ONS Economic Statistics and Analysis Strategy: 2017

The year 2017 is an exciting time for economic statistics, having made a good start in addressing the recommendations of the Bean Review. The Review included some specific recommendations, but it called for a more fundamental change in mindset. The Office for National Statistics needed to be more proactive in responding to changing user needs.

With that in mind, the EU referendum result has inevitably caused us to review our priorities and put more weight on important issues like migration, trade and business investment. Our focus is to transform economic statistics to give decision-makers the information they need as Britain shapes its role in the world, particularly through the use of innovative research and a range of new data sources.

An important element of this new approach is to work with others to identify issues early and to work collaboratively on improving our statistics. This includes working with the academic community more closely through the new Economic Statistics Centre of Excellence to identify emerging challenges and develop methods to address them, or engaging more closely with stakeholders through our expanding London presence. We will use the new ONS Data Science Campus to get beyond producing the same numbers every month, recognising that the questions we need to inform are changing; we have to change to get the data that’s necessary to make sense of those questions in the way they are being posed.

This Economic Statistics and Analysis Strategy is an important tool in this change. This second publication presents a tighter set of priorities and is aimed at stimulating debate about what we prioritise in future years. All statistics can be improved, but we have to recognise the relative cost and importance of the choices we make. While we would love to be in a position to deliver every user’s needs all the time, we recognise that some priorities have to be given prominence. We are keen to engage any and all users to understand your needs and to help us continue to target the UK’s most pressing statistical issues. I look forward to your views.

Jonathan Athow
Deputy National Statistician

This is our second annual Economic Statistics and Analysis Strategy. It provides users, stakeholders and researchers clarity on how we are working to improve UK economic statistics. It highlights the different areas where we wish to work with others, through horizon scanning for new issues to researching new specifications or methodologies for us to use in new or improved statistics.

Vision

Our vision is that we will have inquisitive experts providing users with insightful, innovative economic statistics and analysis. We aim to be an internationally recognised and authoritative voice for the collection, production, compilation, analysis and dissemination of economic statistics.

Achieving our vision will require Economic Statistics Group (ESG) to be:

• insightful – providing our users with the information necessary to make high-quality decisions in the light of the best evidence available
• innovative – constantly exploring, investigating and using the latest techniques, methods and data sources to ensure our statistics remain pertinent and descriptive of modern life
• inquisitive – using strong analytical skills to illuminate our statistics and the stories behind them for our users, ensuring our statistics are scrutinised closely and compared with the best information available to contextualise, calibrate and query our results to assure their quality so we publish accurate, credible, relevant and consistent estimates

Recent achievements

In the past year, we have made significant advances in developing a stronger statistical system. We have:

• developed methods and processes to allow VAT data to be used in the production of the national accounts by December 2017
• launched a new Economic Statistics Centre of Excellence in collaboration with the National Institute for Economic and Social Research
• launched the ONS Data Science Campus at its new Newport home
• launched the ONS Economic Experts Working Group to provide insight and early quality assurance on our latest developments
• delivered the first Economic Statistics Conference at the Celtic Manor in Newport
• created a new London-based economist team to strengthen engagement, while also improving stakeholder engagement around the UK – last year we hosted Economic Forum events in Belfast, Manchester, Cardiff, Edinburgh, York and Birmingham
• redeveloped our publication model to ensure clearer commentary on economic statistics through ”theme days” where similar statistics are published simultaneously to give greater coherence and clarity to users
• launched new flash estimates of productivity and public sector productivity
• published experimental statistics showing country and regional public sector finances

Main issues

The following themes represent the main issues for us to address in the future evolution, measurement and understanding of economic statistics:

1. the modern economy and the national accounts
2. trade and international statistics
3. devolved, regional and local statistics
4. productivity and the supply of labour and capital
5. prices
6. beyond GDP – broader measures of welfare and activity

Future priorities

For the future, the main priorities of our strategy for economic statistics and analysis up to 2021 are in the following areas:

• improving the national accounts and other economic statistics, in line with international reporting standards and the ambition set out in the Bean Review, particularly improving the measurement of gross domestic product (GDP)
• delivering the improvements to reflect the modern economy, in particular investigating means of improving the measurement of services sector activity
• keeping abreast of measurement issues resulting from the rapid pace of change in the economy, including the digital economy, globalisation and the role of intangible assets
• improvement to measuring consumer prices
• development of comprehensive flow of funds statistics to improve coverage, quality and granularity of financial statistics data in the national accounts
• meeting the growing demand for statistics at regional and other subnational geographies
• improving the quality and scope of productivity statistics to deliver a world-class set of statistics to support users attempting to address the “productivity puzzle”
• continuing development of broader measures of social welfare to complement GDP figures, including the regular production of household satellite accounts, development of estimates of natural capital and more timely information on the distribution of incomes
• improving and extending the provision of anonymised microdata to users through the Virtual Microdata Laboratory (VML)
• engaging with academic experts, both in the UK and internationally, through the Economic Statistics Centre of Excellence, the ONS Data Science Campus and the ONS Economic Experts Working Group to enable further improvements to economic statistics

Implementation of these improvements and this strategy have to run alongside the continuing production of a large volume of important regular outputs that are published on a monthly, quarterly and annual basis. We will continue to ensure that these are produced on time and to high quality, that they are clearly explained and that they meet evolving user needs. We aim to deliver greater insight into the economic statistics we produce and to ensure that the economic story behind these statistics is drawn out accurately and clearly.

We published the first Economic Statistics and Analysis Strategy (ESAS) for consultation in May 2016 and in a finalised form in September of that year. We are committed to review and annually update this strategy to reflect changing needs, priorities and availability of resources, to give a clear prioritisation for our development of economic statistics.

This revision aims to continue to make our agenda more transparent to users, stakeholders and experts by providing updates, in relation to economic statistics, on:

• recent developments and improvements
• planned further developments
• concluded research and findings
• commissioned research
• future research commissioning plans

The document will highlight the main priorities of our strategy for economic statistics and analysis to 2021, an outline of progress since last year and provide an overview of how we are delivering the strategy. The aim is to make explicit our assessment of current priorities to allow greater scrutiny and assurance that these are the right ones. ESAS also lays out the research and development priorities to make it easier for external experts to see the areas where we would be particularly keen to collaborate.

We would like to take this opportunity to thank all users of our economic statistics and other parties, including those in government, academia and the private sector who participated in this consultation by providing comments. We have taken these comments on board as best we can.

The following themes represent our main issues and priorities for the future evolution, measurement and understanding of economic statistics:

1. the modern economy and the national accounts
2. trade and international statistics
3. devolved, regional and local statistics
4. productivity and the supply of labour and capital
5. prices
6. beyond GDP – broader measures of welfare and activity

How we are delivering the strategy

We are making significant improvements to our technology, methods, data sources and processes. The Bean Review recognised this was important, but also noted that critical to delivery was the need for us to enhance our economic and analytical capability, both in terms of in-house skills, but also through engagement with external experts.

Improving in-house skills will be achieved by significantly increased recruitment of high-quality analytical staff – and this process is under way – but also by raising the skills of existing staff. Since March 2016, Economic Statistics Group has increased the share of its workforce who have been accredited as part of one of the government’s four analytical professions from 29% to 34%, with more staff awaiting accreditation.

Increased engagement and active collaboration with external experts, both domestic and international, is another critical part of delivering our strategy. Open and effective engagement with all our stakeholders, including the users of our statistics and analysis, is a pre-requisite for understanding needs and deciding how best to meet them. Greater partnership working with the research community and our stakeholders will ensure that collaborative working can be carried out more effectively, drawing on more knowledge and skills than can be provided from within our organisation alone.

We are working more closely with a wider range of individuals and organisations that can help to deliver our strategy. This involves substantially enhanced partnerships with outside academics and other experts through a number of different mechanisms. This includes the annual recruitment of ONS Fellows and ONS Economic Experts, who together form the Economic Experts Working Group and the establishment of the Economic Statistics Centre of Excellence through a consortia led by the National Institute of Economic and Social Research (NIESR). The launch of the new ONS Data Science Campus will greatly strengthen our capability to use new data sources, delivering skills and expertise to augment our existing economics and national accounts knowledge bases. We are prioritising using these skills to analyse and use VAT and trade administrative data, but also to identify data sources that may be available to us but which we are not currently exploiting.

