WITN10500100
WITN10500100
Witness Name: Anthony John de Garr Robinson
Statement No: WITN10500100
Dated: 14 May 2024
POST OFFICE HORIZON INQUIRY
FIRST WITNESS STATEMENT OF
ANTHONY JOHN DE GARR ROBINSON
I, Anthony John de Garr Robinson, will say as follows.
INTRODUCTION
1. I ama barrister practising from One Essex Court, a commercial set of chambers
in London. Its full address is One Essex Court, Middle Temple, London EC4Y
QAR. I acted for Post Office Limited (“POL”) in the group action by which
subpostmasters (“SPMs’”) brought claims against POL in relation to the Horizon
IT System (“Horizon”) and their contractual relationship with POL (the “GLO
Proceedings’) .
2. This witness statement is made to assist the Post Office Horizon IT Inquiry (the
“Inquiry”) with the matters set out in a Rule 9 Request dated 28 March 2024
(the “Request’).
MY PROFESSIONAL BACKGROUND
3. I was called to the Bar in 1987. In 1988, I joined the chambers of Benjamin Levy
at 9 Old Square in Lincoin’s Inn. In 1990, I moved to One Essex Court. I was
appointed Queen's Counsel in 2006.
Page 1 of 68
WITN10500100
WITN10500100
4. I undertake English High Court and appellate proceedings, offshore litigation,
international arbitration (as an advocate and as an arbitrator) and advisory work.
My practice covers a wide range of disciplines, including commercial, commercial
chancery, banking and finance, company law and civil fraud.
THE GLO PROCEEDINGS
5. In paragraphs 2 to 35 of the Request, I am asked a large number of questions
which are organised in quite a complicated way. In the following sections of this
witness statement, I answer those questions to the best of my ability and
knowledge. Before doing so, there are two points that I wish to make clear.
6. _ First, as explained in my email to a member of the Inquiry’s legal team on 28
March 2024, as a result of certain professional commitments, I was unable to
start work on this witness statement until Saturday 4 May 2024. The purpose of
my email was to ask for an extension of time in which to provide my draft
witness statement, until Thursday 23 May. In response, the Inquiry gave me
until Friday 10 May 2024. On that Friday, at my request the Inquiry gave me a
further extension until Sunday 12 May 2024.
7. The Inquiry will have had good reason for specifying these deadlines. But I
should make it clear that, as a result, I have only had nine days in which to read
the documents with which I have been provided and to prepare a statement
answering the many questions I have been asked. This has not allowed me
much time to review the documents in my possession.
8. Second, the questions I have been asked concern events that took place
several years ago (between 2016 and 2019). At this remove in time, my
recollections are very limited. Indeed, in relation to several of the points raised I
have no direct recollection at all. The documents with which I have been
WITN10500100
WITN10500100
provided have refreshed my memory to some extent, and so have some
documents I have found. But my recollection remains incomplete. In the rest of
this witness statement, I have tried to be clear about the points in relation to
which I am giving my direct recollection, as opposed to explaining documents or
doing my best to reconstruct what would have happened.
INTRODUCTION TO THE GLO PROCEEDINGS
9. In paragraph 2 of the Request, I am asked to set out the background to my
instruction by POL and to describe the nature and extent of my role in advising
and representing POL in relation to certain matters.
10. My first involvement in this case occurred in May 2018. As I recall, I was told
that POL was having a series of meetings with several counsel with a view to
instructing one of them to act in a dispute in which it was involved. I was to be
one of the counsel whom POL would be meeting. On 18 May 2018, my clerks
forwarded to me copies of a letter of claim written by SPMs'’ solicitors
(‘Freeths”), a copy of a claim form and a briefing note containing a high level
summary of the dispute prepared for the purpose of these meetings.
11. The meeting with me took place on 20 May 2018. I do not remember much
about the meeting, but I think that Jane MacLeod, POL’s General Counsel, was
present. I also recall a discussion about a point which involved some real
concern for POL. This was the fact that, on the basis of information provided by
Fujitsu, POL had on several occasions formally confirmed that it was not
possible for anyone to use Horizon to alter branch transaction data remotely
(something which became known as “remote access”). POL now knew that this
was possible, and the question was how to manage this problem. I said that
POL should be open about it, and not to try to hide anything.
12.
