2018 Ehrlich Awards For Wrong Predictions

"Labor Senator Murray Watt says Barnaby Joyce shouldn't be Acting PM. Who the hell is Murray Watt? Barnaby will be Acting PM next week. Murray should worry about acting as a real Senator. Like see actual constituents & be relevant. At the moment its Watt by name, who by nature."

George Christensen kicks off this year's Ehrlich Awards for Wrong Predictions with the above piece of banter dated 15 February.  In fact, Barnaby Joyce took leave for the week following the comment, and was not Acting PM, and soon after stood down as Nationals leader and Deputy PM as well.  This one sets the tone for a year in which a common theme was Coalition MPs or sympathisers expecting that pretty much anything would go right for their side of politics.  I'd include more left-wing failed predictions for balance, but this year I just haven't seen that many.  Feel free to add more examples I have missed in comments.



Welcome then to the seventh annual Ehrlich Awards for Wrong Predictions.  Each year I take the mickey out of some of the most instructively or haplessly failed calls of the future that I've seen regarding events in the previous twelve months, in any field of interest to this site.  The Ehrlichs are named for Paul Ehrlich, famed for losing his predictive bet with Julian Simon, but who also made such calls as that "England will not exist in the year 2000".  It's trying hard; perhaps he should have given it more time.  For the broader groundrules see the first edition (a note that any past predictions are considered to be eligible in whatever year they are falsified in) and for previous years click the Ehrlich Awards tab.  One function that the Ehrlich Awards serve is to try to discourage me from letting readers down by prematurely calling seats as certainly done and dusted when electoral events are sometimes less predictable than they may seem.  If I do that and the other side wins, then I have to at least self-nominate.

Christensen's case was not alone as an example of the Coalition not even knowing what its own lineup would be.  Christian Porter was quoted as saying he thought the Coalition would either "undoubtably" or "undoubtedly" (take your pick) run a candidate in the Perth by-election (they didn't).

Christopher Pyne on 5 February said that "This is going to be a very bleak year for Bill Shorten — unfortunately the public have found him out".  While Shorten did continue to struggle in personal polling (the context of the comment) and was subject to a brief outbreak of leadership speculation, a year in which Labor won every 2PP poll and the Liberal Party messily removed its own PM can hardly be described as "very bleak" for Shorten.  Indeed Shorten's personal ratings showed modest and thus far lasting improvement after he saw off yet another PM.

For a predictive failure from the other side of the fence I turn to Tasmania, where ALP state secretary Stuart Benson said, among other things, "We think Rebecca will definitely be the next Premier on [March 3]".  By this stage a bandwagon effect had been seen in public polling. Labor disputed this based on unsupported claims about its internal polling, but the public polling actually underestimated the incumbent Liberal government.  The Liberal Party retained government easily in vote-total terms, albeit with what turned out to be a rather strained seat majority of one.

This year's three medallists are all on a pretty similar level and could have been ranked in any order, with the winner getting the gong mainly for being more originally silly than the other two.  The bronze medal for 2018 goes to a piece by Joe Hildebrand on Greens preferencing decisions for the Wentworth by-election and how they were supposedly going to cause the Liberals to retain the seat.  This piece contained such declarations as:

"This means the Greens will be the first of the four to get knocked out and have their preferences distributed. And if their preferences get distributed to Labor as the party directs then the ALP will overtake Phelps and Phelps will be knocked out."

and

"In short, the Greens losing their sh*t over Phelps preferencing the Libs will result in the Libs getting elected. There might be a word to describe this level of stupidity but it hasn’t been invented yet.

Indeed, the only way that the Liberals won’t win Wentworth is if the majority of Greens voters are smarter than the party they’re voting for — which isn’t exactly a safe bet."

What actually happened with Greens preferences in Wentworth was this.  The Greens recommended their voters preference Labor ahead of Kerryn Phelps.  At the point at which Greens candidate Dominic Wy Kanak was eliminated, Wy Kanak held 8325 votes including 6543 Green primaries.  Tim Murray (ALP) had 9605 and Kerryn Phelps (IND) had 24114.  So even had every 1 Greens vote flowed to Murray, Murray would still have fallen 7966 votes short of overtaking Phelps, falsifying both "the ALP will overtake Phelps" and "the only way that the Liberals won't win Wentworth [..]".  Hildebrand's first error here was to assume (based on seat polling that was rubbish to begin with and that ignored the possibility of strategic voting anyway) that the gap between Phelps and Labor would be small.

