Richard Murphy on GERS: The Perpetually Baffled Man

I'm writing a serious blog on GERS and really don't want it to get bogged down by engaging with the frankly bonkers claims made by the perpetually baffled GERS-denier in chief, Richard Murphy. 

To engage with Murphy on the topic of GERS is as pointless and frustrating as trying to play chess with a pigeon (while wrestling a pig). But I recognise that some of you might think that describes an amusing spectator sport, so let me use this blog to deal with some of his recent - um - "contributions to the debate".

My starting point is his appearance on the "GERS - Scotonomics Special". This is an hour of video and I only made it 5 minutes in before I realised that I simply couldn't sit through any more of it because life, like me, is too short. 

The fact that this entire blog is required to unpick what is basically just one minute of his extraordinarly comprehensive wrongness I hope explains why I will try to avoid engaging in any further pigeon-chess-playing-while-pig-wresting. 

Suffice to say that this blog alone should be enough to convince any rational observer how spectacularly misplaced Murphy's arrogance is when it comes to criticising the work of the Scottish Government's statisticians. If more convincing is needed, see here and here or - if you want to see him taken apart in front of a Holyrood committee by the inestimable Jackie Baillie - here.

The point where I gave up the will to go on appears at about 3:30 in that video - I quote him here verbatim:

"... when you look at the expenditure side, what you actually see [is] the vast proportion of the expenses of the UK are apportioned to Scotland on the basis of population - 8.1% - and yet I put out figures on the blog this morning and show that the vast majority of the expenses actually represent more than 8.1% of the total UK spend and in some cases vastly higher .. in terms of Public Administration costs something like 13 to 14% of the total costs of the UK are apparently incurred in Scotland that's why I think frankly what's happened is they've taken the UK figure and then they've taken the figures for running Scotland and added the two together not allowing for the fact that some things obviously aren't done twice in Scotland ..,. they've done that sort of crass stuff and I'm afraid to say I just think that this is really poor accounting [and on and on he goes ...]"

Honestly it's hard to know where to start with this level of wrongness, but let me try and be calm and forensic.

The blog of his he refers to is the typically boldly titled "Why GERS is wrong - yet again". That blog contains even more nonsense which it's hard not to get distracted by, but I'll relegate the most obvious howlers to a footnote¹ so I can wrestle here specifically with his bizarre assertions about Public Administration (by which his blog makes clear he means the GERS line item "Public & Common Services") cost allocations.

The figure that has him so exercised is £2,410m of Public & Common Services expenditure allocated to Scotland [see Table 3.1 on page 28 of the GERS report] which is indeed 13.7%c of total UK costs.

First of all, the fact that he admits to being baffled by how much higher than the UK average Scotland's Public Administration costs are shows how poorly he understands Scotland.  If he'd read page 7 of the GERS report (remembering that Scotland accounts for just 8.1% of the UK population) he would have seen a broad hint:

"... around 9.1% of UK spending is undertaken for Scotland, slightly higher than a population share. While direct estimates of spend in Scotland are not available, this is consistent with broader indicators of public sector activity in Scotland, which show that the public sector plays a larger role in Scotland than the UK as a whole. For example, around 10% of UK public sector employees are based in Scotland, with regional pay differences resulting in around 9.3% of the UK paybill spent in Scotland." - GERS page 7.

If he'd made it as far as page 35 he's have seen Table 3.8 in GERS shows that £1,683m of the £2,410m of Public & Common Services costs in GERS are directly incurred by the Scottish Government and Local Authorities. So what he's getting exercised about is in fact an allocation of £726m for Public & Common services delivered by the UK Government.

Cross-referenceing Tables 3.8 and 3.7 would have shown him that this £726mn is an allocation of just 4.6% of those UK Government Costs². So his wild assertion that 8.1% of all UK Public & Common Services costs have simply been added on to Scotland's costs is, obviously, wrong. Suggesting that such a basic error could have been made merely advertises the fact that he has no conception of the depth and detail of the work that goes into compiling the GERS figures.

But is that 4.6% reasonable? 

Well intuitively it doesn't sound weird given how much of the UK's integrated machinery of state Scotland still relies upon and we're dealing here with only £0.7bn out of a £36.3bn deficit.

But we don't have to rely on intuition, we can dig deeper. The "supplementary-expenditure-database" which accompanies GERS provides detailed line-level detail to back-up all the GERS spending allocations for the prior year (when £624m of UK Government Public & Common Services costs were allocated to Scotland).

This database shows us that roughy half of those UK government costs³ allocated to Scotland are a simple population share of HMRC costs. Now HMRC is a big department and it's a shared UK resource. This cost allocation is not duplicating anything  it's not something "done twice in Scotland" so - surprise surprise - none of Murphy's alleged double-counting is going on here. 

