SUBS0000017 - Phase 2 Closing Submissions made on behalf of BEIS

Evidence on official site

SUBS0000017

IN THE MATTER OF THE INQUIRIES ACT 2005
AND IN THE MATTER OF THE INQUIRY RULES 2006

THE POST OFFICE HORIZON IT INQUIRY

WRITTEN SUBMISSIONS ON PHASE 2 OF THE INQUIRY
ON BEHALF OF
THE DEPARTMENT FOR BUSINESS, ENERGY AND INDUSTRIAL STRATEGY

Introduction

1. These written closing submissions are provided on behalf of the Department
for Business, Energy and Industrial Strategy (“BEIS”) in relation to the evidence
heard during Phase 2 on the procurement, design, pilot, roll out and

modifications of Horizon.

2. As set out in its opening submissions, BEIS seeks to provide all possible
assistance to the Inquiry and approaches the evidence with a commitment to
openness, self-reflection and co-operation. Consistently with that approach,
these submissions are intended to summarise some of the key aspects of the
evidence heard in Phase 2, insofar as it has been relevant to the role of BEIS (or
its predecessor departments) and government more broadly. The aim is to
draw out 10 important themes which the Chair may wish to consider as part of
his analysis, and to put forward suggestions as to what the evidence has shown

in relation to these themes so far.

3. These themes are:
(1) Policy objectives behind the Horizon project
(2) Problems with the tripartite contractual arrangement and PFI
(3) The role of government within the Horizon project

(4) Reasons why the Horizon project was not terminated in 1998
SUBS0000017

(5) The difference between viability and robustness

(6) Ministerial insistence on field tests

(7) Responsibility for technical functionality

(8) Failure of NFSP to raise practical concerns in the Horizon Working Group
(9) Pressure to roll out Horizon, and the subsequent pause

(10) Assurances given and the continuation of the roll out

1) Policy objectives behind the Horizon project

4. The witnesses were broadly agreed about the objectives of the Horizon project.
As David Sibbick explained!, the aim was to reduce fraud in the benefits system
by moving to an electronic payment method whilst also preserving footfall in
post offices. This was confirmed by Stephen Byers? and Colin Bakers amongst

others.

5. By 1999, when the Horizon project was being reviewed and the decision had
been taken to cancel the Benefit Payment Card aspect, the objective became to
secure the future life of post offices more generally.* It was clear from many
witnesses that there were very serious concerns about whether the post office
network would be economically sustainable without the footfall from benefits
claimants. Bringing in an electronic system for payments, with the potential to
add new forms of financial product as these were developed, was seen as

crucially important to deal with this problem.

6. The aims for the Horizon project throughout its development were thus all
readily understandable and laudable. The difficulty was in reconciling those

aims with each other in the initial stages and, after the decision was made to

123/11/22, 6/15-8/8

2 24/11/22, 4/15-20

3 30/11/22, 4/10-5/15
41/12/22, 71/12-72/21
SUBS0000017

progress in 1999 and the aims had changed, to achieve them within a

reasonable timescale and with a system that worked.

2) Problems with the tripartite contractual arrangement and PFI

7. There was a general consensus amongst all the relevant witnesses that the use
of the Private Finance Initiative (“PFI”) structure did not work effectively for
the Horizon project. For example, Sir Ian McCartney was of the view that the
contract was not a good fit for PFI.5 Sir Adrian Montague noted that IT
procurement projects were amongst the most difficult PFI projects, because
they required great detail and specificity in the sponsor’s requirements, and in
the case of Horizon there were big gaps and much left undefined at the
contractual stage.° The clear suggestion is that this undermined the project from

the beginning.

8. There was also repeated evidence about the difficulties caused by the tripartite
contractual arrangement between the Benefits Agency, Post Office Counters
Limited (“POCL”) and ICL. The Inquiry heard from the ICL witnesses about
this, and from the government perspective the problem was highlighted by
David Sibbick? and Sir Ian McCartney,’ from the point of view of the
Department of Trade and Industry (“DTI”), and Sir Steve Robson, from the

point of view of the Treasury.?

