“[I]f there is to be meaningful debate on this issue then the SNP have a lot of work to do to produce best possible data. The last thing they should do is trust that from London.” – Richard Murphy

Tax expert Richard Murphy, who is currently most notable for exposing the UK’s massive £120 billion per year tax gap, has written an article warning of relying on UK economic data to make the case against Scottish independence.

Before he gets attacked too badly by hacks telling him that the Scottish economic data is produced by Scottish civil servants (Edit: I may already be too late on that) I thought I’d write a parallel piece pointing out what those civil servants have told me about the limits of some of their stats.

The first thing to remember in all of this is that the UK is not a federation or a confederation, it considers itself to be a unitary state of which Scotland is just one region of twelve (plus the “extra-regio” offshore regions). Therefore there is currently no real obligation to even gather the distinct statistics for Scotland and it really only has become important because of the independence campaign.

Tax Revenue

As I’ve pointed out in my paper Beyond GERS, the issue of apportioning tax revenue is fraught with subtle difficulty. GERS itself has updated its methodologies multiple times over the years (particularly since the SNP took the government in 2007. The GERS of today is no longer very closely related to the GERS created by Ian Lang to discredit Scotland in the early ’90’s). There are still differences in the results presented straight by HMRC and the data eventually “Scottishised” [To use the stats folk’s term] and presented in GERS.

Onshore corporation tax is a good example of this. Where an overall UK stat may simply count the location of the HQ of a company for the purposes of assigning corporation tax and this may make sense from a unitary state perspective (albeit this is becoming less true as globalisation increases the ability for multi-national companies to move resources across borders).

For many companies though, the profits one which corporation tax are paid are not generated at the HQ. This is obvious in the case of, for example, a large retail chain which has stores across the country. To correct for this, HMRC and GERS both use different methodologies to apportion the tax more evenly. Various measures (and the weighting applied to those measures) such as estimating volume of sales, number of employees, amount of capital spent in the region and overall population are all used in different ways to reach slightly different estimates. As a result, HMRC estimates that in 2015-16 Scotland produced 7.1% of the UK’s corporation tax compared to 7.3%% estimated by GERS – a gap of  about £100 million.

One can also see possible limits of these methodologies especially if taken individually. For example if one looks at employees then one could probably consider a company (and, it should be stressed that this is a completely hypothetical company) which employs a dozen people in Scotland to make, say, a high value, highly exportable product with a geographic link (call it a similarly hypothetical product like “Scotch blisky”) and then employs a couple of hundred people in London to market it. It may be very difficult to properly apportion the “value” of that product and its profits based on employees alone. It’s possible, after all, to find a market without marketing but a bit harder to drink an advertising campaign.

VAT is another issue where these figures can differ for similar reasons. The UK doesn’t demand point of sale ID to determine the location of VAT spend (If you nip down the road to Carlisle for your shopping, then that results in VAT paid in England but Tesco neither knows nor cares where you came from to get there). Again, various methodologies are used to try to estimate the proportions paid and the estimates are slowly aligning (HMRC claims Scotland paid 8.4% of the UK’s VAT compared to GERS’ 8.6% – a gap of £110 million). There is also a further complication wherein the results between HMRC and GERS are simply presented in a different manner (HMRC measures the cash receipts, GERS measures the accruals)

A third prominent example is Income Tax, and is going to become pertinent now as IT is largely devolved to Scotland and all Scottish residents are to be assigned a distinct Scottish tax code and especially now that the income tax bands in Scotland will soon start to diverge from the UK bands. However, HMRC has been recently criticised for a series of administration issues which is making it difficult to roll out this tax code. As with the difficulties in rolling out devolved welfare, this won’t be nearly so much of an issue once Scotland is independent but highlights the difficulty in trying to run a devolved situation from a centralised unitary setup. This said, both HMRC and GERS arrive at a proportion of about 7.2% of the UK’s income tax coming from Scotland although this may change as the new systems are launched (even if tax rates are kept the same).

