What's this page?

A place where I'll update questions about regional economies I'm chewing over and add in any others (with acknowledgements obv) that I've picked up from reading or chats.

I'll break it down into different kinds of question. Though picking any one question inevitably pulls on other threads, let's see if we can keep it at least moderately sane. I will aim to write self-contained little pieces discussing each and how they connect, where appropriate with some data analysis.

Topics to cover

With exobrain sub-pages linked if made.

• Econ error rates
• Uncertainty. Connects to error rates but is a bit different - some of that is already in Econ error rates. Ranks, the power of the single number.
• National accounts (probably no replacement possible given its status as global accounting system responsible e.g. for regional funding allocation; possible to work alongside it?)
• Wild deflators - telecoms example (see growth point below and links in Productivity digging)
• Imputed rent, the numbers. And why it's a reflection of the 'accounting' part of national accounts / the need to align nations. But then, is that really any use for understanding regional development?
• Public sector output - as an example for thinking through the implications of top-down slicing of output (in this case based purely on average public sector pay times job number, used to slice a nationally estimated pound value up between regions).
• Tradeable vs foundational - thinking about the whole economy.

Misc links

• "Have the UK’s northern cities really experienced a productivity miracle?" Paul Swinney, economics observatory via linkedin.

Intro

A few words on the spirit of this work, going back to a point I was mulling back in 2018. I've just been re-reading a [2023 post by Richard Murphy](GDP are an honest attempt to do something that is conceptually and technically very difficult – and there are honest debates about what to include and what not to include.) about imputed rent (the rent that would be paid if owner-occupier houses were rented). It has - shock! - a really polite, critical, well-informed comment thread. Richard starts off with a GOTCHA claim: imputed rent isn't real! "10% of GDP is made up – it simply does not exist in the real world". Just another example, he says, showing how ridiculous national accounts are.

A commenter replies:

I really do wish you’d dial back the language a bit! GDP are an honest attempt to do something that is conceptually and technically very difficult – and there are honest debates about what to include and what not to include.

I want to keep my own discussions firmly grounded in that respect for work done by much smarter people than me, often under extreme pressure. The ONS, especially in recent years, has taken a bit of a beating despite doing phenomenal work with less money and a much harder survey landscape. Politicians will spin the tiniest percentage point change in GDP, causing stampedes this way and that. None of it's conducive to having those calm, honest debates that commenter wishes for.

But let's try.

I want to be guided by what's useful for understanding regional economies. National accounts are just that - national - and issues arise straight away from re-purposing those tools for subnational thinking.

The thread of questions will circle round - data thoughts lead into "what are we trying to understand" leads into "what theories do we have in our brains? What assumptions can we see, which can't we?" Those circle back round...

Questions

• Top level one: given all this, what should it mean for how UK regions think about their own development? Can we collectively come up with better ways to do this?

The problem there includes: economics' general commitment to spurious accuracy, or more specifically national accounts (economics has econometrics...) Also: ranking is very, very hard to let go of even if you know it's dumb. See: UK universities using THES rankings - dropping in and out of the top 100 is the difference between international student numbers halving or doubling in some fields. It matters. It shouldn't. It does. Cf. Richard Harris trying to push this boulder up the hill...

The data

Data threads

• Revisit Calvin Jones' breakdown of data issues ('measuring what we manage') here on LinkedIn. Via this on using the Welsh IO tables.

What do we want to know and why?

And how does this change the kinds of questions we need to ask, as well as where to look for answers?

A lot of things come under this heading:

• If we want to support development in a region, what's the balance between 'picking winners' (leaning into existing economic strengths a region already has) versus thinking holistically (how to develop an entire region's economy, thinking across sectors and siloes)?
  • See e.g. Giles Wilkes' piece on the Pitfalls of the Sector Method and GVA factory thinking for some common ways to go awry with winner-picking.
  • The whole picture includes the foundational economy (Welsh gov page explaining the concept) - something my Y-PERN colleagues at Sheffield Hallam / CRESR have done amazing work on. One good way to define foundational economy jobs might be "anyone who still had to leave the house to work during COVID". The pandemic drew attention to a harsh, paradoxical truth - those jobs were (a) entirely essential to making sure the country didn't fall over and (b) on average, among the worst-paid jobs (with obvious exceptions). They are also the largest percent of any regional economy (CRESR's work on this has it at 69% of all jobs in South Yorkshire).