9. There can be little doubt that these two factors contributed significantly to the
problems in the Horizon project as it was being developed. A key issue for the

Inquiry to consider is the extent to which they remained a cause of the problems

5 30/11/22, 120

6 25/11/22, 57-58

7 23/11/22, 4/24-7/9

8 30/11/22, 127/6-128/11
9 29/11/22, 7/5-24
with the system as it was eventually implemented, and in particular the effects

on postmasters.

3) The role of government within the Horizon project

10. The consistent picture that emerged in the evidence was that Ministers were
concerned to reach some form of solution for the project but BEIS Ministers and
civil servants were not ever involved in the technical detail of the system or in
the detail of contractual negotiations. The Post Office, and POCL as its
subsidiary, was an arms-length body and the common view was that it was not
for government to engage in day-to-day operational matters. This was put ina

variety of ways by both civil servants and former ministers. For example:

a. David Sibbick noted that “Horizon had become a political issue, with
ministers involved, and ministers were just concerned to get a solution. We
were not concerned to stand over the thing and try to see whether each

individual little bit can get properly resolved.”

b. Sir Adrian Montague commented that from his perspective, when he
was brought in to review the project, he saw no evidence that POCL was

being constrained in its handling of Horizon."

c. Alan Johnson explained the role of government in detail, including by
reference to his sponsorship of the Postal Services Bill, which (whilst
retaining the Post Office within public ownership) made the Post Office
even more arms-length and gave it more commercial freedom than
before.!? He said that it was “seen as being very important that we didn’t

just talk the language of commercial freedom for the Post Office in the public

10 23/11/22, 34/6-11
1125/11/22, 56/5-9
1241/12/22, 55/3-56/17

SUBS0000017
sector; we allowed them to get on with it’, and he did not see it as part of his
role to hold the Post Office to account in the sense of micromanaging
them; except in quite narrowly defined areas such as tariff increases it

was down to them to make commercial decisions about the future.’

11. The essential focus of Ministers is necessarily on high-level strategic and policy
issues. It would quickly become impossible for government to function if that
were not the case. This point is even more important in relation to an arms-
length body like the Post Office. Alan Milburn explained this in his evidence in

the following terms:

“You know, it's difficult, I think, for people to understand, who haven't been in
government, just how much stuff there is. You know, you're getting a lot of
stuff coming at you all the time and there's a lot of correspondence, a lot of,
nowadays, emails, and so on and so forth. So there does have to be some filtering
mechanism, you know, because, otherwise, it just -- you're faced with an
avalanche that it's just impossible to deal with... I don't want to, in any way,
give the impression that ministers are victims in all of this because, in the end,
you're in government, you know, you're responsible for what happens in your
Department, that's the rule of the game, so to speak, even though, very often, to
be perfectly honest, there are things that are happening in your Department that
you have absolutely zero line of sight of, because it's almost impossible to have
any line of sight of it... mean it's a crazy way of life. I mean, you know, there's
a reference to Alistair and Steve and I on Christmas Eve having a conflab about
whatever it was in relation to Horizon, that wouldn't be unusual. There's
another reference somewhere to meeting at 12.30 in the morning, trying to
cobble together a decision. I mean, these are not unusual things. So it's a very
intense thing to do, it's the most purposeful thing I've ever done in my life and

I don't regret a moment of it. But it is -- it's pretty busy.” 14

13. 1/12/22, 59/20-60/12
14 2/12/22, 49/3-50/23

SUBS0000017
SUBS0000017

12. As Mr Milburn also noted, civil servants’ role in deciding precisely how much
detail on any particular matter should be put before Ministers necessarily
involves difficult subjective judgements, which are more art than science. There
is, in the real world, little alternative but to trust officials with the freedom to
make these judgements to the best of their ability, and danger in seeking

prescriptive solutions with the benefit of hindsight.5

13. BEIS essentially saw its role, in relation to Horizon, as helping to set the right
strategic direction, brokering solutions, reviewing overall project viability, and
putting in place a system to provide a reasonable degree of additional
assurance that those whose role it was to ensure that the system worked as
intended - the Post Office and ICL - were doing their jobs properly. As Stephen
Byers put it:

“T think we wouldn't have proceeded had we -- had it been a system that wasn't
going to function properly and deliver what we wanted for the Post Office. Now,
as ministers, we wouldn't be involved in the sort of operational detail of that.
What we could do as ministers was to set up structures to -- as I think I may
have said earlier, to alert us to any problems if they arose. So we had the Horizon
Working Group, which had representatives of the subpostmasters, the CWU
union, and that was meeting and one of its terms of reference was to report on

these matters and how it was being delivered in practice.” 1°

4) Reasons why the Horizon project was not terminated in 1998

14. The Inquiry has heard how the problems with the Horizon project came to a
head in 1998-1999, but after months of deadlock the decision was made to

proceed without the Benefits Payment Card rather than terminate it. The

15 2/12/22, 51/3-52/3
16 24/11/22, 87/23-88/10
SUBS0000017

situation was summarised by Stephen Byers as being that “there were no good
options available to us, so we had to make the best of a very difficult situation, which I

think is what we tried to do.”7

15. Cancellation was clearly an option considered by Government, as the
correspondence between Lord Darling, Peter Mandelson and No. 10
demonstrated. As Sir Geoff Mulgan explained, however, Ministers were
concerned that cancellation would lead to further delay and complication of an
already complex situation, and in the end it was not considered politically

possible.!8

16. Stephen Byers acknowledged that preservation of trade and investment links
with Japan, and the substantial investment that government had already made,
were factors in the decision, but his evidence was that they were not key. He
emphasised that the principal motivation was the realisation that cancellation
“.,,would have had a devastating impact on the Post Office network. You know, we had
postmasters and mistresses who had invested, as we were told, about £1 billion of their
own money in getting the property and goodwill of post offices. That would have been

devalued overnight had we cancelled the Horizon project.”!

5) The difference between viability and robustness

17. Anissue explored with several witnesses was what Ministers understood, from
the report from the panel chaired by Sir Adrian Montague, about the status of
the Horizon project. Sir Adrian Montague himself, when taken to the executive
summary of the report, said that he intended the conclusion that Horizon was

“technically viable” to mean it was capable in practice of delivering the

17 24/11/22, 85/10-13
18 1/12/22, 144/6-20
19 24/11/22, 86/4-11
SUBS0000017

contracted outputs, and that the panel were making a judgment on the ability
of the project over time to be developed in a way that would answer the
specification. He stated that he could not precisely distinguish ‘viable’ and
‘feasible’ in this sense, alluding to the semantic nature of the distinction.” He
did not intend the report to be understood as commenting on its ‘general
robustness’ or to have pronounced the project as being ‘robust’ at that stage.”!
In particular, he noted that none of the technical issues that came to dog the
project were apparent to his panel or brought to their attention by either the
contracting parties or by PA consultants, who had carried out an independent

technical evaluation.

18. On this issue, Stephen Byers accepted the suggestion put to him by Leading
Counsel to the Inquiry that in his mind “ the technical integrity and robustness of
the system” were “one and the same thing” as “what Montague has said about its
technical viability” > He went on to agree that he thought that the Montague

report was understood as meaning that “it was a robust system” .?4

19. He also agreed that the Montague report “did not mean or [I] didn’t take it to mean
that the system was, as it stood, necessarily robust and had integrity”. In his own
words, he put it in terms that what the Montague report said about technical
viability “wasn’t a sort of blank cheque going forward” > He also understood that
the report’s authors had not been able to test how Horizon worked in practice

and that there was a need for live trials of the system.?¢

2025/11/22 69/19-70/5
2125/11/22, 80/5-18
2225/11/22, 77/7-15
2324/11/22, 28/4-11
2424/11/22, 30/2-12

25 24/11/22, 24/2-7

26 24/11/22, 28/22-29/6
SUBS0000017

20. The Inquiry will no doubt carefully consider this evidence, recognising that the
meaning of “robust” or “a robust systent” in this context is ambiguous, and that
witnesses were attempting to recall the sense in which they understood those
words in this context over two decades ago. The Inquiry may conclude that it
is highly improbable that Stephen Byers understood the Montague report to be
signing off Horizon, which at the time remained in development, as being
sufficiently “robust” to roll out across the Post Office network at that time; and
that it is also most unlikely that he understood the report to purport to certify
the Horizon productas being “robust”, ona prospective basis, at whatever stage

it should happen to be rolled out in future.

