Jan McCa
McCanney M
Minister of Ste
Rt Hon Stephen Byers MP
Chief Secretary to the Treasury
Parliament St
London SWIP 3AG
14 December 1998
De. Steeh.
BA/POCL AUTOMATION
Lam grateful
December, in
with Peter, and
differ substantially on
Peter setting out the Post
Overall, Peter and} believe thal
nuinely represe}
e way
key elements ge!
nevertheless lay
forward.
Taking Alistait
commitment to st
performance and
Department's recet
the public sector an
concession on Fujits
single element of unaccep!
{to Alistair Darliny
which he set out his initial reactions
he has agreed that Ishou in the following terms,
number of point
Office’s reactions.
their proposals still fall som
the foundations
1°s points in turn,
tand behind the
funding guarantees,
ntly awarded Elgar
d onto Fujitss several hundred million
u’s past from and removes at 8 stroke prol
tability from the earlier proposal
CBO00000010
B000000010
pepartment of
“trade and tadostry
2 Manors Sireet
Landon
Swit OT
471-248 5008
EeMait
FLO McCarey!
dated 11
ed the letter
actions
letter to
PROJECT: HORIZON
eter Mandel:
to the latest ICL propos:
ag 1
ice Chairman's
son and myself his {etter to you
als, Lhave discuss
efor copying te Pe
ur own initial re
ald write in
the Post
5, You have also seen
‘or move towards OUr position which in all
at least on the main elements. Although
believe that they
in full, we
‘on which to take the project
have now made @ maj
nts their last and best offer,
short of meeting our demands
Aly acceptable basis
TCL
for a commercial
Fujitsu are eed willing to back their
£600 million with legally enforceable
as they have done for my own
t, as We believe it to be. it transfers from
represents & major
ds worth of risk,
bably the largest
our understanding is that ind
project to the tune of some
in exactly the same way
contract. If this is correc’
poun
position,
their previous
Is.
larchives. G°
Second, we are purzied by Alistait’s suggestion that the 8¢% proposals do not “make any significant
change to the 9 November proposals. which we rejected, in the overall balance of expenditure on the
project between the public sector and ICL”. Under the 9 November proposals the NPV gap of £224
million remaining, at the end of the Corbett negotiations Wis ‘extended by « further £80 rillion
contingency fund, to be funded by the public sector partners but held by ICL (and with little prospect
that mach if any of it would in practice prove yefundabie}. The whole of this contingensy provision has
been removed in the latest propusals. tn addition, the £121 million to be funded by Post Olfice
Counters Ltd by price increases within the contract under the earlier proposals has been reduced to
around £90 miflion, with ICL accepting a higher fevel of tisk on volumes, and POCL benefiting from
keener prices within the guaranteed floor. To pat this in perspective. if forecast traffic volumes under
the new arrangements reach only 90% af forecast fevels, ICL's projected pre-tax loss 0° the project will
double from around £260 million to more than £500 million. My interpretation therefore is that on the
commercials” ICL have in practice conceded somewhere between £80-100 million as compared with
the 9 Noveniber proposals, and have also taken 08 significant additional risk.
Third, acceptance testing. Alistair states that “ICL persist jn asking for acceptance on the basis of +
laboratory test of the systems, as opposed 10 8 tive tia... We agree with hin that this is a hugely
important point. It would be unthinkable to sign off acceptance of the system until it has been shown
convincingly to work on 3 reasonable scale in a five environment, But ‘Alistair is, we believe, mistaken
~ {CL are asking no such thing. The company have moved substantially from their 9 Novernber
position and have now conceded that acceptance will follow five trials ‘pased on the NR2 software
release at 300 offices. More specifically, there are ‘24 separate components of the acceptance test
procedare, Some do indeed involve elements of bench and/or model office testing. but all also incl: ude
five taal in, 300 offices. Beyond that, there js contractual provision for any significant fault not defined
in the acceptance process. put which manifests itself during live trial, to be rectified before acceptance
js signed off. Finally, the contracting parties can withhold the release authorisation for national rollout
if they remain dissatisfied at the performance of the system during the live trial phase. Peter and I are
not clear what further reassurance Alistaic requires.
