Written evidence from the Office of National Statistics (ISG 190)

Official Statistics and UK Industry; an Introduction

Official statistics are for the benefit of society and the economy generally.  They allow the formulation of better public policy, and the effective measurement of these policies.

In response to the Business, Innovation and Skills Committee’s call for evidence on the Government’s Industrial Strategy, this note summarises what statistics are available to policy-makers, to assist in the development of an industrial strategy.  It also sets out what high-level trends these statistics currently suggest.

In particular, reflecting the themes highlighted in the Committee’s call for evidence, this note summarises:

• what statistics are available to help us understand trends across different industry sectors, in the context of a continuing reorientation of UK industry toward the services sector;
• how data might help us identify different trends in the performance and concentration of industry in different regions and countries of the UK; and
• what we currently understand about the productivity of UK industry, as well as the ways in which the Office for National Statistics (ONS) is improving the UK’s suite of productivity measures. 

(Please note that this document is not intended to provide an exhaustive account of trends in UK industry; for further detail on any of the statistics in this note, please contact the Office for National Statistics). 

Understanding the performance of UK industry by sector

The Office for National Statistics publishes a range of data to assist decision-makers in understanding how economic activity in the UK is evolving across different industry sectors.

How is industry classified?  

Information on industries is generally presented using the UK Standard Industrial Classification.  The UK Standard Industrial Classification classifies businesses by the primary type of economic activity in which they are engaged. 

Using this system has the advantage of providing a relatively uniform set of definitions which can be applied across all countries, and used to compare trends internationally. The process of standardisation across countries, however, poses some challenges in the measurement of emerging industries. 

What data are available?  
 

• Initial estimates of the quarterly value of the goods and services, broken down by industry sector, are published around four weeks after the end of the quarter, as part of the Preliminary estimate of Gross Domestic Product (GDP). 
• More detailed, monthly data are published on the performance of the services, production (including manufacturing) and construction sectors, approximately eight weeks after the end of the reference month.
• Although the way these estimates are produced varies by industry, for most industries data are collected from ONS’ Monthly Business Survey.  The Monthly Business Survey collects data about the turnover of over 45,000 businesses; this is then adjusted to remove the effect of changes in the price level (inflation).
• Even more detailed industry breakdowns are published based on results from the ONS Annual Business Survey. However, this output only covers non-financial businesses and is produced around 9 months after the end of the year. The business level data from both the annual and monthly business surveys are available to government users for research purposes via the ONS Virtual Microdata Laboratory (VML).
• ONS also publish regularly on the number of workforce jobs by 20 key industrial groups in its regular Labour Market Bulletin. Employer surveys are used to measure the number of filled jobs in the economy (different from the number of people in employment, as some people have more than one job). This information is typically published for the third month in the quarter, around ten weeks after the end of the month.
• Given the prominence of these economic statistics in understanding the development of the economy and detecting turning points in economic growth, the methodologies for these statistics are a key part of ONS’ development programme. In addition, ONS also plans to use administrative data from Value Added Tax (VAT) returns to supplement data from the Monthly Business Survey to improve these estimates.

What do the data tell us?

• As Table 1 sets out, services accounted for nearly 80 per cent of the UK economy in 2013. This proportion has been increasing over time, reflecting a continuing reorientation of the UK economy towards services output. Production accounted for around 15 per cent of the output of the UK economy in 2013; within this, the manufacturing industry group accounted for around 10 per cent of the UK economy’s output.

• As Figure 1 makes clear, the trend of continued reorientation of the UK economy towards services is shared by other similar developed economies. Notably, the services industrial grouping has been a larger share of the UK economy since 2004 than the group of comparable economies.

• Table 2 draws from more detailed quarterly data to show the change in contribution of selected industries to GDP, between 1997 and 2013.

Table 2: Contribution of selected industrial groups to GDP, 2013

1 Contributions are to Output GVA, used as a proxy for contributions to GDP

Understanding the performance of UK industry by region

The Office for National Statistics also publishes a range of data to assist decision-makers in understanding trends in industry by region. 

How are regions defined?  

The principal official statistic for measuring the performance of industries across the regions of the UK is regional Gross Value Added (GVA), which is comparable with GDP, differing only in the exclusion from GVA of taxes (less subsidies) on products (e.g. VAT).

Regional GVA is compiled within the National Accounts framework, which ensures consistency of regional and national statistics.

Currently, regional GVA estimates are produced for areas defined by the European classification system called the Nomenclature of Territorial Units for Statistics (NUTS). In the UK, estimates are published for the top three levels of the NUTS hierarchy, which are:

• NUTS1: 12 areas, comprising Scotland, Wales, Northern Ireland and the 9 English regions (an England aggregate is also produced to allow country comparisons)
• NUTS2: 40 areas, mostly comprising groups of counties and unitary authorities
• NUTS3: 173 areas, mostly comprising single counties and unitary authorities

There is also a category called Extra-regio, which covers GVA that cannot be assigned to a region or regions of the UK. This category includes activities on the continental shelf (oil and gas extraction from the North Sea) and the activities of UK embassies and armed forces overseas.

