Sunday, 29 December 2019

What GERS can and cannot tell us

This meme is frequently shared by independence supporters as if it's some kind of "gotcha!". The quote comes from a radio interview I did with the perpetually confused Richard Murphy. He was indulging in his usual schtick of dismissing the GERS figures as "made up" and criticising them for presenting actual historical figures instead of hypothesising some imagined future scenario of Scottish independence.

Here is the verbatim extract, you can check for yourself

Murphy: "[...] let's not pretend GERS gives any useful information to a politician in Scotland, because it doesn't, it's literally made up information from the the rest of the UK, we have no idea if it's accurate [...] GERS cannot tell us [...] whether Scotland can survive independently or not, this data is not fit for purpose"1
Hague: "Well,  I mean first of all, the key thing is they are fit for purpose - the most important thing is understanding what that purpose is. So Richard finished his piece there by saying they can't tell us what an independent Scotland's finances would look like - that is of course absolutely true; what the GERS figures do tell us is, historically, how do Scotland's revenue and expenditure figures look as an integral part of the UK [...] as long as you understand the data, as long as you understand what the data can and cannot be used for, it is good data [...] But Richard is right on one point, which is all it tells is is where we start from, it tells us how the Scottish economy currently performs as an integral part of the UK - and that is all it can do, that is all it is expected to do"

The fact that the GERS figures are historical actuals that describe Scotland's economy as an integral part of the UK should hardly be a revelation to anybody, but the nature of online debate is such that this observation is somehow considered meme-worthy.

If it needs spelling out any further2, then let me do so again here:
The GERS figures describe Scotland’s historical finances, with Scotland raising taxes and incurring public spending (including reserved expenditure) as an integral part of the UK. From the perspective of hypothesising a possible future independent Scotland, we are looking at what in financial accounting terms would be considered pro-forma accounts.
So of course the figures do not tell us what the future accounts of an independent Scotland would look like - how could they? They do however describe the starting point (the “run-rate”) from where we can start to consider the possible impact and fiscal implications of independence.
Precisely how independence would change Scotland’s economy is of course a hugely complicated subject that require us to consider, amongst other factors;
  • The outcomes of uncertain negotiations around inherited share of UK debt and potential EU membership
  • The challenges of either sustaining money-supply under Sterlingisation or of building the reserves requires to support a new independent Scottish currency [see here]
  • The explicit tax and spend choices that the government of an independent Scotland might make (inevitably constrained by the outcomes and decisions above) which would include choices around wealth redistribution, defence spending, industrial and economic policy, international affairs, debt and deficit management, social policy priorities and much more
  • The costs involved in replicating the delivery of services that currently rely on shared infrastructure with the rest of the UK (e.g DWP and HMRC functions)
  • Depending on the outcome of EU membership negotiations, the possible impact of trade friction between Scotland and  the rest of the UK (the destination for 60% of Scotland's exports)
  • The impact of factors outside the Scottish Government’s direct control such as how businesses and the labour force would respond (e.g. possible capital flight), the global oil price, international credit ratings and the cost of Scotland’s debt
  • The cumulative effect of all of the above (positive or negative) on Scotland’s economic growth

Now look again at the beginning of the Murphy quote above: when he says “let's not pretend GERS gives any useful information to a politician in Scotland, because it doesn't, it's literally made up information from the the rest of the UK, we have no idea if it's accurate" he is insulting: the Scottish Government’s own economists who compile them; the National Statistics Authority who award them National Statistics accreditation; the authors of both the Independence White paper and the SNP's own Sustainable Growth Commission who used the GERS figures as the basis for all of their analysis.

In fact, there are no respected economists who dimiss the GERS figures as "bad data"

I've dealt with the details of Murphy's publicity-seeking confusion over GERS before here but- for completeness - let's remind ourselves what happened when a well-briefed politician challenged Murphy on some of his more extreme assertions in front of a Holyrood committee:



1. His opening ramble is over 2 minutes long - this is very much his style: he chucks out so many daft assertions in one go that it becomes impossible to refute them all in response without descending into a similarly lengthy and confusing statement. His objective is obfuscation and - give him his due - that's one thing he is good at it

2. I really have been very consistent about this - the following from 2016 alone
"Nobody is arguing that things would remain the same. Those of us who argue for rational debate simply ask for those making the case for independence to actually explain coherently what the different "economic strategy" would actually be and provide a realistic assessment of how (and by how much) it would change the figures compared to those of Scotland being within the UK." - Chokkablog August 2016
"We should be very clear about what this analysis of historical fiscal data can and cannot tell us. The figures only tell us how an independent Scotland’s finances would have looked if we had already been independent but were still raising taxes and incurring public spending (including reserved expenditure) as we have been as an integral part of the UK. We are looking at what in financial accounting terms would be considered pro-forma accounts. The figures do not tell us what the future accounts of an independent Scotland would look like. They do however describe the starting point (the “run-rate”) from where we can start to consider the possible impact and fiscal implications of independence." - Chokkablog March 2016
"All I have been attempting to do is ensure we have clarity around our starting point. Were we to be independent or fiscally autonomous now, what would our pro-forma accounts look like? What is our economy's run-rate? This frames the debate, shows the size of the challenge. If we are all honest about this starting point then maybe we can have an interesting and constructive debate" - Chokkablog June 2016
"This onshore deficit gap matters because it is revealed - it becomes real - as oil revenues decline. This is not to say that were Scotland to be independent this gap would remain; it might narrow, it might widen. It merely gives us an idea of the run-rate relative disadvantage we would be starting with if we sacrificed the benefits of UK-wide pooling and sharing (assuming the days of significant oil revenues are indeed behind us). If you like, it's the head-start we'd be giving to the rest of the UK." - Chokkablog February 2016 

Saturday, 21 December 2019

How Much of Scotland's Tax Revenue Does Westminster Keep?


More spending occurs in Scotland than tax revenues raised by Scotland.

