Examination of Witnesses (Questions 1-19)|
25 JUNE 2009
Q1 Chairman: May I extend a very warm
welcome to our witnesses this morning. I hope the other two witnesses
do not mind if I say an especially warm welcome to Karen Dunnell
who has become Dame Karen Dunnell since we last met and richly
deserved. We congratulate you very much. Then we have Glen Watson,
the Census Director and Graham Emmons from the contractor Lockheed
Martin UK. Thank you all for coming. I do not know whether you
were expecting to say something by way of introduction but let
me ask a general question which will give you a chance to say
what you were going to say, if you were going to say it, but if
you were not going to say it, it will not matter because you can
say something else. Here we are talking about the build-up and
preparations to the 2011 Census. You have been working on this
for some time, you have issued a White Paper about it and you
have run test exercises. What I should quite like to start off
with is an overview. Given all that background, where are we now
in terms of this process?
Dame Karen Dunnell:
The Census is a very long-term project and the largest statistical
project that we do in the UK. At the moment it is going extremely
well. It is being planned in great detail. As you say, we have
done the test; we have also awarded all the key contracts for
the work because the contractors are working on the rehearsal,
which is the next big milestone, taking place in three local authorities
in the autumn. Once the results of the rehearsal are assessed,
it is full steam ahead for March 2007.
Q2 Chairman: Let me just ask this
question which is designed to provoke you. I suppose I am looking
at Lockheed Martin in particular. How can we be sure that we are
not going to have a great catastrophe on our hands as we have
had in some other projects? This is a mega project, big money,
hugely intricate organisation. You are doing all the IT stuff,
are you not? It is not unknown for these things to go wrong. What
I want to know is how you can assure us that all that is not going
to go wrong. If things do go wrong, who is going to be responsible?
Mr Emmons: First of all, we have
a huge amount of experience of doing censuses, having done the
US Census in 2000, we did the UK Census in 2001 and we did the
Canadian Census in 2006. We are currently working on the US Census
for 2010 as well as the Canadian Census for 2011. So we have a
huge breadth of knowledge and experience. By far the largest part
of the work is processing all of the paper questionnaires and
in that regard our system has been developed and enhanced literally
over a ten-year period and very, very well proven. That is very
well established technology. In everything we do a huge amount
of independent checking verification. Right now we are working
through very detailed tests and acceptance for rehearsal and the
rehearsal itself will provide a lot of assurances. We have plans
for compliance with business continuity disaster recovery procedures,
again independently vetted and verified so everything that could
go wrong is being assessed and analysed and we get to trial run
that during the rehearsal.
Q3 Chairman: That sounds like a pretty
good guarantee but we will come back to that a bit later because
somebody will want to ask you about the Gateway Reviews. The costings
seem to have gone through the roof, do they not, compared with
2001? Are they not one third higher in real terms? People might
wonder why that could be.
Dame Karen Dunnell: Yes, they
are one third higher in real terms and there are several reasons
for that. One is that of course society has become more complex
and we are putting several things into place which cost more money
to deal with that. The first thing is paying much more attention
to having a really good address register. Because we are adopting
a post-out method for over 90% of addresses in England and Wales,
work on the address register is absolutely fundamental. Also,
because there is a very, very large user demand for extra questions,
we actually have a longer questionnaire this time which adds quite
considerably to the cost. We have four pages rather than three
pages of questions for each individual which cost more. We are
also putting a lot more effort into the field operations, having
a system to monitor what is going on in the field in real time,
which will mean that we can get a better response, particularly
in difficult areas. Also, we are spending more resource this time
on things like engagement at local level with community groups
and local authorities and, of course, because we have very many
more groups now in our society, we have to spend more on supporting
people who do not speak English. There is a whole set of reasons
why the costs have gone up again.
Q4 Chairman: Does it mean that we
are going to get the equivalent additional value in output terms
or is it simply that it costs more to do the same thing for the
reasons you have described?
Dame Karen Dunnell: There is a
little bit of both. It certainly costs more to do because of the
reasons I have given, but last time we were criticised by several
local authorities for not getting their estimates right. We had
to do an awful lot of work afterwards to put that right and that
is exactly what we are trying to avoid. The key purpose of the
Census is to get a very, very accurate population estimate at
local and national level and that is our key goal.
