A conservative British tycoon and aristocrat is not the typical profile of a man of the people. But Lord Michael Ashcroft has done more than anyone else to find out who the people want to represent them in Parliament. In the lead-up to next month’s big national election in the U.K., Ashcroft has surveyed people in more than 100 electoral districts about who they’ll vote for and why. That may not sound radical, but in British polling, it is. Even more radical: Ashcroft isn’t doing it to bolster his own campaign (he just resigned from his unelected office) or because someone is paying him to do it. Instead, he’s shelling out millions mainly, observers think, because he’s curious.
The U.K. has 650 constituencies, each with about 100,000 people — less than one-seventh the population of the average U.S. congressional district. Each one has its own set of candidates running in the May 7 general election. Until now, British pollsters concentrated on regions, not constituencies, because they considered conducting surveys in so many small places to be too complicated and expensive. If the public wanted to know how much support the candidates in a given constituency had, it usually had to wait until the votes were counted. Most election watchers simply assumed that across regions, voters all moved in lockstep.
Ahead of this year’s election, all that has changed. Ashcroft has shined light on this dark corner of U.K. politics by sponsoring polls in roughly 150 of the most closely contested constituencies, from Airdrie & Shotts to Wyre Forest — some more than once. The British press has dubbed him the Pollfather, Lord of the Polls and Pollster-in-Chief. His finding that Nick Clegg, the deputy prime minister and leader of the Liberal Democrats, could lose his seat made headlines in the U.K. this month.
Ashcroft, 69, is a member of the Conservative Party and until last week represented it in the House of Lords.1 Despite his avowed partisanship, political scientists and other pollsters respect his polls and consider them unbiased. And they’ve influenced forecasts, including the one FiveThirtyEight is publishing, in dozens of races. All told, if Ashcroft weren’t polling locally, our forecast as of Sunday would have shown the most likely leader to be different in 46 races. The polls also highlight the surprising resilience of individual Liberal Democrat candidates despite the party’s collapse in national polls. Without Ashcroft’s polls, our latest forecast would have predicted the Lib Dems to win just 15 seats, a dozen fewer than their projected total when incorporating Ashcroft polls. Ashcroft’s biggest impact could come on election day if people make tactical decisions about who to vote for based on the poll findings.
The rise in constituency polling — fueled by Ashcroft, along with a couple of dozen surveys from pollster Survation — “really has changed the forecasting game,” said Will Jennings, a political scientist at the University of Southampton. “But we don’t yet know how much noise they have added as well as the important extra information.”
But so far, Ashcroft’s track record “seems to be very good,” said Robert Ford, who is a senior lecturer in politics at the University of Manchester and is forecasting the election with Jennings. “He uses reputable pollsters, asks the questions in a sensible way, and the polls have performed pretty well in by-elections, which are actually harder to poll than general elections due to low turnout and unpredictable local dynamics.”
Constituency-level polling is expensive — too expensive for most U.K. media organizations. “Telephone polling on this scale does not come cheap; otherwise everyone would be doing it,” said John Curtice, professor of politics at the University of Strathclyde. Chris Hanretty, a political scientist at the University of East Anglia who is helping produce our forecast with electionforecast.co.uk, estimated that a single constituency poll costs £5,000 to £7,000 (about $7,400 to $10,400). Combine that with Ashcroft’s national polling, and “he’s probably spent about £2 million on polling in the past 12 to 18 months,” Hanretty said in an email last month.
That cost is a small fraction of Ashcroft’s fortune. Forbes estimates that Ashcroft, who spent most of his career acquiring and turning around businesses in various industries, is worth $1.5 billion. In return for investing a few million quid on polling, Ashcroft gets broad media coverage, a platform from which to influence U.K. politics, and the opportunity to satisfy his curiosity about the British electorate, including to answer questions such as which actor should star in a hypothetical biopic about each candidate’s life.
“I don’t think there’s any great mystery,” Curtice said. “He’s got a boatload of money, and he’s got the ability to spend money doing it. I think he genuinely wants to understand what’s going on.”
