Australia Election Preview: Can Morrison Pull The Miracle Again?

Finally Here

With the official Australian Election campaign launching this week, I am finally getting to writing this preview, so let’s do this.

In 2019, Australian Labor was cruising in the polls with a narrow but constant lead, with almost every poll showing Labor with a 2 Party Preferred (2PP) of 51, 51.5, or 52, and the Coalition assuming they would lose. Even throughout the campaign, Labor’s lead was never deemed to be overwhelming, but the sheer consistency of the lead made people think that the lead was more secure than it was. And then it wasn’t.

Scott Morrison won again, with an extra seat than the 2016 election and an extra point after preferences compared to Malcolm Turnbull’s performance, but much more importantly, Morrison saved the party in Queensland and Tasmania and lost the party ground in the suburbs (although, not losing seats there). Now, the shine is off Morrison, and the question of whether he can pull off the double upset is the only charm left in this campaign. So, let’s dig into it.

House

If the Coalition hadn’t overcome their polling deficit in 2019 – or, more accurately, had they been up 51-49 on Labor for the last 6 weeks before the 2019 election – there would be no suspense in this election campaign. Now, with the looming error of last time, people are being more circumspect about the prospects of the Labor Party. And while, in a sense, they’re right to raise that prospect, there’s also nothing suggesting that Labor will lose this election based on the current data alone.

That is not to say that Labor can’t lose the election, but based on where we are right now, there’s nothing suggesting Labor will lose. (Also, of note – at the last election, Labor won 68 seats, and the Coalition won 77. Since then, redistributions have created a safe Labor seat in Victoria and eliminated a Coalition seat in Western Australia, meaning the notional starting point is 69-76, with 6 crossbenchers.)

An election “right now” would see Labor win 16 seats against the Coalition on a 2PP basis, per my model – although, 2 of those are also targets of independents, so Labor’s total gain could be only 14. Additionally, I am overriding my model and giving the rural New South Wales seat of Hunter to the Coalition, based on the retirement of the incumbent member and general trends (which I’ll get to later). An election held “right now” would see something like Labor 83, Coalition 61, and 7 crossbenchers, and a Labor 2PP of 53.6. But, frankly, I don’t see that as overwhelmingly likely.

What we have seen in most past Australian elections, especially ones with Coalition governments, is historical recoveries in their polling position, even if it is likelier than not to be enough to win government. Kevin Bonham’s work here is invaluable, and when it was written, it made it clear that government’s tend to do better the closer we get, and I certainly think that applies here. The question, of course, is how much.

Right now, I’m adapting a 1.5% hit to Labor’s 2PP for the expected tightening between now and May 21st, turning their 53.6% to a narrower 52.1%. What would my model spit out? 11 2PP gains for Labor – Boothby, Swan, Casey, Pearce, Robertson, Bass, Flinders, Higgins, Deakin, Brisbane, and Chisholm – while still losing Hunter, for a 79-66-6 Parliament, or 78-66-7 if Flinders goes to an independent. (I am assuming Labor stays ahead of the Greens in Higgins based on the 2019 results, as well.)

Model

This model will be more aggressive on Coalition chances in some seats – for instance, Braddon, Reid, Longman, and Lindsay would all be notionally competitive by uniform swing, and no uniform swing model would have Hunter going Coalition – while having seats like Casey, Flinders, Higgins, and Deakin flipping from decently safe positions. Anyone who is familiar with my work from Canada or the US will know of my belief in the Global Realignment, and the notion that socially liberal areas will trend more leftwing, and socially conservative ones more rightwing, for a given national vote or national swing. It’s a simple enough theory, and with the 2017 postal survey results being broken out by division we have a sense of the social liberalism or cultural conservative of these seats.

The seats where Morrison did better in 2019 than expected – northern Queensland, western Sydney, and Tasmania – all saw lower YES votes on the proposal to allow same sex couples to marry, whereas the places where Morrison did worse – blue-ribbon, wealthy conservative strongholds – were all areas of high YES voting. It’s not quite as simple as a suburbs and cities versus the regions dynamic as it is in the US, because seats like Lindsay in western Sydney – where a large portion of the population are Chinese, and culturally conservative – trended right. It’s about cultural belief, and that informs my projections.

This model was used successfully in the Canadian election, and hopefully will better predict variation between seats than a straight uniform swing model.

Independents

In 2019, former Prime Minister Tony Abbott was defeated in his blue-chip, formerly safe Liberal, seat of Warringah by a pro-climate action independent, a few months after Malcolm Turnbull’s seat of Wentworth, on the other side of Sydney Harbour, fell at the by-election to replace Turnbull to a doctor of similar views. It’s sparked a series of imitators, and it is unclear how many, if any, can win this year.

Predicting these sorts of things is a mug’s game without data, because you’re reliant on reports and intuition, and it’s all a mess. I am going to choose to project Labor-Coalition 2PP in these seats – namely Flinders, Wentworth, Kooyong, and North Sydney – unless and until there’s any actual proof that the independents have made their move to 2nd place. Generally speaking, these sorts of independents do better head to head against the Coalition, as the preferences of those voting for the independent flow weaker to Labor than Labor preferences do to the independent, but that is a general statement about independents in general, not about any specific independent in any specific seat.

Polling

My current polling average is Labor 53.6 2PP, based in the four (not-trash) pollsters in the field recently (I do not use Morgan, because I would spend more time debugging it and applying house effect corrections to it than is worth my time or energy), with the revamped YouGov-run Newspoll getting the plurality of the weight. Newspoll in its current form hasn’t been tested at a federal level but it has been quite good in the states since the revamp, so it gets higher billing than any of Ipsos, Essential, and Resolve.

Kevin Bonham’s average before last week’s Newspoll was 54.6 to Labor, and all we’ve had since then is Newspoll taking Labor from 54-46 to 53-47, so presumably his average would be somewhere in the low 54s or high 53s at this point – a fact I only bring up to make clear that this is not a particularly favourable assessment for Labor.

Could the polls be wrong again? Sure, but most of the time, polling errors go the other way cycle to cycle, as we saw in the UK in 2015 and 2017 (the Tories beating their polls in 2015, way underperforming them in 2017) and Canada 2011 and 2015 (the Tories beating their polls in 2011, the Liberals beating the polls in 2015). Yes, Trump beat his polls twice in a row, but if you just want to quote “but Trump”, you’re probably not reading this site (unless you really hate me).

Prospects

Labor should win the election.

Labor is not guaranteed to win the election.

I’ll get into Hung Parliament dynamics at some other point, but the bottom line is, Labor will govern with any number of seats at or above 74, because there is no chance the Greens or Andrew Wilkie would put a Coalition Government into office over a Labor one in a Hung Parliament – and it’s possible some of the current or prospective independents on the crossbench might prefer Labor to Morrison.

There is absolutely a chance of the kind of swing during the campaign such that Labor loses, or that the polls are just flat-out wrong. Subjectively, Labor losing would be somewhere less surprising than Hillary losing and more surprising than Biden losing (had he lost), so call Morrison’s chances of winning 15%.

That said, though, I don’t see Morrison pulling it off again, and gun to my head, Albo will cruise to The Lodge with the first Labor government since Rudd-Gillard-Rudd.