CBO00100001_072 - Letter from Geoff Mulgan to the Prime Minister (cc: Lord Falconer, Jonathan Powell and others) re Horizon, ICL and the Post Office

Evidence on official site

Prime Minister

RESTRICTED

From:

Date:

ec

CBO00100001_072
CBO00100001_072

Geoff Mulgan
9 December 1998

Lord Falconer

Jonathan Powell
David Miliband
Jeremy Heywood
Geoffrey Norris
Sharon White

Horizon, ICL and the Post Office

1. A decision now needs to be taken on whether to proceed with the Horizon

project. You will recall this is the initiative to automate the Post Office
network, involving ICL. The project is nearly three years behind schedule,
having been plagued with problems. Negotiations have been underway since

the summer to find a way to continue the project.

. The sums involved are big. Total spending, through DSS contracts, will be

around £5bn between now and 2008.

. After a tough period of negotiation ICL has made an offer which represents a

significant move. Acceptance would still cost the government around £200m
more than previously envisaged; equally, ICL would be set to lose £270m in
total on the project. Their offer still leaves a significant gap with the
government: £110m on funding, as well as differences over the practicalities
of implementation. But the Post Office would be prepared to meet the funding
gap out of existing resources, and, despite the continued opposition of the
Benefits Agency, it will almost certainly now be possible to reach a

settlement.

RESTRICTED

CBO00100001_072
CBO00100001_072

RESTRICTED

4. However, the decision is not clear cut. The problems that have beset this
project may well continue; continuation would lock the government in for 10-
12 years to what many see as a flawed system; cancellation on the other hand
would enable the Post Office to take advantage of newer, cheaper and more
flexible technology, while the DSS could move rapidly to paying benefits into
people’s bank accounts. Cancellation would also release around £2-3bn over
the next decade to be spent in other ways to support and automate the Post

Office.

5. The attached paper sets out the main options and contains two alternative
recommendations. Both options involve broadly similar financial implications

for government.

- In making a judgement, the following issues are paramount:

The virtues of the project itself: overall, Horizon, now looks increasingly
flawed. It is centred around a technology, the Benefit Payment Card (BPC),
that is both overengineered — and very expensive - and likely soon to be
obsolete. Indeed, ICL acknowledge that the BPC will have no commercial
value to them at the end of the project. Although they remain underdeveloped,
the alternatives, which involve simpler off-the-shelf banking technology, look
increasingly attractive, offering a route to universal banking, automated Post

Offices and better provision of government information.
The effects on the Post Office network: cancellation would undoubtedly be

destabilising. Subpostmasters fear that without Horizon they will lose their

customer base. Their concerns can be partially, but not wholly, addressed

RESTRICTED

CBO00100001_072
CBO00100001_072

RESTRICTED

Be

through commitments on subsidy and promises that there will be government
support for an alternative automation system. Rural subpostmasters’ fears will
be particularly acute. Subsidies could be structured to keep the great majority
of these open, but, whatever happens, there will be a continuing stream of

closures.

Effects on ICL: Cancellation would directly affect 270 jobs in ICL and more
in suppliers. It would destroy ICL’s prospects of flotation and lead to a
heavy (£200m) loss this year, although even if the project continues the
£250m losses incurred so far may at some point have to be written off (NB the
Jigures in the attached paper are wrong on this issue). Cancellation would
also have a big effect on their reputation. It would probably force the
resignation of the Chief Executive and would put them into a protracted legal

battle with government, in which they would seek to pin blame on government

in general and the DSS in particular. is See le bey
fron F-\h4 Sela I

Effects on Fujitsu: Fujitsu have provided the financial commitment to secure

a deal. This is a sign of how important the project is for them. There would
undoubtedly be a cost in terms of UK relationships with them in particular,
and possibly with Japanese investors in general. They would argue that the
government was guilty of a breach of faith: blaming ICL for what is in effect

a change of policy.
Effects on PFI: continuation would set a precedent for renegotiating PFIs that

are failing; cancellation would make the funding of large IT projects harder in

the future.

RESTRICTED

CBO00100001_072
CBO00100001_072

Effects on the Benefits Agency and welfare reform: continuation would
hold up ACT. DSS estimate that they would save £800m if the project was

cancelled. Government would also lose the opportunity to bank the unbanked.

. Departments remain divided. Alastair Darling remains strongly opposed to

continuing. Ian McCartney for DTI will argue strongly for accepting a deal
(Peter Mandelson has largely kept out of the discussions). The Treasury is
divided at official level, but Stephen Byers will probably, on balance, want to

accept the deal for pragmatic reasons, even though he would prefer to cancel.

. At first glance, most of the factors point towards continuation. However my
view, which Lord Falconer broadly shares, is that although short-term

considerations and expedience point strongly towards making a deal, this will

in the long-run prove unsatisfactory, leaving the Post Office and government

dependent on a hugely expensive, inflexible, inappropriate and possibly

unreliable system.

. In our view, the best outcome would be a deal with ICL to continue with an

automation strategy which drops the Benefit Payment Card but focuses instead

on helping the Post Office to provide banking services. This would enable a

rapid move towards ACT and Post Office automation, but without the many
problems associated with the BPC. It would also enable good relations to be
maintained with ICL and Fujitsu. The costs to government and ICL resulting

from the BPC could legitimately be blamed on the previous government.

10.However, it will not be possible to reach this position without first rejecting

ICL’s offer and making it clear that the government no longer wishes to

RESTRICTED

CBO00100001_072
CBO00100001_072

RESTRICTED

~S-

continue with Horizon in its present form. There would then be a danger that
ICL would refuse to negotiate over a deal which excluded the BPC. In other
words, this option is only feasible if the government is prepared to accept the

significant risk that it would in practice lead to complete cancellation.
-Our view is that the case for cancellation is, just, strong enough to warrant
that risk, although it would need to be announced alongside a strong package

of support for Post Offices and for an alternative automation strategy.

2.Ministers are meeting on Monday afternoon. There are essentially three

options:

Option 1: Attempt to conclude a deal, broadly along the lines offered by ICL

Option 2: Seek a deal with ICL but excluding the Benefit Payment Card

(recognising that in practice this may quickly lead to option 3)

Option 3: Reject ICL’s offer, move to terminate the project and press forward

with an alternative. die )

13. Do you wish to give a steer?

RESTRICTED

CBO00100001_072
CBO00100001_072