Examination of Witness (Questions 1240-1259)
MR JOHNSTON
MCNEILL
15 JANUARY 2007
Q1240 David Taylor: Is that not why
it is working or not working?
Mr McNeill: That is a fair point,
I suppose. The difference was that when the system worked it did
actually work as in you could process claims through the system.
If it fell over there were a number of issues in there, one was
its reliability and availability to our staff and there was a
service credit issue which meant that we could go for breach of
contract, in other words, "The system is here, it does work
when it works but the fact is it is not reliable". We addressed
that. Another issue we had with the Accenture systems, possibly
as a result of the chops and changes that had been made to the
system, was they were incredibly heavy on computer processing
power. At one stage we were considering buying yet more computers,
more hardware, to run the systems so heavy were they which is
usually a sign, I was advised at the time, I think by OGC, of
perhaps poor coding. It was something that would have been addressed
in time by revisiting code structure and discovering why. That
type of issue did arise.
Q1241 David Taylor: What would merit
the description "the system is working" is not a fleeting
moment when the system does as was expected of it but that there
are issues of stability and accessibility and timeliness and all
those sorts of things and they are wrapped in with any reasonable
understanding of what is meant by a working system, are they not?
Mr McNeill: I accept your point
fully. Are Accenture correct in saying the system worked? If you
want to take the view that we put in a claim this end, went through
a number of processes, including business interactions, and a
cheque came out that end, then the answer is the system[17]
worked. I take your point. If you are saying it only did that
for two hours a day or it was not reliable or it was incredibly
inefficient in terms of the amount of processing power then I
can understand you might say it did not really work.
Q1242 David Taylor: I think we have done
to death the process-based and claims-based problems that were
created by that but I do have one particular question which is
at a level of detail. Because it was a hybrid system and, therefore,
10% of the payment was based on the three areas that Margaret
Beckett announced in the April or so, did that not mean that in
essence every claim had to be in before any claim could be paid,
that element of the 10%, so that we knew what the areas were?
Do I misunderstand that?
Mr McNeill: No, you are quite
right. To define entitlements we had a cake and we needed to know
how many people wanted what section of it, I agree. What happened
in the end, and there was extensive legal advice taken on it,
was that we looked at the number of claims, some of which were
fully through the validation process, others which had not quite
made it through, and we took a view that we were able to define
the entitlements.
Q1243 Chairman: In that context,
why was the mapping system such a source of failure? There were
a number of letters, and I was looking at one only today to refresh
my memory. In one case it was taking something like almost two
years to sort out problems. When we had evidence from the Central
Association of Agricultural Valuers they brought us examples of
where they as professionals had worked out the correct maps but
seemingly the system in the example shown to us was utterly incapable
time after time of producing a validated set of maps. Why did
that go so catastrophically wrong?
Mr McNeill: It is an area of difficulty
that I personally spent an inordinate amount of time on. I would
suggest, whilst I try to answer it, that there are people in the
RPA who could give you this in detail that would explain it. There
were a number of issues. The volume of maps was a problem, ten-plus
times the number of IACS 22s or requests for changes to maps that
came through. We had the issue of 40,000 new customers who wanted
land maps, so there was a volume issue. There was also a systems
issue. The RLR[18]
rolled out, as I recollect, towards the end of 2003 and there
was an issue about the clumpiness of the system, the system was
incredibly slow. I remember in that quarter the Ownership Board
was in Exeter and Ian Watmore, who now heads up the Cabinet Office
Performance Unit, noted to the RPA Ownership Board that were it
his system he would put his fist in it, so appalled was he at
the slowness of the response of the RLR system. I do not know
if you remember the TV series Knight Rider, Chairman, but
the staff called it the Knight Rider system because a little
bar would come up whilst they were waiting for the RLR system
to figure out what it was they were trying to map, et cetera.