WITN10500100
WITN10500100
I subsequently learned that I would be instructed on the case. Thereafter, I
received papers. There was a substantial amount of reading in to be done. I see
from an email exchange on 1 June 2016 (POL00140216) between me and
Andy Parsons (a partner of the firm which was then called Bond Dickinson and
which I shall call “WBD”) that my reading in was going slowly and that I already
had a number of questions to ask. I also see that I wanted a junior to be brought
in to help me. In due course, Owain Draper was instructed.
THE SWIFT REVIEW
13.
14.
When I was first instructed, the priority was for WBD to provide a full response
to Freeths’ letter of claim, which had been sent in April 2016. The letter of claim
was a long document. There were numerous points to respond to, including
factual claims, legal arguments and a proposal for the making of a Group
Litigation Order (a “GLO”). My focus at that time would have been on reading
into the case, trying to understand how the relationship between SPMs and
POL worked and trying to work out what POL’s answers were to the various
claims made. However, I see from POL00242402 that, on 8 June 2016, Andy
Parsons asked me for my views on whether POL should carry out further
investigative work in accordance with some recommendations that Jonathan
Swift QC and Christopher Knight had made to the chairman of POL on 8
February 2016.
This email came at an early stage of my involvement in this case. I have no
recollection of this email exchange or of having expressed any views about
these recommendations, either before the email or after it. However, having
read my email exchange with Andy Parsons on 8 June, I see that:
15.
(1)
(3)
WITN10500100
WITN10500100
In his email to me at 2:42pm, Andy Parsons referred to the
recommendations that had been made, indicated that these investigations
would overlap with issues in the GLO proceedings and suggested that, as
these points would need to be investigated in the proceedings in any event
(probably in a more robust way), it would be better for any investigative
work to be done as part of the proceedings so as to ensure that POL had
the protection of privilege.
I responded at 3:19pm. Andy Parsons had talked about providing what he
called “political cover” for the chairman. I said that I was not there to
provide political cover but I was concerned that the client (POL) should
protect its interests as a defendant to a substantial piece of litigation. The
consideration that seemed to me to be important was that privilege should
be preserved. I ended by asking what would happen if (contrary to my
assumption) it was decided not to carry out the recommended
investigations.
Andy Parsons replied to this question by saying that whether the
investigations were carried out would depend on how the litigation went. I
responded by saying that, from a pure litigation perspective, it was highly
desirable that the investigations be carried out.
As I have already said, I do not recall this exchange. Nor do I recall discussing
the point further at a conference on 9 June 2016 (although I believe that I had a
conference at POL'’s offices on that day: see paragraph 25 below). As I have
said, I have no recollection of POL’s response to the Swift review being raised
with me at any time. But it does not surprise me that I appear to have advised
POL that, as POL was facing a substantial claim raising issues which covered
the same ground as the relevant investigations, it would have been unwise for
5
WITN10500100
WITN10500100
POL to undertake those investigations in such a way as to lose the protection of
privilege. It was my duty to promote POL’s interests in the GLO Proceedings by
all proper means, and maintaining POL’s privilege in the GLO Proceedings
would have seemed an obvious means of achieving that objective.
16. The papers with which I have been provided include a letter from Andy Parsons
to POL dated 21 June 2016 (POL00006601). I do not recall seeing that letter or
being told about its contents. I make some comments on the letter in paragraph
174(1) below.
POL’S LITIGATION STRATEGY
17. AsIhave said, when I was first instructed, the priority was for WBD to provide a
response to Freeths’ letter of claim. That was a substantial task, not least
because the letter of claim covered a long period of time and contained many
elements. The letter of response was ultimately sent on 28 July 2016
(POL00110507).
18. My recollection is not clear, but Owain Draper and I would have concentrated
on the legal claims asserted in the letter, including the breach of contract
claims. A wide variety of causes of action were identified by Freeths, although
not always in a way that was easy to follow (for example, implied terms were
said to constrain Post Office’s discretions and powers, but the specific
discretions and powers which they were said to constrain were not identified).
Legal research had to be done on these claims, and sections responding to the
claims had to be drafted.
19. AsI recall, WBD undertook the drafting of the letter (they had knowledge of the
long history). Unsurprisingly (given their knowledge and our recent instruction),
WBD were responsible for the structure, thrust and most of the wording of the
6
20.
21.
22.
WITN10500100
WITN10500100
letter. Having said that, I (and Owain Draper) would have reviewed and made
suggestions about the structure and would have gone through the letter several
times during its gestation, querying points, asking questions, suggesting
amendments, and generally ensuring that the letter and its various schedules
made sense, asserted a properly arguable case and adopted positions that
were reasonable in law.
