HMT00000024
HMT00000024
INFIDENTIAL
Treasury Chambers, Pas ament Street, London, SW1P 3AG
PRIME MINISTER
HORIZON (BA/POCL) PROJECT
This note sets out the current situation on the Horizon (BA/POCL) Project. It
summarises the extensive work that has taken place over the last few weeks to
reach an agreement about how best to proceed. In summarising the position I
have consulted extensively with Stephen Byers, Alistair Darling and Charlie
Falconer.
We have a commitment to give ICL a decision on the way forward with this
project on Monday (but ICL have said they can now wait until Tuesday). They
must file end year accounts on Wednesday. Our policy aim is to move to
Automatic Credit Transfer (ACT) as soon as reasonably practical and to
preserve a national Post Office network, We should keep ICL/Fujitsu on board
if possible. Stephen Byers and Charlie Falconer both prefer Option B1. Alistair
Darling and Alan Milburn favour Option B3 if POCL and ICL can reach a
sensible deal. If they cannot they would favour Option C - terminating the
contract with ICL and allowing POCL to procure a new system that met their
commercial requirements in the light of termination.
Background
We gave an assurance to Fujitsu that the Government will make a decision
HMT00000024
HMT00000024
CONFIDENTIAL
on the way forward for the project by Monday 10 May, although ICL have now
said they can wait until Tuesday.
Economic case
2. The Horizon project was envisaged as a way of reducing benefit fraud and
modernising the benefit system, whilst automating the Post Office network in a
way that would help preserve footfall and therefore maintain a nationwide network
of post offices. It is now three years late. Our view is that continuation with the
project (Option A) is no longer viable, in view of ICL’s failure to deliver and the
irretrievable breakdown in relations between the parties. This view was
effectively confirmed earlier this week when ICL withdrew their offer of 18
December. It is therefore dead.
3. Weare left with three options. First, Option BL - involving the creation of
15 million Post Office benefit accounts (with limited facilities), accessible via a
Post Office smartcard. Benefits would be paid into these accounts by ACT (from
2002). It is the best option to preserve Post Office footfall in the short term, and
the policy value for this cannot be reflected in the figures. It would place the PO
in a position to win electronic Government services by having a base of 15 million
smartcards. It provides automation of counter services. In this respect it has
attractions, but it offers considerably worse value for money in NPV terms than
the alternative options. If Ministers were to decide to pursue Option BI, ICL’s
current position is that the public sector parties must sign an unconditional
agreement on Tuesday, and provide ICL with £180 million.
HMT00000024
HMT00000024
CONFIDENTIAL
4, The second option (Option B3) would involve POCL buying the basic
system from ICL but without the benefit payment application and without the
creation of special POCL benefit accounts. It would provide automation of post
office counter services (from bill payment to postage rates). Benefits would be
paid into conventional High Street bank accounts by ACT. The Post Office would
offer simple cashback facilities (as a minimum) to access these accounts across the
counter. It would also provide a platform for network banking and Modern
Government with a smartcard capability, though it would not provide the certainty
of 15m smartcards as under B1. It would allow BA to roll out the order book
control system (a way of reducing order book fraud). BA and POCL would work
together to market ACT into bank accounts accessible at the post office from
2001, in preparation for a move to ACT as the usual method of payment in 2003.
The NPV figures are sensitive to changes in these dates.
ay Unlike B1, Option B3 would not tie the Post Office to ICL as a long term
business partner. Such a tie in may well prove inflexible if, and when, we take
forward a Public Private Partnership and, therefore, force us to bring in a private
sector partner on less attractive terms.
6. POCL reject option B3 at the moment and ICL have indicated that it is
likely to be expensive. POCL would prefer termination and to obtain a new system
better suited to their needs. In order to make progress on B3 we would have to rule
option B1 off the table and make plain that termination was the only alternative.
However, when this was done over the weekend POCL still preferred termination.
It is unlikely we could force POCL to do B3.
HMT00000024
HMT00000024
CONFIDENTIAL,
a The third option (Option C) would be to terminate the contracts with ICL.
POCL would start afresh. A new automation system would be bought from a new
supplier specifically designed to meet POCL’s automation and network banking
aspirations, including the ability to withdraw cash from bank accounts at post
offices, BA would be given a date to move to ACT (they would like to start the
transfer to ACT from 2001, although POCL say they could not be ready by then
and claim they would lose substantial footfall) working with POCL to maximise
retention of footfall.