Fourth, on slippage we agree that ICL’s failure to hit the mifestone due on 14 December is
disappointing. We believe that the ICL management are NO Jess concerned, and that they will make
every effort to make good the delay. ‘As we understand it, however, the replan of the timetable during
the Corbett negotiations envisaged a window ‘of between early July and early October for the start of
live trials. IOL's management have insisted to my officials that they remain wholeheartedly committed
to keeping withia that window, They algo stress that at no time did the Benefits Agency indicate that
the latter part of the window would involve several additional months of delay to Horizon whilst the
Agency concentrated on last minute Year 2000 jgsuey. Whilst we cannot entirely rule out the
possibility of further delays to Horizon, we at least have the aysurance that it bas been extensively and
comprehensively ‘audited by independent experts, and shown to ‘be in good shape. Starting an entirely
new replacement system would represent a far greater leap inte the unknown.
Fifth, the fraud and administrative savings foregone as a result of delays to Horizon are indeed a cause
for regret. It is however worth pointing out that Alistair’s igure of £800 million savings foregone over
the next 10 years if we choose to continue with Horizon js broadly mirrored by the calculations both of
officials (int last J aly’s report by the Horizon Working Group), and later of KPMG, of the negative NPV
effects of cancellation on POCL (from toss of revenue. compensation to subpostmasters, and subsidy £0
Deparment Te su Inder
CBO00000010
B000000010
CBO00000010
B000000010
PEE BF
RESTRICTED - poLicy AND COMMERCIAL
isk of damages against US of £100 million or
whilst trying tO maintain the
Alistair’s ‘administrative
maintain network at giver size). Added to that js the risk © g
43 of terminating Horizon
more from litigation. In other words, the Cos! ‘
network at something approaching its present size could equal or excee
savings.
id uncontrolled contraction of the network, we may
In other words, unless we are willing to risk major a0 ¢ Sv eeipients
need to deploy all Alistair’s savings ~ and perhaps mor mn it. Tf we try to migrate recip!
quickly, before the Post Office can offer full front end banking facilities, we are fikely to meet song
1{ be lost to the Post Office system
political resistance, and most of those recipients who do migrate wil
vate side of the shop 4% well as for other postal
(together with their associated foottall on the priv
products). The potential savings to the Benefits Agency are greatest, but so are the subsidy
requirements to prevent uncontrolied network collapse, 10 compensate subpostmasters for closures etc.
Fi A ffice’s ability to offer banking
If we move ina more measured way to ACT, it yine with the Post O'
id hence the need for subsidy and
facilities, the damage to the Post Office's customer base ant ‘
compensation will be reduced, but so will the potential savings to BA. Either way We shail have at the
end of the transitional period a network fess able to sustain itself than if we had migrated to ACT via
Horizon and the benefit payment card.
For all these reasons, Peter and [now believe that there js a clear case for contimaing with Horizon on
1, we believe the case
the basis of the latest ICL proposals. ‘When the wider context is given due weight
becomes overwhelming. First, loss of the project would undoubtedly be @ major blow to ICL. Just how
great would depend primarily on the stance taken by Fujitsu, put they have claimed that it could lead to
sid mean ICL would make 4 total loss of around £200m
the colfapse of ICL. Failure of the project wou
in 1998; which would effectively destroy its prospects of flotation in 2000 and might indeed lead
elf of the company. Given the current economic climate in Japan, Fujitsa’s attitude
Fujitsu to divest its
may well have hardened. Even on “least bad” scenario, cancellation would badly damage ICL’s
reputation both here and in export markets, and its future prospects
It {s clear from the recent approach from Mr Naruto, Vice Chairman of Fujitsu and Chairman of ICL, to
‘our Atobassador in Tokyo, Sir David Wright, that cancellation would have & serious effect on our
relations with Fujitsu. Sir David does not doubt it, Fujitsu have been & major inward investor in the
UK, with well over £700 million invested in the last decade, Whatever the justification from where we
sit, cancellation would be seen in Tokyo as a major breach of faith by the UK Government - &
withdrawal from the project because We fad changed our minds on the policy but had sought to put the
blame on ICL. We could expect wider repercussions on inward investment from Japan as Fujitsu's
stary permeated other boardrooms.
Finally, cancellation would clearly damage the credibility of the PFI process generally, but particularly
could be expected to make the funding of future large IT projects on a public/private partnership more
difficult to put together: Interestingly, the PFT funding arrangements for Horizon have been portrayed
as a model for Japan and heavily promoted as such by Fujitsu's Vice Chairman.
‘Depwmnont af Trade so Uedarey
CBO00000010
CBO00000010
RESTRICTED - POLICY AND COMMERCIAL (pe
\
eS)
oner and
ister, Alistair Darling, Jack Cunningham. Charlie Fale
Tam copying this letter to the Prime Mini
to Peter Mandelson
lan McCartocy
‘Deperamant of Trade sd Lndtery