The letter of response contained a robust rejection of the claims made in the
letter of claim. The central point it made was that, although Horizon was not
perfect, it was a reliable system which the vast majority of SPMs used without
difficulty. Freeths suggested otherwise, but they did not have evidence or
particulars to justify that allegation. It seemed to me that POL (and, indeed,
Andy Parsons) believed this to be true. As I understood it, this was not a case in
which POL recognised that Horizon had serious flaws which POL wished to play
down. It was a case in which POL expected to be able to show in due course
that Horizon was at least as good as any other comparable IT system in use.
That remained my understanding throughout the proceedings.
Turning to the legal claims asserted, the letter of response set out POL’s case
as to the rights and obligations of the parties under the relevant SPM contracts,
the breaches of contract which POL was alleged to have committed, and the
various causes of action asserted by the claimants. With the assistance of some
research notes produced by WBD on discrete points, Owain Draper and I
played a substantial role in formulating the legal arguments advanced in the
letter. As I recall, my view at the time was that those arguments were not merely
proper arguments to advance but that they were right.
While the letter of response was being drafted, a separate line of
correspondence was being conducted on the question whether a Group
7
23.
WITN10500100
WITN10500100
Litigation Order should be made, which I would have commented on. In that
correspondence, POL had accepted that such an order should be made. The
letter of response therefore contained some proposals in relation to the scope of
the common issues to be covered by the GLO and also some proposals in
relation to a large number of requests which Freeths had made for immediate
(ie. voluntary) disclosure. I would have reviewed those proposals and
discussed how to respond to them with Andy Parsons.
In relation to disclosure, POL’s position was complicated by the fact that the
claim covered a long period approaching 20 years. At some point (I do not
remember precisely when, but at an early stage) I was informed that, during this
period, POL’s business had changed in various ways and its storage of
documents had also changed. This was going to make it difficult to identify
and/or locate all the documents that Freeths wanted. Similarly, I was informed
that, in relation to the operation of Horizon, many relevant documents were held
by Fujitsu. This meant that POL was dependent on Fujitsu to tell it what
documents were relevant and to retrieve all the relevant documents. In these
circumstances, it would have been difficult for POL to give early disclosure of
many of the requested documents. Andy Parsons and I would have discussed
points of this sort at this time. Such points supported the view that, from POL’s
perspective, it would be better for the disclosure of these documents to be
managed by the Court in due course.
POL’S DECISION MAKERS
24.
I do not recall being specifically told who was responsible for POL’s decision
making in relation to the letter of response, but I note that Andy Parsons’ email
to me of 8 June 2016 (POL00242402) referred to a POL litigation steering group
25.
WITN10500100
WITN10500100
and I believe that the draft letter was shown to and commented on by members
of POL’s senior management. During the GLO Proceedings, it was my
impression that the importance of this claim meant that decisions on significant
issues were considered by senior management, rather than just by POL’s in-
house lawyers. In the normal way, it was WBD who liaised with the client,
although there were occasions when POL’s in-house lawyers attended calls and
conferences with me.
My diary indicates that I had a conference at the client's offices on Thursday 9
June 2016. This was at an early stage of my instruction. I cannot remember
who was present or what topics were discussed at the conference, but I note
that it is referred to in Andy Parsons’ letter of 21 June 2016 (see paragraph 16
above). I also note that the letter describes the conference as having been with
POL’s legal team “and others”. I expect that Jane MacLeod would have
attended the conference, but I cannot say for sure that she did so and nor can I
say whether members of POL’s management were present.
THE ADEQUACY OF MY INSTRUCTIONS
26.
27;
As I saw it, I was being instructed by a defendant to a large and complicated set
of claims, which the client denied on substantial grounds. As time went on,
further documents were provided to me and answers were given to questions
which I asked. In my experience, this is not unusual in a large and complicated
case of this sort. At the time, I do not recall thinking that my instructions were
inadequate or that the documents with which I had been provided were
inadequate.
In Request 12, I am asked whether my views on the adequacy of my
instructions have changed with the benefit of hindsight. My answer to that
28.
29.
30.
WITN10500100
WITN10500100
question is yes. I explain this further below, when dealing with the Horizon
Issues.