We have been proactive in taking forward the legislative agenda to give us access to modern data sources. We are seeking, through the current Digital Economy Act, to gain access to data sources from three important sectors:

• other government departments
• regulatory bodies
• businesses and other private sector organisations

These new data sources will benefit us by allowing the replacement of existing sources with more comprehensive and more detailed datasets. Technical and other advances open up possibilities of compiling economic statistics more efficiently, while offering higher quality by virtue of using a wider range of data sources, including administrative data or “big data” sources. It is crucially important for us to pursue such possibilities, although ensuring appropriate confidentiality is and will always remain a central consideration.

Additional data will also lead to improvements to existing sources by enabling better quality assurance of survey returns and the potential to reshape existing surveys to focus on collecting detail not always available from external datasets. However, the quality of each data source and its relevance for ONS economic statistics, as well as any resulting problems of consistency and coherence, will need to be evaluated on a case-by-case basis.

International collaboration is of increasing importance and we are taking steps to strengthen co-operation with important partners. In some cases, other statistical agencies have made progress from which the UK can learn – developments in flow of funds statistics using administrative data to improve quality that have been accomplished by Portugal and Austria are a case in point. In other instances, it is becoming clear that issues with economic statistics are arising from increased globalisation and can only sensibly be addressed bilaterally or multilaterally. The behaviour of multinational corporations adds to the challenges of reliable measurement of the movement of people and economic activity (for example, digital activities) across international borders.

Finally, this strategy should be considered in the light of the Spending Review settlement announced in November 2015, which laid out the resources available to us during the period financial year ending 2017 to financial year ending 2020 and the subsequent announcements in the Budget in March 2016 following publication of the final report from Professor Sir Charles Bean’s independent Review of UK Economic Statistics.

Future priorities

This strategy outlines the main opportunities for us to deliver improvements to economic statistics. The six themes highlighted each present multiple challenges. In setting our priorities, we have had to make difficult choices to balance the wide variety of demands with the risks of delivering a large change programme.

The main priorities of our strategy for economic statistics and analysis to 2021 are in the following areas.

1. Improve the national accounts and other economic statistics in line with international best practice and meet the ambition set out in the Bean Review. In particular we will improve the measurement of gross domestic product (GDP):

   • addressing identified shortcomings in the system, particularly around trade and construction statistics
   • the introduction of constant price supply and use balancing
   • the adoption of double deflated estimates of gross value added (GVA)
   • the adoption of administrative data sources, particularly VAT and PAYE data
   • the incorporation of data from a new purchases survey
   • improving production systems and processes for the national accounts
2. Deliver the improvements to reflect the modern economy. In particular we must improve the measurement of services sector activity, including through the delivery of a new services enquiry (Servcom).
3. Keep abreast of measurement issues arising from the rapid pace of change in the economy. This includes developments such as the growth in the digital economy, increasing globalisation and the greater role of intangible assets. These areas we will need to engage internationally to understand and shape best practice. In spring 2017, we launched a process to commission research into the stock of intangible assets in the UK.
4. Delivering a project to transform price indices, addressing important issues with our producer price survey collections and continuing to work towards regaining National Statistics status for the Consumer Prices Index including owner occupiers’ housing costs (CPIH) measure. This is alongside meeting other recommendations from Paul Johnson’s 2015 report Consumer Price Statistics: A Review.
5. With an ambitious programme of work, provide comprehensive flow of funds statistics to improve coverage, quality and granularity of financial statistics data in the national accounts (including fuller counterparty “from whom to whom” information). For example, we are aiming to deliver new pension statistics by the end of 2017.
6. Meet the growing demand for statistics at regional and other subnational geographies. This year we aim to implement balanced GVA measures by region.
7. We will improve the quality and scope of productivity statistics to deliver a world-class set of statistics to support users attempting to address the “productivity puzzle”. This year we aim to launch:
   • quarterly multi-factor productivity estimates
   • new capital productivity statistics
   • new regional labour metrics, as used in our growth accounting framework
   • new experimental infrastructure estimates

8. We will continue development of broader measures of social welfare to complement GDP figures, including the regular production of household satellite accounts, development of estimates of natural capital and more timely information on the distribution of incomes, including a streamlined quarterly well-being publication.

9. We will extend the provision of microdata to users by providing access to economic microdata, anonymised as appropriate, through the new Annual Respondent Database (ARDx), available through our Virtual Microdata Laboratory (VML) secure environment, or via the UK Data Service’s Secure Data Service.

Implementation of these improvements and this strategy naturally has to run alongside the continuing production of a large volume of important regular outputs that are published on a monthly, quarterly and annual basis. We will continue to ensure that these are produced on time and to high quality, that they are clearly understood and explained, and that they meet evolving user needs. We aim to deliver greater insight into the economic statistics we produce, and to ensure that the economic story behind these statistics is drawn out accurately and clearly.

Since the publication of the Economic Statistics and Analysis Strategy (ESAS) in September 2016, we have made progress against the themes identified, as summarised in this section.

The modern economy and the national accounts

New data

In the last year, we have been working to better understand VAT data as a prioritised administrative data source for implementation into the national accounts. We have published a number of articles highlighting the results of the analysis we have carried out on the VAT data and these will continue to be published as we work towards using VAT data within the national accounts by December 2017.

We plan to utilise VAT data to complement the Monthly Business Survey (MBS) data. VAT data will be used for smaller, less complex firms in addition to the MBS, which will continue to be used for larger firms. This will dramatically increase coverage of smaller firms, providing both greater accuracy and the potential for finer granularity in the statistics we publish.

We have started to explore Pay As You Earn (PAYE) data for inclusion in the national accounts using the arrangements available under current legislation. The provision of this and other additional data may change how we balance between the three estimates if, for example, it permits us to produce better-quality estimates of the income measure of gross domestic product (GDP) earlier in the cycle.

New concepts

Given the rapid growth of the sharing economy, capturing this sector in official statistics is increasingly important. We have undertaken a feasibility study to explore ways to measure output of this sector through the creation of standalone statistics. We are also taking forward further work including exploring new surveys, administrative data and big data, with the first estimates relating to individuals’ use of sharing economy platforms to arrange accommodation or transport services due to be published in August 2017.

We are looking to obtain access to data from sharing economy businesses either directly from their accounts, or their websites via application programming interfaces (APIs) or web scraping. We are also investigating other administrative data sources, such as self-assessment tax returns from Her Majesty’s Revenue and Customs (HMRC), which could be used to inform sharing economy estimates. We are committed to working closely with sharing economy businesses in order to interpret these data sources and better understand their business models and structures.

Improved processes

We continue to develop a new supply use framework, incorporating double deflation and automated balancing in constant prices. A long-term aim for the UK National Accounts has been to develop annual, volume-based, balanced supply and use tables (SUT) using previous year’s prices (PYP) as a route to balancing different estimates of GDP in constant prices. This is considered international good practice and is supported by the National Statistics Quality Review (NSQR), as well as being a requirement under the European System of Accounts 2010: ESA 2010. We continue to aim to deliver an interim solution for Blue Book 2018.

New publishing model

In January 2017 we improved the way economic statistics are published to provide a more coherent and informed understanding of the UK economy. This included publishing related statistics together (often making data available earlier) and improving the economic narrative that accompanies the data. At the time we committed to looking at whether we could make further improvements by bringing forward the publication of the Index of Services (almost 80% of the output of the economy) by around 2 weeks.

This work has now completed and concluded that we could add the IoS to the short-term economic statistics theme day. However, it would not make sense to announce or implement this without also making changes to the GDP publishing model more generally as 99% of the monthly output of the economy will be available mid-month under the IoS change. Following extensive work to develop a new model we are going to consult on a proposed new GDP publishing model. We will launch the consultation on 13 July at the Economic Forum and close it during September; with a formal response by ONS during October. If the new GDP publishing model is implemented it would be in place for August 2018.

Regaining National Statistics status

We are focused on regaining the National Statistics designation for construction statistics. We have consulted with the Construction Statistics Steering Group to determine the best method available to us in measuring construction prices. Improvements to the current method will be introduced in Blue Book 2017 in September and an article detailing the improvements and the impact on outputs will be published alongside this.

