ONS Subregional Productivity 2022, analysing comparative sub-national productivity in the UK
UK regional productivity, controlling for the “London effect”
This blog is the first of a series of Productivity Lab publications describing the latest UK sub-national statistical productivity data released by the Office for National Statistics (ONS), the ONS Subregional Productivity July 2022 release. This dataset provides information on productivity per job and per hour basis, at three different levels of geographical aggregation (ITL1, 2 and 3), for the period 2002-2020. Based on this dataset, we will focus in this first blog on productivity across UK regions at the ITL2 level (of which there are 40 and include Northern Ireland, counties in England, groups of districts in Greater London, groups of unitary authorities in Wales and groups of council areas in Scotland).
Subsequent blogs in this series will leverage the more detailed regional information at the ITL3 level to delve deeper into the relative productivity performance within sub-regions in the UK (of which there are 174, including counties, unitary authorities, or districts in England, groups of unitary authorities in Wales, groups of council areas in Scotland and groups of districts in Northern Ireland). The latter level of productivity measurement has also been used in the Insights Papers for the Regional Productivity Forums of The Productivity Institute.
UK Sub-national Productivity Level (ITL2) comparisons and the London effect
The UK exhibits some of the highest regional and subregional productivity inequalities among industrialised countries (Carrascal-Incera et al., 2000; McCann, 2020; Zymek and Jones, 2020; van Ark and Venables, 2021). In particular, there is a very wide variation in productivity levels between London and its hinterland and the rest of the UK, except for some parts of eastern Scotland (including Aberdeen, whose productivity record is coloured by the predominance of an oil- and gas-based industry). Moreover, the UK is characterised by significant difference in productivity over very short distances by international standards. These two observations suggest that in terms of the functioning of the urban and regional economic system, there are likely to be severe blockages, constraints, inhibitors, or missing links in place which limit the diffusion of productivity or prosperity across the country. Exactly why this is the case is a key research agenda of The Productivity Institute.
Understanding and addressing regional variations in economic performance and well-being forms one of the cornerstones of the recent 2022 “Levelling Up the United Kingdom” White Paper. It sets out how the very marked interregional ‘core-periphery’ structure of the UK economy traverses all types of places, which has led to high inter-urban as well as inter-regional inequalities alongside significant intra-regional variations, thereby painting a very complex economic geography picture which is largely atypical by international standards.
To illustrate, we present the regional variation in ITL2 productivity levels relative to the national UK ITL2 productivity level average, measured as gross value added per hour worked. The figure on the left shows the regional heterogeneity in productivity levels relative to the UK’s ITL2 regional average for the whole country (£36.95, based on ONS 2022) in 2019. The figure on the right presents a slightly different picture, comparing the ITL2 productivity levels relative to the UK’s ITL2 regional average, EXCLUDING London (IT1 code TLI) (£34.17, based on ONS 2022).
Excluding the “London effect” allows us to consider two main aspects. First, by not including the London region, the map offers a more granular analysis of the sub-national productivity levels outside the capital region, illustrated by the increased variation in the regions’ colours. Second, reproducing the calculation of the UK average, excluding the capital region allows a comparison of sub-national regional productivity relative to an average level that excludes the extreme London productivity values. This results in a more even comparison of average productivity growth between regions that are, broadly speaking, similar in terms of their economic structure.
Changes in sub-national productivity: 2008-2019 UK Productivity Matrix
In order to better understand the relative performance of regional productivity in the UK since the financial crisis (2008), we proceed to plot the UK ITL2 regions according to their nominal productivity level in 2019 and their 2008-2019 change in productivity. Following the analysis performed before, we include the UK ITL2 regional average with and without the capital region (London, ITL1 code TLI) as a reference point. To analyse changes in productivity in the ITL2 sub-regions, we use the four-type taxonomy of UK regions (Zymek and Jones, 2020; TPI, 2021) to build the UK Productivity Matrix. The UK sub-regions are plotted according to their productivity level and change relative to the UK’s average (including and excluding London).
The scatterplot shows the UK’s average reference point of the nominal labour productivity level of £36.95 in 2019 and a 6.1% change in productivity from 2008-2019. Controlling for the “London effect”, the recalculated UK 2019 average level of nominal labour productivity drops to £34.17 and the change in productivity growth increases to 6.4% between 2008-2019 (ONS, 2022 July release).
Productivity growth in this figure is based on the chained volume (unsmoothed) GVA (B) per hour worked indices. Therefore, price changes are excluded from measured productivity growth and expressed in real terms.
In 2019, regions in the South East (Berkshire, Buckinghamshire and Oxfordshire; Hampshire and Isle of Wight) and Eastern Scotland showed a productivity level above the UK’s regional average and productivity growth higher than the average in the period from 2008 to 2019. Hence these ITL2 regions can be described as “Steaming Ahead”.
