INTRODUCTION
AND LIST
OF CONCLUSIONS
AND RECOMMENDATIONS
1. The 2001 Census was conducted throughout the United
Kingdom on 29 April 2001. The Office for National Statistics (the
Department) was responsible for the planning and conduct of the
Census in England and Wales. Responsibility for the Census in
Scotland and Northern Ireland rests with the appropriate Registrars
General.
2. The Department agreed a budget of £207 million
with HM Treasury for the Census in England and Wales. Although
the Census is not yet complete, with further Census results and
reports scheduled for release throughout 2003, the Department
estimates that the Census will be delivered within budget.
3. Planning for this Census began in 1993. The Department
introduced a number of innovations in the 2001 Census to improve
the quality of Census outputs and to achieve better value for
money. For the first time in a United Kingdom Census householders
were asked to return completed Census forms by post, allowing
the Department to reduce the number of field staff employed. The
Department also used electronic scanning and coding technology
to record all the information included in Census returns. It also
conducted a post Census coverage survey to enable it to adjust
the Census results for the estimated 2% of the population not
recorded in the original count.
4. The Department published the first outputs from
the Census on 30 September 2002. These showed an estimated population
of England and Wales of just over 52 million—some 900,000
less than the Department's most recent midyear population
estimate at June 2001. The main national and local Census results
will be published in 2003.
5. On the basis of a Report by the Comptroller and
Auditor General, we examined how well the Office for National
Statistics had managed the outsourcing of the 2001 Census, the
problems with field staff pay, the outcome of the use of post
for returns (post back), and the accuracy of the population figures.
We draw the following conclusions from our examination:
- Outsourcing helped the Department to introduce
some important innovations to deliver the 2001 Census. The post
back allowed the Department to reduce the number of temporary
field staff employed on the Census by some 33% compared with 1991.
The introduction of electronic form scanning and coding meant
that all information submitted on Census returns was recorded.
- In appointing Vogue Consultants as procurement
advisers for the Census, the Department failed to follow its own
procedures by initially engaging and extending the appointment
of the consultants without competition and by entering into a
contract that did not meet its own standard terms and conditions.
It also failed to seek authoritative advice on the requirement
to advertise the procurement and publish details of the award
in Official Journal of the European Communities. The Department
paid Vogue £485,000 in fees before the irregularities came
to light.
- The Department encountered a number of problems
in its outsourcing of the contract for the field staff pay. The
competition was limited and the Department did not sufficiently
assess whether it was procuring the payroll service it needed.
Nor did it assess the ability of field staff to deliver payroll
input forms to the standards required under the contract. It did
not react to warnings by internal audit about potential problems
in the operation of the Census payroll system and the contract
was not let in time to ensure that the proposed approach could
be tested as part of the Census dress rehearsal in April 1999.
As a result of these failings the Census payroll ran into serious
difficulties; payments to 23,000 field staff (30%) were significantly
delayed and 2,800 were overpaid by some £500,000.
- 88% of Census forms were returned through the
Royal Mail, exceeding the Department's forecast of 70%. The Department
had planned on the basis of a single estimate, drawing on the
experience of other countries, and was not adequately prepared
to take advantage of the higher rate of return. The Department's
field information systems proved to be inadequate and prevented
it from assessing the progress of the post back in 2017 Census
districts immediately after the Census. This failure to realise
savings from the post back and the decision to keep the staff
in the field during May contributed to some £11.3 million
being spent which could have been avoided. When managing complex
field operations, Departments should ensure they receive the necessary
information to adjust the project to changes in circumstances.
- The estimated response rate of 98% implies that
the Department may not have obtained responses from some one million
people. The Department undertook an indepth survey to provide
it with a statistical basis for adjusting the results of the count
to arrive at more accurate estimate of the overall population.
But in some areas response rates were much lower than that achieved
nationally. The response rate in ten Inner London boroughs was
below 80%, which has led to some local authorities to question
the results for their areas. A key aspect of the 2001 Census was
to target field staff at the hard to count areas. The Department
should examine why in some areas it nevertheless failed to bring
the response rates anywhere near to the national response rate.
It should also explore alternative approaches in the resourcing
of enumeration teams in Inner London.
6. Our more specific recommendations are as follows:
On the outsourcing of services
(i) Contracts for professional advice on
major procurement projects, including those for specialist information
technology procurement, should be subject to competitive procedures.
(ii) Departments should ensure that those
in charge of procurement understand their responsibilities and
manage procurement activities in accordance with departmental
and Treasury procedures. Senior managers must also exercise sufficient
oversight to ensure conformity with these procedures.
On the field staff payroll
(iii) Where departments encounter problems
in establishing an appropriate bidding field, as was the case
in the contract for the Census payroll, they should examine whether
the planned procurement strategy remains viable. Departments should
not retain bids from those contractors it has assessed as not
having the capability to deliver the required services in order
to maintain the appearance of competition.
(iv) Where the service delivery is subject
to uncertainties, for example in the case of the Department's
ability to deliver error-free forms for input to the payroll,
the assumptions underpinning contract terms and service performance
should be rigorously assessed and tested in advance.
On the post back
(v) The tight timetable set for the conduct
of the Census coverage survey contributed to the pressure on the
Department to complete the enumeration by the end of May. The
Department should assess whether the timetable allowed sufficient
time to complete the post back and conduct any further enumeration
necessary to chase outstanding forms.
On the accuracy of population figures
(vi) When the Department carries out its
review of the 2001 Census, it should consider whether the Census
coverage survey was sufficient to identify underenumeration,
or whether additional resources in the hard to count areas would
have helped to increase the overall response rate and the accuracy
of the count.
(vii) In the light of the Census results
showing that the Department's midyear population figures
for June 2001 were overestimated by some 900,000, the Department
should consider what steps might be taken to improve the systems
to track outward migration from the UK and avoid similar errors
in estimates of the population made in between Censuses.
THE
OUTSOURCING OF
THE CENSUS
7. The Department's budget for the Census included
£84 million for outsourced services. The Department has spent
£69.1 million on Census contracts, of which the contract
with Lockheed Martin for processing Census forms accounted for
£54.3 million. In addition, the Department paid to the Royal
Mail £7.7 million under the service level agreement for the
post back and spent a further £2.5 million on professional
advice.
8. The Department aimed to deliver better quality
data for local communities from the 2001 Census than from previous
Censuses by using electronic data processing technology to capture
and code all the information received. As it had neither the skills
nor the technology to deliver this aim inhouse, the Department
invited tenders from firms who were implementing similar solutions
for the United States Census. The Department estimated that its
contract with Lockheed Martin gave a saving of £15 million
against its public sector comparator. The Department believed
that the procurement and management of this contract had gone
very well, but difficulties in delivering the processed results
had contributed to a five week delay in the publication of the
first population estimates. While the Department had yet to conclude
negotiations on outstanding claims from Lockheed Martin, it remained
confident that the Census would be delivered within the overall
budget.
9. Post back arrangements were introduced for the
first time in a United Kingdom Census. This allowed the Department
to reduce the field force from 115,000 used in 1991 to 73,000
in 2001 and, in its judgement, made a major contribution to the
achievement of better value for money in the 2001 Census. It estimated
that the 2001 Census would have cost 25% more than the 1991 Census
if it had been conducted using the same methods.
10. The Comptroller and Auditor General found that
the Department's strategy for procuring other Census services
was generally sound. Contracts, other than those for professional
advice, were subject to competitive tender. However, the Department's
procedures for appointing external advisers and a contractor to
handle field staff pay fell short of normal public sector standards.