21. The Inquiry may instead conclude that what the Montague report was saying
- and was generally understood to be saying - was that the overall design of
the system was sufficiently “robust” for the purposes for which the report had
been commissioned, namely whether upon satisfactory completion of
development the Horizon system was likely to work as intended. This is not
least because, in the words of Leading Counsel to the Inquiry (and as Stephen
Byers agreed) “It is logical that the report could not offer a view on the existing

integrity or robustness of the system because it had not yet been built” 27

22. Importantly, this accords with the evidence of Lord Darling, Sir lan McCartney,
Sir Steve Robson and David Sibbick. Sir Ian McCartney?§ was clear that he
understood the Montague report to mean “robust” in the sense that there was
a plan to bring the project to fruition, but not in terms of it yet being ready to

sign it off in practice.2° David Sibbick had the same understanding.*° Sir Steve

28 Who as Minister for Posts was closer to the detail of Horizon than was Stephen Byers.
29 30/11/22, 142/1-18
30 23/11/22, 51/17-53/1
SUBS0000017

Robson considered that at the time the focus of Ministers was on the overall

feasibility of the project, rather than the technical feasibility.>!

23. Lord Darling explained that a semantic analysis of the meaning of ‘viable’,
‘feasible’, or ‘robust’ did not really matter because his focus (and that of others)

was simply on “whether [Horizon] would work or not”:

“Q. What did you mean by your use of the phrase " technically viable" ?

A, Well, in blunt terms, whether it would work or not.

Q. Is that right or do you mean that it's feasible in principle?

A. Well, it's both, isn't it? If it's not feasible in principle, it's difficult to see
how it could ever work. But, you know, my interpretation of the "technically
viable" used here, and in Adrian Montague's report -- and you're right that he
does use that term -- as to whether or not it was going to, you know, be delivered
in a way that would be satisfactory to the end user, which is -- it was the DSS

and indeed, you know, other parts of it, to Post Office Counters.” 52

24. For these reasons, it seems that the level of assurance that Ministers and civil
servants who were engaged with the Horizon project in 1998-1999 took from
the Montague report was what was intended: that the planned system could be
made to work properly, not that it did work properly. Whether they should

have sought further technical advice and assurance later is another question.

6) Ministerial insistence on field tests

25. A related theme concerns the insistence by Ministers and civil servants that
there should be effective live trials of Horizon before it was rolled out into the

post office network.

3129/11/22, 6/8-7/13
32 29/11/22, 72/4-22

10
SUBS0000017

26. As Sir Steve Robson explained, by the spring of 1999 POCL were concerned
that more testing would delay the Horizon project, but the Benefits Agency
wanted this and the Treasury agreed, with the solution being live trial in 300
post offices before any contractual acceptance and further roll-out.* The
express reason for this being escalated to Ministers was that this was a major
project that had been going wrong and had potentially widespread
repercussions not just for the public sector but for all the subpostmasters

involved in running the network.+

27. Lord Darling recalled that he thought these live trials were very important and
that even 300 post offices may not be sufficient.3> Sir Ian McCartney also
explained how he pushed ICL to conduct live trials and used the Horizon

Working Group to do this.%°
28. It is clear from this that Ministers recognised the importance of ensuring the
Horizon system worked on the ground and took steps to reassure themselves

that this had been properly tested.

7) Responsibility for technical functionality

29. A substantial issue for the Inquiry concerns the extent to which government
took on - or should have taken on - direct responsibility for ensuring that the
Horizon system had been, or was, developed in a way that ensured it
functioned properly on roll out. This plays into a larger question about the role
of central government in relation to the operational affairs of its arms-length

bodies, to be explored in greater detail in Phase 6 of the Inquiry.

