Examination of Witnesses (Questions 120
- 139)
MONDAY 26 FEBRUARY 2001
MS RACHEL
LOMAX, MR
GEORGE MCCORKELL,
MR DEREK
DAVIS, MR
TONY OPPENHEIM
AND MR
JOHN ROBERTS
CBE
120. At the stage, since the statement made
by Peter Lilley in 1994, did you anticipate that changes that
the minister should have been advising the Secretary of State,
that these changes to the law, were hamstringing your contractors,
and might, given the period of time over which this was being
developed, be presented to Parliament as unnecessary?
(Mr McCorkell) I believe, and I cannot remember specific
examples, I am sure we can let you have a note, we did put through
some changes in order to simplify the system[6].
121. A bit of too little too late. You said,
Permanent Secretary, that you seriously underestimated the scale
and the complexity of the task. I think we have one example of
that. We know from paragraph 134, which you touched on, the maximum
cost to the Department of this failure. We know it already cost
the Post Office some hundreds of millions and £180 million
to Pathway and several, it could be £170 up to £270
million, to the Department. What is savable from this new system?
(Ms Lomax) It is not £270 million, I said £127
million maximum. What is savable is the bulk of the CAPS system,
which is a major redevelopment of our underlying customer accounting
and payment computer systems. It is a very important basis for
improving customer service, for improving our accounting and for
protecting my position as accounting officer and really putting
us on a much better footing for ACT. The CAPS work has been a
major step forward for the Department. Also, the other thing that
is savable to the Department or, you know, a major improvement
is the fact that the Post Office is now automated and we have
been able to put in place the bar coding of order books, which
enables us to save money there.
122. Convince me of one thing has PFI made
much difference to this? It seems to me that under any system
if you brought a private contractor in to use solely the taxpayers'
money and then after some months, or it could possibly be years,
you are negotiating with them and handed them a manual and said,
"We would like to change that, but we need Parliament's permission
and we do not have time". That has nothing to do with PFI,
does it?
(Ms Lomax) The point about PFI is that it puts the
onus on the contractor to work out the "how". What you
contract for is the "what". What service do you want.
The contractor can try different ways of working out how to do
it and, you know, it just enables him to figure it out in an immensely
complicated system like the benefits system. You know it is not
surprising it turned out to be enormously complicated without
very close working with the Department.
(Mr McCorkell) The fundamental thing about PFI is
not what PFI should be, it is how PFI was perceived when this
contract was let, where it was perceived you have a very high
level of requirement and you allow the contractor to go and develop
it. We now know, and we have learned the lesson, as others have
learned the lessons, that particularly with IT projects, what
does not work and cannot work, which is why there is a whole range
of guidance now that advises you not to do that, which is why
we no longer do it. We still let PFI IT contracts, but we start
with the contractor working with us on the business requirement,
that leads through to a technical design study, which provides
a detailed design for the system and a detailed development plan,
at which point you can let that PFI contract.
123. There was obviously a project team
involved in this. They had, presumably, flowcharts and a myriad
of items that had to be ticked and incorporated. One of them was
not checking which legislation had to be changed. That is a problem.
Why did nobody in this project team sit and say, "Look, if
new technology gives us a new way of detecting fraud, countering
fraud that we spend a lot of time chasing paper tails on that
they should get on with that and in the meantime should be briefing
the minister and preparing for legislation to go through Parliament".
As far as I can see, no bid was put in to change legislation.
(Mr McCorkell) Precisely the process you are describing
was carried out, but it was carried out after the contract was
signed and after we were committed to a fixed price and a fixed
timetable for delivery. At present we would be working with the
supplier in defining the business requirements and doing the technical
design study before we got to a fixed price PFI timetable. The
process was carried out, but it was after we were already committed
to a fixed price.
(Ms Lomax) ICL hoped that it would be possible to
make more changes when it came to the point than the Department
was willing or able to make. When that turned out to be the case
you were then forced to develop a more complicated system than
you would need to do. It was that sort of circle, that is why
a bit more pre contract work would have driven out the complexity
that was involved in trying to do this within the constraints
of the existing benefit system.
124. With all due respect, all that seems
to have happened here is a catalogue of previous errors in ordering
new technology was transferred to this one, with an added problem
of PFI and transfer risk, or not a problem, as the case may be.