In fact, more money is spent in Scotland than the sum of tax revenues raised by Scotland plus Scotland's per capita share of the UK's overall deficit (i.e. there is a fiscal transfer in Scotland's favour after allowing for Scotland's share of the increase in UK debt).

Both the statements above are true even before taking into consideration Scotland's share of defence costs, international aid and debt interest.

It is simply not true to argue that Westminster "keeps" any of Scotland's tax revenues

A grievance persistently pushed by nationalists on social media is that "Westminster keeps" a chunk of Scotland's tax revenues - if you've been directed to this blog post, you may even have joined this merry gang yourself

The numbers tend to differ because the people who make such claims tend not to be too interested in actual figures or data audit-trails, but the theme is consistent. For the purposes of illustration let me take one example that appeared in my timeline yesterday and quickly garnered over 1,000 likes1

I'll use the format of the above tweet to walk through the actual figures, always with reference to Scottish Government's GERS figures for 2018/19 . These are figures compiled by Scottish Government economists in St Andrews House and which qualify as National Statistics. In all cases I will reference page and table numbers from this latest GERS report.

"We raise £63bn"This is correct

Scotland (including a full geographic share of oil revenues) raised £62.7bn of revenue in 2018/19 [p4. Table S.3]

"We keep £35bn" - This is incorrect

It's clear this figure is meant to be spending Scotland controls, because the other figures are money Westminster "keeps" or "Spends [..] on our behalf".

The correct figure is therefore £44.7bn, 
this being money spent by "Scottish Government, LAs and Public Corporations" [p33, Table 3.8]

"Westminster keeps £28bn [and] spends a further £12bn on our behalf" - This is also incorrect

The balance of spending allocated to Scotland that is spent by "Other UK Government" (aka Reserved spending) is in fact £30.6bn [p33, Table 3.8]

Sanity check opportunity: above we've accounted for £62.7bn of revenue and £75.3bn [£44.7bn + £30.6bn] of expenditure - that explains a deficit of  £12.6bn which was indeed the GERS deficit in 2018/19 [p2. Table S.1]

So we now have the correct figure for the attempted grievance: the £12.6bn GERS deficit includes £30.6bn spent by "other UK Government" on Scotland's behalf (aka "Reserved Expenditure").

Sanity check opportunity: "Devolved expenditure including housing benefit" [p45, Table 4.8] is shown as £44.7bn "before Scotland Acts 2012 and 2016" - this is consistent with our £44.7bn figure above, hence our £30.6bn "Other UK Government Spending" is what is also referred to as "Reserved Spending" (being that spending which is not devolved). Table 4.8 shows that, of currently Reserved expenditure, an additional £3.1bn will be devolved once these Acts are fully enforced - all of this £3.1bn is related to additional powers over Social protection, as detailed on p43, Table 4.6

All we need to do now is understand how that £30.6bn is spent and we can decide how aggrieved to feel.

I'll run through the detail at the foot of this blog, but of that £30.6bn Reserved Expenditure:

£20.0bn is spent in and for Scotland, largely on Social Protection, but also Transport, Research Grants, Tax Credits etc.

£2.9bn is spent for Scotland on things like DWP, HMRC, the BBC and Nuclear decommissioning costs which also involve spending in Scotland. While £0.1bn less than BBC expenditure allocated for Scotland is actually spent in Scotland, the reverse is true in the case of Nuclear Decommissioning costs and very likely the case for DWP and HMRC costs as well. Scotland gets the "fiscal multiplier" benefit of this spending already and would still need to fund delivery of these services if independent.

Pause for breath: At this point we have explained Scotland raising £62.7bn, having £44.7bn of Devolved Spending and receiving a further £22.9bn of Reserved Spending spent in Scotland - so that's £4.9bn more spent in Scotland than raised in Scotland. Scotland's population share of the UK's deficit is just £1.9bn, so we're already receiving an effective fiscal transfer of £3.0bn by this point

£3.6bn is Public Sector Debt Interest - the SNP's own Sustainable Growth Commission accepted the principle that an independent Scotland would need to service this debt

£3.2bn is Defence Spending - in the most recent year this was 1.8% of Scottish GDP, compared with the NATO member target of 2.0%. Trident is likely to be responsible for just £0.2bn of this figure and of course some of that £3.2bn is spent in Scotland.

£0.9bn is International Services (basically Foreign Aid) and again: the SNP's own Sustainable Growth Commission accepted the principle of maintaining this expenditure.

Check this all adds up: we'd got to a £4.9bn deficit, now we've explained the remaining £7.7bn that gets us to the correct £12.6bn GERS deficit

So people can have a bun-fight about what these figures would look like for an independent Scotland, but it's simply nonsensical to suggest that under current arrangements Westminster in any way "keeps" any of the tax money Scotland raises.

More is spent in Scotland than Scotland raises in taxes and the big sums that are spent for but not in Scotland are (with the possible exception of some defence spending) costs that an independent Scotland would still expect to incur.


The Painful Detail, for the super-keen

The Bad News
To do this properly you need to go through the GERS Supplementary Data: Detailed Expenditure Database. For each of the line-items in Table 3.8 you can see there the detail broken down as per the UK Government Country and Regional Analysis (CRA) Categories - you'll be surfing a database with 4,225 rows. In each case you can then see which items of expenditure are considered "Identifiable" (that is expenditure that can be clearly allocated to Scotland as having been spent for the benefit of Scotland) and which are not (aka "ID" vs "Non-ID"). Just because expenditure is considered "Identifiable" doesn't mean that expenditure is necessarily spent in Scotland - but you can then email the Chief Economic Adviser to the Scottish Government and ask for some help.

The Good News
I've already done this so can now summarise it for you.

The full audit-trail is only available for 2017-18 (because of the timing of CRA data and GERS data publications) - but fortunately Reserved Expenditure in 2017-18 was £30.7bn, so almost identical to the most recent year.

