Fri May 20

Newspoll and Ipsos both give Labor clear leads in final polls; counting of early votes

Written by: Adrian Beaumont, Honorary Associate, School of Mathematics and Statistics, The University of Melbourne

The federal election is on Saturday. Polls close at 6pm local time; that means 6pm AEST in the eastern states, 6:30pm in SA and the NT and 8pm in WA. 124 of the 151 House of Representatives seats are in the eastern time zone ^[1], 12 combined in SA and the NT and 15 in WA.

The Coalition notionally holds 76 of the 151 seats, Labor 69 and there are six crossbenchers. Gains and losses for parties and crossbenchers will be measured against this. This does not include Craig Kelly’s defection from the Coalition to the UAP in Hughes ^[2].

The final Newspoll, conducted May 13-19 from a sample of 2,188, gave Labor a 53-47 lead, a one-point gain for the Coalition since the previous week. Primary votes were 36% Labor (down two), 35% Coalition (steady), 12% Greens (up one), 5% One Nation (down one), 3% UAP (steady) and 9% for all Others (up two).

54% were dissatisfied with Scott Morrison’s performance (up one) and 41% were satisfied (down one) for a net approval of -13, down two points. Anthony Albanese’s net approval improved six points to -5.

The incumbent-skewed better PM measure was tied at 42-42 after a 43-42 Morrison lead last week. Newspoll figures are from The Poll Bludger ^[4].

Labor has a 53-47 lead in both Newspoll and Ipsos, and the Coalition would need Newspoll to be at least as wrong as it was in 2019 to get to a 50-50 two-party tie. It’s not impossible for the Coalition to win, but Labor is far more likely to win this election, probably with a solid majority in its own right.

In the final Newspoll ^[6] before the 2019 election, Morrison had a net +1 approval rating while then Labor leader Bill Shorten was at -8, and Morrison led by 47-38 as better PM. This year’s leaders’ ratings are far worse for Morrison. And Newspoll now weights by education.

Ipsos: 53-47 to Labor

The final Ipsos ^[7] poll for The Financial Review, conducted May 15-18 from a sample of 1,996, gave Labor a 53-47 lead, a four-point gain for the Coalition since last fortnight’s Ipsos. Primary votes ^[8] were 36% Labor, 35% Coalition, 13% Greens, 5% One Nation, 3% UAP and 8% for all Others with undecided excluded.

With undecided included, primary votes were 34% Labor (down one), 33% Coalition (up four), 12% Greens (steady), 15% for Others including One Nation and UAP (down one) and 5% undecided (down two).

By 2019 preference flows, Labor led by 51-44 (52-40 previously) – the headline figure excludes undecided. By respondent preferences, Labor led by 49-40 (50-35 previously). This poll’s result of 53-47 was probably rounded in the Coalition’s favour, but the last Ipsos was probably rounded to Labor.

51% disapproved of Morrison’s performance (steady), and 34% approved (up two), for a net approval of -17, up two points. Albanese’s net approval was up two to -4. Albanese led Morrison by 42-39 as preferred PM (41-36 previously).

I covered the final Resolve (52-48 to Labor), Essential (51-49 to Labor) and Morgan (53-47 to Labor) polls on Wednesday ^[9]. Morgan continued polling this week ^[10], but see no evidence of a further shift towards the Coalition.

In an additional question ^[11] from the Resolve poll, 40% supported Albanese’s proposal to increase the minimum wage by 5.1%, 27% wanted the minimum wage increased by a smaller amount and 16% wanted it kept unchanged.

The question was flawed as it presented Morrison’s argument against the minimum wage increase, but no argument for Albanese, for example that this increase is in line with the 12-month to March inflation rate of 5.1%.

Counting of early votes

With Friday’s data to be added, the ABC’s election analyst Antony Green ^[12] said pre-poll votes cast so far as a percentage of overall enrolment have surpassed that in 2019, even though early voting started a week earlier in 2019. 27% of enrolled voters have voted pre-poll in 2022, compared to 25% at the same point in 2019.

With an increasing number of people voting early, major pre-poll booths have not reported results until very late on election night; these booths often have 10,000 or more votes, and can have a large impact on seat results.

To address this issue, legislation was passed ^[13] last year to allow sorting of votes to begin at 4pm local time in pre-poll booths, two hours before polls close. That means election officials can sort votes into piles for various candidates, but not start a count until 6pm. This change will hopefully decrease the time until pre-polls start reporting.

As well as votes cast at early voting centres, 15.9% of enrolled voters have applied for postal votes, up from 9.4% in 2019. Green says that in 2019 84% of postals were returned and 81% accepted in the count. Postals will not be counted on election night, and their inclusion in later counting will shift results towards the Coalition.

Economic data: wage price and jobs report

The ABS reported ^[14] Wednesday that the wage price index rose 0.7% in the March quarter and 2.4% in the 12 months to March. But inflation ^[15] increased 2.1% in the March quarter and 5.1% in the 12 months to March. So real wages were down 1.4% in the March quarter and 2.7% in the 12 months to March.

The ABC said ^[16] the last time wage growth was above its long-term average of 3.1% annually was in 2013. This is the worst real wage growth since the introduction of the GST. Greg Jericho ^[17] at The Guardian said real wages are 2.2% below where they were at the 2019 election, 1.5% below the 2016 election and near their level at the 2013 election.

The ABS reported ^[18] Thursday that the unemployment rate dropped 0.1% from March to 3.9% in April, and the underemployment rate was down 0.2% to 6.1%. The employment population ratio – the percentage of eligible Australians employed – was steady at 63.8%.

The ABC said ^[19] this is the lowest unemployment rate since 1974. But I believe the reason the government has struggled in the polls in the lead-up to this election despite very good job reports is the decline in real wages.

Seat polls: Higgins, Goldstein, Curtin and Pearce

The Poll Bludger ^[20] reported Thursday that a Redbridge poll of Goldstein ^[21] (Vic, Lib, 7.8% margin) for Climate 200 gave Liberal incumbent Tim Wilson 36.0% and teal independent Zoe Daniel 26.9% with 8.4% undecided. 53% of voters for other candidates would preference Daniel, 13% Wilson and 34% undecided. The Poll Bludger gets 54.6-45.4 to Daniel.

The Poll Bludger ^[22] reported Wednesday that a uComms poll in Higgins (Vic, Lib, 2.6%) for the left-wing Australia Institute, conducted May 2 from a sample of 836, gave Labor a 54-46 lead from primary votes of 37% Liberals, 30% Labor and 20% Greens.

An Utting research poll of Curtin (WA, Lib, 13.9%) for The West Australian ^[23], conducted May 16 from a sample of 514, gave independent Kate Chaney a 52-48 lead over the Liberals. Primary votes were 38% Liberals, 32% Chaney, 13% Labor and 9% Greens.

A YouGov poll ^[24] for Labor of Pearce (WA, Lib, 5.2%) gave Labor a 53-47 lead. This poll was conducted May 15-16 from a sample of 411. It has a similar result to an Utting research poll of Pearce (52-48 to Labor) that I reported Wednesday ^[25].