Over the period following the letter of response, the correspondence with
Freeths continued. In due course, there were several hearings at which
directions were given regarding the GLO Proceedings, covering matters such
as the common issues to be tried and generic pleadings. Thereafter, generic
pleadings were exchanged, including a Generic Defence and Counterclaim
(POL00003340), which was served by POL on 18 July 2017. In the normal way,
the Generic Defence and Counterclaim was settled by counsel (i.e. by Owain
Draper and me). As far as I recall, it was consistent with the thrust of the case
that had been put in WBD’s letter of response.
By July 2017, Owain Draper and I had received more information about the
claim and our understanding of the factual issues regarding Horizon was better.
This included information contained in several reports produced by Deloitte
under the name “Project Bramble”, namely:
(1) areport dated 31 October 2016 was emailed to me on 10 November 2016
(POL00031502); and
(2) adraft executive summary of a further report (drafted by WBD with
amendments and comments from Deloitte) was emailed to me on 19 June
2017 (WITN10500101 and WITN10500102).
It also included matters that would have been discussed at a conference with
WBD, POL and Deloitte in my chambers of 22 June 2019.
At the time of the Generic Defence and Counterclaim, I did not think that my
instructions were inadequate or that the documents with which I had been
provided were inadequate.
31.
32.
WITN10500100
WITN10500100
It would of course have been preferable to have had a comprehensive set of
instructions and bundles containing documents which conveniently contained all
the information we needed. But in my experience, substantial cases are not run
like that. It is normal to ask questions as the case goes on and to form an
overall understanding from a variety of documents produced, emails sent and
meetings tasking place over a period of time.
In Request 11, I am asked about the extent to which I was aware of the
contents of Deloitte’s Project Bramble reports during the GLO Proceedings. In
addition to the reports I refer to in paragraph 29 above), I believe that the
following reports were sent to me, namely:
(1) areport dated 1 September 2017 (POL00041491); and
(2) areport dated 19 January 2018 (POL00028928).
I have found no email to me attaching Deloitte’s Project Sparrow report dated 8
July 2016, the Project Bramble report dated 27 July 2016, the Project Bramble
report dated 3 October 2017, or the project Bramble report dated 15 December
2017.
POL’S CASE ON PARTICULAR ISSUES
33.
34.
In Request 2.3, I am asked to describe the nature and extent of my role in
relation to POL’s case on implied terms, on the existence of bugs or errors and
on remote access.
First, dealing with POL’s case on implied terms, that case was based on the
advice that I (and Owain Draper) gave POL. Our advice was given on the basis
of legal research which was primarily undertaken by Owain Draper and was
reviewed (and approved) by me, both before the letter of response was served
and before the Generic Defence and Counterclaim was served.
35.
36.
37.
38.
WITN10500100
WITN10500100
The sort of factors we took into account in arriving at our views on implied terms
can be seen from the formal joint opinion that was produced subsequently, in
May 2018 (POL00103462). This joint opinion was signed by me, David
Cavender QC, Owain Draper and Gideon Cohen. It was the product of more
legal research than had been done previously, but it is consistent with my and
Owain Draper's thinking when we advised on the letter of response and when
we drafted the Generic Defence and Counterclaim.
As is set out in paragraphs 111 to 153 of the opinion, we took the view that it
was necessary to imply into the SPM contracts 1) a term requiring each party to
refrain from taking steps to inhibit or prevent the other from complying with its
contractual obligations, and 2) a term requiring each party to provide to the
other with such reasonable cooperation as was necessary to the performance
of the other's contractual obligations. On the basis that these terms were to be
implied, we did not think that it was necessary to imply any of the other terms
alleged by the claimants. We recognised that some of these terms were
arguable, but our view was that POL had the better of the argument.
Second, turning to POL’s case as to the existence of bugs, errors and defects in
Horizon, that case was based on instructions from POL, which instructions were
(as far as I was or am aware) based on information provided by Fujitsu.
In Schedule 6 to the Letter of claim, WBD addressed three particular bugs that
had been identified by Second Sight. The Schedule admitted that these three
bugs had in way one or another caused shortfalls in branch accounts and
explained that these bugs had been identified and fixed and that the accounting
problems they had caused been resolved. Paragraph 1.8 of the Schedule made
it clear that these were not the only possible bugs in Horizon and in paragraph
1.3 it stated that the important issue was not whether bugs existed, as they
39.
40.
WITN10500100
WITN10500100
likely did, but whether there were adequate controls in place to identify them
and take any necessary remedial action.
