Methodoleg dadansoddiad o faint y busnesau

This document describes how we calculate estimates for the size analysis of businesses statistical bulletin.

The release provides data on the structure of enterprises active in Wales. It includes: 

• estimates for the very smallest businesses that operate below the VAT threshold
• estimates of the number of enterprises active in Wales by size band (defined in terms of number of UK employees)
• the aggregate employment and turnover in those enterprises

Results are presented by: 

• size band
• broad industry category 
• some of the different geographies within Wales

Coverage of the analysis

This statistical release covers the business sector, which is defined as all private sector organisations, plus all public corporations and nationalised bodies. All central and local government organisations have been excluded. In terms of legal statuses as given in table 1, coverage relates to codes 1, 2, 3, 4 and 7 only. All industries that are covered by these legal statuses are included in the analysis.

Only enterprises considered to be active in Wales are included, namely those with non-zero employment or turnover within their Welsh elements.

Estimates are included for the very smallest businesses that operate below the VAT threshold or that do not operate a PAYE scheme.

Main variables analysed

The analysis in this release is based on enterprises with some activity in Wales, regardless of where the enterprise is registered. In this context, an enterprise is taken to be an entire organisation, including its head office and all the individual UK sites, wherever they may be located. The region of registration of the enterprise is usually based on the location of its head office, although some of the individual sites may be located in other regions.

Three main variables are analysed.

• The number of enterprises which are active in Wales (or one of its sub-areas), which means that an enterprise is counted if at least one of its active sites is located in that area (an active site is one which has some employment or turnover attributed to it).
• The aggregate employment across all the sites within the enterprises that are located in that area.
• The aggregate turnover allocated to these same sites.

The first of these variables cannot be summed across sub-Wales areas to give a Wales figure, as one enterprise may be active in more than one of the different areas and hence be double counted. The other 2 variables can, however, be summed, as the employment and turnover are allocated to individual sites, which can only be present in a single area.

Note that each employee in each enterprise is counted once, whether they are full-time or part time. The employment variable is therefore a headcount measure, as opposed to the number employed on a full-time equivalent (FTE) basis. As a result, the employment variable is a count of jobs, not a count of persons in employment (as the same person could be included as an employee of more than one of the enterprises), and clearly a workplace-based measure (as opposed to a residence-based measure).

Estimates for the turnover variable exclude figures for the financial and business services sector throughout this release. This is because the turnover information available for enterprises in the financial services sector also includes the amounts involved in the transactions carried out by these enterprises and is therefore not comparable to turnover information in the other sectors.

Size bands analysed

In all cases, the size band of the enterprise is based on the number of UK employees in the enterprise. For example, an enterprise employing 10,000 UK staff but only a handful in Wales is categorised as a large enterprise in the analysis.

Summary of sources

The data in this release have been derived from a number of sources. The primary source is a detailed dataset based on extracts taken in March from the Inter Departmental Business Register (IDBR) maintained by the Office for National Statistics (ONS). This dataset gives details of all enterprises active in Wales, including the UK employment and turnover and derived figures for the Welsh elements of employment and turnover (using information about the sites in each enterprise which are located in Wales). It also includes identifiers for industry (using the Standard Industrial Classification, SIC), geographical location and legal status of each enterprise, thus allowing the analysis given in the release.

The IDBR covers a very high proportion of the employment and turnover in the UK. It excludes most of the very large number of smallest businesses that operate below the VAT threshold, with unregistered businesses accounting for 54% of the UK business population in 2025, as reported by the Department for Business and Trade’s Business Population Estimates. In order to correct for this, the figures presented in the release include estimates for unregistered enterprises, together with their related employment and turnover. These estimates for unregistered enterprises are calculated using a combination of:

• IDBR data
• information on persons who are self-employed in their main or second job, taken from the Annual Population Survey (APS) carried out by the ONS
• information on the numbers of sole traders and partnerships from the Survey of Personal Incomes (SPI) data provided by HM Revenue and Customs

Unregistered enterprises are assumed to be UK owned because these sole-proprietors and partnerships are estimated from the APS, which is a survey of UK households.

The APS has seen a fall in sample sizes over recent years and the survey has not been reweighted based on latest population estimates. It is still appropriate to use APS statistics, however users should note the increased uncertainty around estimates derived using the APS. 

At a national level, estimates continue to provide a reasonable indication of trends for Wales. However estimates for smaller geographies or population sub-groups are less reliable.

The analyses of enterprises registered on the IDBR is central to the analysis and is based on two extracts taken in March. One covers all enterprises with some activity in Wales and one covers all Welsh sites (or local units) in these enterprises.

The enterprise extract includes a unique enterprise reference number as well as fields providing information about each enterprise such as: 

• UK and Welsh employment and employees
• UK and Welsh turnover
• industry
• geographical location 
• legal status 

The local unit extract includes a unique local unit reference number as well as a field linking the unit to its unique parent enterprise reference number, together with further fields detailing employment and geographical location.

The enterprise extract is developed into a detailed dataset by adding a field allocating each enterprise to an employee size band (based on the number of UK employees) and excluding all enterprises that are out of scope (see coverage of the analysis). 

Within this detailed dataset, the employee size band for companies (legal status 1) with only 1 employee is changed to the zero employee size band. This is because people who work on their own and register as a company automatically become an employee of the company when it is set up. Without this adjustment, the numbers of businesses with zero or one employee will change purely because of changes in legal status (which can be driven by changes in tax / liability laws for example), rather by any real change in the economy.

In order to carry out the geographical analysis within Wales, additional datasets are created from this first detailed dataset, with separate datasets required for each of the different geographic breakdowns of Wales given in the analysis. Each additional dataset is created by disaggregating those enterprises with multiple local units in Wales into the separate parts of each enterprise that are located in each geographic area being considered, using information about the local units of each enterprise.

