Developing a method for producing annual gross value added at a subnational level

On 13 December 2021 the Office for National Statistics (ONS) published experimental gross value added (GVA) statistics at Middle layer Super Output Areas (MSOA) and higher geographies plus an accompanying article. The publication set Lower layer Super Output Area (LSOA) level data as the building blocks for higher level geographies.  

The ONS ran a consultation from 13 December 2021 to 7 March 2022 to seek users’ opinions on the experimental statistics. We asked users to help us understand:  

1. How they used the experimental statistics 

2. Whether they were going to access the building blocks dataset in the Secure Research Service (SRS) or not 

3. How useful they found the accompanying article 

4. Users’ suggestions about what we must do to improve the methods and procedures for producing subnational statistics 

5. If anything was missing from the accompanying article that would be beneficial to see in the future 

6. Users’ overall view of the experimental statistics 

7. If users wanted statistics broken down by industry, and why 

8. Whether they required any other industry breakdowns for their work 

9. Other subnational data requirements they wanted us to know and/or pursue. 

10. Other comments about the consultation 

The ONS is currently improving the availability of data at subnational level. It seeks to identify key areas of interest of different stakeholders with a view to build towards producing more (experimental) subnational statistics on a regular basis. Further, the ONS is working on a subnational data explorer facility, which is expected to mature into an online self-service platform, as discussed in the Government Statistical Service’s Subnational Data Strategy. 

Professional and personal users of the data responded to the consultation. Of the respondents, 71% were using the data professionally in some capacity, and 56% of the users worked in Government, including local government and public bodies. 

Users were keen to access the experimental data with over 80% showing interest in working with the data. Whilst the disclosive nature of the results meant that the data had to be held in the Secure Research Services (SRS) creating a 50/50 split in the ability in those users to be able to access it. 

You told us that: 

• disaggregating the gross value added (GVA) data to a subnational level helps to identify local trends in data over time, highlighting where economic activity is growing or declining 

• the provision of more granular data (e.g. the building blocks) helps to meet specific stakeholder needs 

• the presentation of data around the specific bespoke areas is clear and well structured 

• that we clearly communicated how disaggregating GVA to a subnational level created issues in disclosure at LSOA, requiring us to hold this data within the SRS due to legal requirements and that you were happy with the restrictions set out to access this data. 

Some users requested the following: 

• further data that can be compared to well-being economies/social systems alongside GVA and gross disposable household income (GDHI). 

• MSOA/LSOA data for hours worked, income tax and income, and employment changes  

• a user dashboard that allows users to visually ascertain data 

• explore the possibility of integrating with 3rd party data providers 

 
Industry breakdowns  

We asked our stakeholders to tell us their views about producing industry breakdowns of GVA at lower levels of geography. Our earlier investigations had shown that industry breakdown at LSOA level is disclosive. There is reduced risk of disclosure at MSOA level, and this can be addressed by grouping some industries together.  

The consultation results show that some users:  

• would prefer Group 1 to be broken down further, which is helpful for rural areas like Cumbria that consist of large agriculture and manufacturing sectors. Combining these two would mask important sector variations. 

• prefer grouping consumer facing businesses (e.g. retail, hospitality) separately rather than combined with distribution and transport. This will be investigated for practicality and the risk of disclosure. 

• suggested that education and health service are separated.  

Respondents to the consultation also made suggestions of what they want us to do in the future.  
Some of the suggestions or requests included: 

•  GVA broken down by environmental impact to inform the net-zero agenda. 

•  more frequent information on jobs and business change. For example, business demography quarterly below regional level – International Territorial Level (ITL) 3. ​ 

• sectoral breakdown of HMRC’s Pay As You Earn (PAYE) Real Time Information (RTI) by full time equivalent job count at ITL 3 level, but local authority preferred. 

• longer and up-to-date time series (less lagged) that will allow them to assess subnational impacts of Brexit.  

We are working on several projects to improve availability of subnational statistics, including: 

• looking to build a repeatable analytical pipeline for the ingestion and processing of data, in order to deliver future results in a timely manner 

• updating the GVA estimates to 2020 

• improving the apportionment methods that we use to breakdown GVA to lower levels of geography 

• break down GDHI to lower levels of geography