The Greens in the 2018 Victorian Election

The resignation of Richard DiNatale from the Senate and need for the preselection of a replacement prompted me to go back and reconsider the 2018 Victorian state election results as it seemed that some unsuccessful candidates there might be candidates for the Senate seat.
As it turned out I got sidetracked and didn’t finish the post until the preselection was over and won by Lidia Thorpe.

The basic story of the election was the the Greens lost a lot of seats, and some have used that so suggest a big loss of electoral support for the Greens, however a look at the percentages does not support this idea, so a more nuanced view is required.

This is not an attempt to paint it as some sort of good result, there was a decline in vote and the Greens failed to turn their gains in 2014 into an increase in vote. However there is some important context. A strategy of trying to win lower house seats has generally been successful and may have come at the cost of upper house seats, and furthermore Group Voting Tickets and preference deals put the Greens at a disadvantage in the upper house.

The lower house overall

The table below shows the vote percentage and number of seats won by the Greens for the 2018 election, as well as two previous elections. (All election numbers in the post are from the Victorian Electoral Commission)

YearGreens Vote (%)Greens seats won
201810.713
201411.482
201011.210

The vote percentage was lower in 2018 than the previous elections but only by a small amount. Looking at this you would note a small dip in the vote but it would hardly look like a disaster, and the number of seats is increasing.
Not shown here is that the Greens actually held three lower house seats going into the election as they won Northcote at a by-election in 2017, so there was no net gain there, but actually they lost one they held (Northcote) and gained one new one (Brunswick) so there was a lower house Greens MP, Lidia Thorpe, who lost a seat.

Over this period the Greens have put considerable effort into winning lower house seats, whereas previously the upper house had been the focus. The lower house consists of single member electorates, so you need to get 50% (possibly after preferences) in order to win a seat, this is why, despite having over 11% of the vote overall in 2010 the Greens had no representation in the lower house. Since then they have no doubt benefited from Adam Bandt holding the Federal seat of Melbourne with a strengthening vote each time, and this has flowed on to strong votes in this area in state seats. The state seat of Melbourne was won in 2014, along with a somewhat surprising victory to unseat the Liberal Party in Prahran (which does not overlap Melbourne and is perhaps more of a sign of the tactical targeting of lower house seats).
In 2018 an additional seat, Brunswick was won, however the seat of Northcote, won in a by-election was lost. These significant wins of seats have been against a background of a slight decline in overall vote, this may well be a result of resources being allocated towards winnable seats and resulting in a slight decline across all of the other seats.
The big picture for the Greens in the lower house has been one of success, going from zero to 3 seats without a signifcant change of votes suggests a successful strategy at correctly prioritising resources.

Lidia Thorpe in Northcote

Lidia Thorpe lost the seat of the Northcote which she had won at a by-election in 2017.

Let’s look at what exactly happened in 2018 and consider it in the context of previous elections.

Northcote 2018

CandidateVotes Percentage
THORPE, Lidia (GRN)1681639.52
THEOPHANOUS, Kat (ALP)1774841.71
MacISAAC, John (LIB)457010.74
Others (4 Candidates) 34188.03

After distribution of preferences the result was

THORPE2054848.29
THEOPHANOUS2200451.71

Northcote 2017

CandidateVotesPercentage
THORPE, Lidia (GRN)1631945.22
BURNS, Clare (ALP)1277935.41
FONTANA, Vince (IND)18645.17
ROSSITER, Dean (LDP)14974.15
Others (7 candidates)362910.07

and after preferences (noting that the count finished with Thorpe reaching 50% and Fontana was still in the count)