Analysing the IDBR datasets

The datasets are analysed to produce the required summaries of those enterprises registered on the IDBR. These analyses give enterprise (or part enterprise) counts and related employment and turnover aggregates. This includes figures for:

• Wales by size band, industry and ownership
• Welsh local authorities by size band, industry and ownership
• Welsh economic regions by size band, industry and ownership

Whilst the employment and turnover totals within the different geographic areas will sum to the figures for Wales, the same is not true of the counts of enterprises active in each area. This is because a number of the enterprises are active in more than one of the areas and is the reason why additional datasets need to be created for each different geographic breakdown considered.

The calculation of unregistered enterprises is set out in the following sections.

Calculation of estimates of unregistered working proprietors

The IDBR covers a very high proportion of the employment and turnover in the UK. As it excludes most of the very large number of smallest businesses that operate below the VAT threshold, its coverage in terms of the number of enterprises is only partial. In order to correct for this, the figures presented in the release include estimates for unregistered enterprises, together with their related employment and turnover. These estimates for unregistered enterprises are calculated using a combination of:

• IDBR data
• information on persons who are self-employed in their main or second job, taken from the APS carried out by the ONS
• information on the numbers of sole traders and partnerships from the SPI data provided by HM Revenue and Customs

It is firstly necessary to calculate an estimate of the number of Welsh working proprietors (i.e. enterprise owners located in Wales) that are not included on the IDBR. To do this the number of Welsh working proprietors recorded on the IDBR is required and can be estimated as the difference between Welsh employment and employees for each enterprise using the detailed main dataset described in developing a dataset for analysis from the IDBR.

This number can then be subtracted from the total number of working proprietors, which are estimated using self-employment data from the APS. The data used are counts of persons (on a workplace basis) who were self-employed in either their main job or second job in the year prior to the reference year (for example, 2025 size analysis uses 2024 APS data). The APS sample is boosted for Wales, this larger sample allows for estimates at a local authority level.

It is assumed that all unregistered enterprises are either sole traders or small partnerships (i.e. with only 2 partners) with no employees, on the basis that bigger enterprises will have turnover above the VAT threshold and hence be included on the IDBR. Therefore, both the IDBR data giving registered Welsh working proprietors and the APS data giving total Welsh working proprietors need to be broken down into these two categories. This is achieved by making reference to the legal status of the enterprise for the IDBR data, but it is more difficult for APS data.

Data from the SPI are used to breakdown the APS data into the 2 categories. A ratio of the number of Welsh sole traders to the sum of Welsh sole traders and partners is derived from the SPI data and applied to the APS data in estimating the number of sole working proprietors from the APS. The remaining working proprietors in the APS are then assumed to be partners. The SPI data used are a rolling 3-year average.

It is therefore possible to calculate raw estimates for the number of sole traders and partnerships not included on the IDBR. Note that this calculation is broken down by industry and can also be broken down by local authority. However, the SPI ratios used to breakdown the APS data are only available at the Wales level by industry, and so in calculating the local authority figures, the Welsh ratios for each industry are assumed to hold for all local authorities. This assumption is probably a reasonable one, based on some analysis of the limited SPI data by industry at a sub-Wales level which indicated that differences in the SPI ratios by industry were more significant than differences by location.

Conversion into unregistered enterprises and related employment and turnover

The calculation of the raw estimates of unregistered Welsh working proprietors yields some negative numbers for the agriculture industry, probably due to some classification inconsistencies in the various datasets in respect of agricultural partnerships. It is assumed therefore that all working proprietors for agriculture are included on the IDBR and so the estimated unregistered figures for agriculture are set to zero.

In order to convert the estimate of unregistered working proprietors into an estimated count of unregistered enterprises in Wales, the number of partnerships is halved and added to the number of sole traders. This is based on the assumptions that there are no employees in either category of unregistered enterprises and that there are only 2 partners in each unregistered partnership.

The number of unregistered working proprietors is equal to the estimated Welsh employment aggregates for unregistered enterprises. This is based on the assumption that there are no employees in either category, the Welsh employment thus being equal to the Welsh working proprietors themselves.

Finally in order to calculate Welsh turnover aggregates for each industry grouping, estimates of turnover per employment for those sole traders and partnerships with no employees which are registered on the IDBR are applied to the total number of Welsh working proprietors derived above. In doing so, the turnover per employment estimates for these registered enterprises are halved (based on the assumption that they will have higher turnover than their unregistered counterparts). Resultant turnover figures are also constrained so that the turnover per unregistered enterprise does not exceed the VAT threshold in each industry.

Note the Wales turnover per employment ratios for registered enterprises are used when calculating turnover aggregates for the unregistered enterprises for the different geographical breakdowns of Wales.

The figures for unregistered enterprises are added to the datasets that emerge from the analysis of the IDBR extracts as detailed in analysing the IDBR datasets. The unregistered figures are added to the zero employee size band for each industry and geographical area to be consistent with the assumption that all unregistered enterprises have no employees.

The figures in this release may not add up exactly as: 

• all enterprise counts have been independently rounded to the nearest 5
• all employment aggregates have been independently rounded to the nearest 100 
• all turnover aggregates have been independently rounded to the nearest million pounds

In order to avoid suppression of data due to disclosure control (which must occur for any aggregates based on less than 20 enterprises), some of the size bands and industries have been presented in an aggregated form. These aggregated groups have been chosen to limit the amount of aggregation whilst ensuring no data are suppressed.

As the estimates of unregistered enterprises and their associated activity are based mainly on survey data, they are subject to a degree of sampling error. The unregistered estimates are included within the analysis as given but are not presented separately.