Examination of Witnesses (Questions 120
WEDNESDAY 21 NOVEMBER 2001
120. It seems sensible. When would you expect
a report on the Annual Report of the National Statistician this
(Sir John Kingman) We are bringing a further draft
of our response to the meeting of the Commission tomorrow. It
may be if the Commission agrees that we have reached a final form
that we shall get it to Ministers after that, but we are anxious
that our report is a well-considered one and that we have had
an opportunity to pick up certain points in the Annual Report
with the ONS before we make the formal response. It is more important
that we do that than we rush out with comment.
121. We were struck by the comment in your foreword
to the Annual Report 2000-2001 where you mentioned that you do
not see the National Statistician's report in advance of publication.
Is that a deliberate policy decision perhaps for clarity of sequencing?
(Sir John Kingman) It is a deliberate decision because
we feel that our independence depends on us being seen to be detached
from the National Statistician. If we saw the report in draft
and commented on it then we would to some extent be implicated
in that report and, therefore, our comments would not be wholly
independent. It is part of the point about openness that we should
see the Annual Report when the rest of the world sees the Annual
Report and we should not have any privileged position that involves
us in the drafting of that report.
122. I follow the argument that a reviewer of
a play should not be involved in the production of the play, but
do you not feel that it might be in the public interest to incorporate
any concerns that you might have before the report is finalised
rather than as a critic in the gallery after it has already come
(Sir John Kingman) I think your analogy is a very
good one. It would be unfortunate if the critic went along to
the dress rehearsal and said, "I will be able to write a
better review if you do not come in at stage left at that particular
point." I would stick to the view that we should see the
report when it is published and we should comment upon it in an
open way without having been involved in any way in its preparation.
Chairman: Can we come back to the issue of the
Code of Practice. James Plaskitt?
123. You said in reply to the Chairman that
this is possibly your biggest frustration. Let's explore that
a little bit further. What, in your view, is holding it up?
(Sir John Kingman) Of course that is something we
do not know, we can only speculate on. We have been told repeatedly
by the National Statistician that he is doing his best to get
the report cleared for publication and open to everyone to comment
on and I believe the National Statistician. We have been told
several times by the Economic Secretary of the Treasury and the
Permanent Secretary of the Treasury that it is not the Treasury
that is holding it up, therefore by elimination it is presumably
other Government departments. We can see the great importance
of getting full consent to the code of practice from all the relevant
departments because it is much more likely to work if departments
are committed to it. At the time when I said we were disappointed
that it had not come, I was convinced by this argument that it
was worth taking time to get that consent. At the time of our
Annual Report when we said that we were extremely disappointed
our patience was, as you can realise, beginning to wear thin,
but I think now we have passed the point at which it is reasonable
to say that things do take a long time in Whitehall and ministers
are very busy. I think we have now reached a stage where the non-appearance
of the code throws doubt on the commitment of the Government to
the new Framework for National Statistics and I hope that the
Government can take rapid action to prove that I am wrong in that
124. Do you think that without the code the
integrity of National Statistics is in question?
(Sir John Kingman) Yes.
125. You mentioned in your report examples of
where information had been trailed in advance of publication.
Could you just give us some examples of that and why it concerns
(Sir John Kingman) Before I do, can I just say a little
more about the non-appearance of the code?
(Sir John Kingman) We had a Green Paper and a White
Paper and a Framework Document and now we are waiting for the
code but we have been told that the code is not going to answer
all of the questions we would want to ask, that there will be
protocols associated with the code, some of which may come with
the code but some may come later. There is a dance of the seven
veils going on and the fourth veil is taking a long time to drop.
It is an unsatisfactory way of carrying the public along with
the need to build trust and confidence. Even when the code comes
we may still have to wait for protocols and, who knows, some of
the protocols may haveI do not know what comes after a
protocolsub-protocols or something. I am waiting impatiently
for the full picture to emerge. As far as disclosure is concerned,
you will know that there have been a number of occasions recently,
as there were in the more distant past, when the Today
programme on the BBC has contained news items that say "Statistics
to be announced today will say such and such . . ." That
does not seem to us to be the right way to build public confidence.
If they are going to be announced how does the Today programme
know what they are going to say and what the Government's response
is likely to be? I think it is important that that be tightened
up and we are looking to the code of practice to contain strong
provisions that will avoid that sort of premature disclosure and
127. Do you think that it is your business to
get involved in the timing of Government announcements? I thought
you were simply supposed to be looking at the integrity of the
statistics. When it is announced does not really imply any question
about the integrity of the statistics, does it?