Of that £30.7bn, £20.1bn is considered Identifiable (spent for the benefit of Scotland):

£17.4bn = Social Protection (Pensions, Tax Credits, Allowances and Benefits)
£1.1bn   = Transport (mainly Network Rail expenditure in Scotland)
£0.5bn   = Science & Technology (Renewable Heat Incentive and various Research Council grants)
£0.4bn   = Enterprise & Economic Development (mainly Tax Reliefs & Tax Credits)
£0.2m    = Employment Policies (Department of Work & Pensions)
£0.5bn   = Other (Lottery Grants, Medical Research Council, NERC, etc. etc.)

But what about expenses that are included in this £20.1bn of "Indentifiable" expenditure that people might challenge as being "fair" or "appropriate" to allocate to Scotland?  Well I've been through the detail and I'm used to the examples most often raised, so let's cover them off here. Of that £20.1bn considered "Identifiable" expenditure for Scotland, the following costs are not spent in Scotland:

£57m = High Speed 2 (based on Scotland's expected share of the economic benefit)
£35m = House of Commons and House of Lords

That's it - if you don't believe me go and look for yourself. So frankly those costs are no more than a rounding error: we can say with confidence that £20bn of Reserved Expenditure is considered as Identifable expenditure for Scotland and is money spent directly in Scotland.

So where does that get us to?

Revenue Raised  = £62.7bn
Devolved Expenditure = £44.7bn
Reserved Expenditure For and In Scotland = £20.1bn

So at this point Scotland has a deficit of £2.1bn - so where does the other £10.5bn of deficit come from?

The big figures are well known - these are UK costs that are allocated to Scotland on the basis of Scotland's population share:

Public Sector Debt Interest = £3.6bn
International Services = £0.9bn (almost entirely Foreign Aid)
Defence Expenditure = £3.2bn

The first two figures were accepted by the SNP's own Sustainable Growth Commission, who proposed an annual "Solidarity Payment" to be paid to rUK by an independent Scotland of £5bn pa. They explained this £5bn pa as being: "debt servicing contributions, 0.7% of GNP for Foreign Aid and a further £1bn set aside for shared services"2.

The defence expenditure figure works out as 1.9% of GDP (1.8% in 2018-19) which should be compared with the 2.0% target expenditure level for NATO members. A common refrain at this point is "but Nukes" or "what about Trident" - well there are plenty of debates to be had there, but focusing purely on the economics: the annual cost of Trident included in this defence expenditure figure for Scotland is likely to be just £0.2bn3.

Of course Scotland could choose to spend less on defence - maybe choose not be a NATO member - those arguing the case for independence are free to make their case for what that £3.2bn cost would be replaced by in an iScotland.

The SNP's Growth Commission was clear that an iScotland would service its population share of UK debt (while trying to build credibility on the international stage and preparing to launch its own currency) and would continue to be a provider of foreign aid. But again: those arguing the case for independenec are free to make alternative arguments.

So what are we left to make up the "other" £2.8bn of non-identifiable expenditure allocated to Scotland in GERS?

£0.5bn = Social Protection: This is DWP benefit delivery and what appears to be some legacy pension costs - as my colleague @staylorish has pointed out before, 12.4% of DWP employees are in Scotland, so there is probably more DWP benefit delivery cost spent in Scotland than allocated to Scotland (based on population share in GERS)

£0.5bn = Public and Common Services: this includes £299m of HMRC costs and - again as @staylorish points out - 14.2% of HMRC employees are in Scotland so again probably more HMRC cost is spent in Scotland than allocated to Scotland (based on population share in GERS). The £0.5bn figure also includes c.£0.2bn of Cabinet Office, DfID etc. costs.

** Correction / Clarification **

Mea culpa: the 12.4% and 14.2% figures above are incorrect. They form no part of the analytical audit trail, but were wrongly used to illustrate my point that "probably more is spent in Scotland than allocated to scotland" in those areas. This is hardly a material issue, but I care about getting things right, so here we go:

As Sam Taylor points out in the thread I link to above, he made a (quickly corrected) error in his intitial calculation which I mistakenly quoted - he has subsequently updated the analysis anyway (see here) and the correct figures for share of employees are DWP 10.3% and HMRC 11.7%.

In both cases these percentatge are still materially greater than 8.2% (the population share of the non-Identifiable elements of these costs allocated to Scotland in GERS) so I think my "probably more is spent in Scotland than allocated to scotland" point still stands - but I'm happy to hold my hands up, admit to the error on those specifc figures and offer the correct figures here - it's what I've always said I'll do if an error is spotted on Chokkablog,

In addition, some confusion seems to have arisen because at this point I am discussing the non-identifable element of DWP benefit delivery costs, which are as I correctly say allocated on a population share basis.  In GERS there are also identifiable DWP benefit delivery costs (as shown in the first of the back-up tables below) of which 10.3% are allocated to Scotland in GERS. This means that of all DWP benefit delivery costs (ID and non-ID), Scotland is allocated an 8.98% share.

I accept that my slip above did overstate the share of DWP and HMRC FTEs in Scotland and could have led a careless reader to believe that all DWP delivery costs are allocated on a population share basis.