Also, anyone who has paid any attention to minor party preferences at all will know that Greens preferences are notoriously undirectable.  Their how-to-vote cards have some impact on their flow, but often 10% or less.  In this case the split of all Wy Kanak's votes favoured Murray over Phelps only slightly (48.95% to 42.92%), meaning that Murray gained on Phelps by only 502 votes.

What clinches the medal for Hildebrand is that in the same article he boasted about how good his past predictions (supposedly) were with lines like "Honestly, it really is tough being right all the time." And yet, in this case, he wasn't, and spectacularly so.  There might be a word to describe this combination of bragging and clueless overconfidence but it hasn't been invented yet.

Overconfidence was also a theme in the Victorian state election, where I tried hard to disclaim all discussion of the Upper House by pointing out that Group Ticket Voting counts are extremely difficult to model and it takes a vast amount of work to be confident about what is going on in them and to be sure one hasn't missed anything.   The broad picture, that the election was going to be a big win for Glenn Druery, was callable off the initial release of the group tickets, and Antony Green was Cassandra-like in spotting the full extent of the oncoming farce pretty much right away.  (It was actually slightly worse than I expected!) However getting the picture right in specific cases is much harder, and there was one notable misfire that I want to highlight here.  Stuart O'Neill of the Aussie Battler Party failed to win in West Metro although "From my modelling of the preferences, Stuart O’Neill of the Aussie Battlers Party is certain of election from Western Metropolitan Region." (source).  O'Neill was supposed to be a sure thing if he could get 0.3%, and in fact got 0.95% but was never in the hunt.

Why did O'Neill not win?  The scenario by which he often won in modelling was taking the preferences of Sustainable Australia, Hudson 4 Northern Victoria and Australian Liberty Alliance and using these to get over the Animal Justice Party, at which point he had such a strong preference snowball that he won.  But it was vulnerable in the early stages.  The AJP polled so high a primary vote that the Aussie Battler Party never even passed the AJP's primary, and it wouldn't have helped O'Neill if they had, because the AJP got more preferences from Reason anyway.

Trying to read the counts correctly on election night was even harder, with limited time to envisage let alone explore new scenarios that had not been canvassed in pre-modelling based on very incomplete and unrepresentative primary figures.  It took a couple of days before a pathway for Fiona Patten (Reason) to retain her seat was spotted, and she eventually did.  But in the meantime The Age had jumped the gun with an article declaring Patten was pretty much toast with avocado.  The article text used a mixture of definite-cally and more appropriately vague language, but the headline and online tag ("Despite losing her spot [..]") displayed no such restraint.

The incorrect call that Patten was gone was compounded by it being in service of an incorrect but exciting-sounding narrative.  That Patten should be beaten by Glenn Druery preference deals after calling out Glenn Druery preference deals made for a dramatic story, but it was never going to be the truth.  Ignoring the quotas won on primaries by Labor, the Greens and (just about) the Liberals, Patten's party started the race for fifth place in fourth position on 3.37%, behind Labor (9.25%), Victorian Socialists (4.20%) and Labour DLP (4.17%).  Patten in fact won by preference-harvesting, using the above the lines of Voluntary Euthanasia, Animal Justice and Vote 1 Local Jobs to jump Victorian Socialists and LDLP and then using Vic Socialist above-the-lines to pass and eliminate Labor.  (I note in passing that the Vote 1 Local Jobs trick, where said party ran dead in North Metro only to siphon votes to Patten in exchange for Lower House preferences, and then fluked the donkey vote as well, was a pretty Drueryesque play by itself).

Labor, not Patten, would have won the final seat without Group Ticket voting.  Had Patten lost, on the other hand, the prime culprit would have been not Glenn Druery, but the Greens. The Greens, outside of Druery's network, had directed preferences to Derryn Hinch's Justice Party ahead of Patten apparently because of something in Patten's Upper House voting record. That would have stopped her passing Labor if the Victorian Socialists had been excluded before the Greens hit quota.  Labor, not Patten, would then have been the victims of the Druery snowball, though that would also have been partly because of Druery parties passing enough preferences to DHJP to get them in front of Victorian Socialists.