33 29/11/22, 15/15-18/19
3429/11/22, 19/22-20/4
35 29/11/22, 126/7-24

36 30/11/22, 159/4-160/3

11
SUBS0000017

30. It is sufficient to say, at this stage, that BEIS Ministers and officials did not
consider that it was BEIS’s role to become directly involved in technical
development aspects of Horizon; having been informed that Horizon was
technically viable, it was left to the contracting parties - POCL and ICL - to

ensure that what was developed and ultimately rolled out actually worked.

31. In any event, the evidence in Phase 2 consistently identified a lack of in-house
technical expertise within government at the time. As David Sibbick noted,
Ministers and civil servants were dependent on what they were told by the
technical subcommittee and the IT experts from both POCL and the Benefits
Agency.” This was echoed by Sir lan McCartney,38 Alan Johnson,3? and Alan
Milburn.

32. This raises two issues for the Inquiry to consider. The first is the issue of
principle - to what extent should government involve itself directly in the

operational affairs of ALBs? - to be addressed in more detail in Module 6.

33. The second is whether, notwithstanding the clear understanding that technical
aspects of the Horizon project were for POCL (and ICL) to manage,
Government should itself have commissioned further expert reports to confirm
whether the system was indeed fit for purpose at roll out. David Sibbick noted
that there was pressure to “simply get this thing on the road and going” after so
many delays, but it may have been better to have done so.*! Sir lan McCartney

similarly accepted that it would have been better, in terms of lessons to be

37 23/11/22, 21/6-13

99 1/12/22, 77/5-6, 98/11-

40.2/12/22, 12/11-19
4123/11/22, 105/17-106/14

12
SUBS0000017

learned, to have had such a review or report.!? The question, of course, is where
to draw the line. In the knowledge that Horizon was in fact unfit for purpose,
it is now clear that government would have benefited from a further report or
reports of a highly detailed technical nature; but the more difficult issue is
whether the information available to government at the time should have
driven it to commission such reports, with the attendant further delays,

uncertainty for subpostmasters and additional public expense.

8) Failure of NFSP to raise practical concerns in Horizon Working Group

34. The Horizon Working Group (“HWG”) was specifically set up as the forum for
subpostmasters and other stakeholders, through their representative bodies, to
raise any practical problems with Horizon directly with government. It was
intended to act as an additional layer of assurance that Horizon would work as

planned, and should have worked in that way.

35. The unanimous evidence of the civil servants and Ministers was that a key
purpose of the HWG, set up initially by Sir Ian McCartney, was to hear the
experiences of actual users of Horizon, the subpostmasters, and to find out
what was happening ‘on the frontline’.“’ Colin Baker, NFSP General Secretary
at the time, also accepted that the HWG was set up to get feedback on whether

there were any problems with Horizon in practice for subpostmasters.+

36. Sir Ian McCartney’s evidence was that his clear belief and expectation at the
time was that the HWG would provide government with a direct line of sight

into any problems within Horizon:

#2 30/11/22, 162/5-17
43 23/11/22, 83/9-85/15; 30/11/22, 160-169; 1/12/22, 76/20-79/15

#4 30/11/22, 21-22

13
“We argued for the setting up of this group, and I argued with it with other
ministers who agreed, in the end, to it, and the reason for that is that I wanted,
not just the voice of the Post Office or ICL, whatever, I wanted the frontline's
voice to be heard. And I put that in writing. I said it verbally, I put it in the
letter of invitation to the first meeting. And so I had no doubt in my mind that
Iwas setting this up so, if anybody did try to hoodwink me, I would have a view
from the frontline about what, in reality, was happening. So that's what -- so

let's be clear about that. I'm no patsy nor would ever be a patsy.”

37. Alan Johnson gave evidence to similar effect.4

38. However, a striking feature of the evidence in Phase 2 was that, despite the
HWG'’s clear purpose, the NFSP never communicated their members’ concerns
about bugs and technical issues in the HWG. Colin Baker himself accepted this,
and all the former Ministers involved with the HWG concurred.“ In particular,
none of the former Ministers recalled being told about Horizon system integrity
issues, and none recalled being told about prosecutions of subpostmasters
involving Horizon data. As Alan Milburn put it, the latter issue was not on his
radar in any sense at all.4” Alan Johnson stated in response to being told about

a subpostmaster being investigated for a Horizon shortfall in January 2000:

