Examination of Witnesses (Questions 60
THURSDAY 8 FEBRUARY 2001
60. This carries the implication that potentially
a very large number of taxpayers may either have paid too much
or too little because of that. Can you quantify that? Is this
something the National Audit Office is looking at?
(Dawn Primarolo) I cannot quantify that at this moment
with regard to the 1.7 million, but the answer would be, yes,
it would be something, as I understand it, the National Audit
Office would be interested in and would seek to quantify at the
relevant time in agreement with our statisticians.
61. Do you think the computer systems that the
Inland Revenue now have, bearing in mind there are two different
ones anyway which do not completely talk to each other, are adequate
for the job they are now committed to?
(Dawn Primarolo) Yes, I do. The average reliability
of the Inland Revenue IT systems is over 99 per centover
99 per centwhich is well in line with any other comparable
organisation. I do not know whether Dave Hartnett remembers but
the scale of the information that we deal with in terms of it
being processed through these computers is truly phenomenal.
(Mr Hartnett) 30 million taxpayers.
(Dawn Primarolo) We are talking about a colossal amount
of information. There were particular problems which are well-documented
with regard to the NIRS2 computer which was commissioned under
the previous Government, was supposed to come live by April 1998,
was then transferred into the Revenue with the Contributions Agency,
and there were particular problems with processing 1997-98 year
returns because of the system not being able to deal completely
with the information. Clearly, my priority as a minister when
the particular system was transferred over to us was, firstly,
to ascertain it would work and it could be made to work, that
it should be stabilised and that we should move as quickly as
we possibly could to deal with the 1997-98 returns and make sure
subsequently the system was working well. We are confident in
terms of the non-matched items that we will be restored to the
previous positionwe always had unmatched items at the end
of every year, it is not just a product of a computer system,
manually you would have those issues as well because it is about
incomplete records, somebody may have done several different jobs
and we cannot trace them, we do not have the right address, or
they have moved on from an employer and that employer does not
know where they are and you cannot match them up, so there has
always been this issue of non-matched items. What we have to make
sure is that that is in a proportion which would be acceptable
and we find ways of driving that down even further, which is what
we are doing. Although incredibly time-consuming and difficult
to do, reviewing cases on an individual basis may help us in developing
strategies for the future on how to monitor this better.
62. When the Government is considering tax changes,
does it also have some kind of analysis of whether the IT systems
which underpin it are robust enough to support the tax change?
(Dawn Primarolo) Yes, absolutely. An example would
be the Working Families Tax Credit when we had to build out of
the Family Credit IT system, and it constrained the development
of some of the policy issues until we had improved the IT system
enough to cope with some of the changes. We absolutely have to
make sure we can deliver the policy, and IT considerations are
63. Do you think the self-assessment system
is now settling down?
(Dawn Primarolo) Yes, I do. If you look at the number
of returns and the proportion of people making returns, the self-assessment
system is bedding down quite well. Clearly there are still outstanding
issues which we need the information from previous years to refine,
for instance, the use of the internet, and our experience from
this year, and analysingI believe it is called "mining",
I do not quite know whythe data in terms of late returns
and why we continue to get late returns. It is in about the same
proportion as before self-assessment, so it is not new. This year
we have tried to focus on whether particular groups were at risk
of sending in late returns, so some of our advertising was targeted
specifically at them. In the last few weeks as we came up to 31st
January, something like 3,000 people were telephoned or written
to, or both, who we thought were at high risk of not getting their
return in on time. Now what we are doing is looking at the groups
which may become late returners, whether there is an underlying
policy issue there, or whether it is presentation and whether
we need to focus more on those groups.
64. What is the proportion of late filers this
year, given that the deadline was last week?
(Dawn Primarolo) It is 10 per cent of the total. There
are approximately 9 million self-assessment returns and about
10 per cent of them come in late.
65. Did I understand you earlier to say that
was about the same as last year?
(Dawn Primarolo) Approximately, and pre-self-assessment
as well. Late returners is not just something which happened with
self-assessment, it was always an issue in the tax system.
66. Are you at all concerned that the proportion
of taxpayers submitting late returns and asking for the Revenue
to calculate their tax bills is increasing?
(Dawn Primarolo) We have no indication of that. Something
like 49 per cent of the returns are received in time for the Revenue
to do the calculation, which is the 30th September date.
Obviously the Revenue will continue to assist and advise, but
the date where the Revenue calculates a liability or refund seems
to be working quite well.
67. Is that not rather a high proportion for
a system which is intended to be self-assessment? If I understood
you rightly, 49 per cent are asking for the calculation to be
(Dawn Primarolo) 49 per cent returned on or before
68. Sorry, I misunderstood. I thought you were
saying 49 per cent were the people who were asking the Inland
Revenue to calculate their tax liability.
(Dawn Primarolo) That is right. After 30th September
we do not guarantee to be able to do the calculation of their
tax in time to tell them how much to pay by 31 January. I was
just indicating that 49 per cent of all returns are in before
30th September, which is the date people know about if they absolutely
want us to do that. The percentage of those returns where we have
calculated the liability, I do not actually have. I do not think
we have that analysis. I can try and get it to see if we can break
the information down. It is a bit like internet filing, 38,000
individuals did it but something like 130,000-odd were actually
logged through internet filing because the remainder were done
by representatives of the taxpayers.
69. 38 per cent filed by internet?
(Dawn Primarolo) 130,000-odd filed by internet, of
which 38,000 were individuals and the remainder were representatives
who were doing the returns for an individual.
70. Is there anything which can be done to make
internet filing easier, because it does seem to be a bit of a
palaver for individuals? They have to register, receive a password
by post, then install their own software. There are quite a few
barriers to them doing it.
(Dawn Primarolo) Yes, all of that
is being changed. What we wanted to do, and we did actually say
this, was to have the internet filing option and we knew we would
have to refine and develop that, as happens with the development
of these sites, and also that we would need to take the site off,
close the site down, on two occasions because we wanted to make
changes. What we are doing now is a review, learning the lessons,
and making improvements, so the sort of points you are addressing
are not issues for next year.
71. Do you have any target for the proportion
of submissions you are aiming to have through the internet?
(Dawn Primarolo) I do not have a target at the moment.
What I have asked Revenue to do for me is a break down of the
self-assessment returnshow many people do it individually,
how many do it through representatives, of that split how many
would, regardless of what we did, prefer still to do it through
the post. So that we try and get a better understanding out of
that population of 9 million who would be likely to want to internet-file
and what we need to do to make that easier for them, rather than
just assuming somehow we could make a one-size-fits-all for the
9 million, because in that 9 million you go from very simple,
straight forward returns to very complex returns because of income
72. The one bit of unfinished business is this
letter which you will be sending us on the closer working report?
(Dawn Primarolo) Yes, and I also promised one to Mr
Cousins on the peer review.
(Dawn Primarolo) And the policy co-ordination in particular.
74. We will give you, if it would be helpful,
our input into the format question on closer working.
(Dawn Primarolo) Indeed, yes, that would be good,
and the annual report, yes. Excellent.
75. Then we will have another meeting.
(Dawn Primarolo) Yes, of course. We will try to get
the notes to you by Monday.
76. Thank you.
(Dawn Primarolo) Thank you very much.
1 Witness correction: the correct figure is 48 per
Witness correction: around 130,000 people registered to file their
self-assessment tax returns over the internet. Back
Witness correction: the answer to this question should read "38,000
individuals filed by internet." Back