Osborne's Brexit cost 'forecast' and regular budget forecast uncertainty
'Leavers' were out in force trying to denounce the Treasury's report on the costs of Brexit, arguing that if Osborne could not forecast the deficit accurately a few years ahead, why should anyone take seriously his forecast of the cost of leaving the EU 13 years ahead? Prime offenders were Tim Montgomerie, John Redwood, Iain Martin [from CapEx] and Dominic Cummings, from VoteLeave.
[Previous version of post fingered Andrew Lilico as offending here, but this was an error, for which apologies].
Giles Wilkes of the FT trumped several tweets I contrived about conditional forecasting with the observation that he might not be able to forecast his own weight in 2030, but he can nevertheless accept the analysis that if he eats butter, that will make him heavier, other things equal.
To make the point much less neatly, the Treasury document is comparing the difference between two forecasts, all else being held equal except the EU membership status. While we might nonetheless be uncertain about the effects of EU membership [we surely are, despite the overwhelming preponderance of analysis that the central tendency is that it is beneficial] all the other uncertainties are cancelled out by us comparing the difference between these two forecasts.
Who knows if there will be another regional war. Of if oil prices will recover before 2030. Or if climate change might accelerate to the point where radical measures are taken. Or if global central banks will escape the zero bound. All these things cloud our ability to predict our incomes in 2030. But they do not cloud our ability to predict the difference between our incomes in and out of the EU. [At least, under the assumption that membership doesn't affect climate change, or the chances of an oil price rise, or a regional war].
I don't know what this tactic says about the Leave campaign: whether it means that they don't understand this analytical point, or they do but are trying to blur it for political advantage.