Jeff Randall and meaningless percentages

Jeff Randall has a piece in the Telegraph today entitled "Hypocrites, lightweights and clones – can't Labour do better?" where he rips into the Labour leadership contenders. I am not going to comment on the entire piece and indeed he does make some interesting points. However the details in the fifth paragraph are not among them:

At the last election, David Miliband's share of his constituency vote fell by 8.8 per cent, Andy Burnham's was down by 6.5 per cent, and Ed Miliband's by 3.8 per cent. Even in seats where Labour could put up mannequins with red rosettes and still win, these would-be champions managed to go backwards. Ed Balls, whose constituency had its boundaries redrawn, suffered a fall of 8.4 per cent and just scraped in. Only Diane Abbott, defending Hackney North, was able to increase her share of the vote (by 6 per cent).

So the strong implication being that those contenders whose percentage share of the vote dropped in their own seats are not worthy of the leadership of their party.

Perhaps Jeff should have checked his historical facts though. In 1970 when the Conservatives won the general election, in the Finchley constituency Margaret Thatcher managed to get 53.8% of the vote. But by the first election in 1974 when the Conservatives lost power, her share of the vote had dropped to 43.7%. So by Randall's own yardstick, the most electorally successful Conservative Prime Minister of the 20th century lost over 10% of the vote in her seat just a year before she became leader (in the election where her party lost power) and then went on to win the general election in 1979. That's a bigger drop than any of the current Labour leadership contenders.

In other words, use of these percentages is utterly meaningless as a predictor of how successful a putative leader will be.

PS: Thanks to Greg Stone and David Boothroyd for helping me find the Finchley data.