A similar approach was adopted in the Generic Defence and Counterclaim, at
paragraphs 49 to 56. Disclosure and expert evidence would plainly be needed
to identify any other relevant bugs (i.e. bugs which caused shortfalls in branch
accounts) and to determine whether the controls being operated were adequate
to identify such bugs, fix them and remedy their consequences.
Third, in relation to remote access:
(1) The letter of response addressed remote access at paras 5.14 to 5.18. In
paragraph 5.16.4, it indicated that a small number of specialist Fujitsu
administrators had edit permissions and that, as far as WBD was aware,
such permissions had not been used to alter branch transaction data and
WBD were seeking further assurance from Fujitsu on this point.
(2) These points reflected my instructions, and I had no reason to doubt them.
(3) Since my first meeting with POL in May 2016, my consistent advice had
been that Post Office should be transparent on the question of remote
access. As Andy Parsons explained in an email to me on 27 July 2016
(WITN0500103), he had included in the draft letter of response a direct
statement that privileged user permissions could be used to change
branch accounts. However, POL were uncomfortable about saying this in
terms, since it might lead to “public criticism”. His email set out the wording
which POL was comfortable with. Andy Parsons sent this email the day
before the letter of response went out. He knew that I was very busy on
another matter but indicated that my views would be welcome.
(4) 1 do not have a direct recollection of responding, but I would have
responded. I would have taken the view that Freeths would understand
13
(5)
(7)
(8)
WITN10500100
WITN10500100
POL’s approved wording to be indicating that privileged administrators
could use their permissions to alter branch transaction data.
The subsequent correspondence between Freeths and WBD had more to
say on this subject. I have not reviewed all the correspondence, but I have
found WBD’s letter to Freeths of 30 November 2016 (WITN10500104), in
which WBD explained POL’s understanding of Fujitsu’s administrator
permissions and how they could be used.
By the time the Generic Defence and Counterclaim was served in July
2017, Deloitte had been investigating the issue with Fujitsu and I had been
provided with a copy of their Project Bramble report dated 31 October
2016, which resulted in my raising some questions on 10 November 2016
(POL00337340). Deloitte continued their work and, on 19 June 2017,
Andy Parsons emailed to me a draft executive summary of the work they
had done in the intervening time (WITN0500101 and WITN10500102). He
informed me that this summary had been drafted by WBD, that Deloitte
had broadly agreed with it. The attached summary contained comments
from Deloitte.
A few days later, my diary indicates that, on 22 June, Deloitte attended a
conference in my chambers.
As I recall, when Owain and I drafted the paragraphs of the Generic
Defence and Counterclaim addressing the risk of Fujitsu privileged users
abusing their access rights in order to amend or delete transaction data
(paragraphs 57(4), 59 and 60), we took care to ensure that these
paragraphs were consistent with what Deloitte were saying.
WITN10500100
WITN10500100
THE PROCESS OF DRAFTING THE DEFENCE
a.
42.
43.
44.
In Request 10, I am asked to summarise the process by which the Generic
Defence and Counterclaim was drafted and my involvement in the same, and to
address certain issues in particular.
The Generic Defence and Counterclaim was drafted by me and Owain Draper.
In accordance with my normal practice, I would have played an active role in the
drafting process. It would have taken us a considerable time to draft and the
draft would have gone through a significant number of iterations, several of
which would have been shared with WBD for comment. We would have a
number of meetings and/or calls to discuss drafts with WBD and possibly also
POL. We would have taken care to ensure that it was based on our instructions
and was consistent with the documents and information with which we had by
then been provided.
In Request 10.1, I am asked to consider paragraphs 43(1) to (3) of the pleading
and to consider the basis on which POL pleaded that “The blocked value is not
(and is not treated as) a debt due to Post Office”:
(1) These paragraphs would have been pleaded on the basis of instructions
from POL, supplemented by the Branch Trading Manual referred to in
paragraph 43(4) (and quite possibly other documents also, although I
cannot now identify what other documents we looked at in this regard).
In Request 10.2, I am asked to explain the basis on which POL denied in
paragraph 48(3)(b) that Fujitsu edited or deleted specific items of transaction
data.
(1) As I read the paragraph, what was denied was the claimants’ allegation
that Fujitsu had managed fixes to coding errors and bugs that had the
45.
46.
47.
WITN10500100
WITN10500100
effect of editing or deleting specific item of transaction data. To be clear,
the claimants’ case regarding remote access was addressed in a different
part of the Generic Defence and Counterclaim (paragraphs 57 to 60).