New measures

Work on improving our measurement through enhanced financial accounts (flow of funds) is continuing. This includes the publication of updated experimental flow of funds balance sheet statistics, consistent with Blue Book 2016, alongside an article on the concept of “enhanced financial accounts”, financial interconnectedness and the UK financial sector in an international context. We also published an article in January 2017 explaining improvements made to the estimation of Defined Contribution pensions and one on understanding the treatment of risk and the calculation and application of financial intermediation services indirectly measured (FISIM) in March.

We have been looking to identify new data sources to underpin the improvements in measuring the flow of funds and have now developed a deeper knowledge of the potential of regulatory data held by the Prudential Regulatory Authority (PRA) for statistical purposes through secondments to important partners, such as the Bank of England. We have also been engaging extensively with a number of potential data suppliers to acquire data and currently are looking to secure data covering consumer and commercial loans and securities and other equity. We anticipate that new data will be available for analysis in the summer.

Improving deflators

We are working to improve and update the price deflators that divide nominal totals (such as turnover) into their price and volume components. Problems with the measurement of prices, or in their application to the national accounts, have the potential to produce corresponding inaccuracies in the measurement of the volume of GDP growth. This important component of the Bean agenda is being tackled through a number of initiatives. These include delivering improvements to the data collection for services sector producer price indices (SPPIs) to feed into the producer price statistics. This will deliver a refined data collection, which will reflect better the changing economy.

Further enhancements of existing deflators are being carried out to support the restoration of the National Statistics badge to construction and trade statistics by providing more reliable estimates of volume.

We continue to work actively with external experts on how to reflect the changes in the modern digital economy in deflators in the telecommunications sector. Other areas of investigation also include looking into new quality adjustments for public services and exploring how to incorporate these into measures of productivity.

Trade and international statistics

In the light of the result of the referendum on leaving the EU, we have significantly expanded the resources devoted to addressing the required improvements in trade statistics. We have published a revised UK trade development plan, undertaken research into the common international measurement issue of trade asymmetries and delivered improvements to the processes and systems used to produce the trade statistics. We have doubled the sample size of International Trade in Services Survey from the first quarter of 2017 to reinforce the granular breakdown of trade in services statistics. The resulting greater detail will be published by September 2017.

In addition, we are continuing our work towards delivering increased frequency and detail of trade in services. Potentially incorporating data from two new sources from HM Revenue and Customs (HMRC) – the EC Sales List, which is a timely record of services traded with the EU, and the new information on cross-border e-commerce.

Devolved, regional and local statistics

We have prioritised the delivery of a balanced measure of regional gross value added (GVA), combining income and production measures, by the end of 2017. This will underpin the dissemination of greater industrial detail¹ at the country and regional² and sub-regional³ levels in both nominal and real terms, and greater detail of the nominal components of income down to the local area level⁴. The emerging results of this work will be published for user consultation in the summer of 2017, with a view to full publication in December 2017.

We have also produced and published estimates of GVA and gross disposable household income (GDHI) for local authority districts in England, as a preliminary output from our development of flexible geographic areas. We updated and expanded the GVA estimates to cover local authority areas in the other countries of the UK in March 2017 and did the same for GDHI estimates in May 2017. Both datasets are now available on the Nomis website, which allows aggregation to other geographic areas such as combined authorities and local enterprise partnerships (LEPs).

In October 2016, we published a feasibility study into the development of regional household final consumption expenditure (HFCE). We aim to develop and publish estimates for the country and regional(2) level by the summer of 2018. If the available data can provide sufficient coverage we will also provide figures for smaller areas, but this may take a little longer.

In May 2017, we published the first experimental statistics on country and regional public sector finances. The publication looked at what public sector expenditure has occurred, for the benefit of residents or enterprises, in each country or region of the UK and what public sector revenues have been raised in each country or region. The analysis also presented the net fiscal balance, which is the difference between total expenditure and total revenue for each of the geographical areas. Work to improve these statistics will continue, subject to the outcome of a public consultation on the results so far.

We have also worked to improve our stakeholder engagement around the UK. Last year we hosted Economic Forum events in Belfast, Manchester, Cardiff, Edinburgh, York and Birmingham. In the autumn, we have plans to go to Belfast, Birmingham, Glasgow, Cardiff and Bristol.

Productivity and the supply of labour and capital

Productivity statistics in the UK are going through a period of fundamental change. In the last year we have embedded a new flash estimate of labour productivity, delivered over 40 days earlier, every quarter, as well as a new flash estimate of public service productivity, which now delivers estimates at a lag of one quarter, as opposed to the previous lag of 2 years.

We are continuing to progress a number of initiatives, including the development of new capital productivity estimates and looking to deliver the Bean Review recommendation of quarterly multi-factor productivity estimates. Other developments are looking at providing productivity estimates by industry by region, enhancing historical productivity series building on the Bank of England “three centuries of data” project, to enhance analysis over the longer-term and new infrastructure estimates. We have also undertaken a pilot survey of management practices in the manufacturing sector alongside collecting new data on management practices more widely. Initial findings have been published and a larger survey of manufacturing and services firms is planned, in collaboration with the Economic Statistics Centre of Excellence for 2017.

To support our productivity work we have developed new microdata resources, which will be available to academics through the Virtual Microdata Laboratory (VML) and Secure Data Service (SDS) imminently. We are also investigating publishing our production code alongside the statistics to enable users to generate consistent productivity estimates for different sectoral aggregations or geographies.

In January 2017, we published the first “theme day” on productivity, covering a range of topics such as labour productivity, regional and local measures, public sector productivity and the measurement of capital services for growth accounting purposes. It also looked at the impact of management skills in the manufacturing sector as part of our agenda to identify both to what extent measurement issues might have a bearing on apparent experience and to advance the understanding of what underlies recent experience.

Productivity measurement is dependent on the measurement of the main inputs, labour and capital, alongside GVA output, which is delivered through the national accounts. We have further developed and quality assured our methods for producing experimental single month Labour Force Survey (LFS) estimates. We plan to publish these estimates during 2017. This work uses a modelling approach to account for survey design effects. The main aim of producing these statistics is to have a better and more timely indication of the latest movements in the labour market than that conveyed by the current rolling quarterly estimates.

During 2016, we increased the range of data regularly published relating to emerging policy areas, for example, children in long-term workless households and NEETs – those not in employment, education or training – who have never worked.

We have also delivered a new gross fixed capital formation system, which delivers more reliable estimates, enhancing an important input into productivity analysis. Alongside this we are redeveloping the capital stocks system to automatically deliver capital services estimates to enable quarterly multi-factor productivity estimates.

Prices

We have announced the decision to move to the use of the Consumer Prices Index including owner occupiers’ housing costs (CPIH) as the lead measure in our publications on consumer price inflation; as part of a smaller suite of inflation statistics, removing some lesser used RPI-related series from our publications. This is as part of a clear future approach to measuring the changing prices and costs faced by consumers and households:

• the CPI and CPIH, which are measures of inflation based on economic principles
• new “Household Costs Indices” (HCIs), which will look at how different household groups experience changing costs through the payments they make, against a benchmark aggregate index; this is currently under development, with plans to produce prototype indices, initially on an annual basis, by the end of 2017
• the Retail Prices Index (RPI) as a legacy measure, which is needed for some existing contracts – we are clear, however, that the use of the RPI is to be discouraged

We continue to address the recommendations raised in the UK Statistics Authority Assessment Report on CPIH and to take the actions needed for it to regain its National Statistics status. The focus is now on strengthening the quality assurance of additional data sources and setting out more clearly our future approach to measuring the changing prices and costs faced by consumers and households.

The Bean Review identified the use of alternative data sources as a priority. We have made good progress in this area through the development of web-scraping techniques and subsequent publication of preliminary price indices using these data. In addition, research indices using an experimental approach to creating price indices from alternative big data sources, clustering large datasets into price indices (CLIP), was published in November 2016.

We will continue to ensure compliance with European regulations for the production of inflation estimates. The UK Consumer Prices Index (CPI) is also the UK’s Harmonised Index of Consumer Prices (HICP), a measure of inflation produced by all EU member states. To ensure compliance in a number of areas, we introduced a lower level of aggregation (called COICOP5) and addressed compliance issues related to the use of a double chain link in the production of CPI and CPIH in March 2017. A temporal sample for certain food items will be implemented in early 2018. We are also researching different approaches to measuring the rate of inflation from airfares and package holidays.

