The LSE report on the increased costs in trade for an independent Scotland is based on unsubstantiated data and absurd assumptions

Posted on

I have tried hard to take seriously the new report from the LSE on the costs of Scottish independence, as measured through the supposed impediments to trade that it will create, but I can’t. It’s available here.

There are numerous issues I can raise. First, and most glaringly obviously, the authors note that:

Unlike independent countries, Scotland does not collect detailed statistics on its external trade. Export Statistics Scotland provides useful data about onshore Scottish exports, but import data is relatively sparse.

In other words, there is quite literally no useful data on which to draw conclusions.

As the authors then add:

Measuring Scotland’s trade is further complicated by the convention that economic statistics are produced separately for the onshore Scottish economy and its offshore counterpart (i.e. oil and gas production).

But, despite the very obvious fact that the only realistic thing to do at this point was, to be polite, to give up, they ploughed on:

However, by merging data sources and combining statistics for the onshore and offshore economies, we can obtain an overview of Scottish trade.

With respect, that conclusion literally cannot be reached from the claims that go before it. The only reasonable conclusion for the authors to reach was that there is no reliable basis on which to work. But they produced data anyway. Let’s be candid, on this basis GERS looks to be reliable, and that that is very far from the case.

What is more, the most cursory inspection of the findings gives clues as to more problems. For example, the authors say:

In European terms, the UK is a large economy, whereas Scotland is medium-sized. In 2017 Scotland’s GDP was £169 billion, similar to Czechia, Greece, Romania and Portugal.

But that assumes GERS is right, and there are good reasons to doubt that.

More significantly,  GDP data is critically dependent upon accurate import and export data, and as just noted, and as the authors acknowledge, that is not available for a Scotland. So in that case these claims have to be unreliable

Exports account for 58% of Scottish GDP and imports for 60% of Scottish GDP. By contrast, in RUK exports account for 35% of GDP and imports for 36% of GDP.

From this they conclude:

We estimate there is around six times more trade between Scotland and the rest of the UK than predicted by a standard gravity trade model. Alternative methods imply there is from 2.6 to 7.8 times more Scotland-rest of UK trade than predicted.

They do at least have the decency to conclude:

This excess trade is partly the consequence of Scotland’s union with the rest of the UK.

This, however, is an insufficient explanation. To have a result like this that so is abnormal (an absolute outlier in statistical terms) suggests one of three things. The first is that the data set is wrong. The second is that the data does not reflect the reality of what is going on. The third is that the behaviour being recorded is actually abnormal. It so happens all three are likely.

First, I have already noted that the data is highly unlikely to be right.

Second, it is exceptionally likely that in the absence of any border  on which to actually measure the flow of products that  what the available data really records is the relocation of the ownership of product. So, for example, it is entirely possible that Scottish product is transferred from Scottish ownership to English ownership prior to export, and the same may well happen in the opposite direction, none of which is surprising. In integrated economies, with companies that have no reason to respect the border between England and Scotland, and with no VAT or tax charge for the transfer of title, the likelihood that English- based import and export entities are being used by large groups of companies (who will dominate trade) to, first, take ownership of Scottish product (or imports) before they then actually move into or out of the UK as a whole, with those transfers across the next border being recorded as English as a consequence, is very high indeed. In other words, the likelihood is that actual imports and exports to England are much lower than reported and that those to elsewhere are much higher, with the data being distorted by what might best be called ‘book transactions’.

Third, the trade is abnormal because there is no border: if there was the flows would be different, as I note below.

The fact that geography also dictates that at present it is convenient to import and export through England just adds to the chance that this possibility is high.

But there are more issues (actually many more, but there are limits to the number that need be noted).

Take, for example, the claim that the cost of trade between Scotland and England will increase by 31%. Ludicrously, this is based on the estimated cost increase in trade between the UK and Ireland in 1922. As assumptions go, I cannot think of anything more ridiculous. The UK and Ireland were pretty much still at war. The UK controlled all the routes in and out of Ireland and could charge what it liked, and technology was utterly different. To pretend that this data is in any way relevant is absurd. Sheer common sense suggests that the comparison makes no sense.

There is a final absurdity to note. The authors seem to assume that there will be no behavioural changes despite these costs. But just look at what is happening in Ireland right now. Literally, almost overnight new freight routes that avoid costs have opened. And so would they from Scotland to avoid the claimed costs of going through England. This would be most especially the case if Scotland rejoined the EU, when import substitution from the UK might be very significant.

Put it all together and the claims in this report are literally not worth the paper they are written on. I suggested the estimates be consigned to that receptacle in the corner of your room, never to be discussed again.