What data are available?

• At the NUTS1 and NUTS2 levels, regional GVA is published with an industrial breakdown to 33 industries. At the more detailed NUTS3 level, 11 broad industry groups are provided.
   
• Regional GVA is compiled using the income approach, and in addition to GVA, estimates are provided for the various components of income that make up GVA, such as Compensation of Employees, Mixed Income of the self-employed, Rental Income, and Gross Trading Profits (GTP) of business corporations. These components allow users to understand in greater depth the nature of industrial performance in an area, as they show which components are driving growth or decline.

• A range of labour market statistics are available which give employment numbers by region and industry and regional unemployment figures. The Business Register and Employment Survey publishes employee and employment estimates at detailed geographical and industrial levels. High level estimates are published on the ONS website and more detailed estimates are published on the Nomis website. Employment data from the survey have also been used by ONS to perform analysis of the geographical concentration of industries, the industrial specialisations of areas and the spatial distribution of industries. Earnings estimates are also produced, broken down by full-time and part-time working patterns. Used with regional GVA these data can also provide regional productivity estimates.

• ONS’s Annual Business Survey (ABS) provides a wide range of data (including historical data) on business enterprises, and can provide estimates for many different levels of geography with far greater detail of industry than any comparable source. The latest ABS regional results show the size and growth of the UK non-financial sectors within the regions of the UK. Data on the regional breakdown of exporters and importers from the ABS are also published on our website. These data are not compiled within the National Accounts framework and can exhibit a high level of volatility, particularly at low levels of geography or industry. They can, however, be used in conjunction with regional GVA to inform analyses requiring a greater degree of industrial precision than is normally available.
   
• The Inter-Departmental Business Register is a further source of regional data. Business counts by region are given in our latest publication of UK Business: activity, size and location, and business births and deaths by UK region are published in our Business Demography release.

What does the data tell us?

• As set out in Table 3, there are some common trends in the growth of industry across regions; with the effect of increasing property values producing widespread growth in real estate, and the equally widespread decline in mining and quarrying affecting a large number of regions.
   
• Table 3 also highlights different trends in the growth of industry across regions; for example, arts and recreation (including sports) show strong growth in the North, and manufacturing is faring badly in the North East, South East and Scotland.

Table 3: NUTS1 regions, growth in nominal GVA 1997 to 2014, highest and lowest SIC industry sections

*growth in the activities of households as employers is greater, but is not considered here as GVA is very small

• The distribution of industries varies across the UK.   If we look at two contrasting industries in England and Wales, manufacturing, and information and communication, we can see the areas with the greatest activity in the more traditional (and generally declining) industries and the newer (and generally growing) technology industries.
               

• The places with the largest GVA in 2014 in manufacturing were Hertfordshire (£3,514m), Cheshire East (£2,997m), and Berkshire (£2,883m). The first two of these have very large contributions from the pharmaceutical and transport equipment industries, while Berkshire has a much broader base across a range of manufacturing industries.
   
• The places with the largest GVA in 2014 in information and communication were Berkshire (£7,834m), Westminster (£7,247m), and Camden and the City of London (£6,961m). This shows the concentration of technology industries in and around the capital, but it is also interesting to see that Berkshire is even stronger in technology than it is in the more traditional manufacturing activity.

Understanding the productivity of UK industry

Given the importance of productivity for firm growth and survival, international competitiveness and living standards more broadly, it seems a natural focus for industrial policy.

As the committee may be aware, ONS’s main measure of labour productivity is calculated by dividing a measure of output by an indicator of labour input. The output measure ONS uses is Gross Value Added (GVA). Labour inputs are measured in terms of workers, jobs, and hours worked, and using these data ONS can calculate productivity metrics for a range of industries and regions as well as for the UK as a whole.

What does the data tell us?

ONS’s preferred measure of labour productivity – output per hour – has grown very slowly by historical standards since the start of the recent economic downturn.

Figure 4 shows the ‘Productivity Puzzle’ in its most simple form. It shows the development of two measures of productivity – output per hour and output per worker – alongside their projected path had they continued to grow at the 1994-2007 rate over the full period. It shows that productivity growth slowed markedly after 2008 and that, as a result, productivity as measured by output per hour 15.5% was below its pre-downturn trend in Q1 2016. ONS produces these headline labour productivity data each quarter.

Figure 4: Output per hour and output per worker

Seasonally adjusted, quarter 1 (Jan to Mar) 2008 to quarter 1 (Jan to Mar) 2016, UK

Source: Office for National Statistics

UK productivity by industry

Alongside productivity statistics for the whole economy, ONS also produce quarterly estimates of productivity at the industry level, enabling analysis of how different industries have been affected by the recent slowdown in productivity growth.