Beyond GDP – broader measures of welfare and activity

We have taken forward a number of developments to progress the Beyond GDP agenda. These include publishing experimental “nowcast” estimates for indicators of the distribution of household income for financial year ending 2016 and contributing to the ongoing work of Eurostat, Organisation for Economic Co-operation and Development (OECD) and other international organisations on development of statistical standards for income, consumption and wealth microdata. We have also identified priority areas for better reconciliation between micro and macro estimates of income and consumption in order to improve experimental estimates of distribution of national accounts aggregates.

We have also engaged with interested parties on opportunities for improving the coherence and quality of our household financial surveys or data and outputs. It is intended that this will lead into work to integrate survey sources as well as make use of administrative data when it becomes available.

Notes for: Latest progress

1. Where possible at the 2-digit SIC level.
2. The 12 NUTS1 regions.
3. The 40 NUTS2 areas.
4. The 173 NUTS3 areas.

6.1 The modern economy and the national accounts

The main issues

The Bean Review identified several challenges to the robust measurement of the UK economy, reflecting conceptual complexity as well as the large number of independent data sources required. Some of these challenges are long-standing whilst other, new challenges continue to emerge. These include the following.

How best to capture new data sources, which bring additional information to bear, particularly administrative data, such as VAT.

How best to balance the information from the three different estimates of gross domestic product (GDP) – income, output and expenditure – at different points in the compilation cycle in order to reflect the relative content and quality of the available data sources.

The absence of double deflated estimates of value added – taking account of differential movements in input and output prices – which is a potential source of measurement error of UK GDP (and therefore productivity).

The absence of a recent Purchases Survey means that changes in businesses’ consumption of goods and services have not been captured over the last decade or more. This may have led to biases in the weighting used to derive estimates of intermediate consumption and therefore estimates of sectoral gross value added (GVA).

Supply and use balancing in constant prices is recognised good practice and users have for some time sought the additional rigour that its introduction in the UK would bring to GDP estimates – including, importantly, the GDP deflator.

Users have been clear about the need for longer runs of consistent time series for the main economic statistics and it is clear from our consultation that historic trends remain a high priority for users.

In 2016, around 80% of total UK output was accounted for by services activities, compared with less than 50% in the 1950s. Whilst the UK is probably ahead of many advanced economies in its measurement of services, particularly in the short-term estimates, there are recognised weaknesses. Economic statistics are still focused more towards things that are easier to measure, such as the output of manufactured products, especially in terms of the degree of granularity of the data. This has been recognised by calls to provide finer breakdowns of services activities to allow better-quality analysis.

As identified by the Bean Review, the growing impact of digital activities in enabling the provision of services, including overseas trade in services, can only exacerbate existing measurement challenges. In essence this requires us working with international bodies to identify how best to measure the digital marketplace alongside the long-standing physical one in an internationally comparable way.

Users are clear about the need for enhanced financial statistics¹. By providing information on the financial inter-connectedness of the economy, the new data could support decision-making in both the UK and international institutions, on matters relating to financial stability and to minimise the risks of a new financial crisis. The work should provide a greater degree of granularity than previously possible, providing information on counterparties to transactions as well as additional disaggregation of financial sectors and instruments. Such information should also improve the quality of the sector and financial accounts, including balance sheet information, thereby aiding understanding of the behaviour of the economy for other purposes.

Value added is increasingly generated not through the production of physical goods or even of services as traditionally understood but through the production of knowledge of intellectual capital and other forms of “intangibles”. The recent capitalisation of research and development spending within the national accounts is a step towards recognition of this issue. But this is only one of a number of steps necessary to move towards adequate measurement.

The Bean Review identified three consequences of the digital revolution, driven by the rapid pace of innovation and technological change in the UK, which have important consequences for measurement of the economy.

Technologically-induced disintermediation, the resulting fall in transaction costs and unconventional business models are blurring the boundaries between work, domestic activity (home production) and leisure – the nature of the resulting transactions means that they may not be picked up in conventional approaches to measurement of the economy, or that it is difficult to observe market prices at which they take place.

The rapid pace of change creates difficulties in accounting for quality changes – for instance, in judging how to cater for replacement of products in price indices.

Classification systems of economic activity, for instance, the Standard Industrial Classification: SIC that is used to define industry groupings, are typically updated only periodically. They therefore cannot keep up with rapid changes in the structure of the economy and they may not provide the granularity necessary to cater for the changing needs of users. This is a long-standing issue, but is especially pertinent in the context of rapid changes as a result of the digital revolution. The importance of defining the digital economy, for example, was raised by users during our consultation, and we are working with Organisation for Economic Co-operation and Development (OECD) to progress this on an internationally agreed basis.

The identification of quality changes in many services activities is potentially important but also difficult in times of rapid change. To address this we are exploring how to expand the number and improve the quality of price indices for the construction, distribution and services industries and develop outputs similar to manufacturing Producer Price Indices such as input price indices.

Planned developments

To enable the development of supply and use tables at previous year’s prices, we will improve the statistical and conceptual quality of our price deflators. This will consider the methodology for double deflation, including appropriate quality-adjustment methods to keep pace with quality changes in the economy, including the ever-increasing digital marketplace and the development of deflators for sectors such as construction and banking. In relation to producer prices, specifically, we are improving survey methodology across business prices. This work is scheduled to align to the ongoing work around supply use.

We have collected new purchases data (reference year 2015), which will enable us to update the supply use framework with new weights used to derive estimates of intermediate consumption and therefore estimates of sectoral GVA. The emerging estimates from the first data will be used to help inform Blue Book 2017 and our original intention for the data to be used in full in the compilation of Blue Book 2018.

The 2018 deadline was challenging as we would only have one year of the new purchases data and only limited time to fully reconcile it against other data sources. As we have to plan our Blue Book changes well in advance, we have concluded the 2018 deadline needs to be deferred to allow additional time to understand better the impact of the new data. Deferring full incorporation by a year (the data will continue to be used to inform Blue Book 2018) will allow us to do this and will also mean there is a second year of data to provide additional evidence. In addition, there are also a number of other improved data sources that will impact on supply use that will become available in 2019.

We have received feedback from users that they prefer us to package such improvements in a way, as far as we can, to single points of implementation so a full as possible evidence base is used to inform the balancing of the economy and to avoid revisions in opposite directions from one year to the next. To help users, though, we plan to publish data from the purchases survey sometime in 2018.

To further improve the quality and range of services industry and product-related statistics, which feed into GDP and productivity estimates, we have prioritised a new Servcom survey. This survey is analogous to the Prodcom survey for the manufacturing industries and will provide a better-quality breakdown of businesses’ output by industrial classification. Servcom will cover the whole services sector (excluding banks and government) – 274 industries with a sample of 40,000 – subsuming the existing Services Turnover Survey (STS), which covers 93 services industries with a sample of 20,000. This will improve the quality of product breakdowns in the supply and use tables, and in turn the quality of balancing and GDP, as well as improving the weights and sample for the Services Producer Prices Index. We have identified a contractor to collect data for 2016 and plan to use the results of the survey in Blue Book 2019.

We have worked to incorporate new data sources to offer a much richer disaggregation of services output, for example, on education-related travel from new sources: for higher education and further education tuition fees from the Higher Education Statistics Agency and for independent school fees from the Independent Schools Council. We aim to implement this in the Blue Book 2017.

We have introduced new questions on the sharing economy in both the Internet Access Survey for households and individuals and the E-commerce Survey for businesses. These will provide data on the proportion of adults and the proportion of businesses that have arranged accommodation or transport services via platforms such as Airbnb, SpareRoom, Uber and Lyft. We are sampling many of these businesses in at least one of our surveys for the next available reference period to provide extra information, which will greatly improve our understanding of this sector. This will include the characteristics of such businesses and enable us to identify further sharing economy platforms, both manually and where possible by applying data science techniques to the Inter-Departmental Business Register (IDBR).

We continue to work to redesignate construction statistics as National Statistics. We aim to achieve this by the end of 2017.

