SNP Growth Commission's GDP Growth Rate Claims

The SNP's long-awaited Sustainable Growth Commission Report is finally available.

I've already blogged on the pre-release headline spin about a "£4100 boost for every Scot" here, but now I have the full report to hand I can offer a more complete analysis.

The Growth Commission report can be looked at as doing three things;

1. It sees what lessons can be learned from looking at their chosen benchmark of "small advance economies"
2. It looks at the growth of those economies to scale the potential upside for an independent Scotland
3. It details their preferred option for fixing the currency problem that would be created by Scotland leaving the UK

The first of these we'll come back to in future blogs, but suffice to say if you read pp 154 - 160 of the report there's an awful lot there that than can be done now, with the powers the Scottish Parliament already has.

To be fair the biggest single recommendation is probably the one about attracting more inwards migration to Scotland, which would require further powers to be devolved to the Scottish Government.  I predict a long discussion to come about the desirability and practicality of that and, for what it's worth, I personally remain open-minded and look forward to that debate.

In this blog I want to focus on the second of the above, namely how they've scaled the economic upside and whether that analysis is robust.

I should be clear that by critiquing the analysis I'm not critiquing the specific growth recommendations. I'm focusing here only on understanding how they've arrived at the "purely illustrative" figures they've used to scale the potential and to ask how reliable those figures are.

So let's start with the figure that has been (mis)used in headlines, the £4,100 per head number that I discussed in my last blog. Having chosen an arbitrary set of 12 countries (we'll come back to that) the report states clearly;

"If Scotland were to be added to the list of 12 benchmark small advanced economies, it would be 12th out of 13 in terms of GDP per capita. The median of this group is 14% higher than Scotland, a gap of $5,500 (£4,100)"

I've recreated the analysis - it doesn't take long, the source data for the benchmark countries can be found and downloaded here and for Scotland the data can be found in the latest GERS report here). There's nothing like recreating a piece of analysis to expose how simplistic it is.

Apologies if this seems condescending, but for those who don't know: the term "median" simply means the middle one in a set of values arranged in order of size. So in this case we simply arrange the 13 countries (the 12 benchmark countries + Scotland) in order of GDP/Capita and find which one lies in the middle. That country turns out to be the Netherlands. The report does show this, but in such a way that  a casual reader might think it's more significant than just saying "let's assume we were the same as the Netherlands":

I know, right? The headline figure from this 354 page report is based on nothing more sophisticated than saying "if we had the same GDP/Capita as the Netherlands we'd have £4,100 more GDP per capita."

To illustrate how flaky this figure is: they could have decided not to include New Zealand and Belgium in this list (it is after all an entirely arbitrary selection) and the median country would become Sweden - if our GDP per Capita matched Sweden's we'd have 25% higher GDP or £7,250 per person. Yay!

I'm sorry, but using this type of "analysis" to scale the GDP per Capita potential of an independent Scotland is pseudo-scientification of the worst kind.

It's not even as if the Netherlands is the economic model the Growth Commission recommends we seek to emulate:

"We recommend a 'Next Generation Economic Model for Scotland', designed to achieve cross-partisan support, which learns in particular from Denmark, Finland and New Zealand" [2.21, p.9]

So why scale the potential by comparing our GDP/capita with the Netherlands? Because if you compare our GDP/capita to the [mean] average of the three countries mentioned above, it would suggest an increase of only 7%, half the amount used to get to the £4,100 GDP per capita figure (used in the headlines.

To be clear: that would translate into an aspiration to achieve additional tax revenues of c.£4.5bn pa, less than half the amount we would lose from the Barnett Formula driven fiscal transfer on day one!

In fact, as the chart above clearly shows, New Zealand's GDP/Capita is lower than Scotland's - which kind of highlights the ridiculousness of this particular "if we were the same as them" analysis. It seems clear to me that this "GDP/capita gap" nonsense was retro-fitted to this report to force out a palatable and tabloid-friendly headline.

***

So what about the analysis that's used to justify the conclusion that, if we can learn from this cohort of countries, we can expect to achieve GDP growth that would be 0.7% superior to that of the rest of the UK?

Note that this assumption is absolutely critical when it comes to determining the timescales involved - the time it would take us to claw back what we know we'd lose from UK-wide pooling and sharing on day one (the net fiscal transfer of around £10bn pa that comes to Scotland as a result of the Barnett Formula).

The good news is that we can find the IMF data used to reach these conclusions here (you can thank me later).

Let's talk first about the 12 countries that have been selected as the list of "benchmark small advanced economies". There doesn't appear to have been an objective criteria applied to arrive at this list, and the extent of its subjectivity is perhaps best illustrated by looking at the list of "small countries used for comparison" when similar analysis was published in the Independence White Paper (page 620) back in 2013..

The Scottish Government analysis in 2013 concluded that the superior GDP per Capita growth rate enjoyed by those countries that had "the bonus of being independent" was just 0.12% greater than Scotland's onshore economic growth over a 30 year period. That's quite a way short of the 0.7% we're now expected to believe we should expect, is it not?

