Featured, Features | 8th June 2015

John Curtice: The man behind the only poll that was right

The only forecaster anyone still trusts is highly sceptical that Labour can win back power, despite the emptiness of Cameron’s plans.

Photo: Getty


It’s “virtually inconceivable that Labour can win the next election by just doing well in England and Wales – it’s just brutal arithmetic,” John Curtice begins when we meet three weeks after the election, on a cloudy afternoon in St James’ Park.

Curtice, 62, is the only forecaster any journalist still listens to. He is the man behind the exit poll, which flashed up on TV screens at 10pm on election night and sent parties, pollsters and pundits into a frenzy. The story it told contrasted completely with more than 700 polls published in the previous year, 99 per cent of which said the Tories had no hope of a majority.

Its accuracy has left pre-election polling on life support. Journalists don’t trust it, pundits ridicule it and newspapers no longer want to publicise it. But the exit poll has never been more revered. Half a dozen men were behind it this year, but it is Curtice who has led the group for more than two decades.

His latest predictions make disconcerting reading for a beleaguered Labour Party. Curtice thinks that unless Labour recover in Scotland, they will need a 12.5 per cent lead over the Tories to win a majority. They haven’t had such a lead since 1997. As significantly, they now need a 3-7 per cent lead in the polls to win more seats than the Tories.

His latest predictions make disconcerting reading for a beleaguered Labour Party.

The electoral system no longer favours them. Nine months ago Labour were set to win 30-40 more seats than the Tories if the parties tied in the polls. Labour’s collapse in Scotland erased that advantage, and two things have now turned the system in the Tories’ favour: the Tory vote has become more efficient – they outperformed in marginal seats – and the Lib Dem collapse has handed Cameron’s party two dozen new safe seats.

But Curtice doesn’t think Tory majorities will now be the norm. “The Tories only just did it. The range of outcomes which will produce a hung parliament is greater than ever before. Book your place for College Green in 2020, because it is getting very hard for any party to win a majority.”

An exit poll predicting that the Conservative Party led by Prime Minister David Cameron will be the largest part with 316 seats is projected onto BBC Broadcasting House, Portland Place in London as voting finishes in Britains general election on May 7, 2015. Prime Minister David Cameron's Conservatives are on course to be the biggest party in the next British parliament, according to an exit poll from the general election showing them winning far more seats than had been expected.   AFP PHOTO / JACK TAYLOR        (Photo credit should read JACK TAYLOR/AFP/Getty Images)

Few people are more likely to be manning election night 2020 than Curtice. He has now been running the exit poll room for 23 years, and has been part of BBC election nights since 1979. Andrew Neil summarised the reaction to his shock poll with an on-air tribute to its “artistry and accuracy”; and he is fifteen years younger than David Dimbleby, who is only just retiring. Curtice may still be explaining elections well into the 2020s, and he is sceptical Labour will challenge the Tories before then.

“The real problem Labour has is, ‘So who have you got that can beat Boris?’ Boris has charisma, he has reach and we know from his success in London that he can appeal to people well beyond the confines of his party. And on many issues he’s a liberal Tory, but the Blue Rinse Brigade love him!

“The real problem Labour has is, ‘So who have you got that can beat Boris?’”

“I wasn’t convinced any of the five [Labour candidates] had it last time [in 2010], and I’m not entirely convinced any of this three have it either.”

And unless Labour’s new leader wins back Scotland, he thinks they have little hope. “The idea that’s what’s required is working out how to deal with the 3 per cent Labour didn’t win in England & Wales, but not working out where the 18 per cent have gone in Scotland…”, is mad, he says, tailing off, exasperated.

“The Labour Party’s battle for the leadership is being played out between and amongst English Labour MPs, whose acquaintance with their position in Scotland is not that great. And, you know, the people that could tell them ‘Bloody hell!’ [we’re in real trouble] are gone.”

If anyone should understand Labour’s generational collapse in Scotland, it is Curtice. He began his career at Oxford’s Nuffield College, deciding early that academia “seemed like a good wheeze”, but has been at the University of Strathclyde since devolution. He spent years doing detailed survey research on Scottish political attitudes, and spent early 2015 stressing that Labour could lose nearly all its Scottish seats.