For the flow of funds, over the next 5 years the main objective will be to improve the quality, coverage and detail of the UK’s financial accounts, including the development of improved counterparty information. Our objectives in these areas include:

• understanding existing technological solutions for processing large datasets, including those run by non-government sources
• publication of experimental statistics and working papers as new data become available
• commence development of a securities database (subject to a successful pilot and funding)
• deliver the International Monetary Fund’s (IMF’s) Special Data Dissemination Standard Plus (SDDS+) and G20 Data Gaps Initiative requirements in full for the UK (by 2021)
• integration of new data sources into the national accounts
• publication of full counterparty information, planned for 2019 on an experimental basis and for full implementation into the national accounts in 2021
• improved coverage of priority areas including other financial intermediaries (OFIs), debt securities, equity and investment fund shares or units

Concluded research and findings

In 2016, we commissioned various projects.

There was a joint project with the Bank of England, in collaboration with interested academics, to reconstruct historical data on the sector and financial accounts and balance sheets in the UK. Time series now run back to the 1920s for some series.

Deloitte were commissioned to undertake a review of mixed income compilation methods. The project to date has identified some data sources that could provide data on a limited range of self-employed industries or occupations but no sources that could be used as a proxy for mixed income. Further work will be undertaken in 2017 to identify further data sources.

We recruited an independent contractor and sector expert in relation to classification of public private partnership. This relationship has delivered high quality advice and guidance to the classifications team.

Deloitte were commissioned and completed three pieces of work covering the financial sector:

• an assessment of the size of the money market funds sector – Deloitte confirmed that this sector is small in the UK
• to validate the estimated service charge for defined benefit pension schemes – Deloitte validated the current estimate of 1% used by ONS
• to provide an assessment of available data sources for flow of funds – Deloitte’s findings validated our research undertaken in this area

We have been working with Professor Diane Coyle, one of our ONS Fellows, on better understanding and mapping the digital economy, particularly in terms of households’ usage of new services and the potential impact on the market economy. A discussion paper will be published by the Economic Statistics Centre of Excellence (ESCoE) in summer 2017, which Professor Coyle is presenting at a Royal Economic Society Conference session jointly organised by us and ESCoE.

We also commissioned Doctor Rebecca Riley of the National Institute of Economic and Social Research (NIESR) to undertake peer review and further analysis of the new ARDx dataset, primarily the Perpetual Inventory Model (PIM) model inherent in this, which generates firm-level capital estimates.

Commissioned research

We have established the Economic Statistics Centre of Excellence (ESCoE) with a consortia led by NIESR. This multi-year project will deliver a significant component of our ongoing research programme. In particular, in relation to the measurement of GDP, ESCoE has been commissioned to deliver research in various areas.

Historical national accounts data – this project will continue the work ONS and others have undertaken in recent years to build consistent time series of important metrics in the UK’s sector and financial accounts beyond 1997. This project will explore the extent to which historical national accounts can be revised to be compatible with modern definitions.

We aim to make available balanced historical data on the components of GDP and on sectoral income and expenditure to users through our website. This addresses an important request from users to provide longer runs of consistent time series for the main economic statistics. This builds on the historic data published with Blue Book 2016 and will give us the capability to match the standards of provision of other statistical offices such as those in the Netherlands and the United States.

Measuring GDP at different time horizons – this project will research the use of administrative data to deliver estimates of the three measures of GDP at different time horizons, including exploring nowcasting VAT data to generate estimates of firm turnover and gross value added (GVA) at Month 1. An important focus for this study is whether this data enhances our ability to identify economic turning points in a timely manner.

Democratic measures of income growth – this project will research different measures of income growth and how this is perceived, taking into account different household characteristics.

Measuring activity in services sectors – the Bean Review highlighted the need for better measures of services activities given their increasing importance as a share of value added and employment. Also, much of the increased productivity growth associated with the information technology revolution occurred in services sectors. The overarching research question to be addressed is: what are the deficiencies in the current measures of services activities for the UK and how might they be improved? Specifically the research will address the following questions:

• how to define and value the activities provided by the services sector in measuring nominal output
• after a fundamental review of how deflators should work to be relevant in the modern economy, identify the most appropriate deflators to use to construct real output measures and how might these be adjusted for quality change
• what additional measures can be used in measuring real output, such as volumes of activity, in the public sector and how might these be adjusted for quality
• what sources of data might be employed (both existing and new) to estimate outputs and prices in the services sectors

Measurement issues in the modern economy – this project covers the classification of digital activities and how these could or should map into an evolution of Standard Occupational and Standard Industrial Classification codes for the services sector. This includes considering the evolution of digital business models and services sector productivity.

The evolution of digital business “as a service” models – are there new categories of service not currently recorded statistically? If so, how could these be captured and what are the balance of trade considerations? Increasingly customised manufacturing is often bundled with services. This bundling is frequently observed in the digital or high-tech sector as a business model choice, for example, smartphone handsets with a service contract.

Services sector productivity – the processes by which “inputs” are turned into “outputs” can be more complex than is the case in manufacturing. This is because the “process” may be part of the service output (in customer experience services, for example, or in any bespoke service where the output is co-designed with the customer). In some other cases, efficiency will be reflected by reductions in time taken to undertake the process to a given level of quality. Critical to these processes is also the role of intangible assets, which are poorly measured.

Future research commissioning plans

We have identified the following research priorities for 2017, over and above those described previously:

• historical reconstruction of household pension assets statistics
• exploitation of a new information portal on micro-businesses (that is, those too small to be registered for VAT or PAYE) to facilitate calibration of adjustments for under-reporting
• ONS, in collaboration with the Institution of Engineering and Technology, OFCOM and academics have been undertaking a review of quality adjustment in the telecommunications industry, initial findings were published in the October productivity bulletin; we have commenced the second stage of this work to identify data sources that could be used to deliver potential improvements to deflators used in the telecommunications industry
• further exploration of potential data sources to improve the quality and coverage of the UK Financial Accounts by continuing to seek administrative, regulatory and commercial microdata (that is, record-level) sources, to replace and enhance existing survey data
• identification of new sources of information on services sector activities
• further research into improved methods and data sources for short-term statistics on services output

We have taken on responsibility for a publication previously released by the Intellectual Property Office, which estimates the value of intangible assets in the UK, using primarily ONS data. We will re-produce this publication this year, refreshing the main assumptions to inform our future production of this report and the national accounts, specifically around artistic originals. This will provide a means to consider the measurement of intangible capital-generating activities not already incorporated into the national accounts, including the role of management and organisational capital in firm and economy performance and productivity, through a detailed experimental breakdown of the “other” category.

A review of how best to implement the new Frascati manual in order to improve estimates of research and development (R&D) spending.

Analysis of how to deliver a long-term software deflator to replace the current interim approach.

6.2 Trade and international statistics

The main issues

In November 2014, UK trade statistics were de-designated as National Statistics. In May 2015, the UK Statistics Authority published 13 requirements that needed to be met for the statistics to be redesignated. In response to this, as outlined in the 2016 Economic Statistics and Analysis Strategy (ESAS), we published the UK Trade Development Plan in March 2016, which set out a range of priorities for trade statistics to meet the Authority’s requirements, but also to future-proof our model for the period 2016 to 2019.

The main aspects of the plan covered improved quality assurance, more analysis of the underlying volatility in trade data and potential improvements to production systems to capture new data sources. Building on this base, we published experimental estimates of exports of services by region and country of the UK, supplementing existing HM Revenue and Customs (HMRC) information on trade in goods by region.

In June 2016, the UK voted to leave the EU, leading to a step-change in the user requirements in this area. As such, we have reviewed the trade development plan and published a refreshed version in February 2017. These new user needs mean that we have had to further review our production system and process to meet the range of statistics required to meet these new user demands.

Planned developments

The February 2017 trade development plan identifies the main steps we have taken to meet user needs, which are:

• undertake design work to ensure new systems and processes meet new operational and user requirements
• increase our analytical capability, both in-house but also through the Economic Statistics Centres of Excellence, to investigate existing and new data sources for innovative uses
• investigate if there are methodological or data issues, which result in co-movement of import and export deflators in the UK and other developed countries
• look to make further use of HMRC EU sales data to improve estimates of trade in services

Concluded research and findings

We commissioned Dr Thomas Baranga (Harvard Department of Economics) to undertake a study of asymmetries between UK and foreign National Statistics Institutes (NSIs) of matching export and import statistics to better understand reasons for any difference between these. This research, which will be published via the Economics Statistics Centre of Excellence in spring or summer 2017, has proposed a methodology for applying adjustments to trade statistics to enhance accuracy.

