Regional GVA (Production Approach)
Regional GVA(P) Project
One of the main recommendations of the Allsopp Review was to compile regional Gross Value Added estimates at constant prices using a production approach (GVA(P)). This project has been set up to carry out that recommendation.
Background
Recommendation 2 of the Allsopp report 'Review of Statistics for Economic Policymaking' (2004) was that:
'...each region and country (at NUTS 1 level) should have annual baseline data for GVA at current prices and in chained volume terms, which would be derived according to the production approach'.
These data will provide estimates of regional economic growth in real terms, which will assist in regional policy-making and support the PSA (Public Service Agreement) target to improve regional economic performance.
Regional GVA at current prices is already published annually using the income method. See ‘Related links’.
Methodology
A methods paper was published in April 2007. This paper provided a review of methodological options and recommended the most suitable method for the calculation of current and constant price regional GVA by industry on a production basis. The paper is available from the ‘Downloads’ section of this page.
Investigation of early results
A prototype system produced some regional estimates of GVA (production approach) at current and constant prices by A31 industry. These estimates differed significantly from the existing published regional GVA estimates compiled on the income approach. The results were presented to the project’s Technical Advisory Group (TAG) which includes representatives from central government departments, devolved administrations and ONS regional statisticians as well as other ONS internal representatives. The TAG initiated some work to investigate at a more detailed level the quality of regional Annual Business Inquiry part 2 (ABI2) data; a key data source in the production approach.
Our analysis of the ABI2 data raised concerns that figures for industries G (Wholesale and Retail) and K (Real Estate) might have a London bias – a ‘headquarters’ effect – compared with the income approach. Given the size of these industries, this has a major impact on the regional picture. Hence the ONS postponed the publication of first estimates and commissioned further investigation.
An internal review of the apportionment model used to produce regional ABI2 (for more information on ABI regional estimation, see ‘Related links’), which reported in August 2008, concluded that, using the current data, no modification of the apportionment model would lead to a reduction of the alleged 'headquarters effect'.
Analysis of several alternative models showed that the current model is robust, in the sense that variation of the underlying assumptions did not lead to significantly different results.
Regional Deflators
Preliminary work has also taken place on the availability of regional deflators and whether any significant differences appear in the rate of price change.
Regional producer-based prices that would be consistent with the national accounts methodology are not available. Therefore, the current approach is to use national deflators to deflate current price regional GVA(P) as per the original project plan, which considered regional deflation currently out of scope. This may be revisited as part of future development.
Future Work
The project so far has concluded that the current methodology makes best use of the data currently available to ONS for the purpose of estimating regional GVA(P). In order to make further improvements to the estimation methodology, in particular to overcome the shortcomings of the apportionment methodology, it will be necessary to incorporate data which have been collected directly at regional level.
Collection of regional level data is taking place as part of wider improvements to ONS data collection, through the Business Register & Employment Survey (BRES) project (for more information, see the BRES Project link below). This survey will collect local unit financial data for a specific month from the selected enterprises, and the returned data will be used to estimate a value for every local unit in the Inter-Departmental Business Register.
It is planned that the GVA(P) project will utilise BRES data and publish regional GVA(P) as an 'experimental statistic'. This will be consistent with the annual ‘UK National Accounts Blue Book’ publication. The timetable for the first publication of experimental GVA(P) is dependent on progress with the BRES project. At present (February 2010) an estimation methodology for local units is being evaluated. The concept of ‘experimental statistics’ is that these are produced and published largely to promote the consultation and involvement of users in the development stage of the production of new statistics.
Following a further period of development, it is planned that the data will then be assessed according to the 'Code of Practice for Statistics' in order to gain 'National Statistics' status. This process is currently being reviewed by the UK Statistics Authority.
The BRES Project
Work is already underway on the Business Register & Employment Survey (BRES). It will replace two existing surveys - the ABI part 1 and the Business Register Survey. This will reduce duplication and respondent burden as well as improving regional estimation more generally. Specifically for the production approach for regional GVA, the BRES project plans to deliver an improved regional apportionment methodology for the ABI2 using Local Unit financial data.
Future Developments in Regional Statistics
The project is now part of a new regional statistical work programme. The Allsopp Programme, which oversaw the regional GVA(P) project, effectively ended on 31 March 2008. Separate development streams for regional statistics and service sector statistics will be run and continue to be developed. Internal and external stakeholder groups for regional and service sector statistics are in the process of being established.
Contacts
Regional Accounts: +44 (0)1633 456878
marc.carlton@ons.gsi.gov.uk
