It’s a bad idea to pick fights with the Institute for Fiscal Studies. It’s an impartial institution which is extremely careful to be fair. And they know what they’re doing. Indeed, the SNP itself likes to highlight IFS work when it is kind about them. But the SNP is having a go at the think-tank for its unhelpful arithmetic.
A standard question I put to the IFS earlier was about whether they had put the figures to the parties. The SNP had seen the IFS figures. They raised no complaints. It's also worth noting that the SNP is hard to follow on fiscal policy.
For example, the party talks about “spending” or "public expenditure" rising by 0.5% a year. That usually implies a figure known as TME, or Total Managed Expenditure. But, actually, that's not what the SNP means.
They mean a much narrower budget - the so-called TDEL, or "total departmental expenditure limits". That budget includes, for example, the cost of DWP officials in their offices but not the cost of pensions paid out to claimants.
It gets odder.
The SNP works out what goes into this budget line by adding three things together: resource departmental expenditure limits (RDEL), capital departmental expenditure limits (CDEL) and depreciation (wear and tear).
But this is problematic. Here's a line from a Treasury document explaining what goes in two of those categories:
Can you see the problem? Resource DEL includes depreciation. But then they add depreciation in on top again. Their sums double count this budget line.
This isn't a fiddle or dodge, it's a straight error. It means the SNP's own maths implies several billion pounds more additional spending than their statements imply. Is this why the IFS and the SNP have different outcomes?
The IFS actually replicate this error so they get to the same figures as the SNP. If they didn't replicate the mistake, it would imply lower SNP spending. They then distribute the extra spending as they think the SNP means it.
Their purpose is to, fair-mindedly, work out what their plans mean. It is not to trip people up on technicalities. That's why the IFS also don't tell anyone. Journalists with spreadsheets have to work out why nothing anyone says adds up.
I've noted some of the SNP's odd accountancy before. It remains slightly baffling: instead of scoring additional debt in the year when it accrues, the SNP has made the baffling decision to score it in the year after. So if they borrowed and spent an extra £1bn today, the spending would be scored in 2015-16 and the debt would appear in 2016-17.
Where are we now with Labour and the SNP? Labour is roughly in line with the SNP if it goes for budget balance in 2018-19. Both are planning historically slow spending increases (see Duncan at 12:35) which will require some budget trimming somewhere. No-one is proposing no cuts.
My estimates, which implicitly accept one point of SNP criticism about the IFS's treatment of tax avoidance, finds Labour has headroom to outspend the SNP in 2019-20 in all scenarios. But it's angels on pinheads.