labor vote with a still-respectable 7.1%. All five Labor candidates achieved the highest Labor vote in at least one polling place.

This map shows which of the five candidates did best in each booth. You can click on a booth to see the vote for all five of the candidates, and you can toggle to see the same for the Liberals in Ginninderra.

Berry performed best in the north-west of the electorate, in particular in Fraser, Holt and Macgregor.

Gordon Ramsay came second across the electorate but only won a single booth (Latham). He tended to do well in the same areas as Berry.

Tara Cheyne did best in the east of the seat, including in Aranda and Belconnen town centre.

Fourth-placed Chris Bourke narrowly won a single booth: Weetangera. He came second in all of the booths won by Cheyne in the east of the seat.

Last-placed Labor candidate Kim Fischer managed to top the poll in three booths in the heart of the electorate, including 16.8% in Florey.

This table shows the vote by sub-area for each of the Labor candidates, as a share of the total Labor vote. These sub-areas are the same as those used in my 2016 and 2020 guides.

Share of Labor primary vote in Ginninderra

Other votes23.822.817.718.517.2

Robson rotation naturally lends itself to evenly-distributed votes, since candidate order is mixed up. If every voter just voted down their party ticket from top to bottom you would expect almost perfect equality between candidates in the same party. Still, it is interesting that the variations follow a geographic pattern. I have heard from a number of sources that Labor encourages their candidates to focus on particular parts of an electorate, which may contribute to this pattern.

It’s also worth noting that Labor did better overall in the west of the electorate than the east, and two of the three Labor candidates to win seats were both focused on the west. Did they win because they were assigned the west to work on, or was the vote higher there because these candidates were stronger? I can’t answer that question, but it’s worth pondering.

The Liberals in Ginninderra had a similarly flat pattern. You can toggle to see the Liberal pattern in the above map. The overall Liberal vote was lower, with the candidates ranging from Ignatius Rozario on 3.6% to Vicki Dunne on 9%.

Dunne won a majority of booths, and she was particularly dominant in the centre of the electorate.

Sweeney came second on primary votes but did not win a single booth. He tended to do best in the centre, where he polled higher than Kikkert, and in the north, where Dunne and Kikkert were neck-and-neck.

Kikkert came third and won four booths, including Macgregor and Fraser in the north-west.

Fourth-placed Fisher won one booth in Aranda.

Share of Liberal primary vote in Ginninderra

Other votes32.420.519.417.410.4


Brindabella was the only electorate where the Liberals won three seats, even though they polled more primary votes in Murrumbidgee. They achieved this result with just 2.5 quotas.

The following map shows which booths were won by which Liberal in Brindabella, and it can also be toggled to show the Labor vote, which was also very evenly distributed.

Zed Seselja had led the Liberals in Brindabella in 2012, winning 1.8 quotas in his own right and dominating the Liberal vote share. His absence in 2016 may have helped contribute to an even field.

Andrew Wall won just over half of the booths and topped the primary vote. He polled 12% while the lowest-polling Liberal polled 3.6% – not as tightly-bunched as we see with the Labor vote in these electorates.

Mark Parton came second, winning five booths, most of which were in the centre of the electorate.

Nicole Lawder won a single booth (Wanniassa Hills) at the north-western end of the electorate, while fourth-placed Ed Cocks won the Gilmore booth.

When you divide the electorate into the same sub-areas I used in my guide, you see that Wall topped the primary vote in the north-east and the south, but Parton outpolled him in the north-west.

Share of Liberal primary vote in Brindabella

Other votes32.420.519.417.410.4


The Labor vote in Brindabella is also very interesting. Their leading candidate, Mick Gentleman, polled 8.5%, while the other elected candidate, Joy Burch, polled 8.2% – almost exactly the same. The lowest-polling Labor candidate managed 4%.

Despite topping the Labor vote, Gentleman didn’t win too many booths. He won just five booths, all in the south of the electorate.

Burch came second but only won one booth.

Third-placed Werner-Gibbings won six booths, mostly in the centre of the seat. Fourth-placed Drake won four booths, all at the northern end of the seat.

How can this be? The following table can help illuminate. Gentleman’s vote was very concentrated in the south of the electorate, while Burch’s vote was much more even. Werner-Gibbings did best in the north-east of the seat, which is also the poorest area for Labor.

Share of Labor primary vote in Brindabella

North West23.923.320.222.110.4
North East18.823.528.620.09.2
Other votes26.728.017.418.29.7

Things could be different in 2020. 2016 was an unusual election, with the Assembly expanding from 17 to 25 members, and the creation of two new electorates. The two electorates I’ve featured were both pre-existing, but they both lost their northern ends to new seats and lost some of their sitting members.

All but two sitting MLAs are recontesting in 2020, so most districts will have more incumbent MLAs running. The redistribution has also been far less dramatic, so most incumbents will be running in areas they have represented for the last four years. This may mean that incumbents do better and have support across an electorate, and it is harder to evenly spread your vote out amongst your candidates.

Liked it? Take a second to support the Tally Room on Patreon!


  1. Thank you for doing this, Ben. It’s very interesting to see. I wonder if the vote is (potentially) more evenly spread across Brindabella and Ginninderra because it’s possible these electorates have more stable populations (i.e. less population turnover)? According to the article below, Brindabella skews older, and Ginnindera has a higher proportion of low-wage earners. As such, if people have been in the ACT longer, could it be that they understand the Hare-Clark system better?

Comments are closed.