Supplementary memorandum by the Statistics
Commission
THE INTER-DEPENDENCE OF STATISTICAL WORK
IN GOVERNMENT
The statistical outputs of government departments
are, in the main, readily available on departmental websites.
Some, such as Census statistics, are well known and widely used
in decision-making. What is less well known is that many of the
thousands of sets of statistics and reports do not arise simply
from within one department. Rather they are the result of agreement
and co-operation between several departments, sometimes involving
shared funding, interchange of data and joint publication arrangements.
This network of interdependencies is not much
discussed in existing official reports although the annual National
Statistics Work Programme provides ample evidence of its existence.
As one example, the interdepartmental group of officials that
co-ordinates health statistics has input from, in addition to
the Department of Health, some 10 other departments and agencies.
NATURE OF
THE INTER-DEPENDENCE
This note illustrates the dependencies through
selected examples. These are particularly relevant at a time when
the Chancellor has announced plans to legislate to make statistical
work more independent from the rest of government.
The Statistics Commission has long argued that
existing statistical arrangements can be improved and we have
welcomed the Chancellor's announcement. We do, however, urge that,
in planning the changes, the statistical interdependency between
ONS and other government departments, and between these departments
and the devolved administrations and some other bodies too (Higher
Education Statistics Agency, Health and Safety Executive etc),
must be factored in to the final structure and reflected in the
roles of any new bodies established.
The nature of the interdepartmental co-operation
varies. Where permitted by law and ethical considerations, one
department may pass individual records (relating to persons or
businesses) to another department so that those records can be
combined with other data to produce a richer basis for statistical
analysisfor example NHS patient lists are used by ONS in
estimating population migration within the UK.
The transfer of raw records is, however, a relatively
uncommon form of co-operationalthough it may grow in the
future. More typically, two or more departments need to agree
on the form of data collection, technical definitions and standards,
the coverage of surveys, the timing of production, the transfer
of aggregate information, the form and timing of publication etc.
There is also some shared funding of statistical activitiessuch
as the Neighbourhood Statistics programme which is funded by both
ODPM and ONS.
Departments also need to co-operate in agreeing
a UK position on statistical matters to be taken forward by ONS
at the EU or wider international level.
Some examples:
Estimating the population:
ONS compiles and publishes estimates of the UK population. The
10-yearly Census provides the benchmark for these. ONS directly
manages the Census for England and Wales but the Census in Scotland
is conducted by the General Registrar Office Scotland; and that
in Northern Ireland by the Northern Ireland Statistics and Research
Agency. Putting together UK estimates requires close co-operation
between the three departments, and with the Welsh Assembly in
some aspects, to ensure a common approach and the compatibility
and consistency of data.
Estimating migration: Data
on migration flowsinternationally and within the UKare
important for tracking population changes, which in turn are fundamental
for resource allocation and policy analysis across government.
They are recognised as being unsatisfactory currently. Statistics
on international migration are published by ONS, which also manages
the International Passenger Survey, a principal source. Other
sources include the devolved administrations and UK departments
such as the Home Office, which compiles statistics on asylum seekers
and estimates of illegal immigration. ONS estimates of internal
migration make use of NHS records from all parts of the UK. Improving
estimates of migration is likely to require data from DH, HO,
ODPM and ONS and devolved administrations to be brought together
to create new estimates.
Analysing demand for social care:
To analyse future demand for social care for the elderly and others
in need, data from the Department of Health, DWP, demographic
estimates from ONS, and data from local authorities (via ODPM
and devolved administrations) need to be drawn together. Coherence
and compatibility of these data sources is an issue at present
(a forthcoming report by Sir Derek Wanless will refer).
Underpinning government targets:
Official statistics are widely used as the basis of hundreds of
top-level targets set by government. There has been criticism
in the past of targets set without adequate data being available
for monitoring. The Statistics Commission's forthcoming report
PSA Targets: The Devil in the Detail (for publication 20
March 2006) points to the need for joined-up planning to meet
current and future data requirements.
