Guidance

Statistical revisions policy

Updated 1 December 2022

Introduction

All Department for Business, Energy and Industrial Strategy (BEIS) Official Statistics publications are produced to the high professional standards as set out in the Code of Practice for Statistics.

As per section ‘T3.9’ of the Code, scheduled revisions (planned amendments to published statistics in order to improve quality by incorporating additional data that were unavailable at the point of initial publication) and unscheduled corrections (amendments made to published statistics in response to the identification of errors following their initial publication) should be released as soon as practicable.

The changes should be handled transparently in line with a published policy. As per section ‘Q3.4’ of the Code, scheduled revisions, or unscheduled corrections should be explained alongside the statistics, being clear on the scale, nature, cause and impact.

The BEIS statistical revisions policies are broken down into those for statistical publications:

(a) with scheduled revisions, for example (1) energy statistics (2) greenhouse gas emissions statistics (3) business statistics

(b) without scheduled revisions, for example (4) other official statistics

Unless otherwise stated below, BEIS will indicate any revisions clearly, indicated by a marker such as ‘r’ whenever data in publications has been revised from that initially published.

1. Energy statistics

Statistics on energy contained in the Digest of UK Energy Statistics (DUKES), Energy Trends, Energy Prices and related publications are classified as National Statistics.

Revisions to data published in the DUKES

It is intended that any revisions should be made to previous years’ data only at the time of the publication of the digest (for example in July 2016 when the digest was published, revisions could be made to 2014 and earlier years). In exceptional circumstances previous years’ data can be amended between digest publication dates, but this will only take place when quarterly Energy Trends is published and would be announced in Energy Trends. The reasons for substantial revisions will be explained in the ‘Highlights’ sheet of the Excel table concerned. Valid reasons for revisions of digest data include:

  • revised and validated data received from a data supplier
  • the figure in the digest was wrong because of a typographical or similar error
  • methodological improvements, which will be pre-announced as they may result in significant back revisions. Full explanations will be provided with back data revised where possible to avoid breaks in the published time series. Methodological changes are normally made in June or July and described in articles in Energy Trends or DUKES.

Section 5 describes the process we will follow if we identify a major error.

In addition, when provisional annual data for a new calendar year (such as 2016) are published in Energy Trends in March of the following year (in this case March 2017), percentage growth rates are liable to be distorted if the prior year (2015) data are constrained to the digest total, when revisions are known to have been made. In these circumstances the prior year (2015) data will be amended for all affected tables in Energy Trends and all affected digest tables will be clearly annotated to show that the data has been up-dated in Energy Trends.

Data revisions occur for the current year until the publication of the digest covering that year. After that, changes to quarterly data would only usually happen at the time of future digest publications. Revisions can occur for a variety of reasons:

  • all validated amendments from data suppliers will be updated when received and published in the next statistical release
  • all errors will be amended as soon as identified and published in the next statistical release
  • data in energy and commodity balances format will be revised on a quarterly basis, to coincide with the publication of Energy Trends

Section 5 describes the process we will follow if we identify a major error.

For the benefit of users of Energy Trends, BEIS has produced guidance to explain which data periods can be revised in any given publication month. Table 1 below shows for each publication month which data periods can be revised for both the monthly and quarterly editions of Energy Trends. Revisions can occur for a variety of reasons:

  • all validated amendments from data suppliers
  • all errors identified
  • data in energy and commodity balances format will be revised on a monthly basis, to coincide with the quarterly publication of Energy Trends
  • seasonal adjustment factors for the previous 3 years will be revised at the same time as the annual re-assessment

Section 5 describes the process we will follow if we identify a major error.

Publication Month Data to be revised
January October data of prior year
February All data in prior year given first publication of provisional annual data
March All quarterly data of previous 2 years, so that both levels and growth rates are as accurate as possible for the latest year
April January data of current year
May January and February data of current year
June Q1 data of current year. All quarterly data of previous years reconciled with revised annual DUKES balances
July April data of current year
August April and May data of current year
September Q1 and Q2 data of current year
October July data of current year
November July and August data of current year
December Q1, Q2 and Q3 data of current year

Note: Monthly and Quarterly editions of Energy Trends produced in March, June, September and December.

In exceptional circumstances significant revisions outside of the data periods specified above may be allowed – if made these revisions will be detailed by BEIS on the ‘Highlights’ worksheet of the relevant Energy Trends table Excel workbook, and in the quarterly editions of the Energy Trends publication.

Quarterly Energy Prices data are published in arrears. For industrial prices, data are published 3 months in arrears, with provisional annual data published in the March edition and final data in June. For domestic prices, provisional annual bills are published in September and final annual bills in December. For international comparisons, 6-monthly data is published 6 months in arrears.