We had appalling problems with the RLR system there which had
to be resolved. Then there was the issue moving forward from that,
Chairman, that we were advised by lawyers, by Defra legal again,
that the SPS scheme EU Policy interpretation required us to send
every customer a set of RLR maps, which we did at the start of
2005, before they made their application under SPS. So we sent
out all our maps detailing our understanding of customer holdings,
and that created an absolute plethora of issues and requests that
came forward about the RLR maps and that added to the difficulties
where people had not seen all of their RLR maps and being aware
that the SPS scheme was land-based farmers wanted to make sure
that every boundary was right and every fence was right, et
cetera. Another issue was the tolerance level we worked at
on the systemwe could have gone for a lesser toleranceso
we ended up with issues which were called "slithers"
and this was where the boundaries were slightly out and you had
a little slither between the two, whose land was it, et cetera.
If we had gone for a broader tolerance we would have probably
got away with a lot fewer problems. The difficulty was this was
a major investment, a national asset, and the view at the time
was that we needed to make sure that the system was as good as
it could be, so that was a problem.
Q1244 Chairman: And it was Defra's lawyers
who set all these parameters?
Mr McNeill: No, I am not sure
about the tolerance one, Chairman, I cannot recollect where that
came from. I am sure there are others in RPA who could advise
you. I cannot recollect that. The issue that the Defra lawyers
were clear about was to avoid disallowance we had to send out
the RLR maps before the SPS claim process started. At the same
time where changes were made to the maps, Chairman, we sent out
an actual note of the changes. We would send a map and say, "These
are the fields that we have changed" but we would not show
the other fields in the way the system operated. That caused great
concern to people who thought their other fields had disappeared,
but if you read the letter RPA it is quite clear that the correspondence
was about the fields you have adjusted as opposed to us sending
all the maps yet again. That created difficulties. A lot of the
queries were coming through the Customer Service Centre which
just was not equipped to deal with mapping issues in terms of
the fact they could not see the map on the screens, they were
working on a Q&A brief, so they were of little help and, again,
you had the cultural problem.
Q1245 Chairman: So this was another
failure of this task-based system not being fit for purpose.
Mr McNeill: I am not sure it was
task-based, Chairman. It is a failure of a move from where you
used to be able to phone up John or Gill or somebody in your local
office perhaps and explore these particular problems, you now
came through this new route into the organisation and they were
not set up to deal with you. We always knew, and I think it was
the case, it was going to be a first year problem and I suspect
as we rolled into the second year of SPS the number of letters
in MPs' postbags was greatly diminished; at least I would hope
so. We have now done it. There was a separate issue from this,
Chairman, but which was related which was there was a linkage
between the actual land and the fields the customer registered.
You may recollect we sent out customer registration forms with
the initial SPS applications and that created difficulties as
well. There were also issues about the customer database and the
land database as well. I suggest, if it helps the Committee, that
the RPA be asked about this and I suggest in the first instance
if you ask Simon Vry, he gave me a contact number to refresh my
own memory on this, and I am sure he can put a note up to the
Committee to assist you.
Q1246 James Duddridge: In late 2004
the RPA identified about 23 changes to the IT system. The question
I was going to askI think I need to ask you a prior questionwas
what discussions did you have with the Permanent Secretary and
ministers about your alternative choices? Given I was quite surprised
that you had only met Margaret Beckett twice perhaps you could
put into context what access you had to ministers and the Permanent
Secretary generally and then specifically in relation to those
alternatives in 2004?
Mr McNeill: I was very pleased
to be able to work closely with Sir Brian Bender in his role as
Permanent Secretary of Defra. I had a series of bilaterals with
him, one a month. Brian was very close to the programme, he took
an in-depth personal interest, as I am sure you have seen when
he has given evidence here, chaired the Executive Review Group
and, indeed, the RPA Ownership Board. I think our working relationship
was very sound. He was particularly keen, and I think it is an
important point which he himself has made, that the policy side
and the operational side should work closely together, hence the
creation of CAPRI Board[19]
which was joint chaired by Andy Lebrecht and myself. I read that,
Chairman, whether rightly or wrongly, as the issues we[20]
attempted to resolve more on a verbal basis and in meetings and
discussions rather than engaging in some form of memo warfare
where "You are doing this and you realise this is going to
do that". Our relationship was more of discussions at CAPRI
Board and frank discussions. I had a number of discussions with
Andy Lebrecht about particular SPS issues and concerns about policy
delays, complexity, et cetera. I do not think you will
find that minuted, Chairman, because the relationship was supposed
to be one of working together to deliver the end product.