He is confident, perhaps too confident, when I ask him why Labour lost. “The vox pop in Scotland goes like this: ‘Yeah, we used to vote Labour around here, but they ignored us and we don’t know what they stand for anymore.’ And that last bit is crucial. It was also the problem in England and Wales. Ed’s problem – we all said it so many times – was he could never articulate his vision in a way that the general public could grab what it was about.”

Some pundits can’t see how Labour can win back both its left-wing in Scotland and centrists in the South, but Curtice thinks ideology is a misnomer.

“The truth is, in part, that it doesn’t matter what the hell you stand for, as long as you’re perceived to stand for something, alright? Everyone knows what the SNP stands for – there’s a sense of excitement, optimism, destiny, they’re going somewhere. And they are perceived to be competent, they know how to run the bloody country.”

Curtice: Labour could never get their vision across. “We had every failed attempt on earth.”

“Labour lacked both characteristics,” he concludes. They were not seen as competent – a fact that he think “cries out” of Ashcroft’s election day research into why people voted as they did – and they could never get their vision across.

“If you read his speeches you got it!” He says of Miliband. “It’s not that the ideas weren’t there – capitalism’s not working for the ordinary punter, here are the ways we’re going to change it – but you had to read the bloody things [the speeches].

“We had every failed attempt on earth – from predators and producers to ‘One Nation’ Labour. And we ended up with them apparently thinking you could persuade people to vote for you on the basis of individual policies. I mean this is nonsense.

LONDON, ENGLAND - MAY 08:  Labour Party leader Ed Miliband and his wife Justine Thornton are greeted by supporters at Labour party headquarters on May 8, 2015 in London, England. After the United Kingdom went to the polls yesterday the Conservative party are confirmed as the winners of a closely fought general election which has returned David Cameron as Prime Minister with a slender majority for his party.  (Photo by Rob Stothard/Getty Images)

“Labour’s fundamental failure,” he adds, “was not to have persuaded people they had some idea of how to run the economy. You had to put 2008 behind you and they didn’t. It was still hanging around like a bloody millstone. Rule number one, you have to make people forget [the less successful parts of] your record in office. We kept on saying, you’re relying too much on protest votes and Lib Dems.”

Explanation comes easily after the fact. Curtice wasn’t as prophetic as he sounds. I overheard him briefing a group of foreign journalists a fortnight before election day. He was focused on how the SNP might influence a Labour government, not on the size of a then unconscionable Tory majority.

But the success of his exit poll has washed away his pre-poll predictions, and made him one of the few stars of the 2015 election. It also marked the fifth time in a row the exit poll has been very accurate. In 1997 it foresaw Blair’s landslide win; in 2001 it didn’t dispute his predicted re-election; in 2005 it predicted his majority exactly; and in 2010 it startled all by correctly foreseeing a Tory-led hung parliament and limited Lib Dem success.

2015 marked the fifth time in a row the exit poll has been very accurate.

The accuracy of the poll has been a repeated vindication of the way it is run. It is reported to cost the broadcasters at least £100,000, and yet is largely self administered, by Curtice and the team he has recruited.

Every election day, Curtice and his men – there have almost always been men – are locked away in a undisclosed location. They have twelve hours to build a prediction model on the fly, with data delivered to them in ‘drops’, or drips, throughout the day.

Clipboard holders are stationed at representative polling stations around the country, and ask every nth person who they have voted for. As the data is collected and sent in, Curtice’s team has to agree which patterns they are seeing, and which parts of the data to trust. Contentious decisions must be made rapidly by academics used to working over months and years, not hours.

The room has to operate like clockwork. It is not a place or time for the unhinged, free spirited or bolshie. Every appointment is considered. Curtice was first recruited by David Butler, his supervisor at Nuffield and the man who explained election nights on the BBC from their inception in the 1950s until 1979.