8. Ag well as the above considerations there are substantial economic and
financial differences between the options. The key figures are (£ million):
HMT00000024
HMT00000024
CONFIDENTIAL
Economic return Additional public sector costs
NPV CSR2 years 10 years
central scenarios
B1 (ACT 2002- 439 666 918 } (eo
2004) 283
B3 (ACT 845 294 361
marketed from
2001,
compulsory
2003-2005)
C (ACT from 791 461 393
2001-2004)
Sensitivities:
B3 (ACT from 684 . 547 ‘559 )
2003-2005) 553 Yeo + Ibo Rerese
Cv(ACT from 577 72 ag =
2003-2006)
All these figures should be seen as indicative rather than precise forecasts. They
depend on various assumptions, particularly the timing of the move to ACT. The
central scenarios above are, in the view of our independent advisers, KPMG,
reasonable and do-able provided BA and POCL co-operate to manage the move
to ACT. It is clear, however, that:
654 {GO Vonunat
§20 [160 VAT
GO \Bp cual
CONFIDENTIAL
- B3 and C offer a better economic return than B1;
- the cash hit under B3 and C are significantly less than under B1 both
in the CSR2 period and over a 10 year timescale.
The figures for B3 and C do not include either the costs or the revenues from
introducing a smart card.
9. The Post Office prefer Option B1 but are not prepared to contribute more
than £37 million to the additional cash costs shown above. They have suggested
that a further £190 million be taken from their customers by delaying the reduction
in the postal monopoly from £1 to 50 pence by three years. The Post Office have
said they would plough the net contribution they expect to make from Government
Direct business under Option A back into B1 and estimate that these will be about
£660m (no allowance has been made for Government Direct revenues in the
figures for B3 or C). But this money is already taken into account in calculating
the additional costs in the table above and so is not relevant. ICL are offering the
possibility of finance to ‘smooth’ the spending profile, but this is simply
borrowing from ICL and the interest payments will add to the project costs.
Political factors
10. All of the options - given that none involve the benefit payment card - will
need to be presented very carefully, given the expectation amongst subpostmasters
that Horizon (i.e. Option A) would secure their future. Option B1 would be the
easiest of the other three options to handle. Option B3 would be harder but would
HMT00000024
HMT00000024
HMT00000024
HMT00000024
CONFIDENTIAL
still have ICL on board and the Post Office would still be getting automation. In
the case of options B3 and option C the Government would need to argue that it
would have been doing the post office and its customers no good by pressing on
with a project that was already 3 years late and couldn’t deliver - and that they
were fully committed to providing one that did. We would make plain that ICL
had withdrawn the existing project (Option A). We have looked hard to salvage
something but unfortunately there was nothing worthwhile.
Positions
11, Stephen Byers and Charlie Falconer strongly believe that B1 is the best
option. They argue that it offers most security of footfall in the short term by
locking in 15 million customers and that it would position the Post Office to win
significant electronic government business, thereby helping to preserve footfall
and reducing the risk of post office closures. Option B] keeps ICL/Fujitsu on
board. They note that there is a funding gap whichever option is chosen. The
figures show this as larger for B1 than the other options but they have doubts
about the robustness of the figures and point out they are sensitive to the assumed
dates of ACT conversion. They believe B1 would also be the surest way to secure
the £2 billion plus of cash savings envisaged under Option A.
12. Alan Milburn and Alistair Darling consider that the larger funding gap with
BI, and the fact that it ties the Post Office into an expensive project over a period
during which we might consider a change in the ownership, present too much of
a downside. They consider that the Post Office’s lack of financial commitment
raises doubts about their commitment to B1. They are concerned that ICL’s failure
HMT00000024
HMT00000024
CONFIDENTIAL
to deliver the Benefit Payment Card on time does not bode well for delivery of a
new and complex system (in contrast B3 would be buying that part of the system
that is ready to roll out and is relatively simple). B3 could also provide a platform
for Modern Government. The Option A savings, largely accruing to BA, referred
to at the end of paragraph 11 are also available under B3 and C.
13. The Post Office favour B1. They have said that they are not interested in B3
and would prefer termination (Option C). However this may partly be a
negotiating tactic (they were reluctant to consider other options while option A
was on the table). If POCL and ICL cannot agree a worthwhile deal on B3, there
would be termination.
Conclusion
14. We have been unable to agree on an option. Option B1 is favoured by ICL
and the Post Office (for whom it offers the most financially), but has a substantial
funding gap and at present unacceptable legal demands on the terms of the
agreement from ICL. It provides most immediate security of footfall but ties the
Post Office into a long term relationship with ICL. Option B3 provides an
automated platform for POCL to develop its business in the future, is clearly more
affordable than B1, but is currently opposed by the Post Office and possibly by
ICL. Option C would provide POCL with a made-to-measure automation system,
is more affordable than B1, but would mean the end of ICL’s involvement in the
contracts and could have more presentational difficulties.
15. We have set in train a handling strategy to ensure the best possible
HMT00000024
HmT00000024
CONFIDENTIAL
presentation from the Government's point of view regardless of which option is
eventually agreed.
16. We would be grateful for your views.
ALAN MILBURN
10 May 1999