(Sir John Kingman) Yes, it does. I think the code
will probably contain the arguments for that. This is something
that Len Cook feels very strongly about and I should let him make
that case, but it is a case that convinces us that it is important
that the statistics should be announced at the time when it was
said they were going to be announced so there is no suspicion
that they are being announced in order to gain some particular
political advantage. It is important that the announcement should
be the time at which everyone knows about it and not several hours
128. Is a statistic which is, therefore, by
your definition sound at three o'clock in the afternoon unsound
at eight o'clock in the morning if it is still the same figure?
(Sir John Kingman) It may be the same figure but it
may have a different gloss put on it if the announcement is made
other than by the professional statisticians.
129. That is all an issue of politics, is it
not, and communication, it is not part of your brief?
(Sir John Kingman) I think the way in which statistics
are communicated and the gloss that is put on them is part of
the issue of integrity. We are not concerned to say that a certain
statistic should be disclosed at this time or that time, but I
think the Commission would support the National Statistician strongly
in saying that there should be a clear protocol about the way
in which the statistics are announced and that protocol should
be adhered to.
130. Are you confident that the code, when it
emerges, will contain guidance on these issues?
(Sir John Kingman) If it does not then it will have
been emasculated at some point because certainly Len Cook will
have written a draft which covers those points. If they are not
there they will have been taken out somewhere and I hope you will
ask people who took them out and why.
Mr Plaskitt: I note that, thank you.
131. Can I pursue this a little. Sir John, you
have been around Whitehall before and when you said that you were
pretty sure the Treasury had not been delaying it, you said it
was other Government departments. Is it a collective lot of Government
departments or one particular department do you know?
(Sir John Kingman) I am sorry, I cannot speculate
132. You have no idea why this delay is there?
(Sir John Kingman) I have no idea where the delay
is occurring or why it is occurring. If I knew which department
was delaying it I might go and talk to people in that department.
133. We spoke earlier about inflation and the
deflators and in addition to that the issue of the average earnings,
an issue you have rightly also raised which is a matter which
concerns us as well on this Committee. Can you tell us how work
on that is progressing?
(Sir John Kingman) We are at an advanced stage in
producing advice to ministers on the way in which the Turnbull
King Report has been implemented and it would be wrong of me to
prejudge what we will say to ministers, but you will know as soon
as we have made the report to ministers.
134. You certainly cannot trail an announcement,
(Sir John Kingman) No, of course not.
135. May we assume it is imminent?
(Sir John Kingman) Yes.
136. Thank you.
(Sir John Kingman) The reason it has not come before
is that we thought it right to check all the facts with the people
in ONS because there is no point in our making criticisms which
are based on a misapprehension of the facts. Therefore, we are
in the operation of clearing all the factual statements in our
response with the people who know the facts at first hand.
Chairman: Thank you, that is helpful.
137. The Office for National Statistics announced
significant data revisions to the National Accounts in September.
Have you had a chance to assess the significance of these revisions?
(Sir John Kingman) It is too early for comment on
the specific revisions but can I perhaps explain what my approach
is to those revisions in general. We have pressed the National
Statistician to indicate when figures like this come out what
is the area of uncertainty in those areas, not just statistical
sampling error but also uncertainty because of the underlying
assumptions. If you were to do that you would find that the area
of uncertainty in the first announcement was relatively large
and that what we call revisions are actually reductions in that
area of uncertainty which would normally stay within the original
area, so the area would reduce but might, as it were, move to
the right or to the left within that area of uncertainty. That
seems to me to be perfectly legitimate but it would be much clearer
to the users if it were accompanied by a clear statement of the
way in which the area of uncertainty was being reduced as successive
revisions were being made.
138. As we have already ventured into the area
of interpretation of statistics, do you not feel that there would
be a risk that those with less pure motives than ours perhaps
would select one end of the range in that case to give a distorted
view and that might be a significant deterrent to doing what you
(Sir John Kingman) Yes, that is one argument against
indicating an area of uncertainty. Another argument is that simply
to admit uncertainty might cause less sophisticated users to query
the value of the statistics at all. Of course, all scientists
know that a numerical result is useless without some indication
of how accurate it is. We have to educate the users of statistics
to understand that there is an element of uncertainty in all statistical
information and that to know how large that area is is an essential
part of understanding and using the statistics.
139. Would you say that the recent revisions
were within the normal range which you would have expected?
(Sir John Kingman) I think that is something we would
have to look at and come back to. I would not give a snap judgment
3 Note by witness: In fact this report will
be addressed to the National Statistician initially. Only if the
Commission is not satisfied with his response will it approach
ministers on the specific issue. Back