Examination of Witnesses (Questions 1-19)|
9 NOVEMBER 2005
Q1 Chairman: Karen Dunnell, may I welcome
you to your first appearance before this Sub-Committee. Would
you introduce yourself formally and your team, please?
Ms Dunnell: Thank
you. I am Karen Dunnell, the National Statistician and the Registrar
General for England and Wales. This is Hilary Douglas, our Chief
Operating Officer; Dennis Roberts, the Deputy Registrar General
and Executive Director of the Office; and Colin Mowl, the Executive
Director of Macroeconomic Statistics and Labour Market Statistics.
Q2 Chairman: Thank you very much and
congratulations from us on your appointment. Len Cook, your predecessor,
set out in the Annual Report we are looking at today six key areas
you are probably familiar with. Do you agree with those or do
you have your own?
Ms Dunnell: No; I am very much
working now with the Office to, what I call, consolidate Len's
inspiration. We are moving forward with the programme, as it appears
in the Report, focusing first on delivering our very large programme
of statistics which has to change constantly in response to new
demands and the environment and then, very importantly, carrying
on with and completing our modernisation programme. We also, of
course, have challenging targets to meet on both efficiency and
relocation and the other most important thing is to do more work
with the Government's Statistical Service to strengthen independence
and trust in statistics right across government.
Q3 Chairman: Right. Just to be clear,
his six challenges are still your six challenges.
Ms Dunnell: Yes, indeed; we have
not changed them. We only changed the way we approach it, but
we have not changed the focus of the office, no.
Q4 Chairman: Fine. Now we will certainly
come a little later to the modernisation programme that you referred
to, but as well as being appointed National Statistician you have
also become Registrar General and you are head of the Government
Statistical Service. Do these three jobs really go together? Are
the other two not going to distract you from the core task of
running the ONS?
Ms Dunnell: I think most people
who are appointed head of government departments have very wide
ranging portfolios. This one has been in existence as it is since
1996, but certainly the combination of a large portfolio of statistics
and the role of Registrar General has been in place for 100 or
more years so I have a very strong team of directors; we share
out the work between us. That was the job I applied for and got
and so we have plans to take it forward.
Q5 Chairman: You are ready to advise
on Royal weddings if required?
Ms Dunnell: With my colleagues,
Q6 Chairman: When we saw the Statistics
Commission, they suggested to us that the code of practice you
operate under is rather ambiguous and aspirational and the day
to day interpretation of it can vary quite widely across government.
Does that reflect your own experience?
Ms Dunnell: The code of practice
was revised, as you know, after National Statistics came into
place in 2000 and some of it has been in use only for a couple
of years. It is a code of practice in that it is guidance and
I believe it probably needs a bit more time to settle down and
for people to experience putting it into practice, but we do accept
that there are areas which are probably difficult or ambiguous
for people to interpret. One of the things I am going to be doing,
starting soon after this week, is to visit all government departments
and Chief Statisticians and get a bit more to grips with the problems
that they are having with the code.
Q7 Chairman: Thank you. I think my colleagues
may well wish to pursue that. But finally, from me, may I ask
you about the classification of PFI liability? It is something
we discussed on many occasions with your predecessor, who wrote
to the Financial Times in May this year saying that the
timescale for the work of analysing all this is determined by
the complexity of the statistical issues, and that, when all the
necessary quality assurance has been completed, you will then
release the figures in accordance with the relevant protocols.
How is that work progressing in terms of calculating the debt
associated with PFI contracts and when do you expect to complete
Ms Dunnell: May I just say something
generally about classifications, and then I will pass over to
Colin? As you know, all of our classification decisions are taken,
along with the rest of our national accounts work, in line with
international guidelines and procedures and all this is done in
a very rigorous and open fashion. However, we do, I think, feel
now that we probably need to develop a protocol, part of the code
which explains the whole process and the criteria and that we
would then incorporate in our Annual Report where we have got
to in the year, which ones we have done, which ones are under
way and which ones we see coming up. The PFI is very complicated
and Colin will tell you a bit more about the progress we are making
Q8 Chairman: It would appear, from that
answer, that you rather agree that the classification should not
be left as an internal process that has no particular timescale,
that the public do need to be told what you are looking at and
when you might expect to finish it.
Ms Dunnell: Sometimes we know,
because we scan the horizon, as it were, what things might be
coming up, so we have some idea about what might be coming up.
Colin will explain why some things take rather longer than others,
but setting this out would probably be very helpful.
Q9 Chairman: Colin?
