Victorian election: What the opinion polls say

By Adrian Beaumont
Honorary Associate, School of Mathematics and Statistics, Faculty of Science, University of Melbourne

Two Victorian opinion polls were released late Thursday.

A YouGov Galaxy poll for The Herald Sun gave Labor a 53-47 lead, unchanged since a late October Galaxy poll poll for the bus industry.  A ReachTEL poll for The Age, also published gave Labor a 54-46 lead, a two-point gain for Labor since early October.

The ReachTEL poll was conducted November 21 from a sample of 1,240.  The Galaxy poll was conducted November 20-21 from a sample of 1,060.

Primary votes in Galaxy were 40% Labor (steady since late October), 40% Coalition (up one) and 11% Greens (down one).

In ReachTEL, primary votes were 38.7% Labor, 35.9% Coalition, 10.4% Greens, 9.9% Other and 5.1% undecided.  ReachTEL asks these undecided voters a forced choice follow-up, but the results of this question are seldom reported by the media.  It is likely that this forced choice gave the Coalition greater support, or Labor would be further ahead than 54-46.

In Galaxy, incumbent Labor leader Daniel Andrews led Opposition Leader Matthew Guy by 47-35 as better Premier (40-33 in the last Herald Sun Galaxy poll in August, when Labor led 51-49).

In ReachTEL, Andrews led Guy by 54.2-45.8 (51.3-48.7 in October).  ReachTEL’s forced choice better PM/Premier questions tend to favour opposition leaders more than polls without a forced choice.

In forced choice issue questions from ReachTEL, Labor led by 56.6-43.4 on cost-of-living (52.9-47.1 in October).  Labor led by 55.2-44.8 on Melbourne’s congestion (54.0-46.0 in October).  Labor led by 53.1-46.9 on managing Melbourne’s population growth (Coalition led by 50.4-49.6 in October).  The Coalition was still ahead by 51.9-48.1 on law and order, but this was down from 53.9-46.1 in October.

Overall, these polls suggest that Labor should clearly win majority government, though there is a small chance that the Greens could hold the balance of power.

At the last Victorian election in 2014 polls overstated the Greens, and this could be an issue again.  However, just because there were polling errors at one election does not mean the same errors will occur at the next.

(Note to readers: The text above (published 23.11.18) updates the text below (published 21.11.18))

The Victorian election will be held on Saturday. There have been no state-wide media-commissioned polls since a late October Newspoll of 1092 voters (54-46 to Labor).

ReachTEL poll conducted for the Victorian National Parks Association on November 13 from a sample of 1530 voters, gave Labor a 56-44 lead, which would be a four-point gain for Labor since an early October ReachTEL poll for The Age of 1239 voters.

I would like to see a media poll before concluding that the Victorian election will be a blowout win for Labor, but the most recent indications are that Labor is likely to win.

The latest state poll

In the late October Newspoll, Labor had 41% of the primary vote, the Coalition 39% and the Greens 11%.

45% were satisfied with Labor Premier Daniel Andrews, and 40% were dissatisfied, for a net approval of +5.  31% were satisfied with Opposition leader Matthew Guy, and 46% were dissatisfied, for a net approval of -15.  Andrews led Guy by 45-29 as better Premier.

State parties tend to perform better when the opposite party is in power federally. With Andrews’ positive net approval, the Labor government will likely be re-elected for a second term.

The same Newspoll had Labor leading the Liberals by 45-37 on managing Victoria’s economy - a bad result for the Liberals who would see economic management as one of their strengths. Labor also led by 43-32 on maintaining energy supply and keeping power prices lower.  On law and order, the Liberals led by just 39-38. Labor led by 33-30 on having the best plan for population growth.

Individual seat polls

According to YouGov Galaxy seat polls for The Herald Sun, conducted November 10-11 from samples of about 550 per seat, Labor leads the Greens by 54-46 in Richmond, leads the Liberals by 56-44 in Geelong, leads the Liberals by 52-48 in Mordialloc and leads the Liberals by 51-49 in Frankston.

However, while state and national polls have been accurate in Australia, individual seat polls have been very unreliable.

These seat polls are all in Labor-held electorates, and do not indicate much swing either way in the Labor vs Liberal contests.  State and national polls are far better indicators of swings than seat polls.  Labor might hold Richmond, but seat polls have been even less reliable than usual in inner city contests.

Micro parties could win several seats in the Upper House

Victoria’s Upper House has eight five-member electorates. A quota is one-sixth of the vote, or 16.7%.

Since the last state election in 2014, no reform has been made to the upper house group ticket voting system. Victoria and Western Australia are the only jurisdictions in Australia where group voting tickets are still allowed. As a result, there are many micro parties who are swapping preferences with each other so that one of them has a good chance of election.

Under group ticket voting, above-the-line voters cannot direct their own preferences. Instead, the party that receives a “1” vote controls that voter’s preferences. Votes can go to parties that are very different from the party that received the “1” vote.

The biggest problem with the group voting tickets is that the near 100% preference flows allow parties with negligible support to pass other parties, and then benefit from those parties’ preferences.

At the 2014 election, the Vote 1 Local Jobs party won the fifth of five seat in Western Victoria region on just 1.3% of the vote, or 0.08 quotas, a seat that should have gone to the Greens who received over 9% of the vote.

According to analyst Kevin Bonham’s simulations of 2018 upper house results, seven micro party representatives could be elected. While the particular micro party that wins could change, the overall numbers probably won’t unless the major parties and Greens do much better than expected, or there is a much higher rate of below-the-line voting.

The Greens in particular appear likely to lose seats that they would win with a more representative system. Labor may well have shot themselves in the foot by sticking with group ticket voting; with a more representative system, Labor and the Greens would probably win an overall upper house majority. Conservative micro party members are likely to stall progressive legislation.

It is easy to vote below-the-line in Victoria, as only five numbers are required for a formal vote (though voters can continue numbering beyond “5”).

I recommend that voters number at least five boxes below-the-line, rather than voting above-the-line, where parties control their voters’ preferences.

If enough people vote below-the-line, the micro parties’ preference harvesting could be thwarted.


politics; election Politics; Election

Election Watch: Past Editions