Thread by @staylorish

Craig Dalzell (@thecommongreen) was in Edinburgh last week discussing the finances of an independent Scotland. His talk included a number of errors, but I’ll focus on just one. (1/18)

Craig claims that Scotland could see a revenue benefit of “up to £1.3 billion” from the repatriation of jobs following independence.

He cites spending allocated to Scotland in GERS which takes place in rUK, and extols the virtues of bringing those jobs to Scotland. (2/18)

Asked to explain the calculation behind this claim, Craig cited his 2016 @Common_Weal paper. (3/18) allofusfirst.org/sites/default/…

The key driver of his calculation is “£2.35 billion per year of non-identifiable spending on Scotland's accounts which cover areas of UK wide governance such as managing reserved taxation and policy, revenue and customs...” (4/18)

The £2.35 billion figure is Scotland’s population share of “non-identifiable expenditure” – that is spending considered to occur on behalf of the UK as a whole and which cannot be decomposed on an individual country basis. (5/18)

A good example of this kind of spending (which Craig in fact explicitly mentions) is HMRC – its staff are administering a UK wide system, and where they happen to be based is irrelevant to the work they carry out. (6/18)

But the UK Government civil service has employees all over the UK. So, as well as Scotland being allocated a share of spending that happens for its benefit in rUK, the reverse is also true: rUK is allocated a share of spending that happens for its benefit in Scotland. (7/18)

When I asked Craig how he had accounted for this very important fact, he said: “It's embedded in the figures as best as could be done within the limits of the data.”

The problem is indeed “embedded in the data”, but Craig has completely overlooked it. (8/18)

The calculation in Craig’s Common Weal paper makes the implicit assumption that none of the “non-identifiable expenditure” allocated to Scotland (on a population share basis) already happens in Scotland. This is an extremely significant error. (9/18)

So, what proportion of UK Government civil servants working on “non-identifiable” matters are actually based in Scotland?

Doing the maths for every department will get you the exact answer, but fortunately just two account for the lion’s share of the issue: DWP and HMRC. (10/18)

This chart illustrates why DWP and HMRC are critical. The Ministry of Justice is completely devolved, and defence is excluded from Craig’s analysis, so the MoD can also be ignored. (11/18)

The Scottish Government publishes statistics on public sector employment, which tell us exactly how many DWP and HMRC employees are based in Scotland: 9,300 and 8,500 respectively. (12/18)

Total UK-wide departmental headcounts at DWP and HMRC are 75,000 and 60,000 respectively, so Scotland has 12.4% of DWP employees and 14.2% of HMRC employees. (13/18)

Scotland’s GERS population share is 8.2%, so it has way more than its population share of employees at the two largest headcount departments for fully reserved spending. (14/18)

Some of the smaller departments are likely to skew in the opposite direction, but overall Scotland probably has at least 10% of UK Government civil servants working on “non-identifiable” matters, comfortably above its population share of 8.2%. (15/18)

Unless UK Government civil servants based in Scotland are paid (on average) 20% less than those based in rUK, there is already more “non-identifiable expenditure” happening in Scotland than the amount allocated to Scotland in GERS. (16/18)

The ONS publishes data on civil service salaries, which reveals that civil servants in Scotland earn 93% of the UK-wide median civil servant salary. (17/18) ons.gov.uk/employmentandl…

All of which means that Craig’s £1.3bn number is entirely fictitious. Worse than that, there should really be a small adjustment in the opposite direction: to account for a net repatriation of civil service jobs out of Scotland and into rUK in the event of independence. (18/18)

Addendum:

Here is Craig’s talk (video should start at the point he begins to speak about this issue).

And an error in the first tweet of this thread needs correcting: the talk was not last week, it was the week before last. (I’m a stickler for detail).

Addendum 2: a small correction

Tweet (13/18) includes a minor error. The departmental totals are actually the number of “full time equivalent” employees, but the figures for Scotland are total headcount (ie part- and full-time).

The data for “full time equivalent” (FTE) employees in Scotland is shown in this table. FTEs in Scotland at DWP and HMRC are: 8,200 and 7,600 respectively.

This means that the percentage of FTEs in Scotland are: 10.9% at DWP and 12.7% at HMRC.

The Common Weal analysis assumed both percentages were zero, and the important point is that both are comfortably above Scotland’s population share of 8.2%.

So, the correction is not material, and does not change the conclusion of the thread. But it’s important to correct errors, and that was an error (albeit a small one).