In turn, Curtice recruited Steve Fisher, now his effective number two, who was first drafted in as “little more than a tea boy” in 1997. He has taken on more responsibility at every election, and this year’s team included two of his former PhD students, Jon Mellon and Rob Ford, as well as a PhD student of Ford’s, Patrick English. There seems to be a decades-old hierarchy to the room. (The sixth member of the team is a statistician, Jouni Kuha.)

Screen Shot 2015-06-08 at 15.15.32

“Shall we say it is not a position that has ever been advertised,” Curtice quips when I put this to him. Like a Guardian editorship, the role seems to have a 20-30 year term. Fisher seems the most likely to take it on next – “Steve particularly gets it”, says Curtice – and, as the man behind Elections Etc, he was the most accurate of the many inaccurate pre-poll election forecasts.

It is hard to argue with this closed shop. It has delivered every year. But it is easy to see how Curtice’s star could one day fall. As Sam Woodhouse, the BBC’s election night editor, put it to me before election day, “If the exit poll is wrong, we’ll be set back thirty years.” It only takes one bad poll to ruin an endless string of good ones.

Curtice is aware. He was in the room on election day in 1992, when the exit poll nearly collapsed before it had been established.

At 9.15pm that night, Curtice was among the team who briefed the BBC’s top journalists on the exit poll’s findings so far. All the data up to that point had agreed with pre-poll forecasts, that Neil Kinnock’s Labour Party would be the largest party in a hung parliament. After the briefing, Curtice and his team were moved from their stowaway in Broadcasting House to the main studio. A final data ‘drop’ was due in the final half hour.

Curtice was in the room on election day in 1992, when the exit poll nearly collapsed.

But, as the exit pollsters relocated, their primitive, pre-internet connection broke down. Minutes away from 10pm, the final data drop still hadn’t arrived. The BBC didn’t have an exit poll. “I can remember Peter Snow shouting, ‘Where’s the exit poll! Where’s the exit poll! I don’t have the exit poll!’ And thinking, ‘Peter shut up – neither do we!’”

Finally the data downloaded. In a shock change, it put the Tories four seats ahead. The poll had finally picked up on some of the shy Tories who would hand Major a most unexpected of victories. Just before he went live, Dimbleby was told the exit poll was changing. As he went on-air, he didn’t know what the exit poll he was about to announce would say.

Recent years haven’t offered such white knuckle drama. Now Curtice just has to deal with regular scepticism and occasional hostility. In 2005 the veteran pollster Bob Worcester thought his Labour majority figure a third too low. In 2010, the BBC spent their first few hours distancing themselves from Curtice’s prediction.

The Prime Minister Introduces His New Members Of Parliament

This year Peter Kellner was one of many dissenting voices, with YouGov’s on-the-day election poll giving the Tories 50 seats too few. But it is Curtice’s methods, advanced at various times by Nuffield statisticians Phil Brown, Clive Payne and David Firth, which have always proven most accurate.

Over the next five years, his pronouncements are likely to be closely followed. And while many Tories will welcome his pessimism over Labour’s chances, he is scathing about Cameron. “Cameron’s only bloody vision is ‘We’ll make you a bit richer’. I thought his victory speech was remarkable. Here’s a man, second term, at the top of his game, and – ah – there was nothing!”

He’s similarly unsympathetic towards Clegg: he “managed to destroy 40 years of progress in 4 years; they’re back to the 1970s.” He thinks Tim Farron will take over the party and just “happily attack the Tories”, go back to “old Lib Dem themes”, and “begin to start again”.

He thinks the Lib Dems’ collapse was “a failure of, above all, bloody tuition fees. Rule number one, do not do the complete opposite of what you say you will do and your USP. They obviously would have hit a bumpy ride from half their voters whichever direction they took, but it just completely destroyed people’s trust in the party. What they delivered for poorer students was [actually] rather more generous, but they didn’t just raise them, they charged £9,000.”

As President of the British Polling Council, Curtice’s attention is now turning to why pre-election polls were so wrong, along with next year’s Scottish Parliament elections, but while academics continue to analyse, the political world’s gaze has begun to turn away.

The exit pollsters have disbanded. In five years, they will meet again, risking their reputations once more with a prediction made on-the-fly behind closed doors – and with 30 million voters hanging on their declaration at 10pm. •