Q5 David Heyes: This is quite a massive
outsourcing as far as ONS is concerned. I think it is true to
say that it is bigger than anything you have ever done before
Dame Karen Dunnell: Last time
we did outsource all the processing to Lockheed Martin and this
time we decided, because of the success of that and because we
regard it as providing best value for money, to outsource rather
more this time.
Q6 David Heyes: Is it true that there
is experience within ONS of managing large-scale outsourced contracts?
Dame Karen Dunnell: Yes.
Q7 David Heyes: Can you describe
to us how that works?
Mr Watson: Yes, certainly there
is experience within the Office and where we are short of experience
we bring in consultants and contractors to assist us and to make
sure that we can manage these contracts very well.
Q8 David Heyes: So you have contractors
Mr Watson: The contracts each
require a contract manager once they are placed and some of those
contract managers at the moment are contractors themselves. That
is just a mechanism for bringing in the best expertise possible.
All of the contracts that we have placed so far have been done
on time roughly and within budget, which is quite remarkable.
So far the contracts which have been placed, including the Lockheed
Martin UK contract are going very well and all of the milestones
are on track and things are progressing very well. Some of those
contracts still have to be placed and there are various procurement
processes which are still underway, but we are confident that
we will have all the contracts in place that we need for the October
rehearsal and then by the end of this year we will have all the
contracts that we need in place for the 2011 Census.
Q9 David Heyes: I guess perhaps the
second largest contract is the Capita contract for recruitment
and training of staff. The numbers have been differently reported,
depending on where you look. What are the numbers of enumerators
who will be recruited and trained through Capita?
Mr Watson: The total number we
expect to pull in is around 30,000 to 35,000. That is down from
2001 for various reasons which I would be happy to share with
you. As you say, the contract with Capita for recruitment, payroll
and training of that temporary field force was placed in March
this year and is progressing well.
Q10 David Heyes: On the face of it,
it looks odd that the numbers should decline so drastically and
that enumerators should be down so drastically in a more complex
environment. When costs are going up significantly, you would
expect that that level of reduction in enumerators, which is potentially
one of your biggest costs, would save money. Those seem to be
inconsistent. Can you explain that?
Dame Karen Dunnell: The key reason
for that is because we believe and we have tested and proved that
the post-out method actually works. Last time all the questionnaires
were hand delivered by enumerators and people were asked to post
them back. That was extremely successful so this time we have
tested the post-out as well, which if you do it avoiding the most
difficult places we have shown in the tests that works very well.
So you do not need to employ lots and lots of enumerators to deliver
the forms. Because we are going to have a much more sophisticated
monitoring system, we can deploy the enumerators we do have in
a much more focused way.
Mr Watson: In 2001 the bulk of
the effort of the 70,000 enumerators was devoted to delivering
the Census questionnaires. In 2011, far and away the biggest proportion
of the effort of the 35,000 enumerators will be to follow up non-response.
In fact, in response to the quality problems which were well reported
after the 2001 Census, what we will be doing is putting in something
like three times as much person power to follow up non-response
in the field. So although the number looks smaller, there is actually
an increase in effort in the important follow-up period.
Q11 David Heyes: I am always a little
anxious when I hear of Capita being involved in projects like
this, especially new projects. Rightly or wrongly they have an
unfortunate record. It is only two months since Capita sent out
a letter to the parents of a child who had died complaining about
their non-attendance at school. It was not Capita who apologised
for that awful thing; it was the school that had to make the apology.
That is a huge leap to the scale of what is going on in ONS but
how are you going to deal with a potential failure of that sort
and the risk. What sort of failsafe do you have?
Dame Karen Dunnell: Because the
Census is a very, very large project, we have identified a large
number of risks which we are managing in the normal way. Obviously
I do not know anything much about your particular episode, but
what we have to do and what I am sure the contractors also do
and we will insist that they do is identify all the possible risks
and we will manage them in the best way that we can. Of course,
when things like that do happen, at the end of the day the National
Statistician is responsible and we have to manage that with the
media. We are working much more effectively, identifying risks,
having much tighter management arrangements, so that we can actually
mitigate those risks much better than before.