There was failure to analyse the scale of the task that underestimated
them. In 1994 we are well into the computer age, that is a time
when many people in their homes have computers and that every
Department and every business of just about any size is computerised.
We are repeating errors in 1994, 1995 and 1996 by not getting
to grips with the specification of what can and cannot be done
and the need for Parliament to change legislation. That is at
the root of this problem, is it not?
(Ms Lomax) I do not think it is. I think that this
was one of the first ever PFI contracts for a large IT software
development, anywhere in the world. I think it involved a different
method of working, between the Department and the contractor,
from anything that either of them had been used to doing in the
past. With IT and software development projects you need collaborative
work, particularly in an area as complicated and as sensitive
as the benefit system. The way in which PFI was being interpreted
at that time got in the way of that necessary collaboration, with
the consequences we see.
125. I am a former welfare rights officer
and I know how complicated it is and I know that the manuals do
exist so that people like me can understand it, and claimants
as well. I have a final question, were the new ministers in 1997
given a categorical assurance by your Department they could get
this back on track?
(Ms Lomax) Were they given a categorical assurance?
126. That could get this project back on
track?
(Ms Lomax) I have drawn attention—
127. I really wanted a yes or no answer
to that.
(Ms Lomax) I do not think I can give you that.
128. In 1997 when new ministers came in
and were briefed on this were they given a categorical assurance
by your predecessors that this project, if it was off track could
be got back on track?
(Ms Lomax) Chairman, you would not expect me to answer
something which reflects advice given to ministers in a previous
government.
Mr Campbell
129. Mr Roberts, can I ask you to clear
up something which confused me. If we had not gone down the route
of Benefit Payments Cards and we wanted to get to the point of
automating Post Office Counters, how much would that have cost
to get us to where we are today?
(Mr Roberts) It would probably cost, without the Benefit
Card, roughly what it will cost now, which is roughly £1
million[7],
because what we have is just that kind of system which we have
now implemented and will finish implementing in the next three
months.
130. Okay. The half billion, or so, which
has been quoted earlier is racked up in that £1 billion,
is it?
(Mr Roberts) Yes, it is.
131. It is not an additional amount of money
on top of that. When we talk about the cost of the failure of
the Benefit Payment Card, Ms Lomax, surely we should build into
the equation the difference between the anti-fraud measures, which
are now possible—one should not underestimate the £68
million saved in bar coding order books—how does that stack
up with the savings that the Benefit Payment Card would have introduced?
(Ms Lomax) We were hoping that the Benefit Payment
Card would virtually eliminate instruments of payment fraud, which
at the time the business case estimated was something like £150
or £160 million a year. We think we now have the instrument
of payment fraud down to something like £100 million through
the order book bar coding and we hope to improve that over the
next few years.
132. While there may have been some benefit
in terms of infrastructure investment from the whole project we
have not arrived at the destination that we would have hoped to
have arrived at?
(Ms Lomax) We have not eliminated all instrument payment
fraud, but we have computer systems within the DSS which are a
value in their own right in terms of improving our accounting
and our customer service and will provide a very good basis for
the move to ACT, which is, of course, a very much cheaper means
of payment for the DSS than the Benefit Payment Card would ever
have been.
133. ACT was not the chosen route announced
in October 1995. Was Peter Lilley still the Secretary of State
then?
(Ms Lomax) Yes.
134. Given, as you described earlier, the
complexity of this project and given the fact there was £1
billion of investment here, and given the fact that this is a
political mine field, to change the benefit system, not only does
one raise the expectations of taxpayers, in terms of bearing down
on fraud, but one also raises the potential questions of recipients
of benefits as well. Was it wise, with the benefit of hindsight,
to announce just three months short of 1996, that the Benefit
Card would be introduced in 1996? Was it a little premature?
(Ms Lomax) Ministers often announce things which have
not happened.
135. To announce in October of one year
that something would be introduced in 1996, and five years later
not only are we waiting for it, but it has not been announced,
is that not over-ambitious, over-optimistic? I am searching for
the right word here.
(Ms Lomax) It turned out that timetable was very over-optimistic.
