The modern history of UK economic statistics goes back to 1941 and the establishment of the Central Statistical Office (CSO) to take charge of developing national income accounts. Following the end of the war, the CSO’s work expanded and the Statistics of Trade Act in 1947 enabled more data to be created from business. However, by 1956 concerns about economic statistics were emerging. In his budget speech, the chancellor Harold Macmillan elegantly critiqued the timeliness of economic statistics.
'I am told that some of our statistics are too late to be as useful as they ought to be. We are always, as it were, looking up a train in last year's Bradshaw.'
Macmillan eschewed the temptation to set up a review and instead made a call for action.
'In a word, I feel sure that we can do a good deal, without embarking on a tedious investigation or setting up too elaborate a structure, to improve the machinery available to us.'
Following this intervention, quarterly national accounts began to be compiled in 1958 and the range of economic statistics improved.
The next external driver for change came in 1966 when the House of Commons Estimates Committee reported on government statistics, reflecting the desire at the time for a more scientific approach to policy making. A key recommendation led to the establishment of the Business Statistics Office (BSO) and the instigation of a range of short period statistical surveys to improve the supply of timely information on the economy.
Government statistics continued to grow and develop until the shock of the 1981 Rayner review. This review, driven by the desire to make government ‘smaller and more efficient’ and to reduce the ‘burden’ imposed on the private sector, established the highly damaging principle that government should broadly produce statistics for its own use. The wider economic value of data was not, at that time, recognised. The review directly led to a cut of 25% in staff numbers and an inevitable decline in quality. Moreover, it would take many years to re-establish the idea that statistics were a public good.
By the end of the 1980s, major concerns were again being expressed about the quality of the statistics. The strength of the economic boom had been misjudged, different economic measures all appeared to conflict with each other and the statistics were subject to large revisions. In June 1988, the chancellor Nigel Lawson set up a review under the leadership of Stephen Pickford, a Treasury official. Pickford's report included a range of recommendations covering the collection and compilation methods, along with further work designed to lead to improvements and organisation changes.
The most significant change to come out of the Pickford review was the absorption of the BSO and the statistical divisions of the Department of Trade and Industry into the CSO. Additionally, some statistical inquiries were given statutory status and increased samples sizes which more closely reflected current economic structures.
Not long after these changes, the ‘chancellor's initiatives’ brought further developments in 1990 and 1991. Instigated by John Major and Norman Lamont, these changes were designed to improve the measurement of key economic variables.
Things went quiet on the review front for a decade or so, probably reflecting the benign economic environment and the absence of pressures on public spending. Indeed, in the early 2000s, the new ONS (following the 2006 merger of CSO and the Office of Population Censuses and Surveys), secured additional funding to improve its systems.
However, in 2003 the chancellor, Gordon Brown asked economist Chris Allsopp to perform another review of economic statistics. There was no specific trigger for this, but the terms of reference included regional statistics and whether the statistical coverage reflected changes in the economy. Allsopp’s most visible result was the creation of the ONS regional office – wound down in 2011 when the funding ran out. Other changes appear to have been frozen as the funding pressures increased with the government looking to improve its administration efficiency. Meanwhile, current users of official statistics remain very critical of the availability of regional data.
At the same time, the ONS was also working out how best to use its budget for ‘re-engineering’. Staff within the national accounts lobbied for a methodological review which I eventually led. This review recommended some high level principles for compiling the national accounts and was partially implemented. Unfortunately, full execution was partly scuppered by the rapid staff turnover and administrative difficulties associated with transferring the work from London to the new office in Newport.
Criticism of the measurement of public sector output led to the national statistician appointing Tony Atkinson to lead a review of this topic. Support from Downing Street for this review gave the ONS some leverage to obtain data from reluctant government departments. But the review also resulted in a potential skewing of internal resources to measuring the public sector.
In the two or three years after 2007, core national accounts were relocated to Newport. This move resulted in the loss of specific knowledge and corporate memory as few experts were willing, or able to relocate. Possibly, but not necessarily as a result, mistakes were made in the production of key statistics. In September 2013, the ONS director general, Glen Watson, invited economist and former Monetary Policy Committee member Kate Barker to ‘review the user needs and methodology underpinning the national accounts and the balance of payments’.
Barker drew heavily on previous reviews – including Allsopp and Pickford – making a wide range of recommendations and suggestions. It's too early to see the results, but the current chancellor clearly sees the need for yet another review designed to further improve economic statistics.
Macmillan may have got it right in his call for action without embarking on a ‘tedious investigation’. The Bean review will doubtless have something useful to say. It should follow the example of Barker and look carefully at previous work. But it must also recognise that producing high quality economic statistics is a complex business.
The quality of these statistics requires a governance structure guaranteeing independence, excellent basic data, good methods and efficient processing systems. However, possibly the most important ingredient is a range of trained, highly skilled and knowledgeable staff.
- Reg Ward and Ted Doggett, Keeping Score: the first fifty years of the Central Statistical Office, London 1991
The views expressed in the Opinion section of StatsLife are solely those of the original authors and other contributors. These views and opinions do not necessarily represent those of The Royal Statistical Society.