Q12 David Heyes: I used that illustration
really because it does fit; people are very sensitive about revealing
this personal information, especially in the light of events over
the last couple of years where large-scale leakage of information
has occurred. That is going to make the operation of this Census
more difficult than any previous one. That kind of sensitivity
or potentially lack of sensitivity could become a major issue.
Another angle on that would be concerns about the involvement
of another major US defence contractor handling this kind of sensitive
Dame Karen Dunnell: There are
two separate things there. On the first one, the Census, we are
not approaching individuals. We do not actually have a list of
individuals, we are approaching addresses and it is up to the
householders at the address to fill in all the details about the
individuals who live there. We are not running the kind of risk
that you identified of actually writing to somebody who has died.
That is not a risk: it is all address based. That is the first
thing. Of course there will still be issues around that. All the
security arrangements and all the communication about how secure
and confidential the information is that people provide are essential
elements of our publicity work that we will have to do, which
of course we will also be rehearsing in the autumn. That is a
very, very important part of the work. Do you want me to come
onto the next part about Lockheed Martin and the Patriot Act?
Q13 David Heyes: Yes.
Dame Karen Dunnell: We have assured
that US Security Services cannot obtain Census data and that has
been done in several ways. No US firms will have access to Census
data. The contract is with Lockheed Martin UK, so that protects
all of that. Also, there are very complex and firm security assurances
that anybody working on the Census will have to sign and those
things are built into all the contracts that we let and of course
we have a 200-year history of doing censuses and making sure that
those data provided are actually only used for statistical purposes.
It is a very, very challenging part of what we have to do.
Q14 David Heyes: Does what you have
just said amount to a guarantee of no leaks of sensitive personal
data during the Census process? Does it amount to a guarantee
that it will not go wrong?
Dame Karen Dunnell: It is impossible
to guarantee that an individual form could not get lost for some
reason. It is impossible to provide a 100% guarantee but we have
put in place everything to hold the information that we have securely
and we will also, (and we are planning at the moment) do an independent
review of all the security and confidentiality arrangements to
reassure the public in advance that everything possible is being
Q15 Chairman: Just on the post-out
issue, because obviously this is a major new component of what
is happening, I think I am right in saying that when you did the
test in 2007 the effect of posting out most forms was that you
said the completion rate would fall by two point eight percentage
points. I should quite like to know why you felt a fall of that
kind was worth it in terms of the process; we are not going to
go into the questions but not least because you rejected the income
question on the basis of that 2007 test. Again, that was only
because there was a drop of two point seven percentage points.
What I really want to know is, if one drop got it out, why did
one drop keep the process in?
Dame Karen Dunnell: The drop in
the postal thing one can compensate for by sending enumerators
round. We cannot compensate for the effect of the income question;
that is lost for ever as it were.
Mr Watson: Yes, the money saved
by adopting a post-out model and contracting with the Royal Mail
for delivery of questionnaires allows us to put more effort into
the follow-up operation and that transfer of resource allows us
the opportunity to compensate for that drop in response rate.
Q16 Chairman: So if you got a really
serious drop in the response rate, even beyond what the tests
suggested, you are saying that all you would do is put more people
in to do follow up work. Is that right?
Mr Watson: Yes. Of course the
test did not test the internet collection option either, so that
is another mechanism which allows us to get some additional response
from people who might not otherwise have done so. Yes, that is
right broadly. Previously, we put three times the effort into
the follow-up operation that we did in 2001, so we felt that was
recoverable. The income question is rather different. It was not
just the overall drop in response to the income question; it was
the drop in quality and the drop in completion for those people
who did respond and also through our testing, through our feedback,
the extent to which people objected to the question. There were
other considerations on the income question.
Q17 Chairman: How much are you going
to save by sending it out as opposed to delivering it?
Mr Watson: I believe that the
figure we calculated at the time was of the order of £20
million. That is a figure we would probably want to go away and
Q18 Chairman: So it is not an entirely
Mr Watson: I am just saying that
I do not have it at my fingertips.
Chairman: We need then to think about
how this sending out is going to work and that means thinking
about this address register.
Q19 Mr Prentice: The census form
goes to individual addresses. Will there be someone, a designated
person in each household who will have to fill in the form, who
will have legal responsibility, as is the case with electoral
Dame Karen Dunnell: Yes, the householder.