Somebody got very excited on Twitter about this, claiming that spotting this slip was "sufficient to defeat your point". This is like dancing on the head of a pin while steel girders are falling about your ears, but for those who might not have grasped the materiality here;
  • The difference between 8.2% (the non-ID allocation) and the 9.0% (the total of ID and non-ID allocation) for DWP delivery costs being discussed here is just £21m
  • The point still stands that a greater proportion of HMRC and DWP employees are in Scotland than the share of HMRC and DWP costs allocated to Scotland in GERS
  • Before considering any of these non-ID costs, per the audit-trail above Scotland is already receiving £2.0bn more in expenditure than it "sends to Westminster"
  • The "we send more to Westminster than we get back" nonsense involves ignoring £20.0bn of Identifiable expenditure and all non-identifiable expenditure - even if the above observations don't convince you that "there is probably more DWP benefit delivery cost spent in Scotland than allocated to Scotland", nobody can credibly argue that none of these cost are spent in Scotland

£0.3bn = Recreation, Culture and Religion: this includes £0.3bn for BBC of which £0.2bn is spent in Scoltand

£0.2bn = Environment Protection: mainly £197m allocated Nuclear decommissioning costs - in fact £57m more than that is spent in Scotland [GERS Table A.6]

£0.7bn = Accounting adjustments - this is just technical stuff to get the figures in comparable format with other National Statistics

£0.2bn = EU Transactions - let's not go there

£0.4bn = "Other" - I'm getting tired and bored of this, so see the tables below if you have the enthusiasm to chase this detail down futher (I have and can assure you there's nothing exciting to see)

The following tables back-up each rows of the table above based on querying the GERS supplementary expenditure database - column headings are Total, ID and non-ID:


1. I replied with a brief thread explaining why that tweet (which had been retweeted by high profile campaigners for Scottish independence including at least 1 SNP MP) was wrong - it has since been deleted so I am not including the twitter handle in the screen cap

2. Sustainable Growth Commission report, p37, 3.139

3. According to Full Fact, the annual running cost of Trident is c.£2bn, of which Scotland's population share would be just £0.2bn

Thursday, 5 December 2019

Dissecting a Deception

Nicola Sturgeon keeps trying to mislead people by repeating a demonstrably false assertion. The truth is that if the SNP's Growth Commission recommendations had been applied to the conditions as they actually were a decade ago, Scotland would have suffered far greater austerity over the last 10 years - and we can prove it


That Ming-bearded mind-bothererTM Derren Brown1 is fond of telling his audiences that he achieves the seemingly impossible via a combination of "suggestion, psychology, misdirection and showmanship".

There's an assertion being repeated so often by Nicola Sturgeon that I'm amazed none of the mainstream fact-checking organisations have bothered to check it. It goes like this:
"if the [SNP] Growth Commission recommendations on spending in an independent Scotland had been applied over the past 10 years, Scotland wouldn't have suffered the austerity cuts to our budget that we have suffered." - Nicola Sturgeon
That quote is lifted directly from her recent interview with Andrew Neil on BBC1 - she made the same asssertion to ITV's Robert Peston2 as well repeating it during the STV leaders' debate3 and, although I didn't listen to her call-in with Nicky Campbell on BBC Radio 5 Live, I'm told she repeated it there as well.

To understand the history of this assertion and how the SNP have tried to defend it is to understand how confident they are that "misdirection and showmanship" can allow them to claim the seemingly impossible.

The first time I heard this claim being made was by SNP MSP Kate Forbes (a member of the Growth Commission) on Question Time back in May 2018:
"Over the last 10 years the Scottish Budget has been cut by 8.5%; in contrast, this report predicts that if we had been an independent country our spending could have increased by 5% over those 10 years. Minus 8 to plus 5." - Kate Forbes
I then discovered that on the same day, during FMQ's in Holyrood, Nicola Sturgeon had told the chamber:
"I've got some analysis here which I'm going to share with the chamber, and hopefully it'll be ... it'll be of embarrassment to the Tories, hopefully it'll be of interest to Labour. If the spending recommendations of the Growth Commission had been applied over the past ten years, the £2.6bn real-terms cuts imposed on the budget of the Scottish government by Tory governments at Westminster would have been completely wiped out, it would have eradicated austerity in Scotland. That is the reality." - Nicola Sturgeon
Then the report's author, Andrew Wilson, used his column in the National a few days later to repeat the claim again:
"If the model we have suggested for reducing the deficit was applied to the last 10 years, it would have eliminated the Tory austerity cuts to the Scottish Budget." - Andrew Wilson
I think we can safely say that the SNP have been committed to this line for some time now - and they've decided that this general election campaign is the right time to crank up the volume.

Back in June 2018 I pointed out that those assertions were completely at odds with the Growth Commission's actual recommendations (I blogged on the subject here and here). This matters, because as a direct result we know how the SNP seek to justify themselves on this.

First of all let's be clear about what the Growth Commission actually recommended. I covered this in great detail in a report that was reviewed and endorsed by senior economists at the time [These Islands Growth Commission Response: The Truth About Austerity], but in essence their recommendation is simple (emphasis mine):
"real increases in public spending should be limited to sufficiently less than GDP growth over the business cycle to reduce the deficit to below 3% within 5 to 10 years" - Growth Commission 3.187
This isn't a selective quote - the recommendation is repeated four times4 in the report and every time it's made clear that spending growth must be sufficently less than GDP growth to reduce the deficit to 3% within 10 years5.

Even before we look at how the SNP attempt to justify their "we'd have escaped Tory austerity2"claim, it should already be obvious that their assertion is false: the Commission's recommendation is to get the deficit under control within a 10 year period entirely through spending restraint6  - so if the current deficit is still above 3%, then over the last 10 years their recommendations would have required Scotland to have reduced spending by more than has been the case.

Scotland's deficit in 2018-19 was in fact £12.6bn or 7.0% of GDP7. For that to have been reduced to 3.0% through spending constraint alone over the prior decade, annual spending now would have to be £7.2bn less - that implies a 9.6% reduction to Scotland's current Total Managed Expenditure8.

It's patently obvious that applying the Growth Commission's rules over the past decade would have led to far deeper austerity than "Tory austerity".

So let's see how the SNP try to justify their transparently indefensible claim.

