National statistics

DCMS Sectors Economic Estimates 2022: Business Demographics - Source Data Change Summary Note

Updated 16 November 2023

Comparison of ABS and IDBR data to produce DCMS Business Demographics estimates

Background

The DCMS Business Demographics statistical release provides estimates of the number of businesses in DCMS sectors, their size and turnover bands, in the UK and regionally.

Up to and including 2021 (2019 statistics) the data source for these estimates was the Annual Business Survey (ABS) published by the Office for National Statistics (ONS). The ABS is an annual survey of businesses in the UK, based on a sample drawn from the Inter-Departmental Business Register (IDBR), and data becomes available on a two-year lag to allow time to collect, verify and analyse the data.

For this 2022 DCMS Business Demographics statistical release, we have experimented with using a snapshot of the Inter-Departmental Business Register (IDBR) to generate estimates of DCMS Business Demographics. This has the advantage of being more timely than ABS, while still committing to most tables included in previous Business Demographics publications. We have used the March 2019, March 2020, March 2021 and March 2022 snapshots from the ONS UK business: activity, size and location release rather than raw data from the IDBR.

Call for Feedback

We are looking for feedback on the approach we have used in this release. We particularly welcome views on:

• Continuing with this approach using the more timely IDBR data source.

• Returning to the previous approach using the ABS as a data source.

• Experimenting with a ‘mixed approach’ where estimates would be produced using the IDBR snapshot, supplemented with further data from the ABS to produce additional tables.

Please contact evidence@dcms.gov.uk before Thursday 9th February 2023 with any feedback.

Summary of ABS and IDBR data sources

In order to help inform users about the impact of the change, we have compiled this short note summarising the comparative analysis we have done between the ABS and the IDBR. A broad comparison of the similarities and differences between the datasets is shown below.

	ABS	IDBR
Frequency of data	Once a year	Once a year
Timeliness of data	2 years behind (2021 we receive 2019 data, etc)	6 months behind (March snapshot published late September/early October)
Methodology	Sample using IDBR to frame sample. Covers the UK Non-Financial Business Economy (UKNFBE), so does not include a minority of sectors (most financial sectors in particular).	Covers all UK businesses that are registered for VAT and/or PAYE taxation.
	Questionnaires are sent out to the sampled businesses, asking questions about the year as a whole	A snapshot is taken every quarter with the March snapshot forming the basis for the estimates of size, regions etc.
Sample Sizes	60,000 registered UKNFBE businesses, sampled from the IDBR.	The IDBR captures registered businesses (numbering around 2.8  million), who make up around 99% of UK economic activity.
	Businesses that have not registered for VAT or PAYE are not covered. Such businesses typically include sole traders and those with low turnover.	Businesses that have not registered for VAT or PAYE are not covered. Such businesses typically include sole traders and those with low turnover

Strengths and Weaknesses

The strengths and weaknesses of the two data sources are generally mirror images of each other. That is to say, the strengths of the IDBR reflect the weaknesses of the ABS and vice versa. Relative benefits of using the IDBR include:

• The IDBR is a much more timely dataset, with a lag of only 6 months (March data snapshot available from late September/early October of the same year) rather than up to 2 years with the ABS (2020 data available summer 2022)
• The IDBR is an admin data source with a comprehensive list of registered UK businesses, while the ABS is a smaller sample of registered UK businesses.
• Some of the data is more accurate, in particular the turnover data, which predominantly comes from HMRC VAT data directly, rather than self-reported as in the ABS.

By contrast, relative benefits of using the ABS include:

• The survey asks about the whole year, rather than the IDBR data used being a snapshot (though the majority of Turnover estimates in the IDBR are based on 12 months of HMRC VAT returns)
• The ABS can answer extra questions compared to the IDBR. In particular, more turnover information is available (though likely to be less accurate than the IDBR) and there are questions about foreign-owned businesses and exporter/importer businesses that the IDBR doesn’t have (though these are contingent on sample sizes being sufficiently robust)

Summary of comparative analysis between ABS and IDBR

Additionally, comparative analysis of estimates based on the ABS and IDBR for 2017 to 2019 show similar values for the overall numbers of businesses in DCMS Sectors, and the trends over the three years are also broadly similar from both sources. As one would expect, there is some evidence of larger differences observed at more disaggregated levels (e.g. small subsectors). However, overall, the numbers and trends are comparable across the two data sources.

The main difference is that ABS records overall higher numbers of businesses in lower turnover bands than the IDBR. This is likely to be due to differing methodologies; ABS data is self-reported by survey respondents, while the IDBR uses HMRC tax returns. Trends from 2017 to 2019 were similar for both sources for medium and high turnover bands, but in the lower bands ABS estimates are more volatile.

Please note that some differences in all of the estimates will be due to the IDBR being a snapshot of the data in March of the reference year, which may mask some of the fluctuations that occur during the year.