Examination of Witnesses (Questions 280-299)
DEPARTMENT FOR
ENVIRONMENT, FOOD
& RURAL AFFAIRS
AND RURAL
PAYMENTS AGENCY
MONDAY 9 JULY
2007
Q280 Mr Davidson: I understand that.
So many of the changes that were introduced, all of which are
expensive as we all know to our cost, came about partly as a result
of stakeholder influence and stakeholder observations being acceded
to by Ministers.
Johnston McNeill: I think this
matter was discussed in some detail in the evidence given by the
NFU to the Environment, Food and Rural Affairs Select Committee,
for example. It is not something I was close to personally, but
I have scanned it and it is obvious that there were a lot of discussions
about what this should really look like in detail.
Q281 Mr Davidson: The NFU helped
to dig the hole in the first place, and that is helpful to know.
On Accenture, we have heard repeatedly how IT
contractors are grossly over-optimistic—they are also over-expensive,
but that is a different matter. To what extent does that over-optimism
apply here? Can we, with hindsight, look back and say that Accenture
should not have said to you that this can be done, because it
was obvious that it could not be done, or were the changes that
were subsequently introduced so complex that no reasonable IT
contractor could have been expected to cope with them?
Johnston McNeill: I think Accenture
would argue with some force, as it has done in previous discussions
in another Committee, that the system was delivered and worked.
The very fact that we were able to process and make payments in
February and March tends to support that.
Q282 Mr Davidson: As I understand
it, the system worked but payments were made at a rate of 10 a
week when 1,000 a week were needed. The system worked, but the
difference was like that between buying a Ferrari or a Model T
Ford. They both move along the road, but it depends on how fast
you want to move, how much you want to carry and so on. I do not
want to say that Accenture is the Model T Ford of such situations,
but is that not a reasonable parallel to draw?
Johnston McNeill: I understand
your point. We had some significant issues with system performance
in terms of how many claims it could get through, its availability—it
had to be brought down at night for work to be done and for modifications
to be made—and so on. We had numerous discussions with Accenture
about improving that. However, in terms of functionality, the
system worked.
Q283 Mr Davidson: This system worked,
but only if you did not use it to the necessary capacity. Then
it was perfect.
Johnston McNeill: The same system
has been used for 2006 claims, and I understand that the bulk
of claims—96.14% or close to it—were dealt with by
the end of June in line with EU requirements, so the system has
worked well this year. It worked for us last year. I have no doubt—I
agree with you—that the system required tidying up and needed
to be less clunky. I accept that it was unreliable in some respects,
but it worked and it continues to be developed.
Q284 Mr Davidson: So Accenture deserves
some of the credit for what happened.
The notes state that when Mrs. Ghosh was with
us, she described: "a conspiracy of optimism" in the
Agency. Were you over-optimistic? Was her comment reasonable?
Johnston McNeill: No, I do not
think so, nor was the comment in the NAO Report. I have asked
repeatedly to see the optimistic Reports that I am supposed to
have sent to Lord Bach.
What does not sit comfortably with me, Chairman,
is when you tell me that we were over-optimistic, and on page
43 of the very same NAO Report you see a consistent line of red
lights. That is our assessment in the RPA to the Executive Review
Group in DEFRA of what we thought progress looked like. As you
see, there is no shortage of red lights. It is a peculiar sort
of optimism from where I am sitting.
You have had a Permanent Secretary before the
Environment, Food and Rural Affairs Committee—Sir Brian
Bender—who has openly discussed 40% confidence levels in
the Agency, and you say that we were over-optimistic. I have asked
repeatedly to see the papers. The papers that went to Lord Bach
came jointly from Andy Lebrecht and me from CAPRI, and Brian Bender
chaired the Executive Review Group, which also Reported to Ministers.
I am not aware that we were over-optimistic anywhere. We were
frank, forthright and explained the difficulties that we regularly
faced.
Q285 Mr Davidson: I understand that.