A different but similar axis:

• Claim: focus on tradeable jobs, as growing these is the quickest route to strong growth. These jobs are almost always the opposite of foundational ones - any job/sector that makes goods and services that can be traded outside the region.
  • The theoretical foundation comes from (I think) this Rice & Venables working paper for the Productivity Institute. From a quick read, it looks like it's re-stating Ricardian comparative advantage for UK regions pretty much exactly. So all the same arguments and counter-arguments for international trade as inter-regional apply (though of course domestic context adds plenty of wrinkles).
• The claim continues: those tradeable jobs are the linchpin of any region; the wage rates of service / foundational sectors are higher in places with better tradeable jobs because there's more internal demand for them. While this is clearly a chunk of the economic story, it's also worth considering the ways non-tradeable sectors feed into overall regional productivity - they aren't just passive beneficiaries of tradeable 'trickle down'.
  • But foundational jobs / sectors trade across borders too. How different are they in terms of cross-regional purchases, if looking at intermediate links? Might be some answers in the industry to industry payment data...

What scale to deepen connections between?

• Strengthen them within a region (e.g. build a West Yorkshire tram system) or between them (cancel that and put money instead into improved services between Northern cities?)
  • Both, of course. But - what's the theory on the relative ROI for each approach, and why? Cf. the arg between Paul Swinney and West Mids.

What is real growth? Why is it all so mad?

See e.g. the telecoms stuff in Productivity digging.

Here's a comment for this Calvin Jones post on digital:

My brain is a bit scrambled by how digital services are measured anyway. Their value shift since all the early 2020s adjustments are boggling - see the plot here: the telecoms sector's chained volume measure increased by a hundred and fifty times (index of 0.7 in 1998 up to 106.1 in 2023). Whereas its proportion of the economy in pounds went from 1.92% in 1998 to 1.45% in 2023 - i.e. dropped by * 0.75ish. I can go through the CV vs CP numbers and I think just about get the quality-adjusted volume / pounds per bit justifications, but it still melts my noggin - mainly along "if growth can be this reliant on model choices and what deflator theory we like, what does it really mean??" lines. Thoughts gratefully received! Cf. the Brookings paper. https://www.brookings.edu/wp-content/uploads/2020/03/WP58_Abdirahman-et-al-1.pdf

Regecon journal

4th-Feb-2026: reflective writing on the whole regecon thing

Trying to get started on this. See Exobrain diary#^4779eb for thoughts on going with the brain grain.

I have a first sentence to put on LinkedIn anyway: "Seems to me there's a bit of a shortage of White Blokes Expressing Strong Opinions about Regional Economics on here, so thought I'd add my tuppence."

What I'd like first, though, is just to shape what I think this is going to be, and some way to maximise the odds it'll progress.

Bits on the uncertainty article 5th-Feb-2026

See here? "Imputed rental is excluded from "Industry L: real estate" because including it would distort productivity measures, since the output is mainly an imputed value rather than a result of labour or market service provision."

I was told it's included. Excluding would seem sensible to me, but... well, that's something we can check against the CP totals.

Note also 'strengths / limitations' in the latest subregional thingyo. Points out it gets volatile below ITL1.

Oh - I asked about regional GDP figures. Let's see if I can check on productivity figures. I'll check totals (as I've done before) and then email.

Oh ah 2: CIs are in the LFS / APS hours worked data - I can grab that from NOMIS. Would be good to check it roughly aligns with hours in the prod stuff. But let's see (or possibly just nab the CIs and apply them).

..

OK, got a reasonable way just with the ABS and linking to region/industry GVA (the usual sector mismatch faff). Just want to list some things to aim for before calling it done:

• Use exactly the same approach for current prices, and put some bounds around locations quotients at ITL1 level.
• Make some assumptions about SEs scaling to smaller geographies and do the same, maybe showing how much uncertainty bounds could increase.
  • This may require thinking through how whole-economy growth numbers come out of its underlying parts, which might be a bit much for a first post.

That's really nice, you can see the COVID signal very clearly in many places.