So what's changed?

Firstly the set of countries used for comparison is different - out go Iceland, Luxembourg and Portugal and in come Hong Kong, Singapore and Switzerland along with Belgium and the Netherlands.  The addition of Hong Kong and Singapore is particularly noteworthy as they significantly raise the average growth rate for this cohort but they are - as the report concedes - countries with appalling levels of income inequality.

It's hard not to conclude that those countries were included to boost the average growth figures of the comparison countries - they certainly aren't countries with socio-economic values consistent with those voiced by the Growth Commission.

The other significant change is that the White Paper looked at onshore economic performance only, whereas the Growth Commission is looking at overall economic performance including Oil & Gas. I can't be bothered to work out how that affects this analysis to be honest, but mention it in case others have the necessary enthusiasm to recreate the analysis separating onshore from off-shore performance.

So can we recreate the 0.7% figure?

The report is surprisingly opaque when it comes to precisely how this number is arrived at (emphasis mine):

"Figure 2-2 shows there has been a a distinctive edge of around 0.7% of GDP growth in small advanced economies over the last 25 years compared to their larger counterparts"

If the number 0.7% jumps out of figure 2-2 for you, you're better at visually interpreting data than I am. I include figure 2-3 above because I presume this shows us (on the right hand side) the countries included in their "large advanced economies" list.

We should now be able to recreate the 0.7% figure and see how sensitive it is to which countries are included or excluded. Notice that the graphs above cover different time periods: 1990 - 2016 (referred to as "the last 25 years") and 2000 - 2016 for the bar charts.

To check that I've understood the methodology, I've recreated chart 2-2 using the IMF data :

Looks spot on to me apart from a superior average growth shown for the SAEs on my graph in 2015. I can't see any problems with the data which I downloaded only today - I suspect the graph in the report was produced over a year ago and the data has subsequently been updated by IMF (but that's just a guess).

So lets see if we can recreate the 0.7%;

• If I use the 27 year period as as used on figure 2-2 and compare these two cohorts (using most recent available IMF data) I get a figure of 0.59%
• If I take the last 25 years as quoted in the text (assuming that's 1992 - 2016) I get a figure of 0.65%
• If I take the last 17 years (as used for the bar charts) I also get a figure of  0.65%

So, at a pinch, we can see how the 0.7% is arrived at.

Sticking with the 25 year time period, let's play with the cohort mix to get back to the cohort used in the White Paper and see how the growth gap changes;

• As used by Growth Commission: 0.65%
• Remove Hong Kong and Singapore: 0.26%*
• .. then add back Portugal: 0.15%
• .. then remove NL, CH, BE & NZ, add back L & IS: 0.45%

---

* [Update 28/05/18] in the detail of the report, Hong Kong and Singapore are pretty much only mentioned in the context of concluding that they are low tax, high income-inequality societies that we don't seek to emulate - which makes it quite bizarre to include them in the cohort for calculating the growth rates we might expect to achieve by pursuing the model the report actually recommends

---

If we change to the last 17 years (as per the bar chart) instead of last 25;

• Using cohorts as used by Growth Commission: 0.65%
• Remove Hong King & Singapore: 0.22%
• Using White paper cohort for SAEs: 0.35%

How about if we look at the performance just of the three countries the report tells us we see to "learn in particular from": Denmark, Finland & New Zealand? Taking the simple average for these three countries we see;

• Over the last 25 years the annual growth gap to the 'large advanced economies' was just 0.06% [just 0.09% to the UK]
• Over the last 17 years these three economies on average actually under-performed the larger countries by -0.02% [during this period the UK was the same as the average for the 'large advanced economies']

Digest this.

The report scales the GDP per capita growth gap by assuming we match the GDP/capita of the Netherlands and scales the rate of growth we might achieve by comparing us to a cohort that includes the high growth, high inequality countries of Hong Kong and Singapore.

But the report actually recommends we seek to mainly emulate Denmark, Finland and New Zealand, countries whose growth rates are not materially different from those of the 'large advanced economies' (or indeed the UK itself).

The Growth Commission's "small advanced economy" cohort is arbitrarily chosen and clearly designed to maximise the "growth gap" claim. Does anybody really think Scotland wants to be socio-economically similar to Singapore or Hong Kong? Just removing those two countries from the cohort reduced the "GDP growth gap" from 0.65% to 0.25%.

If a 0.25% growth gap is a more realistic long term aspiration - and remember the independence White Paper used 0.12% and the three countries the Growth Commission most seeking to emulate achieve at best 0.06% - then instead of having to wait 25 years to deliver the additional GDP per capita the Growth Commission aspires to it would take nearer 70 years. As the analysis above I hope makes clear, that's still being extremely optimistic based on the empirical comparable data.

None of this is to suggest that there aren't good ideas in the Growth Commission worthy of serious consideration - but let's not kid ourselves: the numbers used to scale the upside and indicate how long it might take to get there are backed up by very superficial analysis and have been manipulated to show the most positive case possible.