Ashcroft began sponsoring polling a decade ago. In speeches and rare media interviews,2 Ashcroft says he became a pollster to fact-check Conservative Party leadership on what he saw as unearned confidence ahead of the 2005 election.
“I suspected that the party I supported was going to delude itself into thinking it was doing better than it really was — and might, as a result, do worse than it would do if it had a realistic grasp of the situation,” Ashcroft said in a recent speech. He eventually became the party’s deputy chairman with a special focus on opinion research, a position he held until after the 2010 election, when he stepped down following a scandal over his choice to retain a residency status that sheltered much of his wealth from U.K. taxes. He has since spent even more of his time and money on polling and on his news commentary site, ConservativeHome. In his statement announcing that he would step down from his position in the House of Lords, Ashcroft said: “My other activities do not permit me to devote the time that membership of the Lords properly requires.”3 Soon after, Ashcroft signaled his excitement about focusing on polling via Twitter, announcing a new spate of constituency polls with the words, “OK back to fun!!”
Professional pollsters and political scientists praise Ashcroft for disclosing the details of his polls and for fessing up when he errs. “He regards transparency as an essential element of polling, which is why all his data is published in full,” his spokeswoman Angela Entwistle said.
Ashcroft’s avowed commitment to transparency doesn’t extend to revealing who conducts his polls for him. Several people in the polling industry said it’s an open secret that the U.K. research company Populus conducts most of his polls.4 Most, but not all: Ashcroft’s usual pollster didn’t do three polls that turned out to have major methodological errors and that Ashcroft apologized for while vowing never again to use the pollster behind them.5
Entwistle didn’t answer a question about who conducted the polls. Inquiries to Populus and Survation went unanswered. At a London political event last month, I asked Rick Nye, who directs Populus’s political work, about his firm’s rumored work for Ashcroft, just after Nye had given a speech about the election aided by slides that featured lots of Ashcroft polling. “I really don’t want to get into all that,” Nye responded.
Curtice, who in addition to his academic job is the president of the British Polling Council,6 said it is common practice in the U.K. for the sponsor of a poll not to reveal who carried it out. “He is doing the polling,” Curtice said. “He is taking responsibility.” When Ashcroft found out about the three flawed polls, “he put his hand up and said he screwed up,” Curtice said.
The big question that we can’t answer for at least another month is how good Ashcroft’s polls are. When they zoomed in on voting intention at the local level, did they get the right picture?
Ashcroft himself, like many pollsters, disavows any role as a forecaster. “As he has often said, his polls are snapshots not predictions, and he is not making a forecast as to which party will control the government after the election,” Entwistle said.
There’s an intriguing possibility that the polls could prove to be bad predictors of voting outcomes without being flawed. That depends on whether the act of observing local voters itself changes their behavior — a sort of uncertainty principle for U.K. politics. The hypothetical scenario goes like this: Voters who favor a hopeless U.K. Independence Party candidate in their constituency might decide to back the Conservative candidate instead if polls suggest that the Conservative candidate is trailing her Labour opponent by just a percentage point or two. That could cost Labour a seat — and not punish the UKIP candidate, who had no real chance of winning.
Ashcroft himself has acknowledged the likelihood that his polls will spur tactical voting.
Political scientists aren’t sure how big an effect this could have. Several pointed to studies of how the Observer newspaper affected the 1997 general election when it polled 16 constituencies. The studies showed that the polls affected few races and that many people who tried to vote tactically had their votes backfire — casting their lots with a candidate who turned out to be less viable than another candidate they could have tolerated. Some also pointed out that many voters are dissatisfied with parties they aren’t planning to vote for and therefore are unlikely to find an alternative candidate they’re willing to back for tactical reasons.
“Many voters seem to have taken a pretty clear view on the government and/or other parties,” Jennings said.
Others say the biggest tactical effect that Ashcroft polls will have is on how parties devote resources to individual constituencies — and how they communicate with voters about the importance of backing their candidates to avoid the likely outcomes suggested by the polls. “Will enough learn about the poll results?” Ford asked. “Will enough believe them? Will learning of them change enough minds? These could be critical questions on May 7, but it’s impossible to say right now.”