As an aside: frankly it's highly unlikely that Scotland could replicate the full service offered by HMRC for as little as 8.1% of what it costs the UK today, so this could be argued as a very good value deal - but that's a different question for another day. 

These HMRC costs helpfully illustrate the wrong-headedness of another of Murphy's frequently voiced claims: that GERS is wrong because we're incurring spending outside Scotland and not seeing the tax generating benefits of that spending in Scotland. A quick google would lead him to the UK's civil service statistics which tell us that 7,730 HMRC employees are based (and generating tax) in Scotland.  That's 12.8% of them so, even allowing for salary differentials, more HMRC employee costs are incurred in Scotland than allocated to Scotland in GERS. [A similar situation exists with DWP staff: 8,820 or 9.7% of DWP employees are paying tax in Scotland].

Another aside: that Civil Service headcount data allows us to cross-check the figure quoted on page 7 of GERS - and sure enough 47,590 or 9.97% of all UK-based Civil Servants are based in Scotland. It's almost as if the team who sit in St Andrew's House working on the GERS report have a better understanding of what's going on than Murphy does, isn't it?

So we're now down to explaining just £0.3bn of Public & Common Services costs incurred by the UK Government and charged to Scotland in GERS. The detailed database shows us that £0.2bn of them relate to a population share of HM Treasury, the Cabinet office and the Department for International Development (DFID). Whether Scotland could replicate these services for just 8.1% of the UK's costs charged in GERS is again open to debate (a debate to which I predict the conclusion would be: no). That said, although 18% of UK-based Foreign, Commonwealth & Development Office⁴ employees are based in Scotland, it is of course true that the vast majority of HM Treasury and Cabinet Office employees are in London. 

The £0.1bn that's left (£142m if you really care) is Scotland's share of House of Commons, IPSA, House of Lords and a few other bits and bobs.

So there are some costs which occur in England that are allocated to Scotland - but equally there are some costs that occur in Scotland that are allocated to England (just think of those HMRC and DWP employees or see below to learn about how Scottish ferry costs are treated). 

Murphy blusters that he is sure the amount spent in England which is allocated to Scotland (and could be transferred to Scotland to generate more tax) is "some billions" - needless to say that is a nonsensical number based on nothing more than a cumulation of ignorant assertions.

I've done the analysis in detail here, but the simple answer is it wouldn't be more than a couple of hundred million at most (and you can't both save costs and transfer them!). 

If you can work up a grievance out of the possible fiscal multiplier effect from moving maybe a couple of hundred million of spending within the context of what is a £12 billion fiscal transfer⁵ from the rest of the UK to Scotland ... well good luck to you.

***

Not only could I do this exercise for every expenditure category in GERS, I have done. I know right? What kind of a lunatic would spend time familiarising themselves with the detail behind the figures before opining on how they were compiled?

Having been through that detail with a fine-tooth comb, I do have some disagreements with the GERS methodology. For example I think the EU contribution calculation attributes too high a share of the UK's rebate to Scotland and I think the treatment of Renewable Obligation Certificates understates the effective subsidy from the rest of the UK to the Scottish Renewables sector. But that doesn't mean I rubbish GERS: I make polite enquiries of the team responsible in the Scottish Government to test and build my understanding. Even where I continue to disagree, I recognise the reasonableness of the approach that has been taken.

Not only is it patently obvious that Murphy has never taken the time to check any of his wild theories with the team who actually compile the figures, he clearly hasn't even read the report.

To illustrate my point, here's just one other extract from that YouTube clip which somebody has drawn my attention to. Again I quote him verbatim [from about 5:18]

"HS2 - which is not going to come near Scotland for decades and decades if ever - is being apportioned to Scotland and there's no benefit from it. We know that the mass of public transport investment in the UK is in the South East of England and that's being apportioned to Scotland. This is nonsense..."

He's right, but only insofar as everything he says there is nonsense.

It really couldn't be clearer, it's there in black and white on GERS page 11: "in this edition of GERS none of the expenditure associated with High Speed 2 is allocated to Scotland".

HS2 used to be the exception that proved the rule, because none of any other "transport investment ... in the South East of England" has ever been allocated to Scotland in GERS⁶. 

It's rather useful that HS2 is now completely excluded from GERS (it's been removed from prior years as well) because

1. This saves me having to explain how relatively trivial that HS2 figure was - just 2.3% of HS2 costs used to be allocated in GERS (based on the business case assessment of the knock-on value to the Scottish economy) and last year that would have been just £75m
   
   
2. There is now only one example where GERS allocates transport infrastructure spending in one part of the country to another - and that's where £69m (30%) of the Scottish Government's Ferries costs are allocated to the rest of the UK⁷. I wonder why Murphy hasn't kicked up a fuss about that?