Q1247 James Duddridge: We talked about
the role of Sheila Watson in hopefully keeping the Secretary of
State up to speed, but Margaret Beckett in 2005 was described
in various places as "bloody livid" or in other places
"angry" at the delay that was caused, yet her ministers
and her special adviser had been kept in the loop throughout.
Was this anger or being bloody livid more about PR or was it a
genuine anger and confusion coming out of the Secretary of State's
office in your view?
Mr McNeill: I remember that quote,
I cannot remember what the issue was.
Q1248 James Duddridge: She was telling
the NFU Conference on 21 February that she was "bloody livid
with the situation", namely the two month delay.
Mr McNeill: Yes.
Q1249 James Duddridge: That rather
surprised me because she had all the information coming through
her minister, through her special adviser, what was she livid
about?
Mr McNeill: I think we were in
similar confusion at the time and very concerned that we had staff
who were working their socks off and, indeed, consultants working
very hard, the whole RPA team working very hard to get this done
and it was not very helpful in terms of staff morale.
Q1250 James Duddridge: It strikes
me as a bit of pointing of fingers. The RPA initially developed
a stop-gap claims-based back-up system to make payments and that
cost about £8.4 million. Why was that abandoned at the end
of 2004? What was the rationale for that?
Mr McNeill: We engaged with Xansa
and Sunguard, two suppliers we had worked with on our legacy systems
for some time, and they developed our legacy systems to enable
us as a last resort to come off RITA (the Accenture system), and
do the entitlements calculation and to work through to payment.
We had a particular resistance to doing that because a major problem
was the bulk uploads and downloads of data between systems, particularly
when you had a very old legacy system and a fourth generation
or whatever it was new system. This was one of the reasons why
we took our time in reaching a decision to outsource the mapping
to Infoterra, a similar problem downloading RLR files to them
for them to update with the IACS 22s and then put back into the
new RLR system. Our Director of IT and, indeed, others were very,
very concerned that this might not work for us. As it was, thankfully,
after some testing we were able to make it work. The reason we
decided not to go ahead in 2004 with the contingency system was
that we had had the delivery of the functionality to enable RITA
SPS level one, level two validations working through to entitlements
which had been tested. It was working, we had got them through
and used them. In actual fact, the functionality that was left
in the contingency package, RITA was ahead of it in some respects.
Also we had done more testing on where RITA went from there through
the RITA batch authorisations into customer payments. We had this
contingency horse riding alongside the RITA development but, in
actual fact, the RITA system was out first and we were satisfied
that the contingency did not add any value. As we said, the actual
RITA system did work. Can I just make one point. There was a serious
tension with the contingency because Accenture and others, and
Karen Jordan, I recollect, was giving advice that we needed to
be extremely cautious here. Karen Jordan now chairs the Cabinet
Office Audit Committee and has had tremendous experience with
Transco and other large organisations in programme management.
Karen JordanHelen Ghosh mentioned Karenwas very
heavily involved in the programme of providing assurance to us
and insight to us as to the best way ahead and, indeed, providing
assurance to Brian Bender and Helen Ghosh. There was a concern[21]
that you could deflect excessive effort, resources and expertise
into building a contingency and as such the main programme, what
you are trying to build, suffers. Accenture had that particular
concern and voiced it regularly.
Q1251 James Duddridge: I would like to
turn on to your responsibility as the Accounting Office. I am
fascinated by the lines of accountability. Did you ever speak
to Brian Bender about considering getting a formal ministerial
instruction in relation to the introduction of the Single Payment
Scheme because of the risks of disallowing of value for money?
In retrospect, is that something you wished you had done?
Mr McNeill: No. I have made it
quite clear here, Chairman, and Brian and others have quoted the
same, we never turned and said, "This is not do-able".
As I said, only at the fifty-ninth minute of the eleventh hour
did we discover we had a problem which did not make it do-able.
We thought we could do it with the elements of risk associated,
of course, but we thought it was do-able.
