When Scott Morrison called the election for May 21, betting agencies had the Labor Party at $1.33 to form government, with the Coalition at $3.25. That converts to a 75% probability of Labor winning. And even after a diabolical first week of the election campaign for Anthony Albanese, odds eased Labor only slightly out to $1.45, with the Coalition $2.70 at the half-way mark of the campaign.
Now the current odds may be short for Labor, but they were even shorter in 2019, and we all know what happened then.
It was often touted in the media in the lead up to the 2019 election, that Labor was a good thing to form government because they were so short with the bookies. But how good a predictor of election results are betting markets?
Believe it or not, betting agencies are in the business of making money, not predicting outcomes. Their win quotes will be weighted by dollars held for various outcomes. That Labor ended up such a short-priced favourite just days before the election, would have reflected the greater holding in the win pool for a Labor victory. Sportsbet for example, reported that 70% of total money wagered with them, was held for a Labor win.
Now a quote of $1.15 to win reflects an 87% implied probability of winning. And before placing a bet, the punter should weigh up whether that reflects value or not. If he believed the true probability of a Labor win was greater than 87%, he would gladly accept the $1.15 on offer. If not, he should have walked away.
So was Labor’s true probability of winning anywhere near 87% in 2019? When the TAB opened betting on the election in February 2018, they installed Labor favourite at $1.50, an implied winning probability of 67%, with the Coalition at $2.50.
Labor probably opened favourite for a number of reasons. That the 2016 election result was so unexpectedly close with Labor almost snatching victory. That there was perceived instability within the Coalition with Tony Abbott lurking in the background of a Turnbull government. And that Labor finally had a stable team with a fistful of policies to take to the 2019 election.
Then from that point, things began to get even worse for the Coalition. There was the Barnaby Joyce resignation over a relationship with a former staffer. Then the leadership debacle that saw Malcolm Turnbull ousted and Scott Morrison installed as new PM. Followed by the Wentworth by-election loss, and then the announced exodus of senior Liberal figures, not seeking re-election. Things unravelled fast, and many deemed the Coalition unelectable. Meanwhile, Labor sat back quietly observing the self-destruction.
Consequently, the big bets began to roll in for a Labor win, and its odds tightened with the bookies. Such a trend can often become self-fulfilling, as the prospect of a Labor win attracted more punters chasing seemingly easy money but requiring bigger wagers at short odds for a half decent payout. Meanwhile on the other side of the ledger, the Coalition’s price was blowing out further, becoming attractive only to small time punters prepared to have a nibble, not needing to mortgage the family home for a reasonable reward.
But as the dollar quotes became steadily more lopsided, did the true probabilities shift in tandem? Probably not a lot. Because with the Coalition already at a low base, the probability of them winning surely could not have become much worse, and therefore Labor not much better. In fact to the contrary, things began to slowly turn.
As they steadily unpicked Labor’s high taxing and high spending agenda, the true probability of a Coalition win began to edge upwards. And Morrison began to display his campaigning skills that had largely been unseen by the electorate. He and his party effectively began to make inroads, slowly willing themselves back into the contest. The polls revealed an improved Coalition two party preferred measure, moving from as far out as 56:44 in Labor’s favour, to just 51:49 days before the election. Which was within the margin of error.
So the true probability of a Coalition win had more than likely increased, and that of Labor’s decreased. But the sheer weight of money for a Labor win could not move its short quote. In fact, it tightened even further in the last weeks of the campaign to that skinny price of $1.15, with the report of one punter waging $1 million on Labor to win. Meanwhile the Coalition’s price drifted out as far as $5.50, equating to just an 18% implied probability of winning.
In the end, there simply was not enough money to correct the bias in the win pool. Noting also that punters were unlikely to be representative of the electorate, ensuring persistent bias. Momentum however, continued to swing the Coalition’s way, and the rest as they say, is history.
The thing is in politics, true probabilities can move with sentiment fast, but the betting markets may not have the scope to adjust. Probabilities can be indicative when markets open, but as we have seen, events can quickly cause a disconnect between true and implied probabilities.
Interestingly, we saw similar results with Donald Trump winning at long odds in the US in 2016. And the Leave vote triumphing over the Remain vote in the Brexit referendum also at long odds in 2016, with about 75% of £40 million bet on the outcome, wagered on Remain. Even though it was revealed later that there were far more individual bets on Leave than Remain, which proved prophetic in hindsight. After all, it makes sense intuitively that numbers of bets as opposed to dollars bet, should be a better predictor of what the electorate is thinking.
So when we accuse bookies of failing to read the political landscape, we should realise that that is not their charter.
Tony Dillon is a freelance writer and former actuary. This article is general information and does not consider the circumstances of any investor.