It is not possible to say whether the HMRC or GERS estimate is “better” or “worse” than the other. The Institute of Fiscal Studies has commented saying, especially of corporation tax:

“Neither of these estimates is clearly superior to the other, and both may be some way off. Profits are not necessarily generated in proportion to the number of employees, or their wages. Some employees may be more instrumental in generating profits than others; and profits also arise from capital assets – both physical (such as buildings and equipment) and intangible (such as intellectual property and brand value) – the location and contribution of which may differ from the location and wages of employees. Calculating how much of a company’s profits are attributable to economic activity in different locations is conceptually and practically difficult and is the source of many problems in international corporate taxation”

Balance of Trade

This is the big one that has attracted a lot of shouting in the past few months. Once again, the UK’s status as a unitary state causes much of the furore over the published numbers to be based on false premises and over-massaged numbers. The UK’s balance of trade figures are published here and probably do do a decent job of estimating the UK’s position in the world. What it doesn’t do is show the internal movements of trade within the UK. As a unitary state it simply doesn’t matter to the external balance of trade whether or not Yorkshire is a net exporter to Sussex. The UK does produce figures which try to estimate the trade balance between the regions  with the rest of the world but it only covers goods, not services (hence excludes nearly half of the UK’s total trade) and it does not cover internal trade. For that internal trade, we turn to ESS – Export Statistics Scotland – which surveys exporting companies in Scotland and asks them where they send their goods and services (contrary to a semi-popular belief, these statistics don’t care how the goods reach their destination so it doesn’t matter if they physically leave the UK via an “English port“). There are some limits, again, to this methodology.

First, not all companies know where their goods are going (see the example of Tesco again. If someone from Carlisle buys a crate of beer in Glasgow then goes home then that’s a Scottish export but Tesco wouldn’t be able to record it easily) so they won’t appear in the survey. Goods which are shipped to England then either re-packaged or used as a sub-component before being exported from England to somewhere else (or even back to Scotland) would be counted only as far as their export to England and there may be some cases where service “exports” are caused by, for example, someone in London buying insurance for their house in London from the London branch of a provider who just happens to have a brass plate in Edinburgh. The total proportion of these anomalies in the data is simply unknown at this point and unlikely to be knowable until after independence.

Beyond the Horizon

And this takes us to the most important point in this whole article.  Even if the methodologies above all align and all can capture the full economic picture of Scotland and everyone can agree on the figures produced and everyone agrees that they produce an accurate and complete picture of Scotland’s economy within the Union there is a fact which should be utterly indisputable (and certainly is within the team which put together these stats).

Independence. Changes. Everything.

None of these figures have any validity if you try to use them to project beyond the independence horizon. Corporation tax may change due to the redomiciling of businesses post-independence. Both those seeking to remain within the UK and those seeking to remain within the EU or EEA may shift operations. Trade exports may suddenly become a lot easier to assign (whether there’s a “hard border” or not) and that “extra-regio” oil which is often excluded from stats due to historical and supply chain accounting issues suddenly has to be accounted for. Those tax streams which are simply too embedded to discuss in any terms other than by a population share have to be audited. And all of this is before Scotland starts to make changes to the tax system to optimise it for the Scottish economy or to do things like close the tax gap.

As with everything in science and in economics, statistics are based on models, models are only ever as strong as their underlying assumptions and projections are only ever as strong as the person making the prediction’s understanding of the limits of those assumptions and the models.

I don’t mind discussing the economy of Scotland within the Union. I don’t even mind speculating on the economy of an independent Scotland. But I sense that the next two years of campaigning will get very frustrating if pundits continue to stretch their own models past the point of credibility in a quest to push their political point. This, I should warn, goes for both sides. We need a more meaningful economic debate than we saw last time. Let’s get beyond the headlines to create one.