SESCG 1/1/2004 - REGIONAL ACCOUNTS - ALLSOPP REVIEW UPDATE
This document is also available in PDF format
Purpose
Following the publication of the final report of the Allsopp Review, and the publication of the 2002 NUTS1 Regional GVA estimates by ONS, is based on an ONS report and is intended update the group on future plans for working with ONS, and taking forward the improvements/recommendations presented in the Allsopp report.
Advice is sought from the Group on:
- Whether they are content with the proposed way of working with ONS
- Any outstanding concerns which they would like us to pursue.
Background
As it was known that the final Allsopp Report was not going to be published before the end of March 2004 six work packages were identified to address all the recommendations of the initial report (other than those that were solely consultative), which focussed on regional statistics to support economic policymaking. These work packages were analysis, administrative data sources, business register, business surveys, household surveys and other work, including institutional (e.g. GSS/ONS presence in the English regions). The task of developing these work packages has involved over half the divisions in ONS.
It is because of this tight timetable and the sheer scale of the work involved that ONS has drawn up its SR2004 bid without formally consulting Other Government Departments. ONS gave its Regional and Local Division the responsibility of formulating an SR2004 bid with a base case and some other options. These proposals are to be considered by the ONS Board on 20 April before a submission is made to Treasury by the end of April 2004 - the results of their deliberations will be known by the end of July 2004.
Key principles in addressing the Allsopp Review
- The developments in business surveys, registers, use of administrative data etc are all integrated with other developments and not an "add on" separate development. The aim is for any major systems development to take place after the main ONS Modernisation developments have taken place. Where possible, modernisation work should take account of changes required by Allsopp but not at the expense of delay to Modernisation.
- The recommendations in the final Allsopp report will be addressed at the same time as the more regionally oriented ones in the first report (several key aspects - such as SERVCOM - are covered below)
- In particular, the approach is to (re)design business surveys fully, during 2004-05 and 2005-06 in parallel with statistical modernisation, before initiating any expansion of survey samples. The redesigned surveys will cover Allsopp requirements as well as being implementable in the new CORD (Central ONS repository for data) infrastructure.
- For reliable estimates at NUTS 2 and lower levels, work on synthetic estimation will be required to supplement the changes to surveys intended to enhance the quality of figures at NUTS 1 region level. It would not be a good use of resource to attempt high quality direct estimation at NUTS2 level.
- Work will be undertaken to use administrative sources as replacements for survey sources as well as to aid estimation as above.
Main parts of development programme (NB not a comprehensive list but aiming to cover major parts)
Analysis
1. Regional Gross Value Added from the production approach at current and constant prices will require new uses of existing data, development of new data sources and a new methodology. The production approach has not been used to produce regional estimates before and chained volume measures have not been compiled.
2. Income and expenditure GDP and GDHI at regional level require the continuation of existing activities together with appropriate enhancements to data on incomes. Expenditure sources and methods will need to be reviewed and developed. An approach to deflation will be developed.
3. Integration of national and regional accounts processes will go live after the Spending Review period but some development will be within the period.
4. Annual Population Survey samples in England Unitary Authorities (UAs) to be brought up to the minimum sample size used for UAs elsewhere e.g. in Scotland and Wales.
5. One off development and subsequent running of a new approach to the matching (including additional methodological and fieldwork) of IDBR and LFS to improve the quality of the LFS industry and sector information. In addition, also to improve the coherence of business and household survey jobs data, special surveys targeted at understanding sources of discrepancy better.
6. Development and running of an extension to ONS's enterprise based survey of vacancies with the aim of producing monthly regional level vacancy statistics by industry and sector.
7. Increased samples in consumer price collections, and further development work, to provide regional consumer price deflators, in 2006/07 and 2007/08. This helps to meet the existing commitment from the Chancellor which is only funded to 2005/06 at present.
Administrative data sources
8. Acquire the key sources of micro-level data from IR, DWP and HMCE, (PAYE, Corporation Tax including business accounts supplied in support of returns, Self-Assessment, benefits and VAT) and bring the datasets into ONS as part of extended CORD architecture, including quality checking in ONS, linkage at micro data level and re-engineering costs in IR, DWP and HMCE. This is additional to Neighbourhood Statistics aggregate data acquisition.
Business register
9. Creation of an expanded and merged Annual Register Inquiry/1 and Annual Business Inquiry/1 to provide regional and industrial detail.
10. Redevelop the ABI/2 survey to be stratified at NUTS1 level, rebalance stratification between manufacturing and services etc, and replace it with an Annual Business Survey (ABS). Review and redevelop the apportionment method for Reporting Unit information to Local Unit level. See note above re use of admin data; Given the legal and technical obstacles, it is unknown when administrative data for the ABS (corporation tax etc) will be available. Its quality will also have to be assessed. The preferred option is therefore to assume it will not be available during the first four years of the period of the bid and so the more expensive option is given as the base case.
11. Combine the MIDDS, MPI, RSI monthly surveys into a Monthly Business Survey (MBS), with potential use of VAT data, taking into account potential regional needs raised by the Final Report by the option of boosting the sample size;
12. Develop an annual Business Omnibus Survey (BOS) as raised in the Final Report, which would consist of a telephone survey to ascertain firms' activity e.g. do they do ecommerce or R&D, which would then be followed up by the relevant 'specialist' business survey. This would also be used as the vehicle for ad hoc surveys etc.
13. Develop a SERVCOM survey using the existing PRODCOM survey as the model - and integrate the two fully when the surveys move into CORD;
14. Conduct a feasibility study on regional price deflators (PPI and CSPI) and implement new data collections
Household surveys
15. Extension of sample of Expenditure and Food Survey, as part of the Continuous Population survey and taking advantage of no clustering in CPS. This will improve estimates of final household monetary expenditure and weights used in national and regional indices and deflators.
16. Rely on redevelopment of NHS Information Systems to provide improved source of internalmigration data
Other work, including institutional
17. Establish regional presence of 6 persons per region in each of 9 English regions
18. Establish Public Policy Division in ONS to keep closer track of policy drivers and statistical implications and requirements and work on formulation of government targets.
19. Extend current RLD role on data needs to create and service central/regional/local government data needs, priorities and decision making body, including provision of advice on good practice etc.
20. Build Regional analysis role to cover social data, and Quality Assurance role for Regional Accounts and other outputs
Regional Accounts
The 2002 NUTS1 Regional GVA estimates were published by ONS on Friday 30 April 3004. Thisfollowed a period of 'peer appraisal' which officials form the Scottish Executive were closely involved in. We contributed many comments, and questioned results from the various contributing surveys included ABI, AES, NES etc. We concluded, and confirmed, that based on the existing data sources and methodologies, the results represented the best possible estimates of Scottish GVA. Assuming a very simple GDP deflator, the real annual growth rate for 2002 look fairly similar to what we have published in the Scottish quarterly GDP series.
The Scottish Executive is still recommending a major methodological change for the Regional Accounts, but this is not likely to take place until after the ONS re-engineering project. It is supported by the Allsopp report, and the Regional Accounts Quality Review (Wroe), but each of these recognise that other developments must be put in place first.
In the mean time, ONS have a number of short term development proposal which we will work closely with them on. The first is a review of the Highlands and Islands adjustment. It has been agreed that HIE will be involved in this process, but that the adjustment methodology will be considered across the whole of the UK, not just H&I.
OPS:ASG - OCEA Statistics May 2004