Mr Mowl: Just on that last point
Chairman, there are two types of classification issue we consider.
One is where there is a developing policy proposal, where we are
typically invited by the Treasury, possibly on behalf of another
department, to examine a particular proposal and give them an
opinion and there is a service level agreement with the Treasury
that covers that. The second type of classification issue is where
we decide that we need to look at a particular organisation or
transaction that has already happened, just to make sure that
we have classified it correctly. So there are some where we respond
to a Treasury invitation, some where we take our own initiative.
As regards PFI, this is a very complex area. We have got quite
good coverage in the National Accounts and Public Finances for
most aspects of PFI but, as you rightly say, there should be an
allowance in the government or public sector net debt figures
for PFI. The reason this is a complex area is that there are,
as I am sure you know, about 700 PFI schemes, each with its own
individual characteristics. It is only the on-balance sheet schemes
which should be covered in public sector net debt, and by on-balance
sheet schemes I mean those that are accounted for in the Department's
resource accounts and it is government accountants and the NAO
that decide whether a scheme is on-balance sheet or off-balance
sheet. With on-balance sheet schemes it is quite right that in
principle there should be an allowance in the debt figures. The
reason it is not entirely straightforward is this debt is not
like any other government debt; it is not like conventional debt,
a gilt-edged security; we cannot look at the debt instrument and
read the value of the instrument from the piece of paper. This
is a notional calculation. We have to impute a loan from the private
sector partner to the Government. We then have to imputethis
is a statistical term, which means we make an estimate; it is
an estimate, rather than hard informationa loan in the
first place; we then have to impute a profile of repayments, both
capital repayments and interest. Now, previously there has not
been an obvious data source, a direct data source, we could use
for this. We have now, with government accountants, identified
such a data source. We are going to make use of material that
has been produced for the whole of government accounts. We have
just started a programme of work and we expect to make some initial
estimates by next summer. They will not cover all on-balance sheet
PFI schemes, but they should cover, if this work is successful,
a substantial proportion.
Q10 Chairman: Fine. When will the protocol
Ms Dunnell: We will start working
on it pretty soon.
Q11 Lorely Burt: I want to talk about
the issue of trust. Before I do, I just wanted to pick up Colin
Mowl's point about the PFI off-balance sheet. I think Ms Dunnell's
predecessor said that it would take about six months to get the
PFI figures accounted on the balance sheet. I just wondered whether
that is on target?
Mr Mowl: Yes. Six months was the
time we estimated, or judged, back in May, that we would need
from when we established a method and a data source to completion
of the work. We have only recently, in the last month or so, got
to that stage, which is why we are saying, as a guide, that next
summer is what we are aiming for, so perhaps a little more than
six months from now but, from start to finish, six months; perhaps
seven or eight. So we are still on that sort of track, but that
was six months, not from May when we made that statement, but
from when we established a method and a data source.
Q12 Lorely Burt: Also, talking about
on-balance sheets, on 27 May The Independent claimed
that £21 billion of National Rail's debts would not go on
the balance sheet, and that was done by re-classifying this as
a private organisation. Could you comment on that?
Mr Mowl: Yes. With respect to
that newspaper, that is not a new story. This is a classification
decision we announced in 2002 and, according to national accounts
conventions and guidance, we judged that Network Rail was a private
sector company because we judged that it was not under public
sector control. We have kept that decision under review. If we
get new information, we will look at it again, which is what we
did, I think it was in February 2004 when we received some new
information about the directors' remuneration scheme and, as a
result, for a period of nine months we re-classified Network Rail
as a public sector company. We are now looking at the consequences
of the Railways Act 2005 which received Royal Assent in April,
and if the Committee would like me to, I can summarise where we
have got to, but we propose to write to the Committee fairly soon
to inform you of our conclusions. 
Q13 Lorely Burt: The figures for 2005
show that only 17% of the respondents in your own survey believed
that official figures are produced without political interference,
and that only 14% believed that the Government uses official figures
honestly. One of the examples of this that several people have
raised was the endorsing of Mr Brown's change in the economic
cycle and there is this thought that this was all behind closed
doors. I just wonder whether you could comment about that. Have
you set targets for increasing the trust that people have?
Mr Mowl: I think there are two
questions there. One was about the Treasury and the economic cycle,
which I do not think was directly addressed in the survey you
referred to. I do not think so, but I imagine Karen would want
to respond to your question about the survey and the degree of
Ms Dunnell: Yes. I was very disappointed
that the survey showed such a low level of trust. In fact it endorsed
the findings of the first survey that we carried out last year,
but we must look at it also in the context of other information
that we have about the use and dependence of people on statistics.
