You can sense the true believers starting to get skittish. Despite leading the Coalition on a two-party-preferred basis for most of the last three years, Labor’s lead in the opinion polls has narrowed. Bill Shorten has so far been unable to decisively pull ahead as the campaign enters its frenetic final days.
But the betting markets could give the ALP hope. Some pundits swear by the idea the odds are a stronger indicator of an upcoming election’s outcome than any polling. And, if that’s the case, Bill Shorten looks set for the lodge.
The latest odds
Like opinion polls, betting markets are pointing to a Labor victory. Throughout the campaign, ALP odds have remained between $1.10 and $1.35. Meanwhile, Coalition odds have fluctuated. The first week of the campaign had a Colaition win at $3.50; it’s now sitting around $5.50 on main betting companies like Ladbrokes, Tab and Sportsbet. Sportsbet is now so confident of a Labor win it’s paying out bets ahead of time.
Across the betting sites, Labor is tipped to pick up between 81 and 85 seats. This is consistent with the result you get when you plug in Newspoll’s most recent state-by-state swings into ABC’s swing calculator (which puts Labor on 81 seats). The most recent Newspoll, however, puts Labor on 77 seats.
The betting picture becomes more interesting when you look at marginal seats. The bookies suggest independent Zali Stegtall could narrowly get up over former prime minister Tony Abbott in Warringah ($1.75 to $1.90). Home Affairs Minister Peter Dutton also appears to be in trouble, with Labor’s Ali France favoured to pick up Dickson $2.30 to $1.55.
Ladbrokes and Sportsbet suggest Labor will take NSW marginal seats like Gilmore and Reid, although when the state swing is plugged into the ABC calculator, these are Liberal holds.
Generally speaking, individual seat polls and the odds for those seats seem to line up (although such polls can be quite unreliable). For example, Newspoll recently bumped the Coalition’s chances of picking up the Western Sydney marginal of Lindsay from Labor, and the markets have put them ahead $1.63 to $2.05.
How are odds calculated?
Dylan Leach, content and social media co-ordinator at Ladbrokes, says polls actually have an influential relationship on betting markets.
“Initially before an election is called the polls would play a factor in determining a market, but the traders would consult with those in the know politically before framing a market. Then, once the money comes in from punters, it would be determined from what the book is holding.”
Betting markets consider a number of factors — they fluctuate according to new opinion polls, but also developments in the news cycle and the latest policy announcements, which could perhaps give them the edge on opinion polls as a flexible gauge of electoral momentum.
Can they really predict elections?
Should we be watching Sportsbet more closely than Newspoll? Andrew Leigh, a Labor MP and economist came out in favour of betting markets as a good predictive tool in his 2015 book The Luck of Politics. “Every time, betting markets have been found to perform at least as well, and usually better than, the poll,” Leigh wrote.
Leigh believes that betting markets, like share prices, are better able to response in real-time to the latest developments, therefore giving a live picture of how the contest is faring. In recent decades, they’ve gotten pretty much every federal election correct — even in 1993, when Paul Keating managed to win what was termed an “unlosable election” for opposition leader John Hewson (an ominous signs for Shorten?).
But the markets haven’t always delivered. On the eve of the 2016 election, Malcolm Turnbull was all but guaranteed to form government according to the markets. Even though he did, a substantial swing to Labor saw him only just scrape over the line.
A month earlier, the Brexit vote delivered a stunning victory for Leave, despite all the bookies suggesting otherwise. Later in the year, markets gave Donald Trump a 35% chance of winning the election — his unexpected victory forced betting companies to pay out millions.
Those results suggest that betting markets can be just as fallible as the polls which largely made the same, incorrect call. They also point to structural issues with betting markets that suggest they should be treated with caution.
Speaking to the Independent, Ladbrokes UK’s head of political betting Matthew Shaddick said the Brexit odds could have reflected wishful thinking on the part of punters, who were largely affluent and pro-Remain. He warned that bookies create markets to turn a profit, rather than to forecast results. According to psephologists, this is unsurprising, since betting markets are in the end no better than opinion polls in telling us the future.
Adrian Beaumont, a statistician and psephologist from the University of Melbourne says election betting markets simply follow conventional wisdom. “I don’t think they add anything we would not get from looking at the polls,” Beaumont said.
According to poll-watcher Kevin Bonham, betting sites have been wrong on a number of Australian state elections, and individual seat betting gives them a similar success rate to opinion polls in determining the final numbers in parliament.
So in the post-poll world created by the election of Donald Trump, it’s worth treating “conventional wisdom” with a bit of caution.
Do you make a point of watching the odds? Send your comments and stories to boss@crikey.com.au.
Compulsory voting must play a part. In most of the world pollsters have to consider two questions: Who will you vote for? Will you actually bother to vote? Only the first question is really needed in Australia.
All evidence suggests that betting markets just reflect the published polling. If there was more to them than that, you’d think they would have the opportunity to demonstrate it in a situation like the Victorian state election where the polls were way off Instead, the betting markets simply made the exact same mistakes as the polls on the margin, the number of seats won by Labor, and on individual seat markets.
In the last WA election, the Labor candidate in my electorate (Kingsley) was 7 to 1 against winning standing against a very good Liberal minister in a very safe Liberal seat. As a rusted on Labor supporter, even I thought she was a good local member, responsive and effective in representing constituents.
I was tempted to put a bet on the Labor candidate, but didn’t. Odds of 7 to 1 against are just too implausible for an upset result. The Labor candidate won.
Just one comment re Trump winning in the United States. Hillary Clinton polled 3, 500, 000 more votes than Trump but the intricacies of the electoral college system, (plus her crass and completely unnecessary comment about “deplorables”) was the reason she was defeated.
Well I’m hoping Burt the Psychic Croc is on the money with a resounding win for Labor. And if he is, I think Crikey readers should reward Burt by tossing Bernard Keane into his pool. Go Burt!