Evolution of the United Kingdom statistical system

Issue title: Evolution of National Statistical Systems

Article type: Research Article

Authors: Laux, Richard | Alldritt, Richard | Dunnell, Karen

Affiliations: Office for National Statistics, London, UK

Note: [] Corresponding author. E-mail: karen.dunnell@ons.gov.uk

Abstract: This paper identifies key events in the history of the UK statistical system, and the organisations and stakeholders which make it up. It discusses five distinctive features and their influence on the current and future evolution of the system. The statistical system is decentralised and devolved, reflecting the structure and evolution of government in the UK – and perhaps also reflecting an historical assumption that official statistics existed mainly to meet the needs of government. That view has since been overtaken by a wide acceptance that statistics must serve society as a whole. Statistical co-ordination is managed by the Office for National Statistics but is dependent on institutional and organisational co-operation. Until now there has not been a statutory framework to support the role of the National Statistician, but this is about to change. Despite the introduction of non-statutory `National Statistics' arrangements in 2000, intended to strengthen the quality and integrity of UK statistics, public confidence in the statistical system is low. New legislation – the Statistics and Registration Service Bill – currently before Parliament will address issues of public confidence and central oversight by establishing a non-executive Statistics Board that will monitor the quality and comprehensiveness of UK statistics and report to Parliament. At the same time a range of further non-statutory developments are being planned with the aim of improving the statistical system further and building confidence in it. The five distinctive features identified in Part 2 of this paper are: Decentralised system – for reasons of history, some 80% of the statistical system is the responsibility of organisations – the devolved administrations and government departments – other than the Office for National Statistics. This will continue in the future so the challenge the system faces is one of creating a coherent and efficient system that respects the autonomy of the parts that make it up (see 2.1, 3.1). Co-ordination and management – throughout its existence, central co-ordination and management of the statistical system has been dependent on the co-operation of a large number of government organisations and, more recently, the devolved administrations. However, the creation of the Office for National Statistics (ONS) in 1996 established a central body of sufficient size and authority to make central management a realistic possibility and the new legislation, will take that process a big step forward (see 2.2, 3.2). Low levels of public trust – the UK has a sceptical public and an aggressive press. This environment has fostered widespread distrust of government generally, and statistical outputs have been no exception. The absence of centralised arrangements or over-arching statistical legislation has compounded the challenge of addressing this. Plans are now being made for a range of non-statutory activities to improve public confidence, building upon the new legislation. But it is likely to take several years (2.3, 3.3). Code of Practice – a very detailed Code of Practice was introduced in 2002, presented in some 13 separate published volumes. Under the new legislation, the role of the Code of Practice will be more formally recognised and a revised draft Code is currently the subject of consultation (2.4 and 3.4). Inadequate access to register data for statistical purposes – within the UK, the emphasis of public policy has been on protection of personal privacy rather than the exploitation of administrative records for statistical purposes. The new legislation may start to change this but it is not yet clear whether the arguments in favour of statistical use of register (administrative) information have been fully accepted by Parliament (2.5, 3.5).

Journal: Statistical Journal of the IAOS, vol. 24, no. 1-2, pp. 47-59, 2007