The Organisation for Economic Co-operation and Development (OECD) has just given us its considered opinion on the likely costs of Brexit.
And as the more canny Scots like to insist about the English, we should perhaps send them home to think again. For what its done in presenting their numbers is one of those near unpardonable statistical sins.
It's presenting what will be a small annual cost, even by its calculations, and rolling it up over the years into being a much larger number with which it can scare us all. And whatever your views on Brexit (mine are, "yes, yesterday") we should all at least agree that we're not going to start manipulating the evidence on this most important question.
In its report the OECD says this:
By 2020, GDP would be over three per cent smaller than otherwise (with continued EU membership), equivalent to a cost per household of £2,200 (in today’s prices).
By 2030, in a central scenario GDP would be over five per cent lower than otherwise – with the cost of Brexit equivalent to £3,200 per household (in today’s prices).
Two grand a year? Eeek.
But that's not what is being said at all. For median (the number which 50 per cent of households are above, 50 per cent below) household income in the UK isn't all that high. Three per cent of it is about £700.
So what the OECD has done is say it will cost £700 in 2018, £700 in 2019 and in again 2020. Then it's added those three years together to give us that total cost of some £2,200 (I rounded a bit) - which is more than a bit off really. In fact it stinks. It's supposed to be the impartial international bureaucracy, not political partisans presenting scare figures as rhetoric.
Just to emphasise that rhetorical effect, look at that longer term prediction, the one out to 2030.
That again is not an annual number - it's the total cumulative effect over that period of time. It's also an estimate at the very boundary of our possible knowledge (five per cent more or less on GDP in 14 years' time probably is beyond our ability to predict) and it's not a difference from current numbers, but from what they would be in that unknowable world of our not leaving.
And then cast that same number into the rhetoric of that nice Mr. Farage (who I used to work for).
“It'll cost you less than 300 quid a year to be free of Brussels.”
It's just not so scary, is it? And the very fact it isn't shows quite how much political rhetoric there is behind the way the OECD has presented these figures.
Owen Barder and Jonathan Portes both say that the OECD's trick has been a little different. Instead of measuring the costs against household income, as most will assume they are doing, they are dividing total GDP by the number of households.
But this is a similarly absurd method of measuring these costs and produces the same result. Household costs are over-estimated by roughly three times.