Introductory covering letter from Len Cook, National Statistician and Registrar General for England and Wales

Population statistics in the United Kingdom have long been one of the strengths of national official statistics, and the rich demographic detail that they provide continues to be valued immensely. However, the regular population estimates have been subject to strong criticism, particularly the population Census of 1991 and the estimates for subsequent years that were based upon it.

In common with other countries, we are at a crossroads in population estimation. The One Number Census for 2001 was one move in a new direction, but was only a partial solution to the challenges we face. I am pleased that ONS is today publishing "A demographic statistics service for the 21st century", which provides a full review of our population statistics system.

This review provides a plan for the coming decade that will provide us with a better understanding of the actual and perceived weaknesses in population measures. It will also ensure that we have a planned response to known and emerging limitations, and that the changing context in which information is used and gathered is well reflected in our actions. The review is based on the population statistics system in England and Wales, but will be relevant to the consideration of comparable statistics across the whole of the UK.

Population change over the next decade will be large indeed, as will the policy, community and personal responses to it. We need to anticipate and measure such effects as they occur.

The review identifies a variety of ways that we can improve our current population statistics operation. We need key statistics that are more timely and of better quality. We need a stronger and more definite recognition of the local dimension of these statistics, particularly relating to the accuracy of Local Authority population estimates. We must also take full advantage of opportunities we now have for improving the data sources, methods and systems used to estimate population between Censuses.

The review provides a basis for a broad but clearly-focused agenda, whose cornerstone is the development of data sources, methods and outputs that recognise the dynamic shifts in the population. These enhanced outputs will include greater analysis of ethnic populations and their characteristics, and an improved ability to count the population according to different definitions. We will plan the expansion of our analytical capability in a way attuned to users' needs, recognising the impact of demographic change on policy development and the modelling work that underpins it.

A key part of this work will be to make greater use of administrative data sources in compiling population estimates. We will exploit the potential of parallel developments: the International Migration Quality Review, the Citizen Information Project, the Civil Registration Review, the re-engineering of key ONS processes, and the ongoing development of Neighbourhood Statistics.

The various components of the population estimates system, which includes births, deaths, internal and international migration flows, household formation estimates and demographic parameters, do not have the same levels of quality. The quality of estimates can reduce greatly in moving from national level, to regional or local authority level. Sub-national population estimates are used for the distribution of public funding to local authorities and health organisations, as well as being an integral element of European Union resource transfers to regions with per capita wealth below set national levels. An important goal of our work plan will be to assure the quality of population statistics at local and regional level, recognising the differing enumeration challenges that areas can present.

A short note accompanies this letter, setting out what I see as the key action points to emerge from this review. This is an ambitious programme of work, and particular projects may depend on the future availability of resources (both in ONS and elsewhere) and on political and legislative developments. Nonetheless, I see this as a workable template for making the improvements that the review identified.

ONS recognises that the quality of population estimates in the latter half of the 1990s has been poor, something which was not recognised until quite major revisions were necessary. The revision process generated considerable user discontent and, in some cases, distrust. In retrospect, the lessons from the 1990s must be that we need to regularly reassess whether we have recognised emerging inadequacies, and that coherence among measures needs to be improved between censuses. The inadequacies of the current population measurement system are the result of a great variety of factors, and no one problem, or single resolution of that problem, will necessarily bring an end to it.

I see the publication of this review as a major step in restoring confidence and satisfaction in these key statistics. It provides a basis for a well-defined plan for the advancement of the system of population statistics for the UK. The community of users of population statistics (particularly census counts) will be consulted for advice and evaluation of the development agenda. I hope that this review will provide even the most critical users with an appreciation that we have grasped the substance of their concerns. We recognise that it is for ONS to address the issues raised. We have done so, through this review and through the studies that will conclude later in 2003, and we aim to respond constructively.

This is a time of remarkable challenge to orthodox practice. The growth of migrant flows, transient working and variety in household formation both provide challenges to our existing methods and yet need to be measured with increased reliability. Technology has enabled information to be accessible, and yet contemporary technology is more powerful than our capacity to use it to date. The information revolution has challenged long-established processes for birth and death registration. At the same we are working to develop a population register, as well as already using administrative records to revolutionise information about neighbourhoods. We also engaging with government colleagues working on the development of a national address register.

This review presents the thinking behind the various activities that will bring all this together over the next three to five years in order to deliver the improvements to population statistics which we know are needed.

Len Cook

From the National Statistician: Key action points for the Office for National Statistics

Report: A Demographics Statistics Service for the 21st Century

This page last revised: Friday, 1 August 2003