Commissioned research

As part of the Economics Statistics Centre of Excellence, we have commissioned the following research.

Granularity in trade in value-added data for the main sectors – this research will build on recent work to link a range of data sources to enable the analysis of global trade, particularly contributing to the international debate about the extent of trade in value added (TiVA). This is where value is added by organisations in one country combining parts in a production process that have largely been sourced (imported) from another country and then exporting the combined product at a higher value than the component parts.

Future research commissioning plans

We are investigating the merits of cognitive testing of trade prices survey questions to better understand the denomination of transactions and their reaction to movements in the exchange rates. This is through hedging and other forward contracts to understand the degree to which they mitigate or delay the impact of changes in exchange rates.

Longer-term research priorities include identifying methods to deliver more granular measurement of foreign direct investment, trade statistics, by sector, product and industry and to deliver a better understanding of the impact of global supply chains on trade and GDP statistics.

6.3. Devolved, regional and local statistics

The main issues

The underlying rationale for what we do, including publication of our economic statistics, is that they help decision-makers make better decisions. If they are to continue to fulfil that role, such statistics must keep pace with changes in the requirements of decision-makers. Different users will have different needs, which can only be met by providing information at below the aggregate level.

Improved and expanded devolved, regional and local data will be needed to underpin public debate and decision-making in the context of devolution, decentralisation, local growth (including the new city-regions) and matters such as large infrastructure projects. With any move towards more devolved powers to subnational administrations, there is a need to provide an evidence base upon which policy decisions can be made at that level.

We have always provided information of this kind, although often in a less timely and comprehensive manner than national UK data. We will need to meet changing priority needs by improving granularity, quality, timeliness and flexibility of local area statistics.

There are a number of ways that existing subnational economic statistics can be enhanced. First, there is the potential to increase the range of data available; second, to improve significantly the timeliness of important data and third, to increase the flexibility with which outputs can be used to respond to evolving geographies and policy developments.

Producing more timely statistics at a more disaggregated geographical level is potentially resource intensive. Traditionally many of our surveys are scaled to produce optimum sample sizes for UK figures in aggregate and would therefore be too small to produce viable estimates at more local levels. In addition, local statistics have a wide range of potential uses and typically different users will want different breakdowns to suit their purposes. However, the possibilities opened up by better exploitation of available administrative data in this area are considerable.

Planned developments

We have mapped out the following next steps.

Following the feasibility study we will be developing annual regional estimates of household final consumption expenditure and the household saving ratio, thereby extending the information on the economic impact of households in different parts of the UK (aiming for a first publication in summer 2018).

To address the timeliness issue, we are developing quarterly output indicators for the English regions (including GDP in volume terms and a detailed industry breakdown). Our aim is to publish these by the end of 2018, when they will complement the quarterly indicators already produced by Scotland, Wales and Northern Ireland to provide UK-wide coverage on a coherent basis.

Regional accounts will provide a mechanism for compiling estimates for non-standard, flexible geographic areas by breaking down existing regional measures (mainly at NUTS3 level) to smaller areas, using indicator datasets, and building back up to customised area specifications. The local authority estimates are the first stage of this development and we aim to break both gross value added (GVA) and gross disposable household income (GDHI) down to even smaller areas during 2018. By 2019, we plan to look into combining the flexible geography work with the quarterly output indicators, with the aim of being able to project forward from the annual estimates to provide timely estimates for any areas of interest. It is clear from our consultation that users see this as a high priority. We also now publish industry by region productivity estimates as part of our Productivity statistical bulletin.

Overarching this is the exploration of the use of administrative and other data sources to improve the quality and timeliness of published estimates.

We will work to improve estimates of country and regional public sector finances, that were first published in May 2017, subject to the current consultation exercise. We will continue to work with devolved administrations to provide them with the underlying data to allow them to produce their own country-specific fiscal statistics (as is currently done with Scotland).

We are working with bodies responsible for administering devolved taxes to ensure that as taxation powers are devolved there is no drop in the timeliness or quality of tax data reported in the monthly Public Sector Finances statistical bulletin.

We have combined estimates of regional and sub-regional productivity, previously published in separate outputs, into a single output and this was published as part of the productivity theme day in January 2017. We have also begun producing supplementary analysis exploring regional productivity issues. Regional firm-level productivity analysis for the non-financial business economy was published in January 2017 and provided information on the sources of productivity differences across regions and city regions. A further article on differences between rural and urban areas was published in April 2017.

We will increase the range and timeliness of published small-area data. We have ensured the small-area model-based income estimates for England and Wales are of sufficient high quality to retain their status as National Statistics, following engagement with and feedback from users. We have also started producing the income estimates on an annual basis and publishing them more quickly, to meet user demand for more frequent and timely household income data.

We are now exploring ways to further improve estimates of household income by assessing the feasibility of using more administrative data in the model-based estimates as well as new methods and data sources to produce estimates of income distribution.

We have continued to publish our house price statistics for small areas, adding new geographies such as combined authorities and counties, following user feedback. We are now using these statistics in additional analytical products such as our housing affordability calculator.

Concluded research and findings

Southampton University was commissioned to explore the potential of existing prices microdata to deliver regional price indices and some users, as part of our consultation, confirmed that they felt subnational inflation measures were important for policy development and implementation. The suitability of this data for the production of regional Consumer Prices Indices has been questioned over time, in particular around whether the amount of prices collected would be sufficient to support the production of robust regional estimates. We plan to publish later in 2017.

Commissioned research

As part of the Economic Statistics Centre of Excellence, we have commissioned the following research:

Regional nowcasting in the UK – this project has four aims:

• to produce and disseminate timely model-based quarterly regional estimates of nominal gross value added (GVA)
• to produce historical quarterly estimates of regional GVA and the importance of longer time series has been confirmed by some users as part of the consultation
• to produce real or volume GVA estimates, using our “experimental” real regional GVA data
• to explore the possible use of underlying micro-level and administrative data to produce model-free quarterly regional output data

Improving the quality of regional economic indicators – this project aims to extend economic statistics coverage to regional level and improve the quality of existing data. The research falls into two parts:

• exploration of how to better capture regional inter-linkages within the UK and to develop new methods for measuring and monitoring inter-regional trade flows; helpfully, users continue to confirm the importance of this work
• investigation of how to further improve the quality and robustness of regional fiscal data in the UK

Future research commissioning plans

These projects will inform future research commissioning, primarily in the following areas:

• constructing more robust regional estimates of GDP in volume terms requires the input of regional price information; greater use of third-party data may offer a way forward for producing regional price estimates
• use of “big data”, administrative data and new statistical techniques to provide either new and better or more granular estimates for cities, regions and other subnational geographies
• understanding the drivers of regional and local productivity
• understanding the occurrence of inequalities and deprivation within regions or cities
• developing methodologies and techniques to allow distributions of outcomes to be compiled and published for important statistics

6.4 Productivity and the supply of labour and capital

Main issues

Almost all of the topics discussed in this strategy document have a bearing on the “productivity puzzle” – the unprecedented and persistent weakness of productivity since 2008. This is an important feature of the UK economy but also, to a lesser degree, of other developed economies. It has implications for a wide range of economic policy – including those relating to living standards, the public finances and monetary policy – and is the focus of considerable academic and policy-maker attention.

We have put in place a rigorous new programme to produce a suite of productivity statistics to compare with any in the world. To this end we now produce a flash estimate of labour productivity 45 days after the end of the relevant quarter. We have consolidated all of our other productivity publications onto a single theme day around 97 days after the end of quarter, shortly after the publication of the quarterly national accounts. In order to support policymakers, we have developed new and more granular experimental estimates of labour productivity on both an industrial and a regional basis and we have moved to use “nowcasting” techniques to produce more frequent and timely experimental estimates of public service productivity. We have established a programme of microdata analysis of firm-level productivity and have commissioned research to support third-party and academic research on this topic, as well as developing new and innovative sources of information in-house to explain recent trends.