That is why I am confused. How could we have the situation that
you told us about, and which was described in the NAO Report,
yet Mrs Ghosh can still refer to: "a conspiracy of optimism"?
That obviously implies a lack of communication and understanding
somewhere. You are in front of us, so I am asking you. How do
you think that could have arisen? Given that you are saying that
Mrs Ghosh's statement was not reasonable, how could she have got
into a position where she was able to say that? She obviously
misunderstood.
Johnston McNeill: Perhaps it was
more suitable than reasonable.
Q286 Mr Davidson: That is a reasonable
response.
In terms of feedback, I cannot quite understand
the point about how things just seem to have gone wrong apparently
so suddenly at the end. Can you just clarify whether you think,
with hindsight, that any particular improvements in the system
of feeding back ought to have been made that would have allowed
the Secretary of State or those in the Department who were making
the decisions to be aware earlier; or were they also suffering
from a conspiracy of optimism?
Johnston McNeill: No. I think
the Report makes a fair observation, which I accept and which
we had already come to realise as we got closer to the final systems
development. In moving to a task-based approach to working we
did not have a sophisticated enough understanding of the relationship
between tasks outstanding and actual clearance. The work that
is now being done, which in fact was commissioned with Accenture
some time ago, is on better understanding; if you phone up about
a claim, we want to know how many tasks relate to your claim.
We also need to have the facility of getting those tasks attached
to your claim.
To be perfectly frank, that is something of
a step backward from the original concept from PricewaterhouseCoopers,
which was almost to deskill the work and have tasks so that you
could group the task in various categories, and issue desk instructions
to staff who, with minimum training, could address those particular
issues quickly and clear them off and hence get the claims through.
But there is a need—and it is well recognised now—to
move towards a better understanding of the relationship between
those tasks and individual claims. If we had had that, we would
have been able to say, "There are 20,000 claims here which
have only two tasks each. Now, if we could isolate those tasks
we could clear 40,000 tasks and pay 20,000 claims." We did
not have that facility, and that was an error. I accept that.
Q287 Mr Davidson: In terms of that
error, would it have been reasonable for PricewaterhouseCoopers
to have understood that situation before giving its advice on
the breakdown of tasks?
Johnston McNeill: That side of
things was one of the casualties that resulted when we were starting
to de-scope the work that Accenture delivered. Accenture's line
was quite straight: "If you want this delivered and you want
payments in February, we cannot do all this." Certain parts
of the work and certain parts of the scope, would to have to be
removed. While the sophisticated management and information system
was all very well, the fact was that it was Rolls-Royce system;
it was going to take a lot of work to build. It was going to have
a serious impact and so parts of it, I am afraid, had to be sacrificed.
On reflection, as I have said before, that was an error.
Q288 Mr Davidson: Did they tell you
that it could not be done after they had got the money for it?
Johnston McNeill: No. The arrangement,
as set out in the case study Report that the OGC produced, is
that Accenture was on a fixed-price contract and tightly managed—perhaps
over-tightly in terms of the contractual arrangements.
Q289 Mr Bacon: Mr McNeill, you said
to the Select Committee on Environment, Food and Rural Affairs,
in response to a question from Mr Duddridge, that "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."[2]
I think you meant by that a problem that made it not do-able.
You have said that it was somewhere between 20 February and 10
March when you became aware that it was not do-able. Can you pinpoint
more precisely when?
Q290 Mr Bacon: They were sticking?
Clogging, in other words.
Johnston McNeill: Staff were working
twilight shifts and maximising use of the system. They would clear
a certain category of task only to discover that the impact of
that would not be realised until the next day, because the system
has to run overnight to enable you to understand the impact on
the remaining tasks, or how many tasks would result. It started
to become apparent that, indeed, additional tasks were being generated.
Q291 Mr Bacon: The doing of the tasks
created more tasks: and indeed, paragraph 20 of the Report said
that you were not able accurately to predict how many new tasks
would be created by the performing of tasks. At what date, then,
did it become clear to you, between 20 February and 10 March,
that it was not do-able?