(2) This denial would have been pleaded on the basis of my instructions from
POL, which it would have given after consulting Fujitsu.
In para 10.3, I am asked to explain the basis on which POL denied in paragraph
48(3)(c) that Fujitsu had implemented fixes that affected the reliability of
accounting balances, statements and reports.
(1) This would have been pleaded on the basis my instructions from POL,
which it would have given after consulting Fujitsu.
In para 10.4, I am asked to explain the basis on which POL pleaded in
paragraph 50(4) that, to the best of its knowledge information and belief, there
was no issue in the Known Error Log that could affect the accuracy of a
branch's accounts or the secure transmission or storage of data.
(1) This would have been pleaded on the basis of information provided by
Fujitsu.
In paragraph 10.5, I am asked to explain the basis on which POL pleaded, in
paragraph 57(4), that 1) for Fujitsu’s privileged users to have abused their rights
to so as to alter their transaction data and conceal that this had happened
would have been an extraordinarily difficult thing to do, involving complex steps
which would require months of planning and an exceptional level of technical
expertise, 2) POL had never consented to the use of privileged user rights to
alter branch data, and 3) to the best of POL’s knowledge information and belief,
these rights had not been used for this purpose.
(1) Point 1) would have been pleaded on the basis of the Project Bramble
report dated 31 October 2016, of the Project Bramble executive summary
16
WITN10500100
WITN10500100
which was provided on 19 July 2017 and of the information / confirmations
which Deloitte would have provided in conference on 22 June 2017.
(2) Point 2) would have been pleaded on the basis of instructions from POL.
(3) Point 3) would have been pleaded on the basis of instructions from POL,
which it would have given after consulting Fujitsu.
CONFERENCES WITH POL’S LEGAL DEPARTMENT OR THE BOARD
48.
49.
50.
In Request 9, I am asked to describe any conferences or significant discussions
which I had with POL’s legal department or the board or POL’s decision makers
in relation to the matters set out in Request 2. During the GLO Proceedings, I
generally dealt with WBD. I had a fair number of dealings with POL’s in-house
lawyers, but to the best of my recollection I would not describe them as
frequent. I would describe my dealings with POL’s management as infrequent.
My diary indicates that, between the date of my initial meeting with POL in May
2016 and the date on which the Generic Defence and Counterclaim was served
in July 2017, I had the following conferences or telephone conferences with
POL:
(1) aconference with POL on 9 June 2016 at POL’s offices in Finsbury Dials;
(2) two telephone conferences involving Angela Van Den Bogerd on 28 June
and 5 July 2016;
(3) aconference with POL on 14 November 2016;
(4) atelephone conference with POL on 28 November 2016;
(5) aconference with POL on 7 June 2017; and
(6) aconference with POL (and Deloitte) on 22 June 2017.
I have no direct recollection of these conferences, who attended them or what
they covered. Doing my best to reconstruct:
(1)
(2)
(3)
(4)
(5)
WITN10500100
WITN10500100
The 9 June 2016 conference would probably have concerned the claims
put forward by the claimants and how best to respond to them. I assume
that it will also have included a discussion about the Swift
recommendations.
I suspect that the two telephone conferences with Ms Van Den Bogerd
were Calls to allow me to understand some of the factual points relevant to
the dispute, such as how POL conducted relevant aspects of its business
and/or how it dealt with SPMs.
I have been provided with an agenda for the conference on 14 November
indicating who were to attend on POL’s side and what points were to be
discussed (POL00024971). I do not remember this conference or what
was said on these points but, in relation to the remote access item, I note
that the conference took place a few days after I had received the Project
Bramble report dated 31 October 2016. I infer that this item may have
involved or included a discussion of the progress Deloitte were making in
its investigations into remote access and what further work they could do
in this regard.
The conferences with POL on 7 June and 22 June 2017 would have
involved discussion of points arising on the Generic Defence and
Counterclaim that Owain Draper and I were then drafting. My diary entry
for the 22 June conference is marked “Defence”.
Later on in the GLO Proceedings, I recall attending conferences with
members of the POL board. I discuss these subsequent conferences
below. However, I cannot say whether any of the conferences referred to
above were with board members.
51.
WITN10500100
WITN10500100
I may well have had further telephone conferences which were attended by
POL personnel as well as WBD. I do not remember any specific calls. To the
extