RESTRICTED
Review of the Benefits Agency/Post Office Counters
Automation Project
Working Group Report
July 1998
HMT00000034
HMT00000034
RESTRICTED
Contents
1 Summary and conclusions
2 Introduction
3 Key issues
3.1 The Independent Panel's assessment
3.2 Prospects for paying benefits by ACT
3.3 Prospects for developing banking and other services across PO counters
3.4 Role of PO counters in providing wider government services
4 Options for the future
4.1. Summary of the options
4.2 Feasibility of the options
4.3 Implications for Benefits Agency/DSS
4.4 — Implications for PO counters
4.5 Financial implications
4.6 Implications for ICL
4.7 Presentation and handling
wn
Annexes
Benefit payment
Post Office counters network
The cost structure of Post Office Counters
The project
Recent Ministerial statements
Legal advice
Analysis of financial flows
QAmAMoOaAW>
a
Report of the Independent Panel
Page
11
11
13
17
20
23
23
25
27
28
30
31
33
35:
35
38
40
46
49
31
59
HMT00000034
HMT00000034
HMT00000034
HMT00000034
RESTRICTED
HMT00000034
HMT00000034
RESTRICTED
& Summary and conclusions
1.1 This is the Teport of an interdepartmental Teview of the PFI Project in which ICL Pathway
have contracted with Post Office Counters (POCL) and the Department of Social Security
(DSS) to automate post office counters including facilities for Payment of benefits using
a benefit Payment card (BPC). The review was commissioned because of concerns,
particularly on the Part of DSS Ministers, that the Project was badly behind schedule, over-
budget and indeed might not be viable, We were asked to consider:
- whether the Project is technically viable; and if So, how quickly it can be
completed and at what cost to government;
- the direct and indirect costs of cancellation and of any alternative available to
deliver the Project’s objectives,
14 Against this background we think the choice for Ministers lies between three options:
1 Seeking to continue with the project; negotiating with ICL Over a limited
extension, perhaps to 2007; making plans for a three year transition to benefit
Payment by ACT completing by the end of the contract; and using the platform
HMT00000034
HMT00000034
RESTRICTED
provided by this Project to develop banking and other services across PO counters;
2 Reshaping the project by cancelling the benefit payment card and seeking to
develop the temaining elements of the programme to give POCL a banking
capability; moving to ACT at a rate which takes account of POCL’s ability to
provide banking Services, and so retain as much as possible of the benefit payment
business; and redirecting administrative savings in BA to supporting POCL for the
life of the current contracts; and
3 Terminating the whole project on grounds of non-performance by ICL Pathway;
developing alternative plans for automation of PO counters; moving to ACT ata
Tate which takes account of POCL’s ability to provide banking services; and
providing some subsidy while POCL plans changes to the network.
The decision will have far-reaching Consequences for the way benefits are paid to around
19 million people, and for the viability of the Post office network. Key considerations in
deciding between these options are:
- What is the best way of paying social security benefits? Current Paper based
methods are inefficient, Prone to fraud and hard to audit. The BPC will save
perhaps £110 million a year on fraud and allow Proper auditing, but will cost about
the same to administer as the Current system. Payment direct into bank accounts
by ACT (already used by 28% of claimants) would deliver all the main benefits
would provide a better platform for welfare reform. For these reasons we agree
at some point in the next 10 years it should become the normal method of benefit
payment;
- What is the impact on the Post office counters network? The central dilemma
is that 80% of BA’s administrative cost in paying benefits is income for POCL,
and that income, together with the Steady flow of claimants through post offices,
- Could post office counters provide cash to claimants who get benefits paid
into bank accounts? This would suit BA/DSS because they would like to offer
claimants the option of continuing to collect their cash weekly from the PO; and
HMT00000034
HMT00000034
RESTRICTED
it might help POCL because it would provide some income and a continuing flow
of customers, It is certainly possible: POCL already provide banking services for
could seek new partners, but to do this effectively ona large scale, they need new
technology, either the ICL system or another. But we need to be realistic about the
implication for POCL, Once benefit money is in claimants’ bank accounts they
may well find it more convenient to draw cash from a bank branch, cash machine
(ATM) or retailer’s cashback facility;
What role is there for Post office counters in the long term? POCL cannot rely
On current sources of income to sustain it through the next decade, and in our view
has been slow to develop a business strategy which addresses this challenge. The
PO is pinning its hopes on Providing a range of banking and financial services
across counters, and becoming a ‘front end? for more Government services, We
see some scope for both (particularly for banking), but think it is very doubtful
Tevenue from these sources could offset loss of income from BA, If that is right,
the issue of scaling back the network or providing significant subsidy will have to
be faced at some point whatever happens on this Project. Option 1 Postpones it for
a few years and gives PO a better chance to build new sources of income;
Parties withdrew they would seek to Tecover the compensation the contracts
provide for (up to £335 million in the event of termination ‘for convenience’). The
Programme lawyer’s view is that the Government’s case is a strong one, but any
litigation would inevitably be messy and uncertain;
What are the costs to government of continuation or cancellation? There is
unfortunately no simple answer to this. An early switch to ACT has the potential
to bring forward large efficiency savings for DSS/BA who need to consider
HMT00000034
HMT00000034
RESTRICTED
under each option. There is therefore no clear basis for choosing between the
options on cost grounds, but options 1 and 2 look a bit better than option 3. In each
case there are implications for DSS funding which will have to be resolved with
the Treasury.
What would be the impact on ICL and the UK computer industry? Failure of
the project would undoubtedly be very damaging to ICL and might even bring the
company down. ICL have 270 staff working directly on the project, and hope to
market their system overseas, Fujitsu has plans to float ICL by 2000 which would
be set back.
What are the implications for PFI? This is one of the biggest PFI projects. Any
move to shield ICL from the risks it took on would make it more difficult to
achieve effective risk transfer on other projects, and could risk Challenge from
bidders who lost out in the original procurement process. It could also encourage
ICL to come back for more money later. There are parallels with the National
Insurance Recording System (NIRS2) project which has also suffered delays. In
that case Andersen Consulting has been expected to swallow serious losses, and
indeed to compensate the department for the delays.
How does the project link with wider government objectives? An early move
to ACT would help promote electronic delivery of government services and
services which are part of normal life for most people. Post offices also have a role
in providing services locally, often in villages or on estates where there may be no
other shop; and could play a growing role in future in Providing an access point
for government information and electronic transactions with government.
What are the legal constraints? In broad terms we cannot vary the terms of the
contract in a way that would allow a disappointed bidder to claim they would have
won the competition on the revised terms and were therefore treated unfairly; or
would allow alternative suppliers to claim they have not had the opportunity to bid
for new work. This limits what could be offered to ICL to continue,
How will this look to the public and benefit recipients? There are tricky
problems with presenting each option. With option 1 DSS will need to take 19
million claimants through the upheaval of switching from order books to cards
with the prospect of a second switch to ACT a few years later (although this
could be a switch from the payment card to a similar bank card). Options 2 and
3 will immediately taise fears for the future of the post office network; and for
claimants will mean losing the option of payment in cash sooner than under option
1. Such fears will no doubt be orchestrated by the Federation of Sub-Postmasters,
HMT00000034
HMT00000034
RESTRICTED
CWU and rural lobby. But in each case a reasonable account could be given of
Steps being taken to protect the network.
1.6 To sum up on the options:
- Option 1 will involve ICL swallowing substantial losses. If it can be delivered,
it has the advantage that it secures the PO network until around 2007; and gives
POCL more time to try to build new business and banking facilities to replace BA
income when the contract ends. It avoids the prospect of protracted and costly
litigation which could follow cancellation of the project altogether. On the other
hand it means foregoing for several years the substantial efficiency savings
(around £400 million a year) available to BA from moving to ACT; it involves two
upheavals in the benefit payment system in the space of just a few years; and could
well involve a formal direction to the BA accounting officer.
- Option 2 costs about the same overall as option 1(on our modelling). DSS/BA
believe it would give them more clarity of purpose. It could be presented as taking
forward a range of Government objectives including positive steps toward
universal banking. It would involve only one upheaval in the benefit system and
could deliver substantial efficiency savings for BA a few years earlier than option
1. It could help ICL by making the project more deliverable, however it is not clear
that there is the basis for a commercial deal with POCL on a stripped down
project. The option seeks to protect the PO network from any dramatic short term
dip by re-directing BA savings to supporting POCL, however loss of ‘footfall’
income from claimants would hit sub-postmasters and POCL bringing some
closures. The faster pace of adjustment would worsen POCL’s medium term
prospects,
- Option 3 offers the same benefits to BA as option 2, but complete termination
would be more damaging to ICL who would be more likely to pursue claims
against both POCL and BA. This would be messy and uncertain. POCL would be
starting from scratch in developing counters computerisation, and the delay could
be a serious blow to its business Prospects. It would also be a setback for National
Savings who are counting on counters automation to reduce their administrative
costs. On the other hand POCL could invest in the latest generic banking/retailing
systems which on their own are likely to be considerably cheaper than a system
built around the requirements of the benefit payment card; and limited front end
banking services could be offered quickly.
Recommendations
1.7 No option is without difficulties, but each looks doable. The Working Group, apart from
DSS and the Treasury Social Security team, recommends that:
18
19
HMT00000034
HMT00000034
RESTRICTED
= in line with option 1, ICL Pathway should be given terms for continuing with the
project: there can be no increase in the level of payments already contracted; we
are prepared to discuss an extension of Contracts, but only to ensure a5 year period
of operation following full implementation; and there will be penalties if
completion of roll out is delayed beyond September 2002;
- if ICL agree to these terms, the measures proposed by the Panel should be
implemented forthwith. In particular a ‘troubleshooter’ should be appointed to
work with the three parties; and POCL’s management of the project must be
strengthened;
= if ICL cannot accept these terms, or if the negotiations cannot be satisfactorily
concluded within two months, Ministers should sanction the public sector parties
withdrawing from the contracts on the grounds of ICL Pathway’s non-
performance and we should implement option 3.
- POCL senior management needs to be strengthened, ideally by bringing in
someone with a good track record in Tunning a retail network.
DSS recommend a similar approach with ICL, but based on option 2. Treasury Social
Security team prefers option 3,
Whichever option is chosen, it is important that Ministers should give an early steer. We
can provide more detailed advice on handling once the broad way forward is clear,
HMT00000034
HMT00000034
RESTRICTED
2 Introduction
Background
2.1 The BA/POCL automation project (which POCL refer to as the ‘Horizon’ project) was
initiated in 1993 with the aim of:
= providing a more secure and efficient way of paying benefits, with potential fraud
savings of £185 million a year (of which around £70 million is now being saved
by the Electronic Stop Notice System);
= providing the DSS/BA with the means to account fully for their vast programme
expenditure, now nearing £100 billion a year;
- automating PO counters, to make current business more efficient and help them
win new business. The project also helped to secure the PO network by providing
a secure revenue stream from POCL’s biggest customer until the middle of the
next decade.
2.2 A contract worth around £1 billion was signed with ICL Pathway in May 1996. It isa
huge project, involving automation in 19,000 post offices and 40,000 counter points, and
providing links into DSS/BA and PO systems. The project is described in more detail in
Annex D.
2.3. A revised timetable for the project was agreed by all parties in February 1997. Since then,
it has slipped 21 months behind schedule. The two public sector parties attribute the cause
of the slippage to ICL Pathway in all material tespects (based on an independent review
of the programme by consultants Project Mentors), In September 1997, DSS Ministers
became sufficiently concerned to draw this to the attention of Treasury and DTI
colleagues, and discussions took place between the three parties, facilitated by PA
Consulting, to see what could be done to minimise the delays. On 21 November 1997, a
key.milestone was missed by ICL Pathway for completing an operational live trial, and
the public sector parties placed ICL Pathway in breach of contract. Since then there have
been further delays to the programme, and ICL have sought extensions to the contract or
increases in charges. In May this year, DSS/BA issued a notice of 'cure' to ICL Pathway,
in order to protect their negotiating position.
The review
2.4 Against this background, the Secretary of State for Social Security wrote to the Prime
Minister in February that ‘there is a serious tisk that this project will fail to deliver its
objectives - or will not do so within a timescale that will make it worthwhile’. She
proposed early discussions to find an agreed way forward, emphasising the urgency of a
decision in view of the continuing fraud losses to DSS and the need to agree new
HMT00000034
HMT00000034
RESTRICTED
timetables with ICL if the project were to proceed. DSS was concerned that, since it
guarantees a minimum level of payment to POCL, it in effect carries much of the financial
tisk on the project and feels the effect of delays in a way that POCL do not.
2.5 In response to later correspondence the Chief Secretary proposed that before decisions
were taken, an interdepartmental working group led by the Treasury should be asked to
Prepare an agreed analysis of the options, including an assessment of:
- whether the project is technically viable; and if so, how quickly it can be
completed and at what cost to government;
= the direct and indirect costs of cancellation and of any alternative available to
deliver the project’s objectives.
This approach was agreed by other Ministers. This report is the result of that work.
2.6 The Working Group was made up of officials from the Treasury, DTI, DSS and the
Central Information Technology Unit. The Group set up an Independent Panel under the
chairmanship of Adrian Montague, head of the Treasury’s Private Finance Task Force, and
assisted by PA Consulting, to provide a thorough and balanced assessment of the technical
viability of the project, and its likely timing and costs. Other members were Alec Wylie
(former head of the Northern Ireland Social Security Agency) and Bill Robins (former
Director General of Communications and Information Services at the MOD) both of
whom have substantial prior experience of large IT projects. Their report forms Section
6 of this report.
Structure of the report
2.7 The structure of the report is:
- Section 3 reviews key issues to be considered in assessing the options;
- Section 4 analyses three broad options against a range of criteria;
- Section 5 contains a set of background notes on specific issues;
- Section 6 is the report of the Independent Panel.
10
HMT00000034
HMT00000034
RESTRICTED
3 Key issues to be considered
3.1 The Independent Panel's assessment
3.1.1 The Panel’s teport forms Section 6 of this report. The Panel was asked to consider three
main points:
- whether the project is technically viable;
- when it could be delivered;
- and at what cost.
3.1.2 In addition to examining these questions the Panel identified a number of fundamental
strategic issues (such as the ability of POCL to manage this large project) which are
fundamental to the Programme and says there is a severe risk of failure if they remain
unresolved. Resolution depends on renewed commitment and cooperation from the parties,
3.1.3 The Panel also considered with the parties a number of possible ways forward.
Technical viability
3.1.4 The Panel has concluded that the project is technically viable, although there are some
tisks, in particular around:
=- scaleablility and robustness. The Programme is probably the biggest of its kind and
the system has had to be tested at the level of its component parts. The Panel is
satisfied these risks are being well managed by ICL Pathway, but they nevertheless
remain.
- the system is (necessarily) heavily dependent on a third party ‘middleware’
* product called ‘Riposte’. ICL Pathway have taken Steps to cover their dependency
on this product.
3.1.5 The Panel also believes the basic infrastructure is very robust for the future and is
generally based on industry standard products. It should therefore allow POCL to compete
for new business ina variety of markets, and for example develop new applications based
on smart cards. If significant on-line capacity was required more development work may
be needed.
3.1.6 The Panel has seen no evidence to suggest that the systems being developed by BA and
POCL to connect up to the systems being developed by Pathway will not work as
required.
HMT00000034
HMT00000034
RESTRICTED
Timetable and risks
3.1.7. The original timetable was that roll out of the full system in post offices would start in
June 1997. The February 1997 re-plan put this back to January 1998. The current timetable
would see it start in April 1999. The Panel’s best estimate is
> that there will be a further 9 months delay to January 2000, in the start of roll out;
but to make this possible, a number of outstanding issues on the critical path
would have to be resolved fast (by the end of July). These issues are not
considered to be insurmountable and indeed are expected to be resolvable easily
provided the parties cooperate. The date could be brought forward with
commitment and goodwill on both sides;
_ the roll out programme should be finished within 21 months (by Sepfember 2001).
This requires offices to be brought into the system at a rate of 300 a week. The
Panel believes this has been properly planned and is sustainable but notes there are
some risks around the timetable and the response of POCL’s regionally based
organisation to the demand placed on it.
3.1.8 The Panel has made a number of recommendations on how the project should be managed
to deliver this schedule. These involve early resolution of outstanding issues and
strengthening the Horizon Project Office (set up by POCL) and POCL management. The
Panel also proposes appointing a neutral Government “troubleshooter” directly responsible
to DSS and DTI ministers to help take the project forward. (This would require careful
handling: DSS are concerned about possible conflicts with BA’s accounting officer
tesponsibilities; DTI would want the relationship with the PO Board to be looked at
carefully). The project would effectively need to be restarted with renewed commitment
from all parties.
Cost
3.1.9 The Panel reports estimates provided by the parties: the direct cost of the extra nine
months delay will be £90 million for BA (in the cost of running existing systems etc), £15
million for POCL and £72 million for Pathway on top of costs already incurred. In
addition BA lose fraud savings of around £110 million a year until the card system starts,
3.1.10 However the the Panel does not provide an estimate of the likely cost to Government of
reaching agreement with ICL on terms for continuation of the project. At the end of last
year ICL sought an extension of 5 years or 30% more money. This is a negotiating
position, but the subsequent delays will have worsened their financial position. Our own
very rough estimate is that even if the contract is extended to Testore a 5 year steady state
period of operation, ICL might face a shortfall of the order of £250 million. Our view is
that the cost of continuing will depend on the outcome of negotiations with ICL. We are
under no obligation to ensure ICL makes a return, or even breaks even. But if we want the
12
HMT00000034
HMT00000034
RESTRICTED
project to continue, ICL will need to be in a position where they judge the losses on the
project are bearable and returns from continuing outweigh the likely return from
withdrawing and suing.
Options
3.1.11 The Panel explored a number of options with the parties to see if any could command
broad support. These are set out in an Annex to the Panel’s report.
3.2 Prospects for paying benefits by ACT
3.2.1 To put the project in perspective, and understand the alternatives, we need to look at
different ways benefits can be paid, and in particular at the option of paying direct into
claimants’ bank accounts by automated credit transfer (ACT).
Background
3.2.2 Of the 27 million benefit claimants, around 19 million are currently paid using paper
based methods across post office counters (order books or Girocheques) and 7 million
are paid by ACT. The proportion paid by ACT has been rising steadily from around 16%
5 years ago to nearly 30% today. It is expected to go on rising by around 1 percentage
point a year, since around half of new pension and child benefit claimants opt for ACT.
The project will move the 19 million using paper methods to the benefit payment card
(BPC). Annex A sets out more detail on payment methods.
3.2.3 ACT has a number of advantages:
= it is very much cheaper for BA than other methods, costing less that 2p per
transaction compared to 47p for order books and ‘78p for giros. It is also cheaper
than transactions using the BPC. (It should be acknowledged that it is cheaper for
BA partly because the cost of providing cash to claimants is transferred to the
banking system.)
a it gives BA a full reconciliation of transactions (as does the BPC);
= it cuts opportunities for fraud (as does the BPC);
- it is a widely used generic system which allows DSS to ‘piggy back’ on wider
technological change.
3.2.4 For these reasons, the DSS have long seen ACT as the best long term solution for benefit
13
HMT00000034
HMT00000034
RESTRICTED
Payment. However there have been three main obstacles to adopting ACT as the standard:
= around 2% million claimants do not now have access to current accounts:
- there are 16% million People who do have access to accounts but do not choose
ACT and may resist the change, for example because they find it most convenient
to get their cash from a post Office, some for mobility or health reasons;
- the loss of business would threaten viability of many post offices.
3.2.5 We consider these issues in turn:
The unbanked
3.2.6 Research by DSS, backed up by recent research by the British Bankers Association
(BBA), suggests that around 90% of benefit claimants now have access to a current
account at a bank or building society, leaving only 10%, around 2¥% million people,
without access,
3.2.7. The BBA research Suggests that the great bulk of the unbanked choose not to have
accounts and are not unbanked because they have been refused or would be tefused a
bank account. The unbanked include:
es young people who have not yet Opened an account;
- people with limited funds who prefer to retain the control over their finances that
a cash budget gives them;
- elderly people who may be unfamiliar with banking.
3.2.8 They will also include:
- people who have been, or would, be refused accounts because they cannot produce
the standard identity required by banks, or who have poor credit records;
- people who wish to avoid the banking system for illegitimate reasons - to avoid
tax, to defraud the Benefits Agency etc.
3.2.9 Most banks now offer simple credit only accounts which provide a bank card for
withdrawals but do not allow overdrafts. Some banks only offer these accounts to
particular groups (for example 16 to 19 year olds); other banks apply no such restrictions.
3.2.10 This suggests that if the currently unbanked were given suitable notice that they would
need to open an account in order to receive their benefits, there would be no practical
14
3.2.11
HMT00000034
HMT00000034
RESTRICTED
difficulty for most of them in them doing so. (They may of course be resistant to this for
other reasons.) Those who are unable to open a full account would in most cases be able
to open a credit only account. Inevitably there would be a number of people who could
not open accounts for example because they have no fixed address, or suffer from mental
illness. As under the BPC proposal, special arrangements would have to be made for these
people. 2
However it is not clear how the banks would react to a large number of people - mostly
on low incomes and some of doubtful creditworthiness - being required to have access to
a bank account to receive their benefit. Banks may respond for example by introducing
new charges for withdrawals. Discussions with the banks will be required, but informal
soundings by DSS suggest that at least two major banks would be interested in developing
their services in a way which meets the majority of DSS customers needs. The costing of
options set out below includes as a contingency a sum of £30 million a year to provide an
incentive, although this may well not be needed.
Will claimants with bank accounts accept the switch?
3.2.12
3:2.13;
DSS research indicates that many benefit recipients want a method of payment which:
- allows them to collect payments in cash, and budget over a short period;
= allows local collection, typically from the post office;
- for poorer people, allows frequent (weekly) payment of small amounts;
- does not require them to collect all their benefit due at once (as the BPC does).
Claimants tend to use different benefits for different things eg. using one benefit
order book to "save up"; and do not like the responsibility of handling large
amounts of cash.
Using a bank account to draw social security benefits may not be attractive for some
claimants because:
= most bank accounts provide fewer access points for cash than there are post
offices. So people may have to travel further to get their cash which may be a
problem in rural areas and deprived urban areas, and for the less mobile.
> ATMs (automated teller machines) only provide cash in notes not coin, so a
benefit claimant may have difficulty withdrawing all their benefit without going
into overdraft;
- many people find bank accounts and bank machines intimidating and prefer the
familiarity of the post office and the ease of controlling budgets in cash;
15
HMT00000034
HMT00000034
RESTRICTED
> the timing of payments through bank accounts may be unattractive, For example
pensions paid by ACT are currently paid monthly in arrears, while order book
payments are made weekly in advance.
3.2.14 Each of these issues can be addressed:
<= the number of outlets for cash is growing rapidly. There are now ATMs in over
10,000 separate locations, and many supermarkets offer bank accounts and
cashback facilities. The Post Office itself now offers banking facilities to
customers of three banks. In the near future cashback arrangements are expected
to spread rapidly to shops, garages and pubs. (Terminals are available for rent at
around £15-30 a month witha charge to the retailer of 10-25p per transaction.)
- many people have become more used to using bank accounts with the rapid switch
from cash payment of wages to payment by ACT which took place in the 1980s.
Spreading the banking habit further would fit well with Government’s objectives
for tackling social exclusion, encouraging savings, and encouraging people to use
the most modern and efficient means available to conduct transactions,
- DSS will be looking at the options for adjusting the “periodicity” of benefit
payments to remove any deterrent to switching to ACT.
3.2.15 These changes will tend to make ACT more acceptable, but DSS believe that the most
significant step towards making ACT an acceptable solution for claimants would be to
provide for them to withdraw cash from their bank accounts at post offices, POCL already
provides this service (using inefficient paper based methods) to customers of the Giro
Bank, Cooperative bank and (on a trial basis) Lloyds TSB. But providing such facilities
for millions of claimants would require investment in at least some simple banking
technology. We consider below the scope for POCL delivering this, either through the
Horizon project or some other system.
3.2.16 Clearly if this were achievable, it would offer the Post Office the chance to generate some
income from banking services to replace, at least in part, that lost from the Benefits
Agency when it switches to ACT.
How long would it take?
3.2.17 DSS envisage that the process would involve around three years of preparation and three
years of transition. In the preparation phase DSS would:
= draw up plans for a benefit by benefit migration to ACT;
- tun a publicity campaign to encourage voluntary take up of ACT, explain the
16
HMT00000034
HMT00000034
RESTRICTED
transition and offer help with opening bank accounts;
take the necessary secondary legislation to make ACT the standard method of
payment except where individual claimants can show they cannot open a bank
account or ACT would be for some reason unsuitable;
continue discussion with the banks about availability of accounts for the currently
unbanked. DSS may for example aim to draw up a list of banks prepared to offer
simple credit only accounts.
maintain a dialogue with POCL on a timetable for introducing banking services
over post office counters. (DSS wish to involve POCL fully in the move to ACT.)
3.2.18 In the transition phase, the DSS would:
set a series of dates after which ACT would be the normal method of payment
(other than for exceptional cases) for each benefit;
provide continuing support and advice for claimants seeking to open bank
accounts.
paper based methods (or the BPC) would be phased out as ACT came in.
3.2.19 We agree that ACT is the best long term solution; and DSS should be planning to move
to ACT at some point in the next ten years. The question is when, and whether it is now
sensible to move first to the BPC?
3.3
33-1
Prospects for developing banking and other services across PO counters
The main reason why ACT has not been pursued in the past has been that Ministers have
not been prepared to accept the damage to the post office network that would result. This
was a key factor in the decision to pursue the Horizon project. So a central issue in
assessing the options now is how far could POCL replace income lost from BA by
developing new business such as banking services.
3.3.2 There is an opportunity for POCL to provide:
banking services to benefit recipients who have a bank account so that they
withdraw their cash over PO Counters;
basic credit accounts (in association with a private sector partner) for the
“involuntarily unbanked” - those who would not otherwise be able to obtain a bank
account;
17
HMT00000034
HMT00000034
RESTRICTED
- accessible banking services for the public in general as bank branches close.
3.3.3. These services are already provided to customers of the Girobank (in 18,000 post offices),
Cooperative Bank (15,000 post offices), and (on a trial basis) Lloyds TSB (840 offices).
But current paper based methods are inefficient and expensive, and customers cannot for
example check their bank balance.
3.3.4 To extend the range of services it could provide, the Post Office would need to:
- develop a technology platform; and
- establish commercial partnerships with banks, and get agreement to link their
systems into bank systems.
3.3.5 We consider these issues in turn.
The technology
3.3.6 There appear to be three possible routes through which the Post Office could establish
banking technology;
- building on the Horizon system. The system ICL are installing can be adapted
relatively easily to provide a wide range of banking transactions. The investment
would cost around £5 million for additional hardware (keypads) and up to £10
million for software development. The main constraints on implementation are
resources within ICL and POCL, and getting agreements with the banks to link
into their networks. ICL are naturally focussed on rolling out the benefit payment
card system first. But in principle a banking system based on one partner bank
could be rolled out towards the end of 2001, and with several banks perhaps a year
later. Horizon can support smart card technology, and POCL see development of
a PO smartcard as a central part of their development strategy.
- by using the new on-line debit card terminals. Around 10 million bank
customers who are regarded as too risky to hold a conventional debit card now
hold ‘on line’ debit cards (Electron or Solo) which authorise debits or cash
withdrawals up to the limit of what is in the account. The retailer runs the card
through a machine linked to a telephone line which provides on-line authorisation.
Accounter clerk in a post office could use it in the same way. This would be quick
and cheap to install (£15-30 per month rental per machine plus transaction costs)
but would support only a limited range of functions, and banks are more likely to
charge retailers for such transactions than to pay for them.
= holding a new competition for a counters automation platform. It would take
18
3.3.7
HMT00000034
HMT00000034
RESTRICTED
some time to get a fully integrated system in place, but the Post Office would be
able to acquire more generic banking technology, perhaps more cheaply than the
Horizon system (though with more limited functions).
Each seems a plausible route for POCL to develop the technology for banking services,
Partnerships with the banks
3.3.8
3.3.9
3.3.10
3311
POCL envisage that it would not act directly as a bank itself, but would act as an agent for
and partner with a range of banks. The working group agrees that this is a sensible
approach. The Post Office would be an extension of the banking system, not a separate
‘poor people’s bank’.
What’s in it for the banks? The banks are reducing their branch networks to cut costs but
they still need to provide physical access for their customers for some basic services. The
PO could be attractive to the banks since it is an extensive network, in particular providing
access for those in lower socio-economic groups. (55% of CDEs live within half a mile
of a bank or building society whereas 78% live within half a mile of a post office. Only
9% of rural parishes have a bank or building society whilst 57% have a post office.)
Furthermore it is an organisation that is trusted by the public.
The banks may be willing to put some of their savings from branch closures into the Post
Office to provide coverage for customers that their own networks can no longer achieve,
especially in rural and deprived urban areas, or to discourage customers from visiting their
branches for low value cash withdrawals.
But this is not straightforward. The banks may not view post offices as a suitable interface
with their customers. They may well see the future in the next generation of smartcard
technology and be reluctant to commit to the Post Office. (The Horizon system can
Support smartcards, but it would require considerable design and engineering effort to
enable POCL to offer banking services using the next generation of smartcards.) And there
is a limit to what the banks will pay for the service.
Benefits for POCL
3.3.12
POCL, drawing on work by McKinseys, have estimated that the revenue streams they
could generate from banking might be:
- £220-270 million from providing front-end banking facilities for existing benefit
recipients migrated to ACT accounts with the High Street banks;
- £20-80 million from existing non-benefit recipient customers of the High Street
banks who might choose to carry out some of their basic banking business at post
offices;
HMT00000034
HMT00000034
RESTRICTED
- £15-35 million from wholesale cash distribution for a few major banks using their
current distribution network;
giving a total gross income of £235-385 million a year.
3.3.13 We regard these estimates as extremely ambitious and the timetable for achieving them
uncertain. People have a wide range of alternatives to the post office for cash withdrawals.
Moreover these figures do not include the additional costs POCL would incur in providing
these services.
3.3.14 In short it looks perfectly possible in technical terms for POCL within a few years to offer
a range of banking services with suitable technology, and therefore to offer cash
withdrawals to claimants receiving benefits through ACT if they have an account with one
of a range of partner banks. But there must remain a question mark over the commercial
potential for POCL in the longer term.
3.4 Role of PO counters in providing wider government services
3.4.1 POCL already provides a range of government services including National Savings,
vehicle licensing, TV licensing, passport applications and information about benefits. The
Horizon project will make most of these more efficient and provide clients with better
information. This will help POCL to retain existing clients.
3.4.2 POCL is also exploring the scope for generating new income in the medium term by
becoming the ‘front end’ for:
- information about government services such as benefit entitlement, tax or training
provision;
- registration and notification such as notifying a death to different agencies, or
registering a new small business;
- interactive services such as searching for jobs or training places.
3.4.3 This would build on POCL’s existing relationship with customers and trusted brand
image. As well as existing paper based routes (such as vehicle licensing) these services
could be provided using:
: the Horizon platform: Horizon allows a range of transactions (such as National
Savings) to be done more efficiently, and counter staff to be prompted to ask
security questions. POCL’s preferred route would be to build on the planned
system using smartcard technology to provide stored value and customer
20
HMT00000034
HMT00000034
RESTRICTED
information, and a digital signature for security purposes. The Horizon system can
support smart cards and in this sense is “future-proof” However real time on-line
applications for immediate authorisation would require major changes to the
system.
alternative (or additional) platforms: for example POCL could provide
information kiosks at branches and act as a “call centre”.
3.4.4 POCL believe there are significant efficiency gains to be had by combining information
services under a “one stop shop”. POCL’s advisers estimate that efficiency savings of
around £1,000 million could be achieved across government, and that POCL might secure
£160-£280 million of these savings.
3.4.5 Our view is that there is certainly scope for improving efficiency of existing services, but I
there must be considerable uncertainty about both the timing and the role of post offices I 1.
in future provision of government services and therefore the potential for generating
revenue from this source.
Summary
3.4.6 In summary:
the provision of basic banking services in partnership with the Banks could
provide a plausible alternative income stream for POCL to compensate in part for
the loss of BA revenue from the switch to ACT;
POCL may also be able to generate some income through the provision of “front-
end” central and local government information - though this is much more
uncertain;
these alternative income streams are unlikely, even on optimistic assumptions, to
be sufficient to replace existing benefit income and associated footfall - therefore
(assuming a transfer to ACT at some stage) there will have to be some
restructuring of the PO network over time and/or explicit subsidy.
”
HMT00000034
HMT00000034
RESTRICTED
22
4.0.1
4.0.2
4.0.3
41
HMT00000034
HMT00000034
RESTRICTED
Options for the future
This section of the report looks in detail at the options now available. We have considered
three families of options:
- continuing with the project;
- restructuring the project without the benefit payment card;
- cancelling the whole project.
All options are based on an understanding that Ministers want to find a solution which:
- gives DSS/BA a more secure and cost effective system of benefit payment, and
allows claimants to continue to collect cash from post offices if they wish;
- gives POCL an automated infrastructure which will help it to deliver banking and
financial services, as part of a long term strategy to preserve a viable network.
In each case we identify a lead variant which looks the most practical and goes furthest
towards meeting the objectives of the different parties involved. Annex G models the
impact of each option and sets out key assumptions. In choosing between them the
following issues need to be considered:
- the feasibility of the option; what risks are there around its delivery?
= the Government’s objectives for welfare reform and in particular the need for BA
to implement modern methods of benefit payment;
= the impact on the post office network;
— + the overall impact on public spending;
- the implications for ICL; and
the impact on Government objectives more widely including the wish to move to
universal banking.
Summary of the options
Option 1 involves continuing with the contract with some limited extension of its
duration:
- the project continues, including the benefit payment card, with roll out projected
23
HMT00000034
HMT00000034
RESTRICTED
to start in January 2000.
this would require implementation of the Panel’s recommendations, a renewed
commitment by the parties and agreement with ICL on funding. We do not favour
any increase in charges, but some extension of the contracts to provide for 5 years
of steady state operation after full implementation of the system should be
considered. (Even this will probably leave ICL with a loss on the project overall.)
for this analysis we assume contracts are extended to allow a three year transition
from the BPC to ACT starting early in 2005; DSS’s preferred variant is to start in
2003. In any event there must be a firm date for ending the contract, and DSS
would want to present it as part of a phased transition to ACT;
the BPC would be introduced as the system is rolled out, steadily replacing paper
based payment methods;
POCL would agree with ICL on building banking services into Horizon and seek
partnerships with the banks with a view to at least some partners being in place by
October 2001 and full facilities two years later.
Option 2 involves restructuring the project without the BPC to allow an earlier move to
ACT and parallel provision of banking services in post offices:
BA withdraws from all contracts with ICL for development of the benefit payment
card on grounds of failure to perform.
POCL works with ICL to implement the Horizon technology which is rolled out
between January 2000 and September 2001; and builds on it the capacity to offer
banking services with at least some partner banks by October 2001;
BA draws up plans to migrate to ACT over three years from October 2001. In the
meantime paper based methods of payment continue and BA strengthens anti-
fraud mechanisms;
BA negotiates a new contract with POCL which leaves current arrangements in
place until the migration to ACT is complete;
the Government continues to fund POCL for the equivalent to the ‘floor’ level in
the current contracts with BA (£370 million, declining over time) at least until
2005/6.
Option 3 involves cancelling the whole project on the basis of ICL’s breach of contract
for failure to meet agreed timetable:
24
4.1.4
4.2
4.2.1
HMT00000034
HMT00000034
RESTRICTED
POCL seeks to put in place a separate integrated technology platform to enable it
to offer banking and other services as soon as possible. We assume some basic
technology could be put in place by October 2001 (perhaps based on simple swipe
card terminals) with more comprehensive facilities to provide full banking services
by 2003.
BA draws up plans to migrate to ACT over three years from October 2001. In the
meantime paper based methods of payment continue and BA strengthens anti-
fraud mechanisms, perhaps building on the current electronic stop notice system
(ESNS).
In each option we think it will be necessary to take a long hard look at the viability of the
PO counters network in the medium term and draw up a proper corporate strategy aimed
at cutting the costs of the network while maintaining and where possible improving
services, and developing new business.
Feasibility of the options - what are the risks?
The feasibility of continuing with the project has been examined in detail by the Panel,
and they conclude that it is technically viable and can support banking technology. The
main risks around Option 1 are:
whether the problems which have dogged the project so far can be resolved.
Success of the project on a reasonable timetable depends crucially on the Expert
Panel’s recommendations being implemented, on ICL putting in enough resources
and on the renewed commitment of the parties to successful delivery. A lot will
rest on the capacity of POCL to manage the implementation phase. We believe this
will need new management with business expertise at the top. DSS/BA are very
reluctant to remain exposed to these risks;
whether satisfactory commercial terms can be agreed with ICL. Our legal
advice, summarised in Annex F, suggests that we are tightly constrained in what
could be offered to ICL (and for policy reasons we would want to offer as little as
possible). It is therefore very likely that ICL will make a loss on the project overall
if it continues. As explained below they are not in a strong position to absorb
losses. So there is a chance they will withdraw and litigate to recover what costs
they can. However we think this is unlikely: continuation offers ICL the prospect
of generating further revenues from the exploitation with POCL of the commercial
potential of the system; and from sales of the system overseas. Litigation may well
not succeed and would damage their reputation.
4.2.2 The feasibility of Option 2 raises different issues. Technically it should be much easier
to implement the Horizon system without the BPC, and this should free resources to focus
on developing a banking capability (though, paradoxically, removal of this embedded
25
4.2.3
4.2.4
HMT00000034
HMT00000034
RESTRICTED
feature could now add time and cost). POCL might also be able to cut the cost of the
project by not installing kit in all offices (current plans include automating the 2,000 or
So part time offices, some of which open for only a few hours a week). However there
must be significant risk around:
- whether the project remains a commercially viable proposition for ICL and
POCL without the BPC. Withdrawal of BA from the project would immediately
Temove the major source of revenue from the project, and POCL would be rolling
out kit which had only restricted use and limited revenue generating possibilities
until banking services were operational. Post offices would have fewer customers.
So even with the floor payment from BA it is not clear POCL could afford the
payments to ICL without further subsidy.
= how ICL would respond. This would depend on the outcome on ‘he first point.
They may believe they can recover more of their costs by suing for wrongful
termination. However option 2 could only proceed on the basis of a negotiated
settlement with ICL Pathway. If such a settlement could not be reached then
option 3 would follow.
- whether POCL can get agreement with banks in time to meet BA’s schedule
for moving to ACT and whether BA can complete the transition on schedule.
With Option 3 the key uncertainties are:
- how long it would take for POCL to put in place an alternative banking
platform, if it starts from scratch and conducts a new competition? There should
be little difficulty in installing basic swipe terminals by 2001; but the 2003
projection for a full integrated platform may be optimistic.
- how ICL will respond. Full cancellation would make it more likely that ICL
would litigate to seek to recover its costs. Protracted litigation could make it
difficult to implement a new integrated automation solution for the POCL, and this
could damage POCL’s long term business prospects. But using simple on line
swipe terminals as an interim solution would allow DSS to offer benefit recipients
migrating to ACT the facility of cash withdrawals from a range of bank accounts
across PO counters.
A major area of uncertainty which applies to all the options is whether, in the medium
term, POCL can generate sufficient revenue from banking and other services to
maintain the network in a stable state. The ‘medium term’ comes earlier under options
2 and 3. The impact on the Post Office is discussed in more detail below.
26
4.3
4.3.2
43.4
HMT00000034
HMT00000034
RESTRICTED
Implications for Benefits Agency/DSS
DSS have always envisaged that when the contracts with ICL came to an end they would
move to pay benefits direct into bank accounts. The original timetable would have seen
the BPC introduced at the end of this year and run as the established method of payment
for 7 years before new arrangements were fully implemented. However delays in the
project now cast doubt on whether it makes sense to progress through the BPC to ACT
rather than moving direct.
Option 1 gives the DSS a payment mechanism which is more fraud free and auditable
than current methods by the end of 2001, with cash available to claimants over PO
counters. By giving PO more time to get banking services in place it helps to ensure a
smooth transition to ACT. However:
- it also means postponing by up to 5 years the substantial efficiency savings
available from ACT. DSS are concerned that this option would therefore not
represent value for money for BA and hence be contrary to the obligations of their
accounting officer to seek maximum value for money from public funds. A
Ministerial direction may be needed;
- it involves benefit recipients in two major changes in the space of a few years. The
BPC would be phased in over 20 months from the beginning of 2000, involving
change for 19 million claimants, and major logistical exercise by BA. Five years
later DSS would begin the transition to ACT with the card completely phased out
by 2008. This will complicate the presentation and the administration;
- it leaves DSS exposed to the financial consequences of further delay, and
prolonged lock-in what they perceive as a hostile monopoly supplier.
Option 2 has a number of clear advantages for DSS and their customers:
- it allows DSS to take forward more quickly its proposals for modernising the
benefit system through ACT and payment over post office counters, bringing
forward the potential savings and thus satisfying value for money requirements;
- it involves claimants in only one change rather than two;
- provided the Post Office and ICL are able to deliver, it allows claimants continued
access to cash through post offices, while giving them the wider choice of banks,
ATMs etc. Overall, it looks less risky in technological terms
- it supports a move towards universal banking and combats social exclusion.
Moreover payment book (though not Girocheque) encashment fraud can be tackled
27
HMT00000034
HMT00000034
RESTRICTED
effectively by continuing the Order Book Control Service (OBCS) which is currently
planned as part of Horizon.
4.3.5 The implications for BA under option 3 are essentially similar to those under option 2.
4.4 Implications for Post Office Counters
4.4.1 Option 1 automates post office counters with the technology POCL say they want and
need in order to offer the best prospects of becoming a viable long term business. It
provides a secure income stream, and a flow of largely captive customers, for the life of
the contracts. POCL argue that transition to ACT via the benefit card would help them
ultimately to retain a much larger number of customers than a direct move to ACT from
paper systems. Spending by these customers provides income to both the Post Office and
private sides of the business, and is important to the viability of the network
4.4.2 POCL estimate that under this option there would be an average of 200 post office
closures a year, roughly in line with current trends, up to the end of the contracts; and that
POCL would remain broadly at its current marginally profitable position (£30 million pa
on £1.2 billion turnover) until then before taking account of any income from banking
technology.
4.4.3. Option 2 would give POCL the same automation platform (if a viable deal could be done
with ICL). This would help it to secure income from other clients such as National
Savings and the utilities, and to compete for business from benefit recipients migrating to
ACT. However, whereas option 1 allows POCL to build new services on the base of an
effectively captive market of around 19 million claimants, under this option POCL starts
to lose customers as they switch to ACT. This faces POCL with two major challenges:
- it requires POCL to develop with ICL the appropriate banking technology as part
of the Horizon platform, and establish partnerships with the banks;
- income from BA, its major client will effectively disappear beyond 200S.It is very
unlikely that POCL would be able to replace the loss of BA revenue by revenue
from banking and other new business within that timescale.
4.4.4 There would therefore be a need for a transitional period, possibly involving payment of
a direct Government subsidy to fund the network until it could be made self-financing
and/or slimming down the network, reducing the number of urban offices whilst keeping
the rural network intact as far as possible.
4.4.5 Any significant loss of cash outflows through post offices resulting from an early switch
to ACT under options 2 and 3 could undermine the viability of Girobank’s corporate
deposit business which is extensively used by small firms.
28
4.4.6
4.4.7
44.8
4.4.9
4.4.10
HMT00000034
HMT00000034
RESTRICTED
Cancellation of the project under option 3 would be a serious blow to POCL:
- it would put off full counters automation with the functionality desired by POCL
for some years (but gives it time to reassess what it really needs);
- commercial prospects would be damaged - many of its current clients are already
very keen to have Horizon implemented and may not be prepared to tolerate the
further delay and uncertainty associated with developing a new system and
potential new clients may be lost;
- as BA moved to full ACT, payments to POCL would taper down from around
£400 million in 2001 to close to zero by 2005, representing loss of nearly 40 per
cent of current turnover.
POCL have provided estimates of the impact of the loss of 100% of this income on the
post office network:
- 5,000 to 13,000 post offices could close (depending on whether POCL managed
Crown offices were closed, or more rural and urban offices were allowed to close);
- an annual subsidy of up to £300 million would be required to attempt to stabilise
the network at this level;
- there would be job losses of up to 43,000;
- compensation to sub-postmasters and redundancy payments to Crown office staff
could amount to £350 - 500 million.
However, as explained in Annex C, the mechanism by which a loss of business impacts
on the network is not straightforward. We have not been able to satisfy ourselves on the
basis of the information provided by POCL that these estimates are robust.
POCL would need to respond to this by spending perhaps 6 months developing a serious
business strategy. The impact of losing BA income could be mitigated by winning new
banking business, but POCL could not realistically be ready by 2004 to replace £400
million of revenue from BA with other income streams. So in order to keep the current
network of post offices, a transitional period would be needed which would involve the
payment of a large subsidy to POCL.
If POCL is to be a viable long term business, developing the banking area is essential but
not sufficient. In any case there is likely to be a need for managed reductions in the
network over a period of time, focussing on thinning out the urban network and largely
preserving the rural network, but perhaps with more use of mobile post offices, local
29
45
4.5.1
HMT00000034
HMT00000034
RESTRICTED
agents and kiosks. However, this process will be hard to manage; the network could
become unstable resulting in large compensation to sub-postmasters whose businesses had
become unviable. Continuing the project gives POCL the best chance of a smooth
adjustment path.
Financial implications
The financial implications of each option are analysed in detail in Annex G. The impact
on the public sector has to be seen in the round with losses to the Post Office balanced
against savings for DSS. The net impact can be summarised as follows (net present value
of the options discounted at 6% to 2009/10)
£ billion Option 1 Option 2 Option 3
DSS admin savings I 0.4 1.0 1.3
DSS fraud savings 0.9 0.8 0.8
Net impacton DSS_ I 1.2 1.9 2.1
Net impact on 0.2 -0.6 -1.2
POCL
overall NPV saving I 1.4 13 1.0
exit payment to ICL I 0 -0.2 to +0.3 -0.2 to +0.3
4.5.2
These results should be treated with caution. The estimates of the impact on DSS are much
more robust than the estimates of the impact on POCL. We have identified, but not
included in the totals, a possible ‘exit payment’ to ICL which may result either from
negotiation or litigation; but this is highly uncertain, and does not imply acceptance of any
liability. With these caveats, the results suggest:
- the most cost-effective options from DSS/BA point of view are either option 2 or
3;
- however when the impact on POCL is taken into account option 1 and 2 look more
attractive;
- but the margins of error in this exercise would suggest that caution should be
exercised before ranking options.
Under each option there are questions about DSS funding which would have to be
addressed bilaterally with the Treasury.
30
4.6
4.6.1
4.6.4
4.6.5
HMT00000034
HMT00000034
RESTRICTED
Implications for [CL
We have considered what the impact each option would have on ICL and its parent,
Fujitsu. ICL Pathway has been set up as a subsidiary of ICL which has provided BA and
POCL with parent company guarantees on finance and performance.
The ICL Group balance sheet is not strong. In 1997 it included net assets of around £260
million. We believe around £125 million of this represents capitalisation of the work on
this project, some fixed assets but mainly work in progress. This is likely to be greater
now, perhaps of the order of £160-170 million. Liquidity is worsening, down from £105
million of net current assets in 1996 to £42 million in 1997. There are no obvious areas
where provision for losses on these contracts have been made. There is a possibility that
ICL could strengthen its balance sheet by writing back some of £200 million goodwill
written off according to existing accounting regulations that have subsequently been
revised, but this is not mandatory.
Profitability is also weak. On a like for like basis ICL made an operating loss of £7 million
in 1996, and a profit of £32 million in 1997, on a turnover of £2,477 million in 1997. So
the group is operating at pretty close to a break even position.
Under option 3, cancellation of the whole project, Pathway say they could suffer a loss
of around £250 million. Assuming ICL stand behind Pathway, they will have to bear most
of this. The implications for ICL of cancellation, other things being equal, are:
- it would have to write off a good part of the £125 million or more of capitalised
assets from the project unless, say, the work in progress could be deployed on
another similar project;
- it would bear a loss at Pathway that could amount to £250 million;
- thus reducing net assets by, conservatively, around £350 million; and creating a
situation of net liabilities of up to £100 million before any write back of goodwill;
- it would make a total loss in 1998 of around £350 million;
- and would seemingly put at risk for many years any chance of a successful
flotation (planned for 2000).
In these circumstance it would be possible in principle for ICL to simply wind Pathway
up, leaving Pathway’s creditors, rather than ICL, to foot the bill. However this is unlikely.
ICL are more likely to seek guarantees, or a significant injection of new capital, from
Fujitsu.
31
HMT00000034
HMT00000034
RESTRICTED
4.6.6 In either case ICL’s reputation with Government (one of its main clients) will suffer a
severe setback. But if ICL sued HMG first and secured a reasonable outcome it might be
able to salvage some of its commercial reputation by ensuring blame is seen to lie with the
Government parties and the PFI concept, portrayed as unworkable for large IT projects.
4.6.7 Cancellation would have limited financial implications for Fujitsu, since ICL represents
only around 5 per cent of group shareholders’ funds and 2 per cent of net current assets.
However Fujitsu has underwritten a £200 million loan facility to ICL Pathway, and fujitsu
itself appears to have had a bad year in 1997-98, with group profit after interest and taxes
slumping from £254 million in 1996-97 to just £26 million in 1997-98. This seems to be
due to the economic problems in Japan and south east Asia. So a problem at ICL is one
of many, which may ease any write off necessary from taking action at ICL, or may be
the straw that pushes Fujitsu into loss. Whilst Fujitsu’s survival is not in any way
questioned, there could still be pressure to minimise losses where possible, although ICL,
it would appear, would be one of several target areas.
4.6.8 The implications of the loss of the project for employment within ICL depend on whether
Fujitsu decides to divest itself of ICL subsequent to its inevitable poor profit performance.
It is estimated that some 270 people at ICL Pathway are working on this project and many
more at their contractors. However any IT staff released are likely to be quickly re-
employed by competitors given skill shortages in the IT industry.
4.6.9 Cancellation of the project could affect perceptions of the UK Government as a partner
and hence future inward investment from Japan where Fujitsu are influential. (Japan
accounted for some 9.4 per cent of inward investment into the UK in 1996). There could
well be an impact on Fujitsu’s future strategy for the UK; they are currently planning an
£800 million expansion to their semi-conductor plant in the North East. It is however
difficult to forecast the precise impact.
4.6.10 Under option 2, cancelling the benefit payment card would damage ICL’s reputation but
installing the Horizon system in Post Offices would allow it to salvage something. But to
remunerate the investment it is making it will need to generate income of £100-200
million a year for at least 5S years. POCL could not afford that without subsidy.
4.6.11 Under options 2 and 3 there would be a loss of business for suppliers to ICL supporting
the BPC. These include Girobank who are contracted to provide helpdesk facilities which
will employ around 400 staff.
4.6.12 Under option 1 and option 2 it is clear that on current plans ICL Pathway will make a
large loss. From ICL’s point of view it will want to weigh up, among other factors:
- estimates of income from continuing with the project, including possible contract
extensions and a run-down period and installing banking technology; and
32
4.6.13
4.7
4.7.1
4.7.2
4.7.3
4.7.4
HMT00000034
HMT00000034
RESTRICTED
- the probability the project can be completed successfully with no further delays;
against:
- the likely outcome of legal action.
It is not clear where the balance lies from ICL’s point of view.
Presentation and handling
Annex E sets out recent Ministerial statements on these issues.
Areas of presentational difficulty with option 1 will be around what is said about the end
of the contract period and the future of the counters network. It should be possible to put
out a positive message on the banking services that POCL would be developing. We
would also have to explain how the introduction of the BPC fitted with the Government’s
long term plans for ACT.
With options 2 and 3 an immediate difficulty would be explaining to the sub-postmasters
and others why the Government had reversed its stance on the BPC and to the public in
general what this meant for the future of the network in particularly rural areas. Claimants
may well be concerned if plans for an early move to ACT could not be backed by an
assurance that they would continue to be able to get their cash from post offices. On the
other hand it would be possible to explain how ACT was a more efficient system than the
BPC and would allow DSS and POCL to move forward in partnership.
Potential difficulties would seem to be:
- Government might be accused of wasting money on the project (if legal action or
negotiation leads to large compensation payments to ICL);
- continuing with paper-based systems when a more secure card-based system is
almost ready;
- public promises given to sub-postmasters and others that DSS was fully committed
to the objectives of the project;
- criticism of Government from ICL, with potential knock-on effect on reputation
of PFI projects.
= options 2 and 3 depend on POCL’s having banking capacity for ACT by 2003-04;
the working group believes this can be achieved, but it is challenging for POCL;
33
HMT00000034
HMT00000034
RESTRICTED
it may not prove possible to negotiate an acceptable deal with ICL to take forward
only the Horizon part of the project; ICL might decide the project is not viable in
the proposed form and therefore seek redress through the courts.
34
HMT00000034
HMT00000034
RESTRICTED
5. Annexes
Annex A Benefit payment
A.l DSS pays over £77 billion a year to 27 million benefit customers in - around one billion
transactions a year. There are three methods of payment:
Method of payment % of % of cost per
claimants transactions I transaction
order books across post office counters I 67 80 47 pence
on presentation of giro cheque at post S Es 78 pence
office
via ACT to claimant bank account 28 13: 1.4 pence
= Cost to BA of administering the current system is about £525 million per year, of
which some £400 million is the payment to POCL;
- fraud arising from current means of payment is estimated to be around £110
million a year.
A.2 Features of existing methods of payment are:
- Order Books are a paper based method of payment which used since 1948. They
are well liked by customers and have proved a reasonable reliable delivery system,
giving customers a sense of security, but they are expensive to administer; and
require recall and re-issue if claimant circumstances change - resulting in delays
and sometimes hardship for customers. They are also subject to a high level of
fraud due to accounting and reconciliation problems - it is not possible to reconcile
encashment to an issue. About £85 million a year is lost through fraudulent
encashment - of the total number of order books reported lost or stolen 63% are
subsequently found to have one or more foils cashed.
- Girocheques are the second paper based method of payment and are the most
expensive to administer. They are also subject to high levels of fraud due to
reconciliation problems. About £22 million a year is lost through fraudulent
encashment. Of the total number reported lost or stolen 61% are subsequently
found to have been cashed.
- Automated Credit Transfer is a modern method of payment which is cheap to
administer. Payments are only issued shortly before they are due, so a change of
35
HMT00000034
HMT00000034
RESTRICTED
circumstances does not normally require any recovery or recall - resulting in less
inconvenience or hardship for benefit claimants. There is currently no evidence of
ACT payment fraud. Payments are accounted for accurately and reconciliation is
fully automated.
A.3__ ACT is an option open to all claimants with bank accounts. The percentage who choose
to be paid by ACT has risen steadily in recent years:
Year 1993 1994 1995 1996 1997
% paidby ACT 16.2 22.7 22.2 25.7 28.2
- 47% of new pensioners choose payment by ACT (May 1998).
= 53% of new child benefit customers choose payment by ACT (May 1998).
US and Australian experiences of ACT
A4 Payment of benefits directly into bank accounts has been introduced recently in Australia
and is to be introduced in 1999 in the USA. Some lessons to be learnt from these
experiences are:
- there have to be exemptions for some claimants. Neither Australia nor the USA
has found a way so far to pay benefits by ACT to everybody;
— payment by ACT is far cheaper than paper-based methods such as sending a
cheque through the post;
- the post offices in Australia and the USA have a much more limited role in benefit
payment compared to the UK - only delivering cheques through the post rather
than cash over counters;
- there seems to have been little difficulty making the changeover in Australia - this
is probably because most recipients already had a bank account into which they
paid their benefit cheques.
USA
AS
Benefits are currently paid by sending cheques through the post. These then have
to be paid into a bank account or cashed, with the cost incurred by the recipient.
- Congress passed a law in 1996 which required all payments by the federal
36
Australia
A6
HMT00000034
HMT00000034
RESTRICTED
government after 1 January 1999 to be made by electronic funds transfer (ie ACT).
This includes benefit payments;
recipients of payments are required to designate a financial institution or agent to
which payment should be made (some of which will charge the beneficiary a fee);
the Treasury can grant exemptions for classes of recipients in cases of hardship or
other circumstances;
details of how this will operate are being worked out by the Treasury.
A decision was made pre 1990 that all benefits would be paid by ACT and
customers would not have a choice. Before the change payments were made by
cheques sent in the post;
now over 96% of social security payments are made fortnightly by ACT to banks
and other institutions using a system similar to that used in the UK;
conversion to ACT was not seen as a major problem for customers because a very
high number held bank accounts and Australia Post was not a key player in the
cheque cashing market, largely because most customers already paid their social
security cheques into their bank accounts;
it proved necessary in certain circumstances to allow exemptions which are,
however, limited and difficult to obtain. Those customers not regularly paid by
ACT are still paid by cheque. Payments by cheque are strongly resisted because
of the administration costs, including expenses aligned to fraudulent negotiation;
to make emergency payments Australia successfully trialed a plastic card system
which utilises a‘“DSS Bank”. This holds temporary accounts accessible via the
ATM network - called electronic benefits transfer. It has now been recommended
for national rollout.
oF
HMT00000034
HMT00000034
RESTRICTED
Annex B Post Office Counters Network
Bl
B.2
B.3
B4
BS
Post Office Counters Limited (POCL) provides services through a network of around
19,000 outlets (10,000 in urban areas, 9,000 in rural areas), covering the whole of the UK.
The network provides local communities with a wide range of services - access to postal
services, benefit payments, banking and bill payment services, licensing and other
Government services and a range of other customer driven services (lottery, bureau de
change, insurance).
Local businesses make full use of these services with particular emphasis on corporate
banking, cash provisioning and access to mail services.
These services are universally available through the network, with 96.8% of the UK
population living within a mile of a post office. The services have a particular importance
in rural areas where there may be less alternative provision (for example as the banks
rationalise their branch networks). For example, an estimated 6 to 7,000 rural post offices
serve local communities of less than 3,000 people.
It is estimated that 28 million people per week visit their post office.
Types of post offices
B.6
B7
BS
600 post offices (Crown Post Offices) are operated directly by POCL. The remaining
18,400 are operated on a variety of agency arrangements whereby POCL contracts with
individuals (in a few cases with companies) to provide services on its behalf. Post Office
services are provided in such cases using the agents’ premises and staff - typically in an
operation which also contains a private retail business. This allows cost-effective operation
of the post office through shared overheads. Conversely in rural areas, many village shops
are only able to survive in the face of retail pressures because of the income operating a
post office provides.
POEL pays the agent for work done on its behalf. Payments are generally split into two
areas - a fixed payment which is not dependent on amount of work done, and a variable
payment directly related to work performed. The fixed payment is an important part of
the payment made to smaller offices, the variable payment is more significant for larger
offices.
POCL employs 12,000 staff directly, and the 18,400 agents themselves employ around
40,000 staff. Overall therefore about 70,000 people derive their income from Post Office
Counters activities.
Sources of revenue
B9
POCL mainly offers agency services on behalf of various organisations in the public and
38
HMT00000034
HMT00000034
RESTRICTED
private sectors, including central government departments and agencies, local government,
utility companies and banks, as well as providing the high street outlet for the Post
Office’s Royal Mail and Parcelforce businesses.
B.10 The following chart shows the Post Office’s sources of income in 1997/8:
I Source of Income 1997/98
I
Other
Royal Mail
2%
B.11 This shows that the Benefits Agency (including the Northern Ireland Social Security
Agency) accounts for over a third of POCL’s income. This business is also important
because of the volume of customer visits that it brings and the opportunity to offer to these
customers other services available from POCL. It also provides opportunities to sell them
goods from the co-located private retail businesses. In this context, the benefits business
is particularly significant in that it puts cash into the hands of the customers visiting the
post office.
POCL Income £ million %
Benefits Agency 420 36
Royal Mail 251 21
Bill Payment Services 180 15
DVLA 56 5
National Savings 49 4
Other 209 19
Total 1165 100
39
RESTRICTED
Annex C_ The Cost Structure of Post Office Counters
HMT00000034
HMT00000034
C.1 This annex discusses the cost structure of PO Counters and explains why losses of income
can have a significant impact on profits and the size of the network. The remuneration of
post office agents is an important factor in determining the impact of income losses on the
network.
Network costs
C2 A significant proportion of PO Counters’ costs are the costs of the network: the Crown
Offices and the agencies account for 64% of network costs.
PO Counters Costs 1997/8
Central Serices
finance and other 3%
12%
sales and
marketing
crown offices
19%
4%
regional support
3%
distribution and
warehousing
11%
agency offices
. 48%
Agency income and costs
C.3. Sub-postmasters and other agents are not employees of POCL and are consequently
responsible for providing their own premises and their own staff to run a post office.
POCL will pay them for Post Office work done, but most sub-postmasters will gain an
additional income stream by running a private business (newsagent, general store etc) from
the same premises as his/her post office. This private business can utilise the customer
footfall attracted by the Post Office to create further income for the agent, as well as
40
C4
CS
C6
C7
HMT00000034
HMT00000034
RESTRICTED
enabling the spreading of the costs of running the post office out of the same premises.
It is generally this combination of Post Office and private business which creates the
income levels which, when set against the shared cost base of operating out of common
premises, allow a commercially viable small business to be conducted.
POCL pays its agents for providing services on its behalf. Agents are remunerated through
a system based around a fixed element (paid irrespective of the volume of work done in
the office) and a variable element (based on the specific volumes of work done at the
office and accordingly increasing, or decreasing with volumes handled). The calculation
of the fixed payment for any particular office is determined for most offices through a
formula agreed with the National Federation of Sub Postmasters which takes into account
historic volume and payment levels at the office.
This is determined by the specific contract type the office works under and-the volume of
work handled. The very largest offices have pay which is entirely variable, whereas the
very smallest have pay which is almost entirely fixed. For the largest 20% of offices the
fixed element of pay represents between 20% and 35% of total pay, whereas for the
smallest 20% the fixed element represents between 85% and 95% of total pay.
Fixed Payment as % of Total Pay
100%
90%
80%
70%
60%
50% 1% Fixed
40%
30%
20%
10%
0%
% of fixed pay
x x Pa Pa Pa x x xz Pa
So S$ & &6& &6& 8 8 8 &
> S&S 8 F B® 8B F B GB
deciles from largest to smallest offices
100%
The rationale is straightforward. Smaller offices, often in rural areas, can attract only few
customers. If these offices were paid purely on volumes of transactions the income to the
sub-postmaster would be insufficient to cover the costs of providing and operating the
outlet. Alternatively, larger post offices conduct large volumes of work and therefore
variable payments do create the basis for a sustainable business.
The following chart shows the percentage of transactions which are benefit payments by
size of office. Although the relationship is not wholly straightforward, it shows that the
benefits business is “over-represented” in the smallest offices - i.e. those with the greatest
41
HMT00000034
HMT00000034
RESTRICTED
proportion of fixed payments in total payments.
Benefits as % of Total Transactions
50%
45%
40%
35%
30%
25%
20%
15%
10%
5%
0%
"8
Cost structure
C8 A large proportion of PO Counters costs are fixed: they do not vary with the volume of
business or with the number of outlets. This is for two main reasons:
- the payments to agents: for a large number of sub-post offices POCL’s payments
for business carried out is fixed (see above);
- PO Counters is a network industry. A significant proportion of its support costs -
for example the costs of distributing cash and stock to the national network - are
only partially variable with the number of outlets and the volume of business.
C9 The following chart shows in broad terms the breakdown of variable and semi-variable
costs by volume; costs which are variable and semi-variable by outlet; and costs which are
fixed with outlets and volume. These are broadly as follows:
° variable costs by volume - variable payments to agents; direct support costs
° semi-variable costs by volume - variable costs at Crown Offices
. variable by outlet - Crown Office costs; fixed payments to agents
. semi-variable by outlet - distribution; warehousing; regional support
. fixed - central costs of running POCL as going concern
Only about 30% of PO Counters costs are fully variable with transactions, though the
degree to which these costs remain fixed or only partially variable over time will vary.
HMT00000034
HMT00000034
RESTRICTED
Variable, Semi-Variable and Fixed Costs
Fixed Fs
19% Variable by
volume
32%
Semi-variable by
outlet
14% -
Semi-variable by
Variable by outlet volume
22% 13%
Impact of income losses on costs and the network
C.10 As income from BA falls then:
. ICL loses income from BA in its Crown offices and associated footfall. However
it is able to recover some costs from the reduction in volume-related payments to
its agents for BA and other transactions and other variable overheads. But these
variable costs are only around 30% of total costs - and the benefits business is a
greater proportion on business in smaller offices, which are more dependent on
fixed payments than variable payments (see chart above). Given the small margin
that POCL earns on turnover (£35 million profit on a turnover of £1.2 billion)
POCL moves into loss. A £50 million loss of income would reduce POCL’s profit
to zero.
Li sub-postmasters lose income via the reduction in variable payments from POCL for
transactions carried out - due to both loss of BA business and the impact of loss of
footfall on other PO transactions. This will particularly affect the larger urban,
suburban and rural offices which handle a high proportion of BA transactions, for
which a large proportion of their income from POCL is volume dependent, and for
which PO business is a large proportion of their overall business (including retail).
POCL believe there are around 4,000 urban offices and 500 rural offices which fall
into this category (POCL assume that these offices will close under any scenario
of loss of income from BA). Furthermore the income sub-postmasters gain from
private retail activity will fall as footfall reduces. For many offices the business will
43
Cl
C12
C13
C.14
GAS
HMT00000034
HMT00000034
RESTRICTED
no longer be viable. This will impact on the re-sale value of post offices.
Both of these impacts will lead to network closures arising both from management action
to reduce costs, and from sub-postmasters who no longer found the business viable:
La in order to reduce its losses POCL will seek to reduce its network costs. This will
be via closure of post offices which represent a high fixed cost - i.e. the Crown
offices and the smaller rural offices - where most of the remuneration is fixed and
not volume dependent;
Li the reduced viability of sub-offices (particularly the larger offices dependent on BA
transactions income) and the downturn in the resale market will lead to closures of
agents’ offices.
Network closures will mean that POCL will experience further reductions in income from
BA and other transactions that were formerly carried out at POs which have now closed.
The impact of this on profitability will depend on the management action taken by POCL
and the extent to which this business migrates to other offices.
POCL argue that even at this point the network may not be stable. Other clients may
withdraw business as the network reduces, leading to further losses of income and pressure
to reduce costs.
The extent to which a downward spiral occurs rather than a managed transition to a
reduced network depends on a number of factors:
LJ the size of the loss of income. However the marginal nature of POCL’s business,
its high fixed costs and the importance of footfall income means that even a small
loss of income from BA can reduce profitability significantly.
. migration of business from closed offices to other offices. The greater control that
POCL can have over the offices which close, the more they can ensure that
rationalisation occurs in locations where offices are closely located to one another
so there is a greater chance of capturing migrated business.
Ld] the speed of change relative to the natural rate of turnover of post offices. Post
offices change hands at a rate of 10% per annum, although this rate is dependent
on general economic trends and varies significantly between outlets. The PO can
use this turnover to manage change by seeking to reduce costs by alternative means
to closing offices (or allowing them to “self-close” as they become unviable)
through reducing the size and scope of offices, relocations, reducing the number of
counter positions etc.
The PO argue that the more time they have to manage change the better they can work with
these factors to reduce network costs and avoid disruption. Furthermore reductions in other
fixed costs become possible with more time - for example the radical rationalisation of PO
44
HMT00000034
HMT00000034
RESTRICTED
Counters cash warehousing (from 300 to 20 warehouses) took around eight years to
achieve. Other costs could be reduced in a shorter timescale - for example the re-
organisation of POCL’s field management structure took 18 months.
Agent compensation
C.16 Closure of agency offices is likely to lead to compensation to sub-postmasters. Under
existing arrangements sub-postmasters are compensated by an amount equivalent to 26
months remuneration if their post offices are compulsorily closed. This applies both if the
office is closed while the postmaster is in situ and when closure prevents a retiring
postmaster selling on the post office business to another party. There is no current scheme
for compensation for a loss in value of a post office as a result of reduced turnover
although POCL would come under severe pressure to compensate sub-postmasters for loss
of business arising from a government decision to move to compulsory payment of benefit
by ACT.
Costs of maintaining the rural network
C.17_ The Working Group asked the PO to provide an estimate of the cost of keeping the rural
network open in the face of significantly reduced benefit payment business. POCL argue
that it is impossible to segment out one part of the network in this way - even if a robust
definition for a “rural” post office was available. As shown above, a significant proportion
(around 15%) of the costs of running the network are common costs which are only
partially variable with the number of outlets - for example the costs of distribution. The
distribution system serves both rural and town areas and a loss of volumes would not
significantly reduce costs. At present these costs are spread across the whole network, but
if rural offices were kept open whilst urban offices closed these costs would fall on a
reduced number of offices. So although the costs of maintaining the rural network at
present (with the current network open) are relatively low, the costs of maintaining the
network once urban offices were closed would be much more significant. POCL estimate
the cost could range from £100 million to £300 million a year - although they believe in
practice it would not be possible to target the rural network in this way.
C.18 PO Counters have not provided an estimate of the subsidy that would be required to
prevent any closures of the network (although this was requested by the Working Group).
They argue that such a subsidy would not be desirable, since it would undermine the
commercial nature of the business.
C.19 In our view there is some scope for protecting the rural network by a phased thinning out
urban offices where there is some over-provision. But this would only contribute in a
limited way to meeting a revenue shortfall in the network as a whole.
45
HMT00000034
HMT00000034
RESTRICTED
Annex D_ The project
D.1
D2:
D.3
In May 1996, BA and POCL signed PFI contracts with ICL Pathway for Pathway to set
up and maintain a secure, automated infrastructure for POCL which would.
a allow benefit payments to claimant presenting a special magnetic stripe card, thus
- provide a more secure method of paying benefits; and
- provide an automated platform for other POCL business.
The basic idea of the project is, for benefit payment:
- claimants are given magnetic stripe card instead of order books and giros:
- networked computer terminals are installed at all 40,000 PO counter positions to
automate procedures to assist with delivery of products and administration;
- network is linked to BA computer system (CAPS - Customer Accounting and
Payment Strategy);
- claimants go to their nominated PO; counter staff swipe the cerd through terminal;
- the computer prompts staff to ask appropriate anti-fraud questions and tells them
how much benefit to pay, and;
- beneficiary collects all money due (ie they cannot leave some of the money
uncollected), and there is full automated reconciliation of transactions back to BA.
For other POCL services, the system provides:
- networked terminals with the facility for handling automated payments;
- point of sale information gathering and processing;
- up to date information for counter clerks on a wide range of transactions.
The programme comprises a number of interdependent systems and services, requiring
work to be completed by, and coordinated by Pathway, BA and POCL:
- the benefit encashment service relies on feeds from BA’s CAPS systems,
containing customers’ personal details, records of payments in and out and feeder
systems for each type of benefit.
- Pathway’s systems also rely on POCL providing Reference Data (details of post
46
HMT00000034
HMT00000034
RESTRICTED
offices, products, prices, etc) and need to be integrated with POCL’s existing
automated systems and its accounting and management information systems.
D.4__ Separate testing of systems, followed by integrated testing of ‘model offices’ and live trial.
lead to the full national rollout of the service to all post offices and BA district offices,
subject to acceptance tests at appropriate stages. Benefits migrate to the new system
according to a prepared schedule.
Charges
D.5 Pathway charges to BA and POCL are driven by volume, based on a matrix of unit
transaction prices which fall over time. Some payments are guaranteed in line with rollout
to post offices. BA have a separate contract with POCL for counter services, the payments
including fixed, semi-fixed and variable elements, with a ‘floor’ providing that payments
cannot fall by more than a specified amount in any one year, regardless of volumes.
D.6 The programme is a large and strategic investment for ICL and its parent, Fujitsu. It is also
a major business venture for the sponsors. It has been described as world-class in its scale;
but is now being overtaken by other developments particularly in developing banking
alongside postal services.
D.7 _ Pathway has taken on significant risk - including development, construction, operating,
commercial and funding risks. Pathway has also taken on a degree of risk associated with
encashment fraud (up to a ceiling of £200 million), although the ultimate tisk of secure
service for the £100 billion or so of benefit payment remains with BA/DSS.
Rollout
D.8 For acustomer to have their benefit paid by card requires
- their designated ‘home’ post office to have Horizon installed;
- the appropriate BA district office to be ready; and
- the BA systems which support the benefit(s) concerned to have been switched
across to the card. The timetables are interdependent and need to be planned
accordingly.
D.9 _ The ‘programme complete’ milestone is reached when rollout to post offices and BA local
offices and the switch to payment by card of the main benefits are both completed. It could
be decided not to switch one or more benefits to the card, but the contracts provide for
revenue guarantees to Pathway if transaction volumes on Horizon fall below certain levels.
Short term anti-fraud measures
47
D.10
D.lL
HMT00000034
HMT00000034
RESTRICTED
As an anti-fraud measure within the M25, order books are bar coded and the Automation
of London post offices (ALPS) has provided a bar code reader and an Electronic Stop
Notice System (ESNS) running on a PC on the post office counter. Order books can thus
be checked against a file of ‘stop’ notices downloaded from BA’s Electronic Stop Notice
Control System (ESNCS) and be withdrawn if necessary. ESNS also captures and returns
to ESNCS some information about order book issuance and foil encashment.
The Order Book Control Service (OBCS) extends this facility as part of Horizon.
Progress
D.12
D.13
The competition took place and the contracts were let on the basis of a detailed requirement
agreed by the sponsors which was ‘dropped down’ to detailed service descriptions in the
first few months of the contracts. The process revealed a complexity not previously
properly understood. Timetables slipped. A replan in February 1997 set out new milestones
formally negotiated under the terms of the contracts. Timetables slipped further and BA
and POCL have issued notification of breach of contract by ICLwith BA issuing a legal
‘cure’ notice in response to Pathway’s failure to meet a contractual milestone in November
1997. Pathway has responded by blaming the public sector parties for the delays.
BA and POCL, through their legal advisers Bird and Bird, commissioned an independent
of the programme by consultants Project Mentors. They concluded:
- ‘...Pathway seriously under-estimated the effort and time needed to develop the
services...
- “Pathway have been responsible for the delays to the programme, since the re-plan
in February 1997, that caused the November 1997 breach by not allocating
sufficient resources to complete their contracted obligations within the agreed
timescale...”
- . ‘Asecondary, and ...relatively insignificant, cause of the breach is delays resulting
from slow resolution of issues... We do not believe Pathway allowed sufficient time
for this activity in their estimates...’
48
HMT00000034
HMT00000034
RESTRICTED
Annex E_ Recent Ministerial statements
Manifesto commitments:
The manifesto said:
“ Public services and transport services in rural areas must not be allowed to deteriorate.
The Conservatives have tried to privatise the Post Office. We opposed that in favour of a
public Post Office providing a comprehensive service. Conservative plans would mean
higher charges for letters and put rural post offices under threat.”
Public statements on the Horizon project:
Kate Hoey MP in delivering a speech to the National Federation of Sub-postmasters on behalf of
Frank Field, Minister for Welfare Reform on 12 May 1998 said (extracts):
"We recognise that the Horizon project is as important to you as it is to the DSS... We now
have more than 25,000 Child Benefit customers using the Payment Card to collect their
money at the post office. And 12 million, or 50%, of all benefit customers currently using
post offices to get their money, will be ready to switch to the Payment Card by the time
Horizon begins its national roll-out. This is a remarkable achievement and reflects the
commitment of every one involved in the project to provide a secure, cost effective and
convenient method of payment for our customers.”
“The Payment Card will also help cut down on paper and bureaucracy. Our new payment
systems will consolidate customer details on a single database for all benefits, so in future
a customer will need only report change in personal details, for example, change of
address or post office, only once - the system will automatically update all benefit records.
The Payment Card will also give our customers a more convenient and safer method of
collecting their money, which is what they want. And it will also improve the way DSS
accounts for the massive expenditure on benefits passing through post offices, now
approaching £56 billion a year. The Payment Card is designed to do all these things."
“The project, we know is also crucial to the future of your business. You will understand
the importance of taking great care in planning the next stages of the project to ensure that
we get it right.”
“We are committed to providing a modern, secure payment system which is cost effective,
robust and fully accounted. Our partnership is crucial in making these objectives
achievable for benefit payments through post offices. With your 19,000 outlets you can
reach parts of the country other retailers can't, or don't want to reach. So we need you
as much as the communities you serve need you. But equally, only DSS can bring in 19
million customers who want to spend as well as collect their money.”
“As we move towards the millennium there will be ever greater opportunities to harness
49
HMT00000034
HMT00000034
RESTRICTED
new technology and to expand your business. But whatever the future may hold, you can
be assured that our longstanding association and commitment to providing an efficient and
secure benefit payment service for people who want to collect their money at their post
office is as strong as ever.”
Statements in Hansard:
There have been various written statements made by Ministers in Hansard over the last year
regarding the project. These have tended to focus on the expected completion of roll-out of the
system, and the trial already in place. Statements have also been made about fraud savings. In an
oral exchange in the House of Lords on 1 April, Lord Haskel said:
“My Lords, this is certainly a system which is designed to combat fraud. Unlike order
books and giro cheques, the payment card which will be part of the system has no intrinsic
value and so carries little personal information. The system will be more convenient. Any
changes to the benefit payment cards will be made automatically on the system.
This is a huge project, which is under constant discussion, and Ministers meet to consider
it from time to time. I am not in a position to say what the situation is regarding the
budget.”
50
HMT00000034
HMT00000034
RESTRICTED
Annex F Summary of legal advice
Under procurement law is it legal:
1
to extend the contracts?
The contract term is made up of an implementation period and a fixed 5 year “steady state”
period. The implementation period could be extended if objectively justified. Given the
many delays to this project a 12 month extension looks perfectly possible. Significantly
extending the steady state period would be problematic because a 5 year term was specified
in the OJ notice for the original competition. An extension to the steady state period could
possibly be justified according to the Programme Lawyer if it could be demonstrated that
the additional period of operation was necessary due to “unforeseen circumstances” and
could not be provided separately for “technical or economic reasons...without great
inconvenience to the contracting authority”, or was “strictly necessary” to the performance
of the original contract.
to institute a wind down period after the end of the contract?
The Programme Lawyer says not until nearer the time when it might become necessary
because of unforeseen circumstances (as above). Treasury Solicitor believes the risk of
successful challenge to a 12-18 month wind down period is small and hence this possibility
should not be ruled out.
to increase payments to ICL?
There is no problem about making additional payments to ICL for additional services
provided the additional services have been procured properly. The Programme Lawyer
considers such additional services can be acquired using the formula set out in paragraph
1 above. Treasury Solicitor has some reservations about the scope of the additional services
which can be added - more of the same in the wind down period provided they are of
limited duration and value is probably acceptable. New services could prove more difficult
(see,paragraph 5 below). More could be paid for existing services if it can be shown that
the cost of the renegotiated package is no less economically advantageous than the next
best offer in the original competition (IBM’s). There is some scope for this although it is
limited since in cash terms Pathway’s offer was five per cent more than IBM’s.
to restructure the contracts to remove the benefit payment card, as proposed in
option 2?
Treasury Solicitor believes the answer is yes provided BA withdraws entirely from its
contract with Pathway (as BA proposes) because on present information they consider
POCL not to be covered by the procurement directives. If BA cancels only parts of its
contract with Pathway, Treasury Solicitor believes there is some risk of challenge to BA
and POCL. In this case the Programme Lawyer believes there is minimal risk on grounds
51
HMT00000034
HMT00000034
RESTRICTED
that Pathway are in breach of contract. If POCL were covered by the procurement
directives the likelihood of successful legal challenge would depend on whether POCL
could proceed with the Horizon platform without prejudicing other potential bidders who
could have provided the platform without the benefit payment card.
5 for BA then to renegotiate with ICL for provision of the Order Book Control System
(OBCS)?
EC public procurement law would require BA to hold a competition for the provision of
these computer-related services, unless the value of the consideration for the services was
below the current threshold (£104,435). We believe the value of a project to deliver the
OBCS system to be substantially above the threshold.
6 for POCL to procure banking services from ICL without retendering?
POCL envisaged from the outset that banking facilities would be part of the functionality
of Horizon. POCL believe therefore that there should be no question of retendering for this
work.
On the basis of current information Treasury Solicitor believes POCL is not covered by the
procurement directives and therefore POCL can, in any case, procure banking services
from ICL without retendering.
Under contract law:
7 Would BA/POCL be liable for compensation on termination?
If BA or POCL terminated “for convenience” they would have to compensate ICL
according to a schedule of payment set out in the contract. If termination took place this
year, BA would pay £129 million and POCL £205 million.
If BA terminated on the basis of ICL’s breach of contract, ICL could be liable to BA for
up to £200 million. But ICL would vigorously contest that they were in breach. It is highly
likely that they would sue for wrongful termination and substantial damages. If BA lost
they could be liable for up to the sum available for termination “for convenience” ie £129
million if they terminated this year.
52
HMT00000034
HMT00000034
RESTRICTED
Annex G_ Analysis of financial flows
G.1 This annex analyses the financial impact on the public sector of the three options discussed
in the main report. The figures are subject to varying degrees of uncertainty and should be
treated with caution: they can only be broadly indicative. Key assumptions for each option
are:
Key assumptions underlying options
Option 1 Option 2 Option 3
Horizon system for POCL Continued Continued Cancelled
Benefit payment card Continued Cancelled Cancelled
Roll out of system starts January 2000 January 2000 -
Roll out completed September 2001 September 2001 -
BA commitment to September 2007 September 1998, September 1998
contracts with ICL ends
BA commitment to floor September 2006 September 2006. September 2001. BA
payment to POCL ends negotiates terms for
residual use of paper
based methods after
this date.
