Examination of Witnesses (Questions 200
THURSDAY 16 NOVEMBER 2000
200. Would it be public advice? Would you be
issuing a press release saying the Government are being challenged,
or the Minister is being challenged, on this, that and the next
(Sir John Kingman) Any advice we give is necessarily
public because Ministers have told us that we have to be transparent
in our proceedings. Whether we do it by press release or on our
website or by whatever means, that is a matter of tactics, but
the substance is that if we give advice to Ministers then that
advice will be public.
201. What will the Commission be able to do
to challenge the use of statistics by Government if it does not
think they are being used appropriately?
(Sir John Kingman) Our first priority is to ensure
that the statistical output that bears the brand of National Statistics
is something that we can all trust. It is inevitable that statistics
will be misused, if not by politicians then certainly by journalists,
and we may well find ourselves commenting on the misuse of statistics
but that is not our prime focus. Our prime focus is that when
something comes out with the brand of the National Statistician
on it, this is something that we can all believe. What use we
all make of it is more difficult to control and we certainly have
no power to control anyone. All we can do is to report and advise,
and that we shall do.
202. But if it were a blatant abuse, you would
(Sir John Kingman) I think we might well.
203. What would happen if a department decided
to engage in some survey to determine its policy priorities which
might be considered a dubious survey, it might be one which was
answered before it began? Is there any proposal for the Commission
to give a sort of quality standard mark on these sort of exercises?
(Sir John Kingman) I think you are asking very hypothetical
204. It is not all that hypothetical.
(Sir John Kingman) I have not had any evidence presented
to me that anything like that is going on. If it happened, we
would have to consider what our response would be.
205. Can I press you on that point because I
do not think it is enough for you to say that it is hypothetical?
There will be departments who will initiate polls and surveys
of different kinds as the monitoring of different policies. Now
I am not suggesting any malfeasance but they may not have the
best statistical background, they may not have used the best methodology.
Would you not have a role in scrutinising those to ensure that
the statistics which were then being produced by the Government
as a result of this conformed with the highest standards?
(Sir John Kingman) I think the primary role in that
respect lies with the National Statistician because the National
Statistician has been given a responsibility for the quality of
the whole of National Statistics and a professional responsibility
for the working of the whole Government's statistical service.
It might well be a situation where we might want to comment, positively
or negatively, on the working of the National Statistician in
that regard, but I think you are leading me into areas where as
far as I can see at the moment it is not clear there is a problem.
Of course there always may be a problem but I do not think you
should start from the assumption that there are going to be problems
of the sort you are describing.
206. Her Majesty's Treasury stopped publishing
a series on the tax burden earlier in this Parliament. Could you
envisage the Commission looking retrospectively at such decisions
and giving advice to Ministers on whether they could reinstate
such a series and publishing that advice?
(Sir John Kingman) I could imagine a situation in
which we might offer such advice, yes.
207. Could I ask you to briefly describe to
us how you will carry out your auditing work, perhaps by reference
to your plan to audit the adequacy of the National Statistics
to monitor the Government's National Cancer Plan?
(Sir John Kingman) Obviously the Cancer Plan is a
particular study which may well throw up general questions that
we will want to address in a more general context. We will want
to try to assess the systems by which the various statistical
series are produced and that means that we need to study the workings
of the Office for National Statistics and the various statistical
groups in the other departments and try to assess whether the
systems they are using are robust, whether the methodology they
are using is the best methodology available, whether they have
taken advantage of outside statistical advice and help, for instance
from professional bodies, whether the statisticians who are doing
the work are appropriately qualified and so on. Those will all
be factors that we will take into account in coming to a view
about whether the systems are reliable. Clearly there will always
be risk of failure in any system, just as when one is auditing
a company's accounts one wants to look into whether there are
appropriate precautions in place to manage the risks of failure
or distortion or whatever. That is clearly a demanding agenda.
It is one which we shall carry out to the best of our ability
and we will hope to come to conclusions where we can assure the
public that they can rely on the statistics that are being produced.
208. Do you see any danger of unnecessary duplication
or even harmful duplication in the work that you do and the work
that the National Statistician's Office is doing anyway?
(Sir John Kingman) We shall obviously depend on the
work that the National Statistician is doing both in the Office
for National Statistics and more widely, but we shall have to
form independent judgements on that. We clearly are not going
to look at the detailed calculations of every series. We are going
to look at the systems and try and assess the adequacy of the
systems. That will inevitably be to some extent on a sampling
209. When statistics go out under the National
Statistics brand is it your view that they should always carry
a "health warning" about their accuracy or the limitations
on their use?
(Sir John Kingman) I think I have already answered
that question by saying that we have already talked to the National
Statistician and asked him to think about ways in which the accuracy
and reliability of the statistics can be signalled in the publications.
