Examination of Witnesses (Questions 40
WEDNESDAY 6 MARCH 2002
40. When you say you have increased their risk
score, what does that mean exactly?
(Mr Banyard) When we have received all the Self-Assessment
returns, we perform an analysis on all of them to see which ones
of them present greatest risk, in terms of understating the income.
And one of the factors we have found, that is associated with
risk, is that people who file late also have a greater chance
of underdeclaring their income.
41. Sorry, may I interject, just on that particular
point. You did also give us some supplementary information, yesterday,
I think, to the Committee, about the proportion of returns that
are filed on time, and I wonder if we could just touch on that.
As I understand your letter, you have said, to give a more realistic
view of performance, you have looked at the basis of the figures,
in essence, you have rebased them; and it so happens that your
result comes out at 90.6 when the target was 90.5, for filed on
time, and comes out at 96.4 for the ones filed within 12 months
when the target was, in fact, 96.4?
(Mr Banyard) I think the main thing is, from the run
of the figures, if you look at column G, which is the figures
on the original basis, whether you look at that for the new basis,
the main result is we have effected an improvement, improvement
is there on either basis. But what we did also was to look at
the way we defined this, to see whether we were being realistic
in the figure that we were declaring, and we have talked to our
own internal audit, we have exposed it openly to you, and we have
rebased the figures slightly; it makes a small difference. But
the main conclusion I draw from those figures is that we have
reversed a four-year downward trend, and we have some indicators
of what we can use in later years to encourage better filing.
42. It must be good news that you have reversed
the downward trend, but we are bound to observe that the rebasing
of the figures has brought them incredibly close to the original
(Mr Banyard) It means we have hit our target, but
I am confident that the rebased figures are realistic.
43. I just wanted to follow up this very interesting
issue of risk scores. Does everyone have a risk score?
(Mr Banyard) Not in the sense that we attach a number
to them, but what we do is, when the information from your Self-Assessment
tax return is received, we want to investigate only those cases
where we think tax is at risk, we do not want to worry people,
the overwhelming majority of people, who comply. So we try to
see which cases are most at risk of being non-compliant; we do
that by looking at the figures that people have declared themselves,
we add into that data from third-party information that we receive,
and then we have
44. Third-party information?
(Mr Banyard) Bank and building society interest figures.
And then we have our own local risk and research teams which add
in local information, and from that we make a judgement on which
cases are most at risk, and therefore to which ones we should
devote our enquiry resources.
45. So these risk scores, or the people who
have risk scores, are they generated largely by computer, electronically,
or are they generated by, as it were, individual assessment, by
(Mr Banyard) Initially, the computer looks at the
data that is on the Self-Assessment return, and, from our past
experience in investigations, predicts those cases which are more
likely to be at risk. We then feed into that third-party information,
and we then also feed in local information, and from that our
staff make a judgement on which cases they best enquire into.
The aim of this is to make sure that we investigate the cases
that should be investigated and leave people alone when they are
46. How many people have risk scores?
(Mr Banyard) I am not sure, I am looking at my colleagues,
as to whether we actually attach a number to people so much as
we do a risk analysis and identify the most risky cases. I do
not think we attach a score to people, per se.
47. So when you say a risk score, you do not
actually mean a score, you mean they are either risky or not risky?
(Mr Banyard) There are degrees of risk, are there
not, and it is that degree of risk.
48. How many people do you regard as having
a degree of risk; how many cases?
(Mr Banyard) We look at the population as a whole,
we score the tax return, not the person, and the risk score is
49. We have dealt with the risk score, it is
not a score. I am asking how many cases are risk scores; how many
people are in this category of being deemed at risk?
(Mr Banyard) It varies across the population as a
whole, and it depends where you see the level; it is smallest
amongst pensioner groups.
50. Yes. How many?
(Mr Banyard) I have not got the number, in terms of
51. Is it millions?
(Mr Banyard) A proportion of the community has a risk
of underdeclaring their income.
52. But what is that proportion?
(Mr Banyard) It depends on groups. A pensioner group,
for example, we would estimate, there is a 2 per cent chance of
their underdeclaring. And I would not want you to understand,
from that, that we assume automatically that pensioners . . .
Perhaps I should explain that we do a random audit each year and
we investigate 0.1 per cent of the Self-Assessment population,
and from those we can form a picture, proportionately, of where
risk lies, and it is from that that we produce the figures.
Chairman: Alright; well perhaps if there
is any more specific information you can supply it afterwards.
53. In July 2000 you introduced a system for
filing tax returns on the Internet, and, in the first year, you
estimated that there were would be 350,000 people using it; and,
in fact, 39,000 used it, a big gap. Can you explain it? The same
was true last year as well.
(Mr Banyard) Internet projections are notoriously
difficult to do. We did customer research, we looked at trends
in the industry and we looked at take-up of Internet services
elsewhere, and formed the estimate to the best of our ability.
It was meant to be a maximum figure, because we had to size our
system so that actually it would cope with the demand.
54. It was not 10 per cent out, was it, it was
nearly ten times out?
(Mr Banyard) Yes.
55. So it is a fairly big error?
(Mr Banyard) It is a very difficult area to make projections
in; it is a territory where it is very difficult to predict, and
other people have also either underpredicted or overpredicted
the proportion of people who take up.
56. But you predicted that there would be very
big advantages to people who were filing by Internet; do they
not trust your account of things?
(Mr Banyard) We have found slow take-up of the Internet
service. We have talked to people and done research with people
who have used it, and, as a result of that, we will be introducing
changes this coming year, as in fact we did in the past year,
to improve the service, to try to address areas where people wanted
us to make it easier, and introduce them. So, for example, next
year, we will be making the service simpler to use and we will
be introducing more facilities to it, so that it is more useful
to people. People had some difficulty with the registration process,
and we are streamlining that for them next year; we are also simplifying
the return itself, so that it will be done on a question and answer
basis, so that if you do not need to use part of the return you
answer a question and it will take you through to the next part.
In those ways, we hope to make the service more useful to people
and that the take-up then will increase. Perhaps I should say
that, although the take-up of the Internet service itself was
39,000 in the first year, 75,000 in the second, we also have an
Electronic Lodgement Service, which 350,000 people use, and if
you add the two together fully 5 per cent of our Self-Assessment
57. I am sorry, I missed that?
(Mr Banyard) It is an Electronic Lodgement Service.
It is a different way of submitting your return electronically,
but it is not over the Internet.
58. What does it mean; what is involved in it?
(Mr Banyard) It is a commercial service, where you
submit your file electronically, agents, in the main, use it,
and you do not have to send us paper, you send it on 'e', but
it is not an Internet service. If you add that into the Internet
service, about 5 per cent of our taxpayers are coming in electronically.
59. When you said part of the difficulty was
through registration, what is involved in registering, what is
(Mr Banyard) The process really is around security,
but in the coming year we will be writing to all of our unrepresented
taxpayers in the spring, and we will be giving them a password,
so that if they want to have a go at the Internet service they
can log on, use their password and submit the return, in one go.
Last year, you had to wait, you had to send off for your password
and the password came back, and then when you registered there
were one or two steps to go through. This year, we are cutting
through all that, and it will be a very much simpler process.
6 See Ev. 22, para 2. Back