But all this stuff is very, very difficult, making errors more excusable.  For the silver medal for 2018 I stay with the Victorian state election but head for the simpler terrain of the Lower House, and this gorgeous piece of Twitter bluster by Chris Kenny:

A rare case of a tweet being "ratioed" eight months after it was sent!  Generally in covering the Victorian election, the right-leaning side of the commentariat told voters what the voters thought instead of getting out of the bubble and commissioning proper research to find out.  Law and order is usually a niche issue at state elections, and was particularly so at this one where it came across as a front for thinly disguised xenophobia and even racism reinforced by misleading statistics.  (Discrepancies in crime rates between some ethnic communities are explainable overwhelmingly by age profile.)  Voters didn't care about the Red Shirts "scandal", partly because they didn't understand it ("Inside Baseball" as one ALP operative put it).  Voters respected that Daniel Andrews "gets stuff done" (as Liberal doorknockers reported) and didn't mind if he spent money or made the odd mess on the way.  To call Andrews "finished" at a time when he was ahead in the polls, and given the track record of first-term state governments not losing and often winning big while the other party is in power federally, was brave to say the least. 

Also on Victoria (thanks to Stephen Spencer for this one), we have John Ferguson (The Australian) tweeting on 16/11/18 "Libs talking a big game. Convinced the swing is coming their way. Happy to listen. Reckon it will be very tight".  I'm not sure how many were saying that but whoever it was, he would have been better off not listening, or at least not taking it seriously.  No poll taken after August showed a swing to the Liberals.

And William Bowe has reminded me that Michael Kroger was quoted as saying the following re Labor's chances:

They don't have many people on the ground, the crime issue is biting hard, cost of living is biting hard, Andrews is unpopular and will be lucky to win his own seat

Saying that someone will be lucky to win is tantamount to declaring that their chance of victory is somewhere below 50%, which was a clearly ludicrous assessment at all times in Andrews' case.  (As it happens he won with 62.71% 2PP, a well-above-average swing to him, which is no surprise given that he was a first-term Premier).  The factual basis of the claim was also partly objectively spurious as Andrews' previous two net satisfaction ratings at the time had been positive.

Finally, the 2018 gold medal.  Back in 2015 a really bad Prime Minister was rolled, and the nation breathed a sigh of collective relief.  As the new chap raced up the polling charts, some people got a little bit carried away, ignoring both the history of honeymoon effects in polling (something also seen in op-eds of the early Morrison days) and the extent to which Malcolm Turnbull had sold out piece by piece on issues such as same-sex marriage on his way back up the Liberal greasy pole.  But unrealistic expectations continued all the same:

" And yet. And yet. Malcolm speaks to us not as a rabble of blithering chimps wanting their buttons pushed but as grownups, capable of considered argument, reasoned reflection and conscientious decision. For Australia, this is huge."

"But this is more than a prediction. It's a judgment. Malcolm's political longevity will be a Very Good Thing. Not because he'll necessarily manage to repurpose the crazier cowboy fringes of the Coalition. But because – far more importantly – the explicitness of Malcolm's intelligence makes it OK for us to be intelligent too.

Not just OK. Intelligence is almost expected. And expectation, as we know, is the best single predictor of performance"

Frankly I can't think of a PM (not even Abbott) whose tenure did more to harm the idea that it was OK to be smart than Turnbull's.  Under Turnbull, intelligence was a fringe skill that one used to craft debating-school style or lawyerly put-downs of opposing arguments, or to argue a position for the sake of trying it on whether one actually believed in it or not.  Where intelligence was to be used for the purposes of factual debate it was to be spread widely but with no great depth of study (resulting in instant-expert style embarrassments like "The High Court so shall hold").  Any intelligent purpose behind Turnbull's Prime Ministership seemed to be lost beneath a semi-justified belief that if he tried to do anything much he would be rolled, meaning that in practice he was just a smart-sounding vessel for what were often bad ideas.  But all this is in the field of opinion and debatable.  Where our winner stood out objectively is in the conclusions that she drew, based on something as obviously marginal and wishful (in terms of the views of the average voter) as Turnbull's supposed pride in being smart, about Malcolm Turnbull's political longevity:

"So here's my prediction. Malcolm – who like Beyonce is known universally by his first name – will be the longest-serving prime minister since Menzies. Possibly ever."

Malcolm Turnbull became Prime Minister on 15 September 2015.  To become the longest-serving PM since Menzies, Turnbull would have had to beat John Howard's spell of 4284 days, which would have seen him still in office on 8 June 2027.  To get there he would have almost certainly had to win four elections, in the process breaking Robert Menzies' record as the oldest PM ever re-elected.  To stretch Turnbull's tenure out to "possibly ever" would have seen him still in office in 2034 at the age of 79.

In the end, Turnbull's tenure, which ended this year (I feel safe in assuming permanently) was shorter than not only Howard's, but also Hawke's, Fraser's, Keating's and even Gillard's.  Elizabeth Farrelly's prediction of a long and glorious Turnbull reign fell the small matter of eight years, nine months and sixteen days short of target and wins the 2018 Ehrlich Award for Wrong Predictions.