This is something which is going to be done in different ways
depending on the sort of statistics that are being considered
and the sorts of uses to which they are put. For example, with
something like the average earnings index, clearly the Monetary
Policy Committee and other bodies that use those figures need
to know just how accurate it is. Is it to one decimal place or
two decimal places or no decimal places, for instance, whereas
with a lot of the social statistics that are produced it is much
more important to know whether the trends that are apparent in
the statistics are real trends, whether if something is increasing
and then starts to level off, that levelling off is a real effect
or is some statistical blip. That may be expressed in a quite
different way from the way in which you express the accuracy of
one of the key economic indicators. That is something we will
want to discuss with Len Cook and his colleagues to find a way
forward which gives the best information to the users, both the
professional users who may understand some deep statistical statement
and the more general user who will want something much more general
and much more easily understood without the benefit of a statistical
210. So to my question, is it your view that
they should always carry a "health warning", is your
answer, "It is a matter of horses for courses. Some do require
some help and some do not"?
(Sir John Kingman) I think my answer is yes, but the
sort of "health warning" will depend on the sort of
output and that is something we shall want to discuss.
211. In a submission we had from the Society
of Business Economists they commented that with the exception
of the financial sector the membership of the Commission does
not include representation from business and so they were worried
you would not have a sufficiently high consciousness of the needs
of business and the kinds of statistics that they make use of
in their businesses.
(Sir John Kingman) The White Paper made it very clear
that the Government had opted for a Commission which was not representative
but was a small Commission of people who had the ability to carry
out the functions required of Commissioners. Therefore the Commission
is not actually representative of anyone and is not intended to
be. There are only eight of us all told but the Commissioners
do have a very wide variety of experience and in my view the Commission
is a very strong one in terms of the personal qualities of the
Commissioners. The fact that we are not representative means that
we have to listen very carefully to the views that are expressed
by the sort of body you are now quoting, the various user groups,
but also groups that do not think of themselves as statistical
users at all but really are.
212. The Monetary Policy Committee faces similar
criticisms from time to time and does its best to overcome the
criticisms by going out and talking to businesses. Is that part
of your plan?
(Sir John Kingman) We shall certainly want to be sensitive
to the views of businesses. The exact mechanisms by which we do
this are still under consideration and we are very open to suggestions
from anyone who is interested in our work.
213. I suggest you go out and meet them.
(Sir John Kingman) Thank you.
214. Finally, on the question of wider public
opinion about statistics and whether there is confidence in them,
are there any plans for the Commission to try to measure public
confidence in official statistics?
(Sir John Kingman) Yes. This is one of the most interesting
questions that we face and it is something that we are discussing
with the National Statistician and his colleagues because they
have a similar interest in measuring the public confidence and
trying to get a feeling of how it changes over time. It is not
an easy problem. As measurement problems go it is a very complex
one and the last thing we want to do is to have unreliable statistics
about the public confidence in statistics and getting into a rather
ridiculous logical conundrum if you are not careful. It is very
important that we do it and we have this very high on our agenda,
to find a way which can be relied upon to measure the public confidence.
There are certain anecdotal indicators. When people quote what
is supposed to have been said by Disraeli that is a negative sign
of public confidence in statistics, but we need something much
more reliable as a measure of that confidence and whether we are
managing to improve the confidence over time.
215. Sir John, are there circumstances when
you might wish to comment on the adequacy of the resources in
(Sir John Kingman) Yes, I think it is quite likely
that we will find ourselves commenting on resources both for ONS
and more widely for the production of National Statistics, and
we may well want to comment on the cost effectiveness of statistical
activity and on the priorities that are being set. I do not think
we will necessarily always be arguing for spending more money.
We may be arguing for targeting the available resources more effectively
but we have not yet reached the stage where we have substantive
comments to make on resources.
216. Have you had some discussions on the impacts
of the Spending Review on ONS?
(Sir John Kingman) When we look at the systems within
ONS one of the aspects that of course will inevitably arise is
whether the systems are being starved of resources or whether
the money that is available is going in the right way in the right
217. So you have had some initial discussions?
(Sir John Kingman) We have had some initial discussions
on that and it is clear that there are very difficult issues,
for instance, around the provision of proper IT support for National
Statistics. These are certainly issues on which I would not want
to make comment without a lot more study but it is certainly within
the purview of the Commission.
218. So that is a priority?
(Sir John Kingman) It must be a priority to look at
the way the systems are working and if, when we look at the systems,
we think that there are resource problems we shall say so.
219. Back in March the Minister described the
Commission as guardians of integrity, independence and professionalism,
so should the Commission be auditing and commenting on the quality
of professionalism of the staff at the ONS and government statisticians?
(Sir John Kingman) I think that is certainly a material
factor. I have already included that in the factors we will be
looking at when we look at the systems in ONS and more widely
in the departments that produce National Statistics.