Monitoring agriculture: DEFRA
are responsible for compiling the UK agricultural statistics required
to comply with European regulations. The devolved administrations
collect and provide DEFRA with the relevant data for their countries.
Compiling National Accounts:
ONS put together the UK national accounts drawing in data from
other departments. For example, construction output comes from
DTI; agricultural output from DEFRA and devolved administations.
Monitoring public sector finances:
ONS and Treasury work together to produce, and jointly publish,
monthly data for public sector finances. HMRC (tax receipts),
ODPM (local authorities borrowing) and the Bank of England (banking
data) are also involved as data suppliers. Pulling the monthly
data together quickly requires close co-operation.
Tracking cancer incidence:
Cancer statistics for the UK are published by ONS. Data come from
nine regional cancer registries in England and from cancer registries
in Wales, Scotland and Northern Ireland. The registries need to
co-operate both with each other and with ONS and the devolved
administrations to ensure that the information is consistent.
Cancer statistics are an example of sharing confidential personal
data for statistical purposes.
Evidence for local areas:
The Neighbourhood Statistics service provides a wealth of small
area information for a wide range of indicators, covering topics
such as education, health and care, crime and safety, housing,
the physical environment, population and migration and aspects
of deprivation. The website is managed by ONS, but maintaining
it requires the active co-operation of many other government departments
including DfES, ODPM, DH and Home Office.
Monitoring the labour force:
The Labour Force Survey is conducted by ONS for Great Britain
(there is a separate LFS for Northern Ireland). Together the surveys
provide information on the UK labour market. In order to improve
the robustness of data for particular areas, ONS make a number
of targeted "boosts" to LFS sample size. These boosts
have been agreed as partnership projects with the relevant departments
which provide funding support in return for more robust data.
Partnership projects are in place with DWP and DfES for England,
with the Welsh Assembly and with the Scottish Executive.
Maintaining the Inter-Departmental
Business Register: This is a core index of UK businesses maintained
by ONS, which is used as a base for selecting samples for the
statistical surveys that underpin UK economic statistics. Information
for the IDBR comes from a number of sources including HMRC (VAT
payments and repayments, and PAYE collection). A number of departments
besides ONS use the IDBR for their surveys, including ODPM, DTI
and DEFRA and the devolved administrations.
Measuring government output:
The work to follow up the Atkinson Report on measuring government
output and productivity is led by ONS but involves a number of
departments and the sharing of key staff and expertise.
Addressing deprivation: The
Index of Multiple Deprivation is commissioned by ODPM, but constructed
by a research institution. Corresponding indices are commissioned
by the devolved administrations. This work involves the co-operation
of a number of government departments (including ONS, DfES, DWP,
DH, HMRC, ODPM, HO) in supplying local data covering income, employment,
health and disability, education, skills and training, housing
and services, environment, crime. The indices are used extensively
by local and central government to identify deprived areas. The
future of such indices will require ongoing central co-ordination.
English Longitudinal Study of
Ageing: This statistical survey is jointly sponsored and funded
by a number of government departments (DH, DWP, ONS, ODPM, DCM,
DfES, Treasury), as well as the National Institute on Aging in
America. Several departments have drawn on ELSA data in their
research activities eg ODPM in their work on social exclusion
and older people, ONS and DWP.
Work and Pensions Longitudinal
Study: The WPLS is an example of detailed data sharing between
two departments. It links benefits information held by DWP, with
employment records from HMRC. It is governed by the data-sharing
provisions of the Employment Act 2002. The individual data records
are confidential so there are restrictions on access and on the
purposes for which access is allowed. Examples of uses include
providing information on the effectiveness of Job Centre Plus,
and understanding the links between savings held and the benefits
in retirement.
Regional statistics: Implementing
the Allsopp Review of statistics for economic policy making will
require close co-operation across all government departments and
the devolved administrations.
May 2006
|