Figure 5 shows the average quarterly rate of output per hour growth in a range of different industries between (a) Q1 1994 and Q1 2008 and (b) the period since Q1 2008. It suggests – consistent with a large academic literature on this matter – that the slowdown of productivity growth has been broadly based. While specific industries account for a large portion of the slowdown in productivity, output per hour growth has weakened markedly in almost all industries in recent years.

Figure 5: Average quarterly productivity growth across services industries and manufacturing, Q1 1994 to Q1 2008 and Q1 2008 to Q1 2016.

Output per hour, average quarterly growth, UK, %

Source: Office for National Statistics

To enable international comparisons of productivity, ONS also produces a set of biannual articles examining labour productivity in the UK and a range of other countries. While the focus of these articles is on headline labour productivity, the most recent edition also included a set of experimental estimates of labour productivity in different industries across countries which may be of interest to the committee.

Productivity by UK region

ONS also publishes annual estimates labour productivity for the high-level, NUTS1 regions, as well as experimental estimates for the smaller NUTS2 and 3 sub-regions, Local Enterprise Partnerships and selected city regions. These enable a comparison of output per hour across different parts of the UK, which reflect both the mix of industrial activity and the quantity of labour input in each region.

Figure 6 shows two alternative measures of labour productivity in each region expressed relative to the UK level, and indicates that productivity in London was almost 30% higher than the UK average on both metrics in 2014. The only other region with productivity above the UK average was the South East of England (9% above the UK average), while output per hour in the north of England and across the midlands was between 9 and 13% below the UK average. However, unlike ONS’ other estimates of labour productivity, these data are expressed in nominal terms, which means that they cannot account for variation in price-levels or price-growth across regions. This is an area in which ONS is planning to make improvements in the future.

Figure 6: Labour productivity by NUTS 1 region or country, 2014

GVA per filled job and GVA per hour worked, UK=100

Source: Office for National Statistics

More detailed measures and analyses of productivity

Alongside these estimates of labour productivity, ONS also produces a range of more detailed statistics and analyses of productivity. ONS’ estimates of multi-factor productivity (MFP) offer particularly detailed insights into the performance of the UK economy and its constituent industries. These experimental estimates – published annually at present – divide output growth in to the contributions from changes in the (a) quantity and quality of labour and (b) the volume of capital services applied to the production process. These analyses enable ONS to disentangle output growth which is due to:

• Changes in the number of ‘effective’ hours being worked in a given industry – accounting for both changes in the quantity and quality of labour used in production (‘Labour composition’)
• Changes in the amount of capital being applied to the productive process – accounting for differences in capital: labour ratios in different industries through time (‘Capital deepening’)
• A measure of technical progress – capturing the efficiency with which different factors are combined in production (‘Multi-factor productivity’ (MFP)).

Of these, the last measure of MFP growth is considered particularly important for changes in living standards. Figure 7 shows the contribution of each of these effects to labour productivity growth over the last four decades. It shows that MFP fell sharply during the recent economic downturn, as it did during previous downturns. However, the failure of MFP to rebound from this period (as it has done following earlier recessions) is one of the defining characteristics of the UK productivity puzzle.

Figure 7: Decomposition of labour productivity growth, 1971 to 2014, UK

Source: Office for National Statistics

Finally, ONS also produce a range of analytical pieces on productivity which are designed to enable users to better understand developments in the headline statistics. The newly formed ‘Productivity Bulletin’ now provides the vehicle for these outputs. In the past, these analyses have included an examination of industry contributions to labour productivity growth and the distribution of firm level productivity, as well as stand-alone articles on productivity puzzle, including 'What is the productivity puzzle?', ‘The productivity conundrum, explanations and preliminary analysis’, and 'The Productivity Conundrum, Interpreting the Recent Behavior of the Economy'.

Planned developments

As the committee will be aware, the measurement of productivity was at the centre of the recent Review of Economic Statistics conducted by Professor Sir Charles Bean. Bean The review made a range of recommendations on the existing suite of productivity outputs, which ONS are now working to implement.[1] In broad terms, these include:

1. Enhancements to ONS’ Multi-factor Productivity estimates: This work will involve developing a greater degree of industrial detail available in our MFP statistics, and include a shift from producing them on an annual cycle to the same, quarterly cycle as the labour productivity estimates. These data will enable ONS to support the policy-making process by producing a conceptually stronger measure of productivity in something closer to real-time. Quarterly MFP statistics are planned for publication from January 2018 onwards.

2. Improving estimates of labour productivity: ONS will also be working to enhance the robustness and coverage of existing measures of labour productivity. This will include changes to existing estimation systems, analysis of the accuracy of our estimates of hours worked by industry and changes to market sector measures of labour input to enhance consistency across the National Accounts.

3. Developing micro-data resources: ONS is also undertaking work to improve the availability of survey- and other micro-data resources for productivity research. This work – carried out in concert with academics, the ONS Fellows and the National Institute for Economic and Social Research – will enable more detailed analyses of firm-level productivity.

October 2016