Where is all the investment in the UK!?


Where is all the investment in the UK!?

By Michael Becker

It’s no secret that there is a productivity puzzle in the UK and that business investment has stagnated in recent years. But assuming this pattern holds true across the whole of the UK may be wrong. With the economic geography of the UK gaining a lot more attention in the policymaking space (did somebody say “Levelling Up”?), new data to shine a light on the subnational landscape of capital investment in the economy is a welcome development.

A new ESCoE Discussion Paper from Josh Martin and I looks at the regional distribution of capital investment in the UK from 1997-2019.

What we did

This paper used new experimental estimates of capital investment by region from the Office for National Statistics (ONS) which introduced asset-type breakdowns for the first time and provided greater subnational detail than previously available. We first documented how the new and previous measures were put together, describing the data sources and assumptions underpinning each, and the dimensions of each published dataset. We then explored how investment is spread across the UK, along with exploring which types of assets each region is investing in. The paper explores correlates of regional investment with output, productivity, hours worked, firm size distribution, and Foreign Direct Investment.

The findings

The paper has a number of insights and presents some helpful visualisations in understanding the patterns of investment in the UK.

Capital investment is shown to be unequally distributed across the UK, in a similar fashion to economic activity and employment. There is significant variation between both high-level regions of the UK, and within those regions. Capital investment is unsurprisingly highly correlated with several macroeconomic indicators such as GVA, hours worked, and population. The paper shows the regional patterns of investment across the UK in 5 major asset sub-categories.

We find that regions with higher labour productivity tend to have a higher share of their total investment in intangible assets (think R&D and software). The share of investment on intangible assets tends to be higher in the South East and East of England, and to a lesser extent, London. Investment per hour worked is also typically higher in areas that have higher levels of labour productivity, and that is strongest for ICT equipment and intangible assets.

Asset share of total investment, by productivity decile, ITL3 regions, Great Britain, 2004 to 2019 average
Source: ONS, authors’ calculations.
Notes: Decile 1 is lowest productivity, decile 10 is highest productivity. See Figure 7 in the paper for more details.

As the estimates from the ONS are in current prices (nominal terms), and so don’t take into account inflation, we developed price indices (deflators) for the estimates so we could explore real (constant price) regional investment growth in the UK. This highlighted that the distribution of investment intensity across regions was converging somewhat over time, with regions that had high levels of investment in the late 1990s seeing slower investment growth over the subsequent 20 years on average, and those with low initial levels of investment seeing faster growth since then.

The deflators we constructed in this paper are available to download from The Productivity Institute’s Productivity Lab. More details are in the paper.

Why it matters

There are several reasons why understanding the subnational landscape matters and the insights it can offer has potential policy implications which can be used to fuel future economic policy.

Understanding where the investment in the UK takes place allows for policymakers to understand the economic geography better. With the expanding portfolio of UK economic statistics at a granularity of geography not previously available, understanding the geographic economic inequalities at a more localised geography is getting easier. There is also the angle that understanding the economic inequalities that exist in the UK allows for policy makers, through agendas like ‘Levelling Up’, to address these inequalities through targeted policy.

More data at a lower geography can hopefully help unlock the productivity puzzle of the UK. With aggregated national estimates generally being used to investigate productivity, this approach can mask significant differences in the economy. Subnational data allows for us to explore productivity locally and show what the differences are, what asset investment is preferred, how has it changed over time, and what might be driving the productivity differences in the UK. It might even offer a blueprint to addressing the UK’s productivity woes. For example, as more productive regions have a higher share of intangible investment, encouraging investment into software or R&D in areas which have previously had a lower share, (e.g. through subsidies or tax breaks) might be an approach to boosting productivity in these regions.

Finally, from a measurement perspective, having more detailed regional capital investment estimates is the first in-road towards regional capital stocks, capital services and total factor productivity estimates. These require detailed information on the types of investment, such that the appropriate price deflators and depreciation rates can be applied. The asset breakdown introduced in the new estimates, and explored in our paper, are an important step.

More work needs to be done before regional capital stocks, capital services, and total factor productivity estimates are possible. The paper contains an account of the developments in data for regional capital investment in the UK in the past, with the new experimental estimates from the ONS being the latest addition. For more detail, please explore the paper below.

To download the Discussion Paper click here.

Michael Becker is an Analyst in the Public Policy Analysis team at the Office for National Statistics

ESCoE blogs are published to further debate. Any views expressed are solely those of the author(s) and so cannot be taken to represent those of the ESCoE, its partner institutions or the Office for National Statistics.

About the authors

Michael Becker

Related publications