A new feature: me commenting on the things that come through my letterbox - magazines, brochures, secondhand books, etc. Today, it's the new issue of the London Review of Books. As usual, the first issue of a new year carries Alan Bennett's diary for the old year (though this year it's been demoted to the back of the magazine: a healthy development, I think, since from its old vantage-point at the front it tended to overshadow everything else). À propos of the Obama inauguration, Bennett writes:
"...I don’t read any official estimates of the numbers though it’s to be
hoped they estimate more accurately in the US than they do here, where
any demonstration of which the police disapprove – the Stop the War
marches, for instance – is routinely marked down whereas demos on which
the police look kindly, the Countryside Alliance, say, are
correspondingly inflated."
The idea that the police talk down numbers at left-wing (broadly defined) protests and assemblies of all kinds is regarded as a truism by many on the left; but what evidence is there that it is true?
There have been many occasions when organisers of protests disagree with police over numbers present; but I can't see that the police have a strong incentive for talking the numbers down - indeed, I'd have thought that, like any other bureaucracy, they'd be wanting to talk up the difficulty of the job they face and hence boost their claim for resources. On the other hand, protesters themselves have a strong incentive for talking up the numbers.
I'm not aware, either, of any evidence of the police overestimating numbers at right-wing (broadly defined) protests: as far as the Countryside Alliance goes, one thing I recall from being, briefly, a spectator at their 2002 "Liberty and Livelihood" (blecch!) march is that they made everybody walk past a counting-point to give their numbers credibility; I don't know that the fuzz had much to do with it.
None of this is to say that Bennett is wrong: just that I have no idea why he's so sure he's right. This article suggests that police methodology may routinely underestimate crowd numbers, but not that there is any political bias: the readers' comments set out the arguments nicely.
The picture is from the Grunwick strike of the 1970s - the decade to which, according to the Daily Mail, we are now returning.