Wednesday, January 8, 2014

2013 Ehrlich Award For Wrong Predictions

Not much going on in the world of pseph with all the polls still asleep (though Morgan is showing some signs of waking up soon) so not much to report lately.  I am hoping to soon publish a review of Cory Bernardi's oxymoronically-titled "The Conservative Revolution" but this depends on me being able to obtain a copy without paying for it. (If any readers can help out with this, that would be great.  I promise to take appropriate medical precautions prior to opening its pages.)

===================================================================

This article unveils the second annual Ehrlich Award, which at or about the start of each year will be given to the "wrongest" published prediction I observe relating to (or made in) the previous calendar year. 

There are a few groundrules - for instance the predictions need to be meaningful (in terms of being able to assess objectively whether they have happened), and secondly predictions that carry a stated assessment of chance of falsehood are not included unless that assessed chance is ludicrously low.  After all, even odds-on favourites do get beaten sometimes.



This year's nominations are heavily stacked towards the left, but that is partly because a lot of Coalition promises that appear to have already been broken were vague ones (like being a government of "no surprises" or stating they would "stop the boats" without being completely explicit when.)  If I have missed any outstanding candidates feel very free to nominate more in comments. 

The Ehrlich Award is named in (dis)honour of Paul Ehrlich for reasons stated in last year's article here.

I should disclaim that my own predictive game is not always perfect.  I made a number of predictoid statements in or concerning 2013 that turned out well, but there were a few duds in the mix as well.  The most notable of these errors were (i) in late 2012, giving Tony Abbott at most a 20% chance of recording a Newspoll netsat of better than -10 before the election (he just beat the target with -7 and -6 in the third-last and last Newspolls to go) (ii) never giving PUP any more than a "remote" chance of winning the last Tasmanian Senate seat, which they then did (albeit precariously).  I'm undecided whether prediction (i) was overconfident or just unlucky, but the failure of my predictive efforts for the last Tas Senate seat was at least partly caused by not getting around to re-running simulations before the election.

In general, while the Australian pseph community turned in fantastic efforts in terms of late-count projection of final Senate results, the business of predicting the results of these contests before election day has a very long way to go.  It may even be that predicting the results of the current Senate system before polling day is too complex for good results to be possible, with the good news being hopefully that we'll only get one more shot at it before the current system is disposed of for good.

Anyway, I don't think either of the above are in the class of those that follow, so here goes:

(Dis)honourable Mentions

* Kevin Rudd via spokesperson, March 22, "Mr Rudd wishes to make 100 per cent clear to all members of the parliamentary Labor Party, including his own supporters, that there are no circumstances under which he will return to the Labor Party leadership in the future."

* Everyone in the Gillard camp who then agreed that the above was so.

* Kevin Rudd on election night, stating that Labor had not lost any seats in Queensland.  That was effectively a prediction about the course of post-counting, and one that failed, since Labor actually lost two seats.

(In a similar category, we might include Liberal federal director Brian Loughnane's claim that the Coalition had “received a majority of the two-party preferred vote in all six States for the first time since 1977".  In the final results it lost the 2PP in Tasmania and Victoria.  But it is possible Loughnane just didn't realise the 2PP count was ongoing and would change.)

* A slightly dubious inclusion is Clive Palmer's claim his party could win 100 seats (it won precisely one, and that by a very thin whisker).  The only quote I can find for this is "We're going for 100 seats in the House of Representatives the way the mood is".  "Going" might be read as "aiming for".

* Malcolm Mackerras for predicting Kevin Rudd would lose his seat.  This one deserves inclusion because it's a typical Mackerras forecast - attention-grabbing, bold, high-reward but also (unnecessarily) high-risk.

* Ian Botham, for repeatedly predicting a 5-0 win to England in The Ashes, although he did employ the ancient excuse of hedgers and trolls everywhere that this forecast was made just to annoy the Australians.  Right margin, wrong team.

* iinet, for almost everything they said would happen to my phone service during the transition to the NBN.

