Morgan gives Labor election-winning lead

The first national poll to be taken since the official campaign began has produced a shock result. Morgan has Labor leading by 52.5-47.5 on respondent allocated preferences, and 52-48 on previous election preferences, representing a 1.5 point gain for Labor on both measures. Unlike most Morgans, where the fieldwork is taken over two weekends, this Morgan was taken over only last weekend. The sample size was 2320, conducted by face-to-face and SMS.

We have also had an unchanged 51-49 to Labor result from Essential, with 1790 sampled from fieldwork done last week and the week before. Essential has the major parties' primary votes much higher than Morgan. In Essential, the Coalition is at 42%, Labor 38% and Greens 9%, while Morgan has Coalition 36.5%, Labor 33%, Greens 15.5%. In my opinion, Essential’s primary votes are much more likely than Morgan’s.

Essential has a reputation for excessive stability, so a real change may not show up in Essential. Morgan gave Turnbull some of his best results late last year, but under Abbott it leaned to Labor by about a point. It may be that the Labor lean is reasserting itself now that Turnbull is unpopular with the left.

We do not yet know whether there has been a genuine move to Labor, and need to wait for more polling. Newspoll and Ipsos are scheduled for next week.

In the last few days, we have had several individual seat Galaxy polls, and a ReachTEL poll of all five Tasmanian electorates. If these polls are accurate, they show that the Coalition is holding most of its marginal seats. This is what would be expected given roughly a 50-50 Two Party Preferred (2PP) national vote, as Labor needs to win at least 51% 2PP to gain enough marginals to govern.

Individual seat polling was poor at the 2013 election, and it was biased towards the Coalition. I trust the national polls far more than the individual seat polls.

A Galaxy Queensland-wide Federal poll has the Coalition ahead by 54-46, which is a 3 point swing to Labor, both since the 2013 election and the February Federal Queensland Galaxy. A 3 point uniform swing would only net Labor two extra Queensland seats, so they need a larger swing to have a chance of winning government. Galaxy’s Federal Queensland polls have appeared to lean a little to the Coalition.

More on Essential

In this week’s Essential, Turnbull had a net rating of -2, down 2 points from April. Shorten’s net rating was -9, up 5 points. 48% approved of internships for young people, and 32% disapproved. 52% thought the 8-week election campaign was too long, 32% about right and only 5% too short.

81% thought social classes exist in Australia, with 8% disagreeing. 34% considered themselves working class, 48% middle class and 2% upper class. 39% thought Labor best represented the working class, 17% the middle class and 10% the upper class. For the Liberals, these figures were respectively 4%, 15% and 53%.

In last week’s Essential, 20% approved of the budget, and 29% disapproved; after last year’s budget, it was 34-33 approve. 21% were more confident in the government’s ability to manage the economy, and 32% less confident. The budget was thought to be bad for most, with the exceptions of businesses (both small and large) and well-off people.

31% trusted Scott Morrison most to handle the economy, while 20% trusted Chris Bowen. In April this was 26-23 to Morrison; the increase for Morrison is mainly because Coalition voters now back him by 68-4.

Respondents were informed that a court had ruled that the detention of asylum seekers on Manus Island was illegal. However, by 48-30, they still opposed bringing the Manus Island asylum seekers to Australia. 39% thought conditions for asylum seekers on Nauru and Manus Island were poor, and 32% good.

Impact of a Liberal-Greens preference deal

There is much speculation that the Liberals and Greens will make a deal on preferences. Unlike the old Senate system, in the House parties cannot direct their voters' preferences, but they can make a recommendation on their How to Vote cards. Voters decide whether to follow that recommendation.

The Greens would have by far the greater benefit from such a deal. In the seat of Melbourne, now held by Green Adam Bandt, 2010 Liberal preferences flowed 80-20 to Bandt over Labor. In 2013, when the Liberals changed to putting Labor ahead of the Greens, Bandt only received 34% of their preferences.

If the Liberals win about 20-25% in Labor vs Greens contests in inner Sydney and Melbourne, the difference in preferences caused by their How to Vote recommendations is worth about 10 points after preferences.

The Greens have never preferenced the Coalition parties ahead of Labor, but they often use open How to Vote cards, with no recommendation on which major party to preference. However, the difference in flows between an open ticket and a recommendation for Labor is only about 3% of all Greens ballot papers.

In Coalition vs Labor marginals, the Greens would expect about 10%. So the benefit for the Coalition from the Greens issuing open How to Vote cards in marginal seats is about 0.3% after preferences.

If the Coalition were to lose any votes to Labor because of a deal with the Greens, it could more than cancel out any benefit they win on preferences from such a deal. The Greens are the only clear beneficiaries.

Antony Green has much information on the Greens most likely seat targets. In summary, the Greens have a reasonable chance of winning the Victorian seats of Batman and Wills in addition to Melbourne if they receive Liberal preferences ahead of Labor, but are long shots in other seats.

High Court rejects challenge to Senate reforms

Senate voting reforms were passed in March, but Family First’s Bob Day challenged these reforms in the High Court. Last Friday, the High Court unanimously rejected the challenge. As I wrote here, the Senate reforms are a huge improvement on the old system, where parties decided the preferences of their voters, leading to bizarre outcomes.

ABC’s Vote Compass is opt-in, and thus not credible

The ABC’s news page often reports findings from Vote Compass. People opt to do Vote Compass surveys, so it is not representative in the way that genuine polling should be. In particular, those who visit the ABC website, and those who are politically engaged are over-represented. Although the data are weighted to reflect ABS demographics, I think we should be very sceptical of Vote Compass findings.