While much of this development programme has been deliverable in isolation, the changes outlined previously to the measurement of output in the national accounts clearly have a bearing on our work, as do changes needed to deliver stronger measures of labour and capital inputs. On the labour market side, there have been a number of structural changes, for example, an ageing workforce, a growth in self-employment and non-standard working models, such as the gig economy. We will consider how these statistics could be developed, in particular using new, administrative data from HM Revenue and Customs (HMRC) as it becomes available.

How we capture income and earnings also has an important role to play. In recent years, employee pay growth as measured by average weekly earnings (AWE) has remained subdued and has shown some discrepancies with the annual structural earnings survey (Annual Survey of Hours and Earnings (ASHE)). This may be due to the changing structure of the workforce and we have produced analysis on the impact of sex and job tenure on pay. However, further work on the impact of type of job and qualifications could provide more insight.

Developing new and appropriate measures of capital is also important – including both conventional measures of capital and so-called “missing capitals” – for example, human capital, or natural and environmental capital. These are widely considered to be increasingly important in driving output and productivity growth in a modern economy and good statistical information is essential to underpin public debate about aspects of sustainability.

Planned developments

During 2017, we will:

Continue to develop our suite of labour productivity measures. This will include completing the switch to our new production system; extending the time-series available for several of our new experimental outputs on a detailed industry basis and improving our historical labour productivity metrics. These changes will be complemented by a set of regional labour productivity metrics on a constant price basis when these become available.

Review the current methodology for producing our International Comparisons of Productivity release, exploring the use of new production-side public-private partnership (PPP) estimates in international comparisons of productivity.

Develop a suite of publications on infrastructure statistics. This will include publications on the measurement of investment in infrastructure assets, the stock of these assets and estimates of the services that these provide. The objective of this work is to better understand how infrastructure influences productivity growth – ideally within a growth accounting framework.

Re-develop the multi-factor productivity methodology, accommodating improved inputs – such as more granular estimates of quality-adjusted labour inputs (QALI) and capital services – as well as new inputs, such as natural capital, infrastructure and management capital. We will also examine the feasibility of separating tangible and intangible capital within this framework and look to include any available estimates on non-national accounts intangibles. This will involve the development, as identified in the Bean Review, of a quarterly volume index of capital services (VICS) and quality-adjusted labour indices (QALI). We will look to collaborate with Economic Statistics Centre of Excellence (ESCoE) to conduct further research into intangible assets.

Continue our programme of microdata analysis to explore the productivity puzzle and the longer-term productivity gap. This will examine the levels, growth rates and dispersion of productivity at the firm-level using matched and linked data on a range of different firm-level characteristics. Of central importance to this work will be the exploitation of more administrative data in assessing firm-level trends and the completion of work on firm-level capital stocks attached to the Annual Respondents Database (ARD). Early targets for this work will include the association between foreign direct investment (FDI) and trade status and firm-level productivity.

Put the pilot Management and Expectations Survey (MES) into field to explore the impact of both management practices and firm-level uncertainty on productivity performances. In collaboration with ESCoE, this work will explore how firm level management practices are related to productivity across a wide range of industries and builds on the earlier pilot Management Practices Survey in the manufacturing sector. These new data will be the focus of research over the coming 2 years, but the publication of preliminary results is planned for the Productivity statistical bulletin in early 2018.

Work with colleagues in Organisation for Economic Co-operation and Development (OECD) on their DynEmp and MultiProd programmes to ensure that the UK is represented in these multi-national research efforts.

Build on our earlier work in measuring public service productivity and efficiency, carried out by UK Centre for the Measurement of Government Activity (UKCeMGA) between 2005 and 2011, and reinforce its activities in this domain. This will start with a comprehensive review of methodologies used to produce public service output and productivity. In addition, work will be carried out on new and improved measures of quality and developments of new productivity measures for areas previously measured using the inputs equals output approach particularly in relation to criminal justice, adult social care and education.

Work with ESCoE to understand the impact of quality adjustment, particularly in relation to services and the digital economy.

Work with HMRC on analysing data from the Real Time Information (RTI) tax system pending the implementation of the Digital Economy Bill. Linked to this, ONS will publish a reconciliation of AWE and ASHE and introduce methodological improvements to the AWE.

Investigate how administrative data could be used to supplement our current labour market surveys to provide more granular analysis. This will initially be Pay As You Earn (PAYE) Income Tax records but ONS will also work with the Department for Work and Pensions (DWP) to exploit the data available from the developing Universal Credit system. We are scheduled to receive PAYE data in late 2017 and have begun to identify opportunities to use this data in a timely fashion. A priority for this data will be the construction of an employer-employee linked dataset to facilitate analysis and research.

Concluded research and findings

ONS is concluding work with the University of West of England, King’s College London and National Institute of Economic and Social Research (NIESR) to improve the documentation and metadata labelling of microdata databases held in the Virtual Microdata Laboratory (VML). In particular, ONS commissioned the University of the West of England to review, update and create a new version of the Annual Respondent Database (called ARDx) – a principal dataset for microdata analysis of productivity in the VML. This work has focused on the following areas:

• updating the ARD and ARD Register Panel to include the most recent business and employment surveys and establishing a simplified process for regular and timely annual updates in the future; this involved clarifying data sources for the labour variables currently in the ARD/Inter-Departmental Business Register (IDBR) and better describing these in the metadata
• improving the metadata, labelling and documentation for the ARD, especially for derived variables
• making marginal changes to gross value added (GVA) measures, removing some identified coding errors

The final part of this work is the completion of the capital stocks estimates attached to the ARDX, which are expected to be delivered this year. This will enable robust productivity analyses based on the growth accounting framework. We will explore and encourage the potential uses of the ARDx for use in projects on productivity and other economic issues.

Commissioned research

As part of the Economics Statistics Centre of Excellence, ONS has commissioned the following research.

Measurement issues in the modern economy

This project seeks to address the measurement challenges presented by digital technology and other drivers of structural economic change. The main aims of this project are:

• investigate how new digital business models map into existing statistics and how measurement approaches might evolve
• explore the implications of the digital revolution for the measurement of economic welfare

Measuring activity in services sectors

This project will investigate the deficiencies in the current measures of services activities for the UK and how might they be improved. Specifically this project will:

• define and value the activities provided by the services sector in measuring nominal output
• identify the most appropriate deflators to use to construct real output measures and how these might be adjusted for quality change
• identify additional measures that can be used in measuring real output, such as volumes of activity, in the public sector and how these might be adjusted for quality
• identify which sources of data might be employed to estimate outputs and prices in the services sectors

Sectoral productivity estimates

This project seeks to examine if, and the extent to which, measurement error contributes to the productivity puzzle. This will include revisiting the sector level productivity estimates in the light of advances in measurement and examination of both growth accounting and productivity levels relative to our major competitors.

Developing microdata for the analysis of productivity

This project will develop new micro-level data on firm-level characteristics to explore their role in determining productivity outcomes. In particular, this project has led the commissioning of the Management and Expectations Survey (MES), which will gather information about firm-level management and uncertainty. The project team plan to link the resulting data to existing information from the Annual Business Survey to examine the role of these factors in explaining productivity growth.

Using administrative data to develop new labour force and migration statistics

This project will address a number of research questions relating to labour market dynamics, in particular in relation to the non-UK born population:

• mapping the EU resident population
• short-term and circular migration
• length of migration or return migration
• mapping changing migrant populations at small area level

Using administrative and “big data” to improve labour market statistics

This project aims to explore opportunities offered by government datasets to achieve a fuller understanding of labour market flows, including the potential for combining datasets to allow a more complete observation of individuals’ life cycles to be observed and to enrich the data currently used for analytical purposes. An important part of this project is to understand the implications of problems associated with the use of such datasets and to identify ways of mitigating them.

The project will also:

• explore the potential to supplement the Annual Survey of Hours and Earnings surveys with DWP or HMRC administrative data in order to give a panel dataset and thereby the ability to allow the period between survey interviews to be covered
• explore whether online job ads data can be used to produce improved data on occupations, skills and wages

We have also commissioned a consortia of London Economics and DIW Econ to deliver a review of international productivity statistics to identify best practice in terms of frequency, quality, range and granularity of productivity statistics.