The Minor Medals

Bronze

Steve Biddulph, "author, activist and psychologist" published this classic following the 2007 election in which he predicted not only that the Greens would surge to 20% in 2010 but that by the following election (which as it turned out was held last year) they would have become the Opposition to Labor. He said this would take "probably two" electoral cycles, but meant by that that it probably wouldn't happen in just one!  Biddulph hung these predictions off forecasts of catastrophic climate change and peak oil disasters that have more or less entirely failed to eventuate.  The whole piece was triumphalism: people had moved in Biddulph's greenish political direction at one election, therefore they would do so to an even greater degree at subsequent polls. Any number of similar pieces portraying the imminent total demise of the ALP have been written both before and following the Coalition's win in this election, and I'll be back to rub their noses in those once they too have definitively failed.

Silver

Bob Ellis - a yearlong if not lifelong achievement award really, but especially for a piece in which he grotesquely converted a Newspoll showing a 53-47 Coalition lead to a "certain" Labor victory, even indicating its scale as "56.8 percent to Labor, and a majority of sixty-two", throwing in a token "but I might be wrong."  It's been deleted but thankfully the Wayback Machine snapped the cult classic "Today's Newspoll: Labor Victory Certain, Possibly Huge".  Oh, and I've kept a text copy in case it gets taken down from there too.

Gold

One of the main political battlegrounds in the early leadup to the 2013 election was the condition of the upcoming Budget. A budget surplus within three years was repeatedly promised by Labor's leadership team going into the 2010 federal election.  These promises (couched initially in terms like "no ifs no buts" (Gillard, August 2010) and "Come hell or high water." (Swan, August 2010) continued to be made long after the party's victory in the 2010 federal election and were still doing the rounds well into 2012 although the writing was on the wall in terms of global economic conditions.  Even in Wayne Swan's 2012 Budget Speech the prediction of a surplus in the 2012-13 financial year was unequivocal.

It is not quite clear why Wayne Swan and Julia Gillard persisted so long in making these inaccurate forecasts of a budget surplus before Swan conceded in December 2012 that it was not going to happen, but the final outcome of an intended $1.5 billion surplus was described thus:

"In 2012‑13, the Australian Government general government sector recorded an underlying cash deficit of $18.8 billion (1.2 per cent of gross domestic product (GDP)). The fiscal balance was in deficit by $23.5 billion (1.6 per cent of GDP)."

Some think (example here) that Labor never really intended to deliver a surplus in 2012-13 and that the surplus pledge was just an attempt to disguise the government's plans to continue spending in deficit.  It's also possible that the persistence of the surplus pledge way past its use-by date was an attempt to improve Labor's polling in order to forestall leadership challenges from Kevin Rudd.  But whatever the case, the former government was left looking clueless, having so many times stated that it was going to deliver a budget surplus, but having not actually done so. For those still making up their minds as to whether the government had been economically skilled in its response to the GFC, or simply beneficiaries of the mining boom and the previous regime, here was a strong-sounding argument that the ALP government had no idea and had just been lucky (whether that was actually true or not).

Furthermore, people respond more negatively when they are promised good news and delivered bad than when they are tentatively promised trouble that turns out not to occur.  (This is probably why some commercialised weather networks deliberately overpredict rain - eg see Silver's The Signal and The Noise for a review of this sort of thing.) 

There are no direct signals from which we can judge the polling impact of the rescinding of this pledge.  Wayne Swan's admission that the surplus probably wasn't going to happen was buried on 20 December 2012, when the polls were on holidays.  There was not much to see in the few polls released in January 2013 (most of which were only Morgan Face to Face and Essential anyway) and from February onwards there were far more things driving the government's polling down than just the state of the budget.  But given the extent of political fetishisation of the idea (or even the ideal) of the budget surplus, it's hard to see anything easier to damage a government's economic credibility with than saying that that government had promised a surplus and failed by some margin to deliver it.

If Labor was simply relying on external advice that a surplus could credibly be achieved, review of the pretty average record of budget forecasting under the previous government (among others) should have been enough to convince them that any reliance on budget forecasting needed to be placed with more caution than they displayed.  "Trusting the experts" is not much excuse if there is evidence their field is a difficult one and one in which overconfidence sometimes happens.

I therefore cannot go past the 2012-13 budget surplus pledge jointly owned by Julia Gillard and Wayne Swan as the most significant and glaring predictive failure relating to the year 2013 that I observed, with extra points for the stubbornness and frequency with which the false prediction was repeated.  They are the co-winners of the 2013 Ehrlich Award for Wrong Predictions. 

1 comment:

  1. Kevin I'm predicting a Green landslide with Nick McKim being comfortably sworn-in as Premier and only giving a 45 minute acceptance speech.

    ReplyDelete