Nevertheless, one of the things I did in my first months in office
was to meet with all my senior colleagues, the chief statisticians
of other government departments, and we spent most of our meeting
working out how to take forward some work with the aim of improving
trust and we set up several groups to address five main areas
where we think we can improve the situation.
Q14 Lorely Burt: Thank you. I wanted
to put it a bit wider now and talk about the wider public and
their perception of national statistics. In fact, your later survey
on public confidence found that trust was higher amongst those
who thought statistics were an important basis for decision-making
in society than amongst the wider society, if you like. How important
do you think it is to increase trust amongst those who are not
involved in using statistics?
Ms Dunnell: I think it is very
important because what we know, from schoolchildren upwards, is
virtually everybody needs to understand statistics in order to
get the most out of their lives and so we feel it is very important
and we start with schoolchildren, by supporting the special web
site we have, Stats for Schools, and that is very important because
the earlier in people's education you can get them to appreciate
statistics and how to use them and how to interpret them, the
better. Also, one of these groups that I mentioned just now is
going to look very carefully at what we can do to improve communication
about statistics generally. So, for example, when we put out statistics,
are they expressed in the clearest way? Is it clear why we collect
them, why they are important, how you should interpret them and
so on? We feel there is probably an awful lot more we could do
as statisticians to help people to understand why something that
we have produced is so important to the whole democratic process.
That is what we want to be doing much more generally across the
whole of society.
Q15 Lorely Burt: Thank you. Finally,
I just wanted to pick up on the trust of statistics again. Do
you think that this very low figure is simply a problem of perception?
Do you think there is a genuine problem of misrepresentation here?
Ms Dunnell: I think it is perception,
and one of the things, for example, is that you have somehow to
distinguish between what people understand about the basic statistics
that are produced, for example by the Government Statistical Service,
and the interpretation that ministers, the media, charitable organisations
or whoever put on them. Because we have such a huge range of statistics
you can almost pick one to match your point and I think it is
almost in the interpretation that the risk arises, which is why
it is very important that we probably place more emphasis on our
Q16 Lorely Burt: Do you think that is
a route that you would want to go down, as well as producing the
statistics, to increase the amount of interpretation that you
can offer to the public as part of the service that you offer?
Ms Dunnell: Yes, very much so
and, indeed, that is the way that, particularly, the Office for
National Statistics has been going and one of the main reasons
for modernisation and efficiency is to make the collection process
cheaper, to be able to spend more resource on interpretation so,
for example, very well established products, like Social Trends,
have been enormously successful in providing a coherent, comprehensive
picture about what is changing in society. Another examplewe
have just produced a compendium of statistics about pensions which
hopefully will have the same impact. It brings it all together
and points up which kinds of things you need to look at from one
point of view and which from the other. So we very much are trying
to increase the amount of that kind of work that we do.
Q17 Lorely Burt: I can see that will
be a really useful way to try to reduce the amount of misrepresentation
as well. Are there any other ways that you think you can stop
the misrepresentation by media and politicians of national statistics?
Ms Dunnell: One of the things
that will be happening fairly shortly is a review of our Framework
document and one of the things that will be reviewed in that is
the whole issue of pre-release access to statistics because having
pre-release access, of course, allows ministers and departments
to prepare statements about the statistics in advance and that
is possibly one of the areas where we might be able to get our
interpretation in first.
Lorely Burt: Thank you very much.
Q18 Peter Viggers: I think I heard Mr
Mowl say that there was a period during which it was thought that
Network Rail was not under public control.
Mr Mowl: That is correct.
Q19 Peter Viggers: Whose control did
you think it was under?
Mr Mowl: The Government's, public
sector control. The classification of whether an organisation
is in the public or the private sector depends on whether the
organisation is deemed to be under public sector control or not.
A decision we announced in 2002, the original decision, was that
Network Rail was under the control of the members, not of the
government, and that is why it was classified as a private sector
company. We subsequently discovered that, in the early days of
Network Rail, the directors' remuneration scheme was influenced
by the government in a way which it was not subsequently, and
in a way which we were not aware of at the time we made our original
decision. It was because of the difference in the directors' remuneration
scheme where, for this nine-month figure I was talking about,
that remuneration was influenced by the Strategic Rail Authority,
which is an arm of Government, that led to this re-classification.
1 Ev 13, further correspondence is available on the
ONS website, www.statistics.gov.uk/about/Methodology_by_theme/rail_network Back