THORPE1838050.93
BURNS1441039.93
FONTANA32989.14

Northcote 2014

CandidateVotePercentage
McCARTHY, Trent (GRN)1410136.28
RICHARDSON, Fiona (ALP)1592840.98
D’ANGELO, Anthony (LIB)640716.48
Others (4 candidates)24356.25

and after preferences

McCARTHY1708843.96
RICHARDSON2178356.04

So what can we see from these figures? Back in 2014 the Greens were still well off winning it, trailing by almost 5% on the primaries and then well behind after preferences. The notable change in the by-election is that the Liberals did not have a candidate. The number of formal votes in was 36088, and in 2018 was 42552. Even though she lost, Thorpe actually got more votes in 2018, but with the change in total votes this went from a hard to beat 45% down to below 40%. The ALP also got a big increase which may have been from non-voters in the by-election or from the people who voted for some of the other candidates in a large field. It is not unusual for the major parties to do worse in by-elections. In a general election there is a lot of of media on the major party leaders and who will form government but in a by-election the focus is entirely on the local candidate and they are much more winnable by non-major party candidates.

So my take is that Northcote was not yet winnable for the Greens at a general election, but was at a by-election, while there may have been some benefits of incumbency, Thorpe had only been in place for a short time, and the big story of the election was Labor beating the Liberals, so it is not surprising that the seat could return to Labor and is not particularly a poor reflection on Thorpe as a politician. In particular in the final count it would take just 729 voters to change preferences from Labor to Green in order for Thorpe to win, so any attempt to paint her as being massively rejected by the electorate would be poor analysis of the results.

The Upper House

This is where things really went badly for the Greens as their representation was decreased.

YearHouseGreens Vote (%)Greens seats won
2018Upper House9.25 1
2014Upper House10.755
2010Upper House 12.013

If you consider that the number of seats went from 5 to 1, it looks like a disaster, on the other hand if you look at the vote percentage going from 10.75 to 9.25 it is a slight decline.

It is clear from these numbers that the voting system plays a big part. It may seem like a cop out to blame the voting system, but to say that a drop in representation means that the electorate has rejected them when the vote dropped by 1.5% is hardly a credible argument.

There are a couple of factors here. One is that the Greens did very well out of the 2014 election. You can see that their vote went down on 2010 but they won 2 extra seats. The upper house consists of 8 regions which each elect 5 members. This means that a quota is 1/6 = 16.76%.

The only region where the Greens achieved this quota was Northern Metropolitan where Samantha Ratnam was elected. The same was true in 2014 (when Greg Barber was the candidate). In all regions the last 1 or 2 positions are decided by preferences.

In 2014 the Greens were lucky in some very close results, basically every close count that could have gone their way in the upper house did, as a result they got a number of representatives that would not necessarily be guaranteed to be maintained at a subsequent election even with the same or a modestly increased vote.

Even so, with the small reduce in vote they might have been expected to keep 2 or 3 seats rather than just one. The other important factor is Group Voting Tickets (GVTs) and preference deals.

In Victorian state elections there are still GVTs which have been abolished in Federal elections as well as in NSW and SA (Tas has an entirely different system and QLD has no upper house). These allow parties to control the preferences of voters, and as a result make preference deals very important. The increasing impact of these deals explains why the small decrease in Greens vote resulted in such a big decrease in seats. Most of the small parties have signed up to deals to distribute preferences amongst each other, regardless of whether they actually agree on policy, in the hope that they’ll win out and gets the seat. The result is that large numbers of votes move in ways determined by deals rather than by voters.

For example in the Southern Metropolitan Region the Greens were on 13.46% or 0.8 of a quota. Labor and Liberal each had just over 2 quotas so it was just down to one remaining seat and it ended up going to Sustainable Australia who started with just 1.32% or 0.07 of a quota. Sustainable Australia stayed in the count thanks to preferences from a bunch of small right wing parties (Country, Shooters, Battlers). Without GVTs I can’t see these preferences going this way to such an extent, the likely result would be that Sustainable Australia would not stay in the count as long and their own preferences would most likely go to helping the Green candidate get elected.

Sometimes people claim that this system is fair because the total vote for small parties is over a quota, but evidence from elections where voters control their own preference shows that this is not how it works. Common sense also suggests that many of these parties are very far apart from each other on policy and the genuine choice of many voters would be for one of the larger parties rather than another small party with completely opposite views to their own.