136. What effect would that timetable have?
Presumably clocks were set ticking here. We have heard a great
deal about the nature of early PFIs and the Committee is not unfamiliar
with those concerns, it does seem to me, from answers which were
given earlier, that the Post Office and the Department had a view
about early PFIs, which was about transferring risk to the supplier,
about specifying what you want and how you want it and it is up
the supplier to deliver on that. Yet, if I heard Mr Oppenheim
correctly earlier, he said, "They did not really understand
the complexities of what they were being asked to do". That
is not just in the nature of PFI, is it, it is in some decisions
that were made by people at the time?
(Ms Lomax) The point we are making about PFI, is that
this was government policy at the time, this was the way the government
wanted to conduct an investment at the time, and not something
that was just sort of invented around this project. The point
about the PFI is that it matters more within the context of a
fixed price, fixed timetable contract if you have not thrashed
these things out before you start. I think that the PFI context
is absolutely pitiless if you rush into things, but it also gives
people a tremendous incentive to race ahead. ICL was very anxious
to move ahead to contract as well. All of the people who were
bidding for this contract were spending money without remuneration.
The private sector supplier does not get any money until the service
is delivered. It was an enormous up-front commitment for them.
They too want to move ahead as fast as they can. It was not a
question of them asking for more detail and being frustrated by
the Department. There was tremendous time pressure both on the
Department and on ICL to move this thing along as fast as possible,
ICL for commercial reasons and the Department because they wanted
the fraud savings, and the Post Office because they wanted automation.
Everyone was under a lot of time pressure.
137. Presumably at the end of the day somebody
has to decide, somebody did decide, to call a halt to this when
it became obvious it was not going to deliver what they hoped
to deliver. Let me ask you a question about that, paragraph 1.11,
the "Initial Go-Live" in October 1996 dealt with, as
I understand it, only Child Benefit, it dealt with a limited volume
of transactions and it also came up against difficulties about
the collection of benefits only by nominated individuals. For
anyone who knows much about general benefits, there are all sorts
of other issues, it is also quite common for people to collect
benefits on behalf of other people. Why did the Initial Go-Live
take place and what purpose did it serve?
(Mr Oppenheim) I think it served as a marker. It was
possible to put service into place very quickly, to give an impetus
to delivery. Following the thrust of your question, it actually
was a distraction from defining the eventual service and it probably
put back the eventual service, to some extent. It was treated
as an off line special project and ring-fenced so as not to do
that to the maximum, but it was an additional delivery that needed
attention.
138. Was it true that this was a distraction
and it really had not served a useful purpose? Was that a decision
which was reached quite quickly or a view which was taken quite
quickly?
(Mr Oppenheim) It is a personal view I am expressing
here, based on the amount of work of delivering something which
did not lead anywhere because it was a special piece of software,
it was an intercept of the eventual solution. The interim solution,
1C, which was released in 205 post offices doing, admittedly,
just Child Benefit was dealing with foreign encashments and did
deal with proxies, and so on. It did have those features. It did
have the EVP security system, but it was hard wired as opposed
to being flexible. It had all of the headline features, barring
one or two support and security features that the DSS considered
to be very important for the higher risk benefits, such as income
support. That is why the DSS decided not to put income support
through the system. Our system did not mind which benefit it was,
it made no distinction whether it was Child Benefit or Income
Support, because it all came from CAPS. It was part of the decision
not to release any other benefits on us, but that is just a technical
point.
139. We heard earlier, Ms Lomax used the
phrase, "A proper review in September 1997", and then
the decision by the new government not to go ahead with the Benefit
Payment Card. There was an earlier opportunity, was there not,
in February 1997, which was the replan. What is the difference
between a review and a replan?
(Mr Oppenheim) The replan was something that we did
between the three parties to break up this very large, single
block of deliverables into three or four discrete elements of
delivery. Progress was, indeed, made down that track and it also
reset the timetable by mutual agreement. Progress was made down
that track, but roundabout July further major problems were identified
which caused a review. Those were essentially around, we would
certainly argue, the agreements to agree territory of security
and other very difficult areas. At that point we had largely the
functionality, as perceived by the customer in place, but we did
not have some of what we call the nonfunctional requirements.
6 Note: See Evidence, Appendix 1, page 20 (PAC 2000-2001/158). Back
7
Note by Witness: The figure is, in fact, £1 billion,
not £1 million. Back
|