“That's crazy. It’s should have been something we knew about, if it's connected,
as it seems very much to be connected, with the software. That's precisely what
the Horizon Working Group was there to hear, those kind of issues. That, in
particular, I mean, would have been -- set alarm bells going with anyone

remotely concerned with the trade union world and employment relations, and

45 1/12/22, 94/9-96/2
4% 30/11/22, 21-22; 30/11/22, 185/14 - 185/7; 1/12/22, 5/17-7/5

47 2/12/22, 23/6-17

14

SUBS0000017
SUBS0000017

who, as I did, understood some of the kind of overreactions you sometimes got

from what we used to call the Post Office Investigation Division” 8

39. The HWG had an essentially important objective, and it singularly failed to
fulfil it. The Inquiry will no doubt wish to consider why.

9) Pressure to roll out Horizon, and the subsequent pause

40. There has been some suggestion that the emphasis that Ministers, particularly
Sir Ian McCartney, placed on avoiding further delays put pressure on POCL to
agree to roll out Horizon without taking reasonable steps to ensure it was fit
for purpose. When asked about ministerial time pressures (in the context of his
own report), Sir Adrian Montague explained that “with great respect, ministers
often have ideal notions of how long tasks are going to take” and part of his job was

to say “We need the time we need” 4°

41. Sir Ian McCartney's own evidence on this issue emphasised that the pressure
to avoid further delays was “part of the politics of discussions and negotiations”
and the 2001 date was believed to be “doable”®° but “wouldn't be easy”®!. He
rejected the suggestion that he was giving the impression that “speed was now
more important than technical concerns” 52 He was categoric that “It wasn't - why
would I be involved in a situation where I wanted to see it by a certain date, and as soon

as you acknowledge it, it wasn't fit for purpose? That's just not feasible.”5>

48 1/12/22, 94/9-96/2
4925/11/22, 67/11-19

50 1/12/22, 50/2-17
5130/11/22, 176/11-17

52 30/11/22, 176/24-177/12
53 1/12/22, 50/22-25

15
42. Alan Johnson stated that “I didn’t differ from Ian McCartney in the importance of
the timescales and deadlines, but my view, and I think it would have been Ian's view as
well, was that we had to sort out the problems’. For that reason, Mr Johnson
intervened and the roll out was in fact paused, at his initiative, for an extended
period pending resolution of the issues that had by then been brought to his
attention. The Inquiry may consider the fact of this extended pause to be clear
evidence that ministers did not consider speed to be more important than

technical concerns.

10) Assurances given and the continuation of the roll out

43. Having intervened to delay roll out in this way, on 11 January 2000 Alan
Johnson attended the latter part of a Post Office board meeting.®® At that

meeting, Mr Johnson was informed*” that:

“The rollout of Horizon was due to recommence on 24 January. A great deal of work
had been undertaken to rectify difficulties identified in three areas:

- system stability;

- accounting integrity; and

- _ the provision of support to offices

Although as yet uncertain, it was not anticipated that these issues would not

prevent rollout recommencing.”

44. The Minister was therefore informed, by the Post Office, that the technical
issues that had been identified as being sufficiently important to pause rollout

had been (or would be) properly addressed so as to enable the roll out to

U/12/
55 NFSP00000458
56 POL00000336

57 1/12/22, 101/16, 102/20-23

16

SUBS0000017
continue. His evidence on this point was “there were no alarm bells ringing that
[there] might be a huge problem with this [...] issues were raised, they were raised with
the Post Office, they raised them with ICL, we got a thumbs-up that those were resolved,

and rollout carried on.”°8

45. The Inquiry will no doubt wish to consider why the Minister was provided

with this assurance.

Conclusion

46. It is hoped that these submissions are of some assistance to the Inquiry. BEIS
stands ready to assist the Inquiry in whatever ways it can, and looks forward
to its findings and recommendations concerning the issues addressed in Phase
2.

7 December 2022
NICHOLAS CHAPMAN
Temple Garden Chambers

ALASDAIR HENDERSON
1 Crown Office Row

58 1/12/22, 93/1-7

17

SUBS0000017