We are aware of the lack of information relating to the self-employed (particularly around timely measures of income) and those participating in other “non-standard” forms of employment (for example, the “gig” and sharing economies) and are looking at how to fill this gap. A project has been commissioned with Deloitte to help take this work forward.

Future research commissioning plans

We will commission a review of the relationship of investment, access to finance and productivity from the Annual Respondent Database (ARDx) in the Virtual Microdata Laboratory (VML).

6.5 Prices

The main issues

Price statistics are an important indicator of how the economy is behaving and this is reflected in the attention attracted by the suite of price indices we publish. The National Statistician, in November 2016, published a statement outlining decisions to shape the future of consumer inflation statistics in the UK following a consultation in 2015.

The consultation followed a comprehensive review of how inflation is measured in the UK by Paul Johnson. Inflation statistics were also considered further in the Bean Review of economic statistics.

Planned developments

Delivering the Transforming Inflation Statistics (TraInS) Project to address the main issues with our producer price survey collections and continuing to work towards re-badging the new Consumer Prices Index including owner occupiers’ housing costs (CPIH) measure as a National Statistic, alongside meeting other recommendations from Paul Johnson’s 2015 report Consumer Price Statistics: A Review.

Whilst good progress has been made in the development of web-scraping techniques (including cleaning and classification) and publication of preliminary price indices, further work is required to secure access to point-of-sale scanner data and to develop our methods and capabilities to utilise these sources of data in a production environment. Further articles using web-scraped data and developing the methodology to be used will be published during the spring and summer of 2017.

We have published a discussion paper and a timetable for the development of Household Costs Indices (HCIs). Feedback has been received on the discussion paper and a summary of responses published. This will culminate in the publication of experimental statistics in late 2017.

Concluded research and findings

Doctor Paul Smith of Southampton University was commissioned to undertake an exploration of the potential of existing prices microdata to deliver regional consumer price indices. A feasibility study will be published by summer 2017.

Commissioned research

We have commissioned a review of the stock of residential housing to inform housing inflation estimates from Professor Huw Dixon at Cardiff University. This research will report in summer 2017.

Future research commissioning plans

We will review quality-adjustment methods used in the production of inflation statistics and provide more detail on how quality-adjustment is monitored. The treatment of quality change was raised in the Johnson Review, the Bean Review and by users in response to the 2015 consultation.

We will research the HCIs further, making use of advice offered by the two advisory panels on consumer price statistics and considering similar work undertaken by other National Statistical Institutes and academics.

We continue to investigate options for improving the measurement of construction output prices, leading to improvements in the current interim indices.

6.6 Beyond GDP – broader measures of welfare and activity

The main issues

While gross domestic product (GDP) is an internationally accepted measure of the size of the economy, its limitations are widely acknowledged. In particular, for a number of reasons it is flawed as a measure of broader social welfare. For instance, headline GDP estimates include the impact of population changes on aggregate economic activity, but they do not capture what this means at the individual or household level. The Bean Review has raised challenges around capture activities where no market transaction takes place, for instance, time spent by households on “production” activities such as childcare and transport.

We have been working with Professor Diane Coyle, one of our ONS Fellows, on better understanding and mapping the digital economy, particularly in terms of household usage of new services and the potential impact on the market economy. A discussion paper will be published by the Economic Statistics Centre of Excellence (ESCoE) in summer 2017, which Professor Coyle is presenting at a Royal Economic Society (RES) Conference session jointly organised by ourselves and ESCoE.

There is, therefore, value in compiling and publishing a wider range of information rather than GDP alone. We now publish a suite of measures to complement the GDP data on a quarterly basis, including GDP per person, net national disposable income per person and information about the distribution of income.

Conventional GDP estimates make no allowance for the depletion of natural resources that may be inherent in many forms of economic activity. We publish regular environmental accounts for the UK and the information contained therein can be developed to form the basis of an estimate of GDP that is adjusted to reflect the consumption of natural resources.

A project is in place to produce innovative ecosystem accounts, for incorporation into the environmental satellite accounts. The accounts value and record changes in the extent and condition of the UK’s natural resources and also the services provided by nature, such as carbon storage and climate regulation. By valuing natural capital, it is possible to measure the gains and losses in nature that would otherwise go unrecorded and thereby support economic decision-making. The accounts aim to provide consistent and comparable datasets, which allow economic and environmental information to work together.

Finally, whilst the national accounts tell the story of movements in aggregate living standards across the whole population, policymakers wish to take decisions on the basis of more granular information, for instance, how the headline changes are distributed between different types of households.

Data are of most use for decision-making if it is timely, rather than being compiled as a document of record well after the event. We will respond to this need by greater use of “nowcasting”, as seen in the recent experimental release of provisional estimates of the distribution of household incomes for 2014 to 2015.

Planned developments

We will update recent household satellite account estimates making use of time-use survey data (from the Centre for Time Use Research) to value time spent on home production (autumn 2016).

Working in partnership with the Department for the Environment, Food and Rural Affairs (Defra) natural capital accounts and supported by the Natural Capital Committee, ONS will look to develop natural capital accounts with the aim of inclusion in the environmental satellite accounts by 2020.

Commissioned research

We have commissioned the Economic Statistics Centre of Excellence (ESCoE) to deliver research into the following.

Democratic measures of income growth – this project will research different measures of income growth and how this is perceived, taking into account different households’ characteristics.

Measurement issues in the modern economy – this project covers the classification of digital activities and how these could or should map into an evolution of Standard Occupation and Standard Industrial Classification codes for the services sector. This includes considering the following.

The evolution of digital business “as a service” models – are there new categories of service not currently recorded statistically? If so, how could these be captured and what are the balance of trade considerations? Increasingly customised manufacturing is often bundled with services. This bundling is frequently observed in the digital or high-tech sector as a business model choice, for example, smartphone handsets with a service contract.

Services sector productivity: the processes by which “inputs” are turned into “outputs” can be more complex than is the case in manufacturing. The “process” may be part of the service output (in customer experience services, for example, or in any bespoke service where the output is co-designed with the customer). In some other cases, efficiency will be reflected by reductions in time taken to undertake the process to a given level of quality. Critical to these processes is also the role of intangible assets, which are poorly measured.

We have also commissioned a review of discount rates used in various areas of economic statistics to ensure coherency between different rates used in different parts of the national and satellite accounts. We have also commissioned studies to expand pollutant absorption estimates and to estimate the value of floodplain defences for the environmental accounts. We are also conducting research into estimating measures of climate change.

We are carrying out research into household childcare activity to improve the household account and into mapping which national accounts industries may have been affected by changes in household production. We have also commissioned work to look at the valuation of social transfers in kind and benefits in kind for the household account.

We are also looking to see how we can produce Gini co-efficients more efficiently.

Future research commissioning plans

Building on the recently published household satellite accounts (HHSA), we want to investigate the inclusion of internet-based activities, for example, booking holidays at home rather than at the travel agent, in the HHSA.

We wish to carry out research to better understand household distribution. This will include the identification of data and methods that would enable more timely statistics on the distribution of household income, including the continued development of early estimates of the main indicators through the use of “nowcasting” techniques. We also want to investigate reconciling micro and macro estimates of household income, consumption and saving, and using microdata to produce distributional analysis of national accounts aggregates.

We also intend to look into the identification of opportunities for developing data and analysis of income, consumption and wealth together, including the harmonisation and integration of survey data sources, as well as the potential for greater use of administrative data.

Notes for: Annex: Economic statistics; issues, priorities and plans

1. Sometimes referred to as “flow of funds”.

Better statistics, better decisions: strategy for UK statistics, 2015 to 2020, UK Statistics Authority, October 2014.

National Accounts strategy, 2015 to 2020, Office for National Statistics, July 2015.

National Accounts medium-term work plan 2016 to 2018, Office for National Statistics, November 2015.

Economic statistics and analysis strategy 2016, Office for National Statistics, September 2016.

Economics at ONS: Increasing openness, improving capability, Office for National Statistics, December 2015.

Independent review of UK economic statistics, Professor Sir Charles Bean, March 2016.

UK trade development plan: 2016, March 2016.

UK Statistics Authority Business Plan, April 2016 to March 2020, UK Statistics Authority, April 2016.

National Statistics Quality Review, (NSQR) Series (2) Report No. 2: Review of National Accounts and Balance of Payments, Office for National Statistics, July 2014