BA moves to ACT Over a period of 3 years I Over 3 years from Over 3 years from
starting leading up to I October 2001 to 2004 —_I October 2001
end of contract.
POCL introduces banking Building on Horizon Building on Horizon POCL commissions
services platform - technology in I platform - start by new system - limited
place by October 2001; I October 2001; fully in services from October
services up and running I place by October 2003 I 2001; fully in place
by October 2003 by October 2003
Fraud control system for Order book control Order book control Some equivalent
paper based transactions system built into system built into system based on that
Horizon system Horizon system now used in London.
DSS savings
G2. Net administrative savings are based on BA modelling and are judged to be reasonably
robust. Changes in administrative costs are driven by the number of transactions paid by
ACT. However a large proportion of the current administrative costs of paying benefits is
the fixed payment to POCL for handling order books. As ACT increases, variable
payments to POCL reduce but the fixed payment remains. When ACT reaches 100% of
benefit transactions total savings against the baseline are £440 million a year.
53
HMT00000034
HMT00000034
RESTRICTED
G.3 Programme savings are expected reductions in fraud associated with paper-based payment
systems. They therefore accrue both with the benefit payment card and with ACT. Fraud
savings are estimated to reach £135 million a year when order books and girocheques are
fully phased out.
G.4__ The payment for the “unbanked” assumes that about half of those benefit recipients who
do not currently have a bank account would be very unattractive to the banks (due to a
variety of reasons such as low incomes, likelihood of immediate withdrawal of benefit or
a (perceived) high risk of default). BA start from the position that they would not pay the
banks to set up simple credit only accounts for these individuals, but we have assumed that
an upfront payment of £10 per account with an on-going charge of £20 a year would be
required as an incentive to the banks. The analysis assumes action is taken to provide bank
accounts for this group early on in the process of transfer to ACT.
Impact on POCL
G.5 In contrast to the BA figures, the modelling of the impact on POCL of each option is
judged to be less robust.
G.6 The first round impact on POCL net profits is based on POCL modelling of the impact of
reduced BA income - both directly and indirectly through loss of footfall. POCL’s figures
suggest that the total loss of BA income of around £400 million would turn net profits,
currently £30 million a year, into losses of around £270 million a year. A large proportion
of POCL’s costs are fixed, and so a reduction in income is not matched by a significant
reduction in costs. In order to reduce. costs POCL must close offices, which in turn reduces
income. The more time that POCL has to manage changes to the network, the less likely
this process becomes a spiral of decline. This is discussed further in Annex C.
G.7 _ It is assumed that POCL losses increase towards this figure as ACT increases - although
recognising that under options 1 and 2 POCL is still protected by the floor payment from
BA until the contracts end (it is not, however, protected against footfall losses). Once 100%
ACT is achieved it is assumed that POCL’s net profitability remains negative over time but
gradually improves as the network is restructured and costs are driven out.
G8 As income earned from benefit transactions and associated footfall reduces then the
commercial attractiveness of sub-post offices will decline. Legally POCL is required to
compensate sub-postmasters by an equivalent to 26 months remuneration if their post
offices are compulsorily closed. It is likely that POCL would be placed under significant
pressure to compensate sub-postmasters if the Government announced a switch to payment
of benefit to ACT. The options therefore include a sum for compensation (based on POCL
modelling) which broadly reflects the number of post offices that would close under each
scenario.
G.9 The potential income from providing simple banking and financial services is highly
speculative. It is assumed that POCL generates income from three sources:
54
G.10
G.12
HMT00000034
HMT00000034
RESTRICTED
- the cash handling business (about £25 million a year);
- by capturing a small share (between 5 and 10%) of the £2 billion cash withdrawals
market;
- by providing access to bank accounts and cash withdrawal facilities for benefit
claimants who wish to continue to obtain their benefits across PO counters. It is
assumed that POCL is able to negotiate a fee of £20 per account handled. The
ability of POCL to retain the benefit business as ACT increases varies under each
option. POCL believe that they would be able to retain around 65% of those
transferring to ACT when their banking services are fully operational - but that in
the earlier years the retention will be less. The analysis assumes that in option 1
retention rates are higher than in option 2 and 3 (since POCL argue they will be
able to retain more business with the benefit payment card).
These estimates are, in the Working Group’s judgement, optimistic about the possibilities
for income generation from banking services - given the highly competitive environment
in which POCL will be operating.
Set up costs under option 1 reflect POCL assumptions (confirmed by ICL Pathway). Set
up costs in option 2 assume that POCL will need to pay around £75 million a year to
Pathway to remunerate the investment (since payments for the benefit card have been
removed, but costs to ICL are not expected to reduce significantly). Set up costs under
option 3 are based on the cost of a replacement technology for POCL (expected to be
somewhat cheaper than Horizon) plus an interim banking facility to deliver banking
services by 2001/2.
The impact on the network of post offices was provided by POCL modelling, and are
against a baseline of urban and rural closures over time. The figures are highly uncertain.
POCL argue that about 4,000 urban and 500 rural offices could close whether they loose
40% or 100% of BA income. This is discussed further in Annex Cc.
Exit payment to/from ICL
G.13
BA and POCL’s starting point is that ICL are in breach of contract due to their failure to
meet a key project milestone. If the Horizon project had to be re-scoped (option 2) or
cancelled (option 3) they would argue that ICL was liable for costs of up to £335 million.
However either of these options would be dependent ona negotiation with ICL who would
argue forcefully that the delay was due to the contracting parties. It is assumed for the
purposes of the analysis that no payment is made to ICL, but the summary table below
shows a range of possible payments.
55
RESTRICTED
Results - NPV of the options discounted at 6% to 2009/10
HMT00000034
HMT00000034
£ billion Option 1 Option 2 Option 3
DSS admin savings I 0.4 1.0 L3
DSS fraud savings 0.9 0.8 0.8
Net impact on DSS _ I 1.2 19 Bil
Net impact on 0.2 -0.6 -1.2
POCL
overall NPV saving I 1.4 13 1.0
exit payment toICL I 0 -0.2 to +0.3 -0.2 to +0.3
G.14 These results should be interpreted carefully. The estimates of the impact on DSS are much
more robust than the estimates of the impact on POCL. With these caveats, the results
suggest:
- the most cost-effective options from DSS/BA point of view are either option 2 or
3;
- however when the impact on POCL is taken into account option 1 and 2 look more
attractive;
- but the margins of error in this exercise would suggest that caution should be
exercised before ranking options.
56
HORIZON: SUMMARY OF C)
(RAL OPTIONS
Baseline
DSS administrative costs
CAPS costs
total administrative costs ine. CAPS
ne savings (ESNS)
9 of claimants paid by ACT
6 of transactions made by ACT
POCL. profits Em
POCL network
DSS prograr
rural
urban
total
Continue with project on present forecasts then start to move
1998-99 1999-00 2000-01 2001-02 2002
530 525 $25
100 20 20
630 S45 545
70 10 30
3% 32%
13% 15%
30 to
9000 8400
10000 9290
19000 176901
DSS net administrative savings (£m)
contingency
DSS programme savings (£m)
payment for “unbanked"
‘Total DSS savings (£m)
6 of claimants paid by ACT
of transactions made by ACT
pact on POCL profits Em
compensation to subpostmasters
cost of banking technology for POCL
profits for POCL from banking ém
Total impact on POCL,
impact on PO network
rural
urban
total
-30 385 90 5
0) 0 0
§ 110
0 0
-80 20
30° 31°)
14% 15%
0) 10
0 °
0 10
0 D)
0 0
0 0) 0) 0
0) 240 480 710
0) 240 480 710
-03 2003-04 2004-05 2005-06 2006-07 2007-08 2008-09 2009-10
8200
9050
7250
1s
135
520
20
20
0)
135
0
155
32%
16%
40
°
0
9
49
1190
1190
20
0)
135
3
147.5
55%
16%
50
°
0
7
o7
0)
1430
1430
7600
8330
15930
to compulsory ACT from 2004/5 before the end of the contract in Sept 2006
45 95
-10 10
135 135
20 30
150 190
100%
26% 53%
0 0
° 0
0 0
98 130
98 130
0 0)
1470 1510
1470 1510
510
20
$30
30
Me
18%
-60
7200
7860
15060
1550
1550
1590
1590
NET IMPACT ON PUBLIC SECTOR
50S
20
BS
-10
135
32
100%
NPV «id 6%
to 2010
494
-28
868
86
1248
-142
24
-16
182
1430
HMT00000034
HMT00000034
NPV id 6%
to 2005
-145
490
3
340
108
“16
109
449
2 Continuation of Horizon without BPC, move to
DSS net administrative savings (Em)
contingency
DSS programme savings (£m)
Payment to banks POCL for "unbanked"
Total DSS savings (Em)
9 of claimants paid by ACT
6 of transactions made by ACT
first round impact on POCL profits Em
‘compensation to subpostmasters
cost of banking technology for POCL
profits for POCL from banking Em
‘Total impact on POCL
impact on PO network
rural
urban
total
3 Horizon ¢: sMled: or ain
DSS net administrative savings (€m)
contingency
DSS programme savings (£m)
Payment to banks/POCL for "unbanked”
Total DSS savings (£m)
9 of claimants paid by ACT
94 of transactions made by ACT
first round impact on POCL protits &m
compensation to subpostmasters
cost of banking technology for POCL
profits for POCL from banking m
‘Total im pact on POCL
impact on PO network
rural
urban
total
Notes
1A negative saving means an addition to costs
2 Baxeline assumes "business as usual" with no Horizon
1998-99 1999-00 2000-01 2001-02 2002-03 2003-04 2004-05 2005-06 2006-07 2007-08 2008-09 2009-10 NPV
iT ulsory from 10/01), floor paid until end of contract
+10 “15 EB 1S 15 435 435
+10 “10 10 -10 -10 -10 -10
0 5 90 130 135 135 135
8 -20 330 332 332 “32
28 15 168 528 528
30% 88 100% 100% 100%»
14% 34% 99% 100% 100%
0 20 80 210-210
o 0 -50 -50 -30
0 85 15 “15 0 0
0 10 63 3 63 63
0) 0 “55 18-282, 197-197-187
0 0 0 0 0 0 0-300 -1500-2500 =
0 ) 0 0 o 0 0-800-100-1500 -1500
0 ) 0 0 ) 0 0 -1000 — -2500 4000-4500
)
0-400
0-760
b=
§ Each option is expressed as a change relative to the baseline
and 3 assume that no compe
4. NPV figures for optic
snsation is paid cither by or to ICL.
-2100
§ Note that the modelling assumes a contract end date of September 2006 in Option 1, rather
twas not possible to reflect this change to the spe
ation of Option 1 in the modelling i
NET IMPACT ON PUBLIC SECTOR
435
-303
4200-4500 A750. 4750
2490-2650 -2650 2820
$890 © 6850-7150 -7570
435 435 435
-10 -10 -10
135 35
32 2 32
528 528 528
100% 100% 100%
100% 100% — 100%
168-147-120
0 ° °
“67 67 “67
49 48 48
186-166-139.
4750-4750 4750-4750
-2340 2110-1870 -1530
-7090 6860-6620 -6280
NET IMPACT ON PUBLIC SECTOR
+ than September 2007 as in the main report (the timing
i the time available. However the impact on the financi
of the switch to ACT is unchanged).
ial flows is expected to be minimal.
1306
84
an
-561
1303
1576
84
347
-205
2135
681
-393
-306
-1162
972
NPV a
6%
358
356
469
“115
685
-195
-383
+119
2
“624
ED
HMT00000034
HMT00000034