The “Losing Ground” regions show a productivity level above the UK’s regional average but productivity growth less than the UK’s regional average. In 2019, several Inner (East, West) and Outer (West and North West; East and North East; South) London boroughs, Surrey, East and West Sussex in the South East, and Cheshire in the North West, fell into this category. Despite being highly productive, these ITL2 regions show below-average productivity growth.
Compared with the UK ITL2 average excluding the capital region, North East Scotland; Gloucestershire, Wiltshire, and Bath-Bristol area in the South West; Bedfordshire and Hertfordshire and Essex in the East of England and Kent in the South East also belong to this category.
Although the “Catching Up” ITL2 regions have productivity levels below the UK regional average, they show productivity growth rates higher than the UK regional average from 2008 to 2019. These include the Scottish regions of the Highlands & Islands and West Central Scotland; the Welsh areas in East Wales and West Wales and the Valleys; Northern Ireland; the parts of Tees Valley and Durham, Northumberland and Tyne and Wear in the North East; Greater Manchester in the North West; Herefordshire, Worcestershire and Warwickshire, and West Midlands in the Midlands; Leicestershire, Rutland and Northamptonshire, and Lincolnshire in the East Midlands; and Cornwall and the Isles of Scilly and Devon in the South West of England. It is interesting to point out the high 2008-2019 productivity change of Cornwall and the Isles of Scilly (17.8%) and West Central Scotland (19.5%) compared to the UK regional average of just above 6%.
A good number of the ITL2 areas are described as “Falling Behind”. The productivity level and growth rates from 2008 to 2019 in these UK regions are below the UK’s regional average. Moreover, the majority of these regions are falling behind the average, both when calculated with respect to the inclusion or exclusion of London. Many areas in the Midlands (Derbyshire and Nottinghamshire; Shropshire and Staffordshire), North West (Cumbria; Lancashire and Merseyside), Yorkshire & the Humber (East Yorkshire and Northern Lincolnshire; North; South and West Yorkshire), East Anglia in the East of England, Dorset and Somerset in the South East and Southern Scotland remain behind the UK regional averages for both productivity levels and productivity growth rates. A small number of regions, all in the South of England and North Eastern Scotland, switch between the “falling behind” category and the “losing ground” category when London is removed. These regions are highly productive but somewhat less dynamic than others with similar productivity levels.
The figure below pictures the UK regional productivity matrix based on a revised version of the four-type taxonomy (Zymek and Jones, 2020). In this revised version, the change in ITL2 regional productivity for 2008-2019 is compared to a UK regional growth performance that does not include the capital ITL1 region of London (TLI). The maps above indicate that the UK exhibits a strong core-periphery structure to its economic geography of productivity, with London and its hinterland plus parts of Eastern Scotland representing the core, with the rest of the country reflecting an economic periphery. However, many aspects of the core-periphery structure remain even when we remove London from the analysis. In particular, most of the areas which are falling behind are clustered in the Midlands, the North of England and the South of Scotland.
Yet, within these overall patterns, there are some interesting nuances. We can group the falling behind and steaming ahead categories as broadly reflecting interregional divergence processes and also group the catching up and losing ground categories as broadly reflecting interregional convergence processes. From the matrix scatterplot, we see that when London is included in the calculations, there are almost exactly the same numbers of regions within the convergence and divergence categories. This suggests that allowing for the strong core-periphery structure of the UK’s economic geography, the interregional convergence and divergence are largely similarly-sized processes. Meanwhile, when London is removed from the calculations, we see that the number of regions in the broad convergence category is some two-thirds higher than in the broad divergence category. This suggests that outside of London, many of the regions are slowly becoming relatively more similar to each other. These results provide some nuanced observations to consider when formulating spatial policies, depending on whether the economy of London is or is not a consideration in terms of the policy-design and policy intentions, such as setting labour productivity targets or considerations around infrastructure and land use or the key priority of maximising the diffusion of knowledge from London and its hinterland across the country.
Carrascal-Incera, A., McCann, P., Ortega-Argilés, R., and Rodriguez-Pose, A., 2020, “UK Interregional Inequality in a Historical and International Comparative Context”, National Institute Economic Review, 253, R4-R17.
McCann, P., 2020, “Perceptions of Regional Inequality and the Geography of Discontent: Insights from the UK”, Regional Studies, 54.2, 256-267
ONS (2022). Subregional productivity in the UK: July 2022.
Bart van Ark and Anthony Venables (2020), A Concerted Effort to Tackle the UK Productivity Puzzle, International Productivity Monitor, Centre for the Study of Living Standards, vol. 39, pages 3-15, Fall. (also published as Working Paper 001, The Productivity Institute)
Zymek, R. and B. Jones (2020), UK Regional Productivity Differences: An Evidence Review, Research Paper, Industrial Strategy Council, February.