A myth that will not die
The brief media fuss over Michael Fabricant's paper on some sort of Con-UKIP pact has included various references to old canard that UKIP cost the Conservatives 21 or more seats at the last election.
"At the General Election in 2010, UKIP cost the Conservatives, on a realistic estimate, around 20–40 seats, just by standing and taking a mere 3-4 % of the vote."
This is nonsense of the highest order, but however often the argument is knocked down it refuses to lie down and die. At the last election there were 21 seats where the number of votes that UKIP got was larger than the number of extra votes the Conservatives needed to win. Where the argument that UKIP cost the Conservatives up to 40 seats comes from is beyond me, if every single UKIP voter has instead voted Conservative they would have got 21 more seats, no more.
However that is something of a big ask... every single UKIP voter voting Conservative? UKIP support has historically come disproportionately from the Conservatives, but not exclusively so, they take some votes from Labour, the Liberal Democrats and people who would otherwise not vote. Neither would those people miraculously go back to the Conservatives in the absence of UKIP - they were, presumably, unhappy enough with the Conservatives to want to vote against them, many would have found an alternate route to express their unhappiness, or just stayed at home.
Back in 2010 YouGov did a survey of 706 people who voted UKIP in the 2010 general election. It included questions asking which parties they would consider voting for in the future, and which parties they had voted for in the past.
Looking forward only 27% said they would consider voting Conservative in the future (compared to 9% who said they would consider voting Labour and 8% who'd consider voting Lib Dem). Looking back 20% of them had voted Tory in 2005, 12% were people who had backed Labour in 2005.
If we use that data, we can make an estimate of how many seats UKIP might have actually cost the Conservatives. Say UKIP had vanished, the 27% of 2010 UKIP voters who said they'd consider voting Tory had done so and the 8% and 9% who said they'd consider voting Lib Dem and Labour had done that...the Conservatives then would have won an extra five seats. If instead we look at how people voted in 2005, and assume people went back to their old voting habits in the absence of UKIP, then the Conservatives would have won one or two extra seats.
The idea that UKIP stopped the Conservatives winning 21 seats is lazy nonsense based on unsustainable assumptions (delightfully called "the politics of not understanding data" by Justin Fisher yesterday), a myth that deserves putting to bed for once and for all. A better estimate is around five seats.