Data in Energy Consumption in the UK (ECUK) are revised annually based on consumption outputs from DUKES, revisions to DUKES in turn flow through into ECUK. An overview of likely revisions is provided in the ECUK User Guide and details of revisions relating to the specific year of publication will be included in the commentary.

In general, sub-national data are not revised in conjunction with the UK based data as the sources and timings differ. However:

  • all validated amendments from data suppliers will be updated when received and published in the next statistical release
  • all errors will be amended as soon as identified and published in the next statistical release

Section 5 describes the process we will follow if we identify a major error.

Revisions to sub-national consumption statistics

On occasions, previously published data will need to be revised. These revisions are usually due to forecasted values being replaced with actual data, where actual figures were not available at the time of publication. In particular, annual revisions are made to the road transport, residuals and total final energy publications, and these revisions have been explained in further detail in chapters 7, 8 and 9 of the sub-national consumption statistics methodology note respectively.

Changes to historic data will only be made as part of a new publication. Where a large revision has taken place reasons will be provided. In cases where entire historic datasets have been revised, this will be clearly marked in the dataset. Changes to methodology would be pre-announced and impact of revisions explained when changes are made (with at least one year of data produced by both methods if appropriate).

Section 5 describes the process we will follow if we identify a major error.

Revisions analysis

BEIS undertake revisions analysis of some key energy series. Revisions to DUKES are shown in Table 1J of Chapter 1 of DUKES. Analysis of revisions, including revision trends over time, to quarterly production, supply, demand and consumption are published as part of Energy Trends on the last Thursday of every quarter.

2. Fuel Poverty statistics

In the annual fuel poverty statistics from 2023, the headline figure is a one-year ahead projection based on survey data. This change has been made to present more timely fuel poverty estimates using an established methodology used for fuel poverty projections. The one-year ahead projection uses many observed changes in date since the survey data for energy prices, energy efficiency and household incomes.

The headline figure and supporting detailed and supplementary tables are therefore provisional figures that will be revised approximately one year after publication when actual survey data for that period is available.

3. Greenhouse gas emissions statistics

Definitions

The UK greenhouse gas inventory for year X is the time series of UK greenhouse gas emissions from 1990 to year X. The inventory can be broken down into a range of sectors and gases. The inventory for year X is normally published in February of the year X+2 (so for example National Statistics on the 2013 inventory were published in February 2015).

Annual revisions to the UK GHG inventory

In order to comply with UN reporting requirements for greenhouse gas inventories, the time series of UK greenhouse gas emissions going back to 1990 is revised on an annual basis when the new greenhouse gas inventory is published. Any methodological improvements or new sources of emissions that have been applied to the latest year of data are also applied to the full time series going back to 1990.

This means that all of our annual National Statistics on greenhouse gas emissions (Final UK Greenhouse Gas Emissions, Provisional UK Greenhouse Gas Emissions, and Local Authority Carbon Dioxide Emissions) will have a fully revised historical time series of emissions every year.

The annual Final UK Greenhouse Gas Emissions statistics contains a section discussing the annual scheduled revisions to the historical emissions time series.

Revisions to data published in greenhouse gas emissions statistics

It is intended that any minor revisions to UK greenhouse gas emissions statistics will only be made at the time of publication of future editions of those statistics.

Valid reasons for revising data (including routine revisions at time of next publication) include:

  • annual revisions to UK GHG inventory data, as explained above
  • amendments to UK GHG inventory data between publications as National Statistics in February each year and final submission to the UN in April
  • errors identified with numbers in the publication, either due to inaccurate data being provided by data suppliers or to data processing errors
  • revisions to underlying energy data statistics in the Digest of UK Energy Statistics or Energy Trends; revisions made on an annual basis to the historical greenhouse gas emissions time series will not be marked in Official Statistics publications, as this affects the whole time series each year

Section 5 describes the process we will follow if we identify a major error.

Provisional and final data

UK greenhouse gas emissions data are considered to be final once the corresponding greenhouse gas inventory has been submitted to the UN. Provisional estimates are calculated based on UK energy data for up to 7 quarters of more recent emissions than are included in the latest UK GHG inventory. This means that:

  • all data in the Final UK Greenhouse Gas Emissions and Local Authority Carbon Dioxide Emissions are considered to be final (though the historical time series from 1990 onwards will continue to be revised annually)
  • in the Provisional Greenhouse Gas Emissions publication, the latest year of emissions data presented is considered provisional, while earlier years are consistent with the latest greenhouse gas inventory and are considered final

Data tables will be clearly labelled to indicate any data that are provisional. The Final UK Greenhouse Gas Emissions includes a section comparing the provisional emissions previously published for the latest inventory year with the final emissions estimate for that year.