Justification Attempt No. 1

The first justification was offered directly by Kate Forbes herself, who had the good grace to reply to me when I challenged her about what she'd said on BBC's Question Time. She DM'd me the following 9  (remember: she was a member of the Growth Commission):
"its v simplistically based on what the GC recommends in terms of increasing public spending in Scotland over a 10 year period (by 0.5% per year (5.1% compounded))."
The obvious problem here is that the 0.5% figure is arrived at based on two key assumptions
  1. The starting deficit position - to determine by how much less than GDP growth spending would need to grow to get the deficit to 3% within 10 years (let's call it the "spending lag")
  2. Real annual GDP growth - to determine what real spending growth would be if it was suffciently less than GDP growth
To get to the 0.5% figure the Growth Commission assumed
  1. An onshore deficit starting position of 5.9%10 which - as they modelled it11 - meant that spending growth would have to lag GDP growth by 1.0% to get to the 3% target after 10 years.
  2. Real GDP growth at a "long-term trend" rate of 1.5% - so spending lagging this by 1.0% would mean real spending growth of 0.5% pa.
Now it's surely obvious that you can't apply the 0.5% figure to the last 10 years, because the starting position 10 years ago was a worse deficit and GDP growth was of course slower than "long-term trend" - that's why austerity was deemed necessary in the first place.

So Justification Attempt No. 1 is clearly nonsense - and although it's not really our job, for fun let's try and see what happens if we apply the actual figures
  1. 10 years ago, Scotland's onshore defict was 8.8%12. Using a recreated version of the GC's model suggests that spending growth would have to lag GDP growth by 1.5 - 1.9% to get to the 3% deficit target after 10 years13.
  2. Actual real GDP growth for Scotland over that 10 year period was 1.0%14 - so spending lagging this by 1.5 - 1.9% would mean a real spending decline of (0.5) - (0.9)% pa

Depending how you choose to define "the last 10 years", average annual real spending growth (TME8) for Scotland within the UK has been 0.9% (06/07 to 16/17), 0.6% (07/08 to 17/19) or 0.4% (08/09 to 18/19).

So however you cut it, the reality of the last 10 years hasn't been anything like as bad as a 0.5 - 0.9% real annual spending decline. Of course it hasn't - if it had been Scotland's deficit would now already have been reduced to below 3%.

Justification Attempt No. 2

Freedom of Information requests are funny things. I stumbled across this one a while back and it is extremely revealing. It includes emails that were being exchanged with the SNP on the same day that our These Islands response to the Growth Commission was released. The following caught my eye:

It's nice to know that "Special Adviser to the First Minister" Callum McCaig was paying attention (*waves*) - I wonder who the redacted recipient of this email could possibly have been?

But what should interest us most here is what this tells us about the analysis they used (internally) to justify their claim - presumably this is the "I've got some analysis" that Sturgeon referred to in the Chamber.

First of all, we have the familiar SNP trick of playing fast and loose with the concept time.

We've recently (November 2019) seen Ian Blackford referring to "the last 40 years" when he's actually talking about a 32 year period up to 2011-12. Now we discover that in July 2018 when Sturgeon referred to "the last 10 years" she was actually referring to a period that included 2 years into the future!

But the SpAd is right when he says that the starting year makes a difference - what they've done is ignore 3 years of generous spending increases which, needless to say, had they been applying the Growth Commission's recommendations then Scotland wouldn't have seen. Over the three years they conveniently ignore, average real expenditure (TME) increased by 2.4% pa which was an average of 3.4% pa more than GDP growth (because recession)

["bailing out the banks" is a red-herring - GERS only sees a per capita share of the "P&L" impact of the bail-outs (see here) - the impact in 2010 GERS was just £0.7bn or 0.5% of GDP, hardly significant in the context of the numbers we're considering]

What they apparently have now done is applied the 1.0% "spending lag" to the actual GDP growth rates over their "last 7 years plus the next 3 years forecast" definition of the last 10 years (i.e. instead of blindly applying the 0.5% real growth as per Justification Attempt No. 1) .

But there's a huge problem here - they've applied that spending lag of 1.0% which was calculated based on starting deficit of 5.5%. The actual onshore deficit in their chosen base year of 2010/11  - precisely because of those years of spending growth they've conveniently ignored - was 13.2% 14

If you're following what's going on here the issue is obvious: with that starting deficit, to get the deficit down to 3% within a decade (the Growth Commission's First Fiscal Rule) would require a "spending lag" of far greater than 1.0%.

So we've shown that Justification Attempt No. 2 is also nonsense - and again, althought it's not really  our job, for fun we can have a quick go at estimating the "spending lag" required to get from a 13.2% deficit to a 3.0% deficit in 10 years. I have had a go (using the Growth Commission's modelling methodology) and I reckon the figure becomes something like a 3.0% spending lag required.

So even taking the "long-term trend" GDP growth rate of 1.5% pa, I estimate that applying the Growth Commission's model to the 2010/11 starting conditions would imply real terms spending cuts of 1.5% pa for a decade. Even allowing for "through the business cycle" caveats, it's clear that this would have been far more extreme austerity than Scotland has actually suffered.

Now whether my quick attempts to correct the SNP's assertions are right or not could be open to debate, but that's not really the point - what is clear is that the justifications they've offered (or that we've found out they're using internally) are clearly nonsense.

Nicola Sturgeon is frequently repeating the same false assertion - in the Chamber; on BBC1 and ITV; in interviews and in debates - and she should be held to account.

No amount of suggestion, psychology, misdirection, showmanship (or even bending of the concept of time) should be allowed to distract from this simple fact: Sturgeon's claim that applying the Growth Commission's recommendations on spending over the last decade would have spared Scotland from austerity is as shonky* as hell.

*shonky: dishonest, unreliable, especially in a devious way.



1. I'm a fan - that description is his own from the dust-jacket of "Tricks of the Mind" (which I highly recommend)

2. "If [the Growth Commission's] recommendations had been implemented retrospectively we'd have escaped Tory austerity over the last decade"

3. See exchange with Richard Leonard at 47:20: "In fact if we'd had the Growth Commission's recommendations implemented over the past decade, we wouldn't have had the austerity that we've suffered under the Tories, that's the reality"

4. Directly from the Sustainable Growth Commission report:

page 42
Page 45
Page 92
Page 101

5. The report used the phrase "within 5 to 10 years" which is just the same as saying "within 10 years" (unless anybody is going to suggest that they're implying they wouldn't want to have the deficit below 3% of GDP in less than 5 years?)