Johnston McNeill: The payments
went out on 20 February, and it became clear that it was not do-able
when we started to monitor payment performance.
Q292 Mr Bacon: After how many more
days did you become aware that it was not do-able?
Johnston McNeill: I do not recollect
that detail, but it was a relatively small number of days, after
which the Agency went back and began investigating in detail what
was happening.
Q293 Mr Bacon: Was it within a few
days? Presumably, the situation became worse and worse?
Johnston McNeill: The period between
20 February and 10 March is 18 days.
Q294 Mr Bacon: I am not clear about
why it took so long. You said to Mr David Taylor that the: "Department
understood that this was a high risk, one of the top high risks"
in Government. The Chairman of that Committee said that: "the
scheme was high risk at the beginning, and it got riskier"
and you answered that you "had a very sophisticated risk
assessment model". Later, you said: "I have to say it
was a particularly sophisticated approach to assessing risk."
I am grateful that you pointed out the Gateway
Reviews, which showed that there were 16 reds in a row between
December 2004 and March 2006, but no double reds. The Committee
has an arrangement with Sir John Bourn that when there is a double
red, a Permanent Secretary will notify him, and he will then notify
the Chairman of the Committee. There were no double reds, but
16 reds on the chart on page 43 of the Report.
Johnston McNeill: That is not
the OGC's Report of risk, but the Agency's. If you look, you will
see that it states that it is the "Executive Group's risk
matrix".
Q295 Mr Bacon: I am sorry—you
are absolutely right; the gateway review is on page 44. In fact,
your assessment was still more bleak than the OGC's.
Johnston McNeill: Absolutely.
Q296 Mr Bacon: You said that you:
"had a very sophisticated risk assessment model" and
also that: "it was a particularly sophisticated approach
to assessing risk". However, your assessments, made over
two years, marked the risk with red after red, which was a bleaker
assessment than that of the OGC's external Gateway Review. How
was it that you began for the first time in late February or early
March—very late in the day—to realise that it was
not do-able?
Johnston McNeill: The use of the
traffic lights is confusing. Red does not mean stop. Red means
that there are significant issues that need to be addressed, which
means that one has to take mitigating action. For each red, we
would have told the Executive Review Group what the business case,
scope, schedule and resources issues were. We would go through
that, and say what we needed to do about the red. We would address
the issues for the month marked red, and move on.
Q297 Mr Bacon: I am aware of the
fact that red does not mean stop. In fact, that is one of the
criticisms that has been levelled at the OGC process, because
a double, triple or quadruple red does not mean stop. No one will
ever say that a situation is sufficiently bad that the risks of
continuing are greater than the risks of not continuing. You are
not responsible for the fact that red after red does not mean
stop. However, given that your own internal assessment was so
consistently bleak for so long, one might have thought that you,
as Accounting Officer, might have said: "This is too bleak
and it is not good enough, so my recommendation is that we should
not continue."
Johnston McNeill: We are talking
about the Executive Review Group. My point is that I was sitting
with the Permanent Secretary, as the Chief Executive of the Agency
and the senior person responsible for the programme, and saying
that we had such a level of risk. I also said that if we were
to deliver, we must do this and that, and sought his and the board's
agreement to proceed.
Q298 Mr Bacon: Having got agreement,
would you then proceed?
Johnston McNeill: We would push
on and make changes.
Q299 Mr Bacon: And then meet three
months later and you would have another series of reds.
Johnston McNeill: Absolutely,
because the programme was high risk, and incredibly complex and
difficult.
2 Environment, Food and Rural Affairs Select Committee,
15 January 2007, HC 107-II Johnston McNeill: Well,
it became apparent that claims were moving through level 1 validation,
which is a basic check on name, address and whatnot, and then
level 2 validation was more complex. They were sticking in level
2 validation. Back
| 