You do have to wonder why, every time he's "baffled" or "simply does not understand" GERS, Murphy leaps to the conclusion that the figures must be wrong and the Scottish Government's own economists must be lazy and/or incompetent. Surely it couldn't be that he's a grifter who's spotted a chance to build a profile by saying what cybernats and reality-denying nationalists want to hear?

As a parting thought: imagine being part of the team that has spent years refining the GERS methodology and compiling the reports, only to have this man - on the basis of transparently superficial "analysis" and without bothering to gain even the most rudimentary understanding of the data he's dealing with - to accuse you of doing "crass stuff" and being guilty of "really poor accounting".

He really is beyond the pale.

Chokka-bloke, checking out.

***

Footnotes (the first one's a doozy)

1. Some other howlers in Muprphy's blog

• "Scottish data is inflated compared to the UK by including local pension fund costs, which are not in the UK data, which you might say is odd" - I would say it is odd, because his assertion that those costs are not in the UK data is simply and obviously untrue! The figures are shown in the supplementary-expenditure-database which accompanies GERS (something Murphy has clearly never troubled himself to look at), where we can see that Scotland is allocated £2.0bn of the UK's £17.2bn of public sector pensions debt interest.
• He's goes on to be baffled by debt interest allocations - "why has Scotland got a disproportionately large share of interest payments, picking up almost 15% too much?". It's not 15% "too much", as he would know if he bothered reading the GERS report. Below are direct quotes from the main GERS report which explain clearly: why Scotland picks up a higher share (9.3%) of debt interest payments than our population share (8.1%); that this is entirely explained by a correction to the historical treatment of public sector pension fund interest expenditure; that this is consistent with the way interest income is apportioned (and so the net effect on the stated deficit of these public sector pension fund interest expense and income allocations is roughly zero); that these are local government pension funds where data are available for Scotland.  
• "GERS includes two categories of interest spending. The first is reserved UK debt interest, and Scotland is allocated a population share of this, amounting to £2.3 billion in 2020-21. The second is interest spending associated with public sector pension funds. These funds also generate interest income, and in 2020-21 Scotland is apportioned £1.5 billion of interest expenditure associated with public sector pensions, and £1.4 billion of interest income" GERS page 7
• "In previous editions of GERS, Scotland has been allocated a population share of all UK public sector interest expenditure. In this year’s publication, this has been changed to separate out interest expenditure which is associated with public sector pension funds. These are local government funds which generate both interest revenue and expenditure, and where data are available for Scotland. This change brings the treatment of expenditure associated with these funds into line with the treatment of their revenue, which was already allocated to Scotland using Scottish data" GERS page 11
• Consistent with the CRA, interest expenditure by public sector pension funds is shown as spending by HM Treasury." GERS page 30
• "The main methodology change in GERS 2020-21 relates to the treatment of public sector interest expenditure, where interest expenditure associated with Scottish Local Government pension funds is separated out. All this expenditure is treated as Scottish, rather than apportioning Scotland a population share." GERS page 25

[Given Scotland's larger public sector, the fact that public sector pensions costs are higher than the UK average is singularly unsurprising]

2. For those who care about following the audit-trail

• Numerator: £726 [Table 3.8]
• Denominator: Total UK P&CS costs of £17,524 [Table 3.7] less Scottish P&CS costs of £1,683 [Table 3.8]= £15,841

3. £294m out of £624m to be precise - so 47%

4. In September 2020 the Foreign & Commonwealth Office (FCO) and Department for International Development (DFID) merged to form the Foreign, Commonwealth and Development Office (FCDO)

5. This will be covered by my other blog, but it's an easy enough number to calculate:

• Scotland's 8.1% population share of the UK's deficit was £24.26billion - this is the implied increase in Scotland's share of the UK's debt liability
  
  
• Scotland's actual contribution to the UK's deficit (i.e. the GERS deficit) was £36.34 billion - this is the implied contribution Scotland's economy actually made to that debt liability
   
• The difference - the benefit to Scotland of only having to assume a population share of the UK's deficit (in the form of a population share of the UK's debt) was £12.1 billion

So in 2020-21 the GERS figures show an implied fiscal transfer between the rest of the UK and Scotland of £12.1 billion - that's £2,210 for every man, woman and child in Scotland.

6. at least not in the 8 years that I've been studying these figures

7. per the "supplementary-expenditure-database" which accompanies GERS: 30% of Scottish Government costs associated with Caledonian Maritime Assets Limited and Ferry Service Department are apportioned to rUK (reflecting an assessment of the share of the value of Scottish ferry services enjoyed by travellers from the rest of the UK)