6. It's important to note that when modelling deficit reduction over a 10 year period, the Growth Commission simply use "trend GDP growth" (which given the inevitable economic disruption that separating from the UK would cause - leaving the UK single market, leaving the Sterling currency union, losing the fiscal transfer from rUK - is arguably a very optimistic assumption). More importantly: as we see in the two attempts to justify these assertions, at no point do the SNP present a case that suggests they'd have changed historical rates of GDP growth.

7. Source GERS (the same base data the Growth Commission used) for 2018-19

8. The sound-bites offered by the SNP tend to refer to the "Scottish Budget" - of course the Growth Commission's recommendations relate to Scotland's total expenditure ("Total Managed Expenditure" in GERS terminology) which includes rather important costs that are outside Scotland's Budget, not least Social Welfare costs including pensions. It's impossible to read across from the Growth Commission's spending restraint recommendation to specifically what that would mean for only those costs in the Scottish Budget, as the GC make no specific recommendations about Social Welfare spending (for example).

9. I considered sharing a screen cap of the DM but that feels like crossing a line - but if any journalists or fact-checkers want to see confirmation of this I would happily provide

10. This is in itself a hugely optimistic assumption based on some highly questionable analysis - but I don't want to get bogged down in those arguments which were covered in These Islands Growth Commission Response: A Reality Check. Similarly the target of a deficit of 3% of GDP within a decade is insufficiently aggressive on the Growth Commission's own terms, as covered in the These Islands Growth Commission Response: Aiming Too Low, but I'm trying to avoid this blog re-running too much ground that's already been covered

11. There is a wrinkle in their model which (see B12.19) in that they assume "savings of 0.3% GDP associated with investment in best-in-class institutions realised over a three year period from year 5" - that basically makes it easier for them to get to the 3% figure in 10 years. Bar that wrinkle and the impact of assumed debt interest on new debt, the model is incredibly simple: revenue grows with GDP, Spending grows more slowly (by the "spending lag") and hence the deficit must come down.

12. As we'll come on to see, the definition of "10 years ago" will become important. Here I have used 2007/08 as the starting position: this is the year the SNP came to power and - given these initial assertions were being made in May 2018 so we'd just completed year 2017/18 - that seems like 10 years ago to me. The second definitional issue we have is "onshore deficit" at that time (the onshore deficit is the figure the Growth Commission consistently use) - per most recent GERS (so dealing with any historical restatements) the 2007/08 onshore deficit was £12,774m. If we divide that by onshore GDP of £120,929m we get a figure of 10.6%; if we divide it by total GDP of £145,372 we get a figure of 8.8%. To err on the side of caution, I have taken the lower figure.

13. The previous conclusion we drew when modelling this was that a lag of 1.5% would be required - the analytical challenge relates to what one assumes in terms debt costs for the counter-factual historical case. The higher figure comes from simply taking the forward model (as per our recreated Growth Commission model), changing the "legacy deficit" assumption to be as per the historical deficit and then working out what the lag would need to be to get to the deficit to 3% after 10 years - that gives the higher end of the range (basically because we're having to service more incremental debt in the future model because we're no longer sharing our deficit with the rest of the UK)

14. This gets a little tricky as at the time Kate Forbes and Sturgeon first made their assertions, the latest GERS figures available were for 2016-17; since then 2017-18 and 2018-19 figures have been published and of course these will include historical restatements / corrections. Taking the most recent 2018-19 version of the figures, over either of the 10 year periods from 06/07 to 16/17 or 07/08 to 17/18 gives the same answer of 1.0% annual GDP growth [the 10 years to 18/19 would now give us a 1.2% average]

15. Again using the latest (2018-19) GERS report: the onshore deficit in 2010-11 was £19,607, onshore GDP was £124,611m and total GDP was £147,983 - again I'm erring on the side of generosity by using total GDP as my onshore deficit denominator

Tuesday, 3 December 2019

The SNP: Living in the Past

The SNP's Ian Blackford recently made the following statement (presumably to Adam Boulton) on Sky News:
"But the simple fact of the matter, Adam, is that if you look back over the course of the last 40 years, UK Government figures make it absolutely clear that Scotland has subsidised the rest of the UK in most of the last 40 year period.”
This set alarm bells ringing for those who pay attention to this sort thing, because it is a statement dramatically at odds with the Scottish Government's own data.

Following a minor Twitter storm, Channel 4's @FactCheck had a look at it and published a Fact Check which concluded that it was "impossible for us to stand up Mr Blackford’s claim at the moment". Some of us thought that was an extremely generous conclusion and this blog will, I hope, explain why.

An aside on Channel 4's FactCheck: I have spoken with one of their team who took the time to call me and explain why they had written what they have. I appreciate that they have a policy of not doing analysis, but rather following the audit-trail offered to them for any claims made. Whilst I will come on to show why the audit trail they were sent down was perhaps a dodgy path, I want to be clear that this blog is a criticism of Blackford's claim, not of Channel 4's FactCheck. Heaven knows, given the breadth of issues they have to cover, we can't expect them to be as familiar with the intricacies of Scotland's economic data as sad souls like me who have buried ourselves in this topic for more than 5 years now.

What's particularly useful about Channel 4's FactCheck is that they have clearly spoken with the SNP (as of course they should) and been told how they attempt to justify the claim. FactCheck quote an SNP spokesman (*waves*) as saying:
"We have demonstrated to you using official government statistics that Ian Blackford’s comments are correct. The evidence shows that tax revenues per head in Scotland have been considerably higher than the UK – that has unquestionably subsidised UK public spending over the period."
We'll come back to this statement, but first let's clarify what's going on here.

