Is the IFS right about the SNP's spending plans?

http://www.bbc.co.uk/news/election-2015-32480910

Version 0 of 1.

The Institute for Fiscal Studies is a quiet, careful body. I have quibbled with some of the central London think-tank's assumptions in the past, but it's decent, assiduous and unbiddable.

Indeed, that is why the Scottish National Party itself likes to highlight IFS work when it thinks they are helpful, as it did last Tuesday and Wednesday. But, as I have been noting over on Newsnight Live, the SNP is having a go at the think-tank for its unhelpful arithmetic about the party's general election plans. The IFS believes Labour's plans imply more spending than the SNP's.

But the SNP is on rather weak ground here. Alex Salmond, the former SNP leader, weighed in against the IFS analysis over the weekend. He wrote: "In reality, it doesn't take an Einstein-type analysis to understand that, given the SNP (with our Green and Plaid allies) are the only major party pledged to increase rather than decrease public spending."

I cannot follow this: "public spending" is usually taken to mean the overall spending figure known as TME, or "total managed expenditure". Every major party's plans imply a real-terms increase in spending from this year to 2019-20. Even the Tories' projections. Labour and the Liberal Democrats would increase it in each year.

It is possible he is using a different definition of spending. The SNP has a bit of form on this. For example, the SNP manifesto says it opposes spending cuts and proposes "modest spending increases - of 0.5 per cent above inflation - in each year of the next Parliament." That does not mean what you think it means.

The party does not mean TME. The SNP is talking about increasing the national budget by the equivalent of increasing one narrow budget line by 0.5% - so-called TDEL, or the "Total Departmental Expenditure Limits". That includes the cost of DWP officials in their offices but not pensions or benefits paid out to claimants.

Even on this measure, though, the SNP is not the only party proposing more spending. The Lib Dems would increase it over the parliament, and Labour would increase it every year.

Double depreciation

This mistake may have come about because the SNP persistently miscalculates how large the TDEL budget line actually is. They double-count one dull-sounding but quite big budget line: depreciation (wear and tear). The SNP's internal estimates of what its plans would mean for government spending plans are often bigger than their external claims actually imply.

That is not the cause of their arithmetic problems with the IFS. The think-tank, since they are kindly sorts, actually replicate this error and do not tell anyone about it. It's only because I was trying to replicate the maths of the SNP manifesto that I spotted it. And the IFS runs its sums past the parties before publishing. The SNP saw and signed off their estimates.

So what's the truth about the comparison with Labour? Well, the SNP has one legitimate complaint about the IFS's treatment of income from cracking down on tax avoidance. And, taking that into account, if Labour was to balance the current budget by 2018-19, the two parties would be basically tied: the SNP would have room to spend a little more than Labour in 2018-19.

But in 2019-20, Labour would have room to out-spend the SNP in the last year by about £9bn. Furthermore, if Labour was to aim to balance the current budget in 2019-20, which its manifesto also allows, it could out-spend the SNP throughout.

It's not clear what Labour actually plans to do. It is, however, arithmetically true that it has the fiscal headroom to out-spend the SNP every year. The big picture is the two sets of plans are roughly the same, and their plans could easily be married off. As for the SNP, if it published a table of its proposals using conventional accounting standards, these disputes could be avoided.

An accounting mystery

I note "using conventional accounting standards" for a reason. As I've pointed out on Newsnight Live, I've recently spotted something quite odd. A recent Scottish Government note sets out a novel way of accounting for debt. They report the "Treasury" method:

And the Scottish Government method:

What does this mean? If I borrow and spend £1bn today, their approach would mean the spending happens in 2015-16 and the debt doesn't appear until 2016-7. It is delaying the moment at which it scores debt for a year. But you owe debt when you owe debt. There's really no good rationale for claiming otherwise. The SNP is using this approach.

I am quite surprised that a civil servant allowed this methodology to make it out into public. No auditor (nor the Office of National Statistics) would permit it.