Experimental feature

Listen to this article

00:00
00:00
Experimental feature
or

The opinion polls may be close but as far as the betting markets are concerned, the result of the Scottish referendum is clear: the No camp is going to win.

One punter with William Hill is so sure that over the summer he put £800,000 on the Scots rejecting independence, thought by the bookmaker to be the largest wager on a political event in modern history. Even at the short odds on offer – his first bet of £400,000 was at 1-4, meaning he would only win £1 for every £4 placed – he stands to win more than £173,000 in total.

“We initially thought this would be a niche market and if we took a five or six-figure sum [in wagers], fair enough,” said Graham Sharpe, spokesman for William Hill. “Now we have already taken over £2m and this is the biggest political event we’ve ever been involved with – we’ve taken as much as the last UK general election and last US presidential election combined.”

Alex Donohue, spokesman for Ladbrokes, said: “This is on course to be the biggest political betting event in history with an estimated £40m bet across the industry. We took over £100,000 worth of bets this Monday alone.”

On the comparison site Oddschecker, the best price on a No victory is 1-4 while the equivalent for a Yes win is 10-3 (meaning that a £3 stake would earn you £10). Broadly speaking, those odds have held for several weeks.

Even the September 7 YouGov poll showing the Yes camp had edged into the lead – which sent hordes of Westminster politicians scurrying north of the border – had only a temporary impact.

“Up until and including August, the price for No was around 1-9 or 1-10 – much shorter than now,” said Andy Lulham, Oddschecker’s head of marketing. “When that poll came out you saw odds like 7-4 for Yes and as big as 4-9 for No. But we never saw Yes become the favourite – in betting terms, that is still a significant differential.”

Betting markets have a good record at predicting elections. “Betting tends to differ from, or is more polarised than, opinion polls,” said James Midmer of the betting exchange Betfair. “In polls people are asked, who they are voting for, whereas with betting people are making a more calculated financial decision in saying who they think will actually win.”

Several bookmakers reported a clear geographical divide in betting patterns. Ladbrokes have seen a 70-30 per cent split in favour of Yes in Scotland but the reverse in the rest of the UK. Yes bettors tend to place smaller stakes, perhaps reflecting a perception that there is more value in that side; by overall UK volume, therefore, Yes bets are in the majority, but No stakes tend to be much higher.

Aside from William Hill’s high roller, one Ladbrokes customer wagered £200,000 while a Betfair punter recently placed £55,000 – both on a No victory. About 85 per cent of wagers placed on Betfair are for No, even at the unattractive prices on offer.

According to Leighton Vaughan Williams, professor of the betting research unit at Nottingham Business School, political bettors are looking at previous elections in concluding a No result is looking likely. “History tends to suggest that polling overestimates the likelihood of a big change, and that voters tend to back the status quo at the last minute,” he said.

William Hill’s Mr Sharpe was similarly dubious about the tight polls. “If the polls putting Yes in front are to be believed, people should be banging our door down and helping themselves,” he said. But even with the more favourable odds, “we’re not seeing the sort of money that would reflect that”.

As with any bet the usual caveats apply. “Polls and betting markets are in new territory because referendums are rare,” said Mr Midmer.

But for more straws in the wind, Ladbrokes have 6-1 on Alex Salmond quitting as first minister of Scotland within 48 hours of the referendum result, and 4-1 that Mr Cameron is replaced as prime minister this year.

Copyright The Financial Times Limited 2017. All rights reserved.
myFT

Follow the topics mentioned in this article

Follow the authors of this article