According to FactCheck, the claim is based on a publication in April 2013 which relied on "An Experimental series of historical fiscal balances for Scotland 1980-81 to 2011-12". We'll refer to this data source from here forward as the SNAP data (as it's colloquially known1).

So here's the first point, as rightly made by FactCheck: "The stats only cover 1980/81 to 2011/12 (that’s 32 years), and recent years are missing".

Pause for a moment.

So when Blackford said "the last 40 years" he was actually referring to "the 32 years up to 2011/12"?

That's a 32 year period that ended 7 years ago!

Fun fact: in 2011/12 that SNAP data showed income from North Sea oil for Scotland of £7.9bn. Anybody with even a passing familiarity with what's been happening to tax revenues from the North Sea will immediately see the issue that's coming here - but let's step our way there carefully.

Here's another area where I have sympathy with Channel 4's FactCheck - I've followed enough audit-trails offered by the SNP to be familiar with their shoddy approach to data sources.  In fact those SNAP figures were updated to cover the 33 year period to 2012/13 [here] so from here forwards the most recently available SNAP figures are the ones I'll use. I presume the information FactCheck link to was as given to them by the SNP - and it's perhaps not unreasonable for them to assume they'd been pointed at the most recent data by the people attempting to justify Blackford's claim.

OK, so let's see what that SNAP data shows us (before we come on to the SNP's attempted debt-dodge).

First we need to understand how we can identify when Scotland could be considered to be "subsidising" the rest of the UK.

Scotland is allocated a population share of the UK's debt costs both in the SNAP data and in the more up-to-date and accurate GERS data that we'll come on to consider. This means that Scotland is considered a net contributor to the rest of the UK when it's per capita deficit is lower than the UK's per capita defict (and vice-versa).

A corollary of this is that - assuming Scotland is accruing responsiblity for a population share of the UK's debt -  when Scotland is responsible for less than its population share of growth in that debt, it's net contributing to the rest of the UK2.

So let's graph that SNAP data and see what's going on - this is the difference between Scotland's SNAP deficit per capita and the UK's deficit per capita (from the same source):

[note: "1980" = 1980/81 (and so on) and all figures are deflated using the UK GDP deflator]

When that line is above the axis, Scotland was a net contributor to the rest of the UK, when it's below it was a net beneficiary.

So immediately we can see a profile which will be familiar to regular visitors to Chokkablog: Scotland was a massive net contributor to the UK when oil boomed in the 1980's but - according to the experimental data as presented by the Scottish Government in 2014 - from 1998 onwards we pretty much just paid our way.

So far so good. If we accept setting the clock running just as we strike oil and if we accept that Scotland gets to "keep" oil revenues from oil fields in Scottish waters (which is what these figures assume) then the SNAP data shows Scotland being a material net contributor to the UK over the 32 years between 1980 and 2012.

In fact if we accept these assumptions and add the annual numbers together, it works out at a cumulative net contribution from Scotland of nearly £16,000 per person. It's Grievo-Max time and perhaps one can see why Ian Blackford wishes that's where we still were.

But let's remind ourselves what Blackford actually said: "UK Government figures make it absolutely clear that Scotland has subsidised the rest of the UK in most of the last 40 year period."

Well it's only "absolutely clear" that Scotland "subsidised" the rest of the UK when that line is above the axis  - and you'll find that's the case in just 15 of the 33 years shown. So even using this extremely out-of-date data (that conveniently stops just when oil revenues crashed) the statement "in most of the last 40 year period" is not supported.

But more importantly, we're not still in 2012.

Time doesn't stand still, oil reserves are finite and we have both more recent figures and a better handle on the figures for some of the years covered by SNAP - we have GERS.

GERS [Government Expenditure and Revenue Scotland] figures are compiled by Scottish Government economists in St Andrew's House and - unlike the experimental SNAP data - they qualify for the National Statistics Kitemark. Thankfully we don't need to re-run the tired old arguments put forward by GERS-deniers to be able to agree that GERS is better data than SNAP - that point at least should be beyond dispute.

Now FactCheck baulked when they realised that the GERS data only goes back as far as 1998/99. I guess that's fair enough; they wouldn't necessarily know how or if GERS figures reconcile with SNAP figures  - and you'd need to be familiar with the reporting history to understand how to explain the apparent inconsistencies.

But fortunately we can deal with those issues because we've been down this road (so many times) before.

It's easy to plot the two sets of data and see how they overlap:

So immediately we can see that where the data overlaps the more recent and more accurate GERS figures show a (slightly) worse picture for Scotland than the SNAP data.

The simple explanation is that revisions made by the Scottish Government's own economists to the historical figures reduced the amount of North Sea oil revenue that they believe should be attributed to Scotland. The most dramatic adjustments were made in the 2016-17 GERS figures - I highlighted this at the time. To avoid getting bogged down in the detail of that here, see note 3 below.

Of course even more significant than the correction downwards in the overlapping years is the dramatic shift to Scotland becoming a large net beneficiary of the UK in the last 7 years (as North Sea oil revenues have declined).

So if we use the best available data (GERS after 1998, SNAP before) we can create a full time series against which we can fairly judge Ian Blackford's claim:

Remember: when that line is above the axis, Scotland is a net contributor to the UK, when it's below Scotland is a net beneficiary.

You can see why Ian Blackford wishes the world had stood still in 2013 (and that he could ignore Scottish Government data corrections), but I don't think that makes it acceptable that he makes assertions assuming that were the case.

The more eagle-eyed among you will have spotted this is in fact a 39 year time period, but that's by-the-by (it's closer to 40 than 32 is!). Now let's check Blackford's statement against this best available data:
"UK Government figures make it absolutely clear that Scotland has subsidised the rest of the UK in most of the last 40 year period"
In fact these figures (which come from the Scottish Government, albeit they include UK Government figures) show that - to adopt Blackford's own terminology - Scotland has subsidised the rest of the UK in only 12 of the last 39 years.