Q1252 James Duddridge: Can I ask
about the nature and number of informal meetings you had with
the Permanent Secretary just in the final months of 2004 and the
early part of 2005?
Mr McNeill: When Helen Ghosh took
over from Sir Brian Bender and, indeed, when Mark Addison stood
in for a brief period we continued with the monthly bilaterals
which lasted about an hour. They were particularly frank discussions
about where we were and what was happening. There was useful feedback
from what was going on at the top of the office because, after
all, we were based in Reading and were not as close to things
that were happening in Nobel House in London, so it was very useful
in that respect and also gave me an opportunity to give a very
frank and thorough briefing as to where we were.
Q1253 James Duddridge: Did special
advisers sit in on those meetings?
Mr McNeill: No.
Q1254 David Taylor: It is the sixth
anniversary this month of your appointment as Chief Executive
of the predecessor body to the Rural Payments Agency and shortly
after your appointment you appeared before our predecessor committee,
the Agriculture Committee, and acknowledged that you were persuaded
against your instincts into the job and one of the concerns you
had was the IT background that you did not have. Would that be
fair?
Mr McNeill: I think that was the
situation, yes.
Q1255 David Taylor: The high priority
for you at that time, February 2001, was to avoid flying blind.
In the land of the blind the one-eyed man is king. Was your one-eyed
man Alan McDermott, appointed on a very generous contract shortly
afterwards? Was he the person that you saw as your backstop on
IT matters, an area that you say was not a solid part of your
CV?
Mr McNeill: Yes. We went to quite
significant effort to find the best person to fill the post of
IT Director for the business. We went through a large recruitment
organisation. As I recollect, we had some 30-plus applications.
I had those analysed by a consultant from PA Consulting to split
down those applications. This was personally, such was my concern
about getting the right person. I looked at those in some detail
with the consultant who was very experienced in IT and development
programme management. We went through that and what became clear
was that of that batch of applications, I do not know if it is
standard in the industry, and this was after extensive advertisement,
I seem to recollect, in Computer World and all of the international
magazines, et cetera
Q1256 David Taylor: But he was there
for five years and what surprised me about this whole saga is
that is a bit like the poem Macavity the Cat, that whatever happens
he is not there, and yet I would have expected that he would have
been at the core of what was happening in relation to the SPS
system. Were you satisfied with his performance in regard to this?
Mr McNeill: Sorry, yes I am rambling
on about the appointment. The point I am trying to make is that
it certainly seemed to me and, indeed, the panel that appointed
him with external advice that he was the best candidate and certainly
he had the experience. The point I was trying to get to was he
had the experience of actually taking a programme from conception,
in our case the enabling model, et cetera, and delivering
it, and had done that in TNT, in an international environment,
and from that point of view certainly appeared to be the strongest
candidate. Alan joined us and played a full and, I felt, extremely
valuable role in the organisation. He was not the only expert
we had in IT, we had Glenn Rogers who came from California, a
senior partner in Accenture, brought in not through Accenture
but brought in separately to assist us.
Q1257 David Taylor: But if he was
the only expert in the IT area, something that was at the very
core of the do-ability of SPS, I would have expected his views
and recommendations to have been more apparent in the evidence
that we have heard and the records that we have inspected. His
presence does not seem all that obvious to me.
Mr McNeill: I think I have explained
the process of considering what the impact of the various policy
options were, that was where Hugh Mackinnon and Bill Duncan would
come back to our RPA Executive Board.
Q1258 David Taylor: I am thinking
about the IT side now.
Mr McNeill: Alan sat on the RPA
Executive Board, he was a RPA director, and he would input. As
Hugh would talk about it from an operations perspective and Simon
would talk about it from a programme management perspective, so
Alan would talk about the IT consequences and what they meant.
Q1259 David Taylor: So Alan would
have said, "Johnson, it is do-able"?
Mr McNeill: Yes.
17 Note by witness: RITA. Back
18
RLR: Rural Land Register. Back
19
CAPRI: Common Agricultural Policy Reform Implementation. Back
20
Note by witness: Andy Lebrecht and I. Back
21
Note by witness: "expressed by Helen Ghosh, the OGC
and Accenture". Back
|