Archiving of greenhouse gas emissions statistics

Because the entire greenhouse gas inventory time series is revised each year, only the most recent edition of each publication will contain up to date and relevant data. Previous editions will be archived on the GOV.UK website, but should not be used unless the user is specifically interested in data from old greenhouse gas inventories.

Note that the first publication of data from each new greenhouse gas inventory is in the publication of Final UK Greenhouse Gas Emissions in February each year. Other statistics will not be made consistent with the new inventory until their first publication after that date.

UN reviews of the UK GHG inventory

The UN Framework Convention on Climate Change (UNFCCC) carries out an annual review of the UK’s greenhouse gas inventory each autumn and publishes a report, which can include making adjustments to the UK inventory. UK greenhouse gas emissions statistics will not normally be revised to reflect any such adjustments, as these adjustments are normally not finalised until the next year’s greenhouse gas inventory and corresponding National Statistics have already been published. However, the UNFCCC reports are made publicly available on the UNFCCC website.

4. Business statistics

Revisions to building materials and components: monthly statistics

BEIS publishes monthly statistics of building materials and components covering the statistics on a range of selected building materials. This section sets out a revisions policy to revise or correct the publication.

It is our policy to be open and transparent with users about the reasons for revisions and how they are handled. We will pre-announce revisions where practicable.

There are 3 main reasons that will cause revisions to the publication.

  • receipt of more data
  • methodology changes
  • errors

Scheduled revisions will occur when BEIS receive more accurate information or complete data for a reference period following more detailed validation of source data as planned. Some of building material data series are subject to scheduled revisions. They are initially published as provisional and will be clearly marked by a ‘(P)’ marker next to each provisional figure.

At a later scheduled date, the provisional statistics will be finalised and the ‘(P)’ markers will be removed. The table below provides a revision schedule for the data series subject to scheduled revisions. Data series not included in the following schedule don’t contain provisional data.

Section 5 describes the process we will follow if we identify a major error.

Table 2: Building material statistics revision schedule

Data series Revision timing after the publication of provisional information
Table 1: Construction Material Price Indices 3 months
Tables 2 and 3: Price Indices of Construction Materials Usually 2 months
The latest 5 months data are subject to revisions
Table 4: Sales of Sand and Gravel in Great Britain 1 month
Table 9: Bricks – Production, Deliveries and Stocks 1 month
Table 11: Concrete Blocks – Production, Deliveries and Stocks Up to 3 months
Data are finalised at the end of each reference quarter
Table 14 and 15: Value of Overseas Trade in Selected Materials and Components for Constructional Use Up to 5 quarters

Revisions to Business Population Estimates (BPE)

In all editions, when re-weighted (for updated population estimates) ONS Labour Force Survey (LFS) data are made available BEIS will revise the headline estimates of the number of private sector businesses in the UK for the relevant years. Updated estimates by size band will be provided in Figure 3 and Table B of the statistical release and Table 25 of the detailed data tables, to reflect the impact on the unregistered business population estimate.

ONS did not re-weight LFS data in 2015, but has re-weighted in 2016 and 2017 to bring the outputs in line with improved mid-year population estimates. BEIS has consequently used the improved data to revise BPE estimates for 2013, 2014, 2015 and 2016.

If possible, BEIS will also revise previously released estimates if new sources of information become available that are shown to improve the robustness of the BPE. Any future substantial methodological changes of this nature will be consulted on and pre-announced before they are implemented.

Section 5 describes the process we will follow if we identify a major error.

5. Other official statistics

Revisions to BEIS official statistics publications that do not have scheduled revisions may be caused by a range of factors: the inclusion of data received after the cut-off point for the original release, a change to coverage or methodology, or unscheduled corrections due to errors.

In the case of implementation of unforeseen changes in the methodology or the unexpected acquisition of new data, it will not usually be possible to pre-announce the changes, but as far as possible we will produce a consistent headline back series to enable robust comparisons over time.

If an error is discovered, the process will be as follows for minor or major error.

Minor error

If it is minor and does not materially affect the data, particularly headline messages, then this will be corrected and data revised in the next scheduled release.

Major error

If it is major and does have a material impact that could affect decisions made on the basis of the data, then an unscheduled correction will be made.

A corrected version of the publication will be placed online alongside the statistics as soon as possible, together with an explanation of the reason for the change, being clear on the scale, nature, cause and impact. Depending on the timing and with agreement by the Head of Profession for Statistics, the statistics may be withdrawn or a warning may be placed online in advance of the corrected version being published.

Decisions relating to the dissemination of the announcement of any unscheduled corrections will be made by the relevant BEIS statistician with agreement by the Head of Profession for Statistics. For example, they may alert key users.