So not "in most of" them then - that's simply wrong.

But more importantly, unless we believe a time-machine is going to be an option in any future referendum4, we really should look at the more recent run-rate. That shows us that in every one of the last 10 years Scotland has been a net beneficiary of UK-wide pooling and sharing, in recent years to the order of £2,000 a year for every man woman and child in Scotland.

That figure shouldn't suprise anybody paying attention: £2,000 x a population of 5.4m = £10.8bn - that is the effective fiscal transfer that people like me have been banging on about ever since 2014/155, when we could see that this would be the result of oil revenues falling away.

What about that cumulative figure that showed Scotland having net contributed £16,000 per person between 1980 and 2012 (per the SNAP data)?  Well that figure now (using the best available an dmost up-to-date data) works out to now be just £2,000 per person as of 2018.

Given we're well into 2019/20 and we can be confident that the fiscal transfer (I don't like "subsidy") this year is going to be similar to the last few years, as we sit here today it's fair to say that, in real terms, Scotland has got back from pooling and sharing as much as it has paid in since 1980.

Isn't that great? Pooling and sharing since 1980 has worked out to be pretty much neutral for Scotland and - as long as Scotland remains in the UK - the fact that going forwards Scotland looks set to continue to be a net beneficiary is nothing to be embarrassed about: what goes around comes around.

It's not a "subsidy" - its UK-wide pooling and sharing in action.

But wait: remember I said all this was before the SNP's debt-dodge?

Well this is based on an extremely dubious interpretation of a thought-experiment that Prof Brian Ashcroft proposed way back in 2013. At that time he hypothesised that (again: if the clock had started in 1980, if Scotland could have kept it's geographic share of oil revenue) then Scotland could have been debt free - so what if Scotland didn't pay for a population share of the UK's debt?

Apart from the obvious point I've already made (we don't have a time-machine option) the flip-side of that assumption is that now Scotland is contributing far more than it's population share to the UK's debt, you'd need to be allocating more than a population share of the interest costs of that incremental debt. I'm sorry but this really does start to get a bit silly.

But you know what?  Let's see what the figures would  have looked like if Scotland paid nothing towards the UK's debt costs over that time (i.e. this is a debt-free Scotland vs a debt-burdened UK):

So even with this frankly outrageous assumption [i.e. that we're running a massive deficit and paying no debt costs because we've somehow jumped in a time machine and created an oil fund instead of pooling and sharing with the rest of the UK] we'd still be in a situation where Scotland is now a significant beneficiary from UK-wide pooling and sharing, even if we pay nothing towards the UK's debt costs.

Now of course you could start arguing that in the time-machine scenario Scotland would have invested in an oil fund which by now would ..... ahhgggghhhhhh make it stop!

We can't change the past: decisions we make now should be based on where we are now and reflect what we can see is happening to revenues going forwards not looking backwards.

But let's finish where we started: with Ian Blackford's claim and the statement made by the SNP spokesman:

"We have demonstrated to you using official government statistics that Ian Blackford’s comments are correct."
Well given this was by reference to 32 years of SNAP data that stopped in 2011-12 and did include a population share of UK debt costs ... no you haven't.

"The evidence shows that tax revenues per head in Scotland have been considerably higher than the UK ..."
For what it's worth that is true, but only "considerably higher" because of oil revenues (which is why it hasn't been true for the last 5 years) ...

[note: green = onshore revenue only, black = inc geographic share of oil revenues; for SNAP/GERS overlap see note 6.]

... and of course that observation fails to consider the impact of higher per capita spending levels in Scotland, which have "been considerably higher than UK" for every one of the last 39 years (and that gap in spending is actually growing significantly - thank you Joel Barnett)

[note: for SNAP/GERS overlap see note 6.]

"... that has unquestionably subsidised UK public spending over the period."
Well I'm sorry, but if anybody has read all of the above and believes that statement is an accurate representation of reality ... well I give up.



1. Among people who are sad enough to need colloquial short-hands for different Scottish Government data sources - in this case it refers to the Scottish National Accounts Project

2. This is not a controversial assumption and is consistent with base approach taken by Prof Brian Ashcroft whose analysis from 2013 is cited by FactCheck (I presume having been pointed in that direction by the SNP's spokesman).  It's perhaps only fair to point out that Prof Ashcroft has reviewed and approved my analysis in this area before (see here).

3. Extracts from GERS 2016-17: there are of course other adjustments every year, but this North Sea revenue adjsutment was the one that had the most dramatic impact on the historical view of Scotland's deficit

4. It isnt - at least not one that allows people to travel back in time, otherwise they'd be here now telling us about it

5. Readers of this blog will have seen this graph in April 2015 - it shows how the "onshore deficit gap" (the difference between the blue line of relative onshore per capita revenue generation and the red line of relative per capita spending ) was fairly constant at around £9bn pa ... and that this gap was sometimes "washed away"by North Sea oil revenues  (the black line adds oil revenue per capita difference to onshore revenue)

We can bring this right up-to-date now (with GERS data that now corrects the historical over-statement of oil revenues) and we can see that even including current oil revenue the deficit gap has grown to over £11bn (because of the widening gap between per capita spending in Scotland and rUK).

I know people get confused by the difference between the deficit, the deficit gap and the effective fiscal transfer, so here's a simple summary:

6. For completeness, these graphs show the overlapping SNAP and GERS data, so you can see how the SNAP data compares when more accurate GERS figures are available.

On the revenue figures it's clear that the big correction is the (Scottish Government economists') downwards revision in assumptions about oil revenues attributable to Scotland - it would be interesting to see the same corrections applied back to 1980, but I suspect there isn't much political will to get that done

On the cost figures (note the left-hand scale is much smaller than on the revenue graph) the effect of the SNAP/GERS revisions are fairly marginal