• Lack of broadcasters’ exit poll may mean hours of uncertainty
  • Use of polls before voting ends raises questions of legality

Hedge funds and banks are seeking to cash in on the outcome of next month’s referendum on British membership of the European Union hours before the first results are known.

Staff at two opinion-polling companies, speaking on the condition of anonymity because discussions are private, report approaches from financiers looking to glean insights into June 23’s so-called Brexit vote and any possible last-minute shift in public opinion.

The need for information is great because no official exit poll is being conducted for release once voting ends at 10 p.m., it’s unclear how soon early results will show a clear trend, and a final tally isn’t due until the next morning.

Given that the pound is predicted to rise on a vote to stay in the EU or tumble if Britons back leaving, any hint as to whether the margin of victory is closer or wider than expected could provide a trading advantage in the 24-hour foreign-exchange market.

“It’s widely anticipated that one of the financial houses will be running something on the day,” John Curtice, the professor of politics at Strathclyde University in Glasgow who ran last year’s general-election exit poll for broadcasters, told reporters this week. Curtice won’t be involved in any such effort; He said he’ll be asleep during the day on June 23, in preparation for staying up all night to analyze the referendum results for the BBC.

Unpredicted Victory

Polling companies and the $2.9 trillion hedge-fund industry could both do with a break. Last year’s election victory by Prime Minister David Cameron’s Conservatives went largely unpredicted, while hedge funds are suffering their worst start to a year in performance and investor withdrawals since the financial crisis.

While major banks including JPMorgan Chase & Co. are virtually unanimous in warning against Brexit, hedge-fund managers are divided. Manny Roman of Man Group Plc and David Harding of Winton Capital Management endorse staying, while Crispin Odey of Odey Asset Management and Paul Marshall of Marshall Wace want out.

Four weeks from referendum day, polls are split over how close the vote will be. Some show a clear lead for “Remain,” while others suggest the race is neck-and-neck. For investors such uncertainty spells a money-making opportunity.

Still, even if cash were no object, there are many difficulties with producing a proper exit poll. Such surveys typically work by returning to the same carefully chosen polling stations for each election, interviewing a sample of voters, and measuring how vote shared have shifted over the years.

Extrapolating Data

The aim is to predict the results in electoral districts where the result is typically tight and to use that data to estimate how each party will fare nationwide.

Predicting the result of a referendum is more difficult because every vote matters regardless of where it’s cast. There’s also no way to measure the change in vote share at individual locations, because it’s 41 years since the last plebiscite on EU membership.

That doesn’t mean there’s nothing polling companies could do for those willing to sign them up. A simple on-the-day poll would give an indication of the result. Even better would be a callback poll, in which a group of perhaps 5,000 people are asked for their opinion a week before they vote and then again on voting day to detect any change in sentiment.

Callback Quote

Such a survey on the day of the 2014 Scottish referendum allowed YouGov Plc to declare that turnout was high and that voters were shifting against independence. One company quoted 15,000 pounds ($22,000) as the approximate price it would charge for a callback poll this time around, without saying how large a sample that would buy.

“This gets rid of some of the problems with polling -- you’ve only got actual voters in your sample,” said polling analyst Matt Singh of Number Cruncher Politics. “But you still have the problem that you’re dealing with a regular opinion poll, with all the problems of whether you have a representative sample.”

One polling company said it hadn’t yet decided what it would offer, and that potential clients were still debating whether the referendum result was likely to be close enough for them to profit from a private poll.

Scottish Experience

Two years ago, as Scotland debated breaking with the U.K., pollsters said hedge funds would call them for clues into what they were finding, especially after one poll showed a lurch toward a vote for independence two weeks before referendum day.

The ideal money-maker would be a private poll that correctly predicted the opposite result from what published reports had suggested ahead of the vote. To profit from that, investors would have to be confident that their information was better than the public data.

There is a question of legality. Section 66A of the 1983 Representation of the People Act makes it illegal to publish information based on how people have voted while polling stations are still open. The question is whether passing information to a client in a financial institution would count as publishing.

The client would also have to be careful about who they passed the information to -- anyone breaking the law is liable for a 5,000-pound fine and six months in prison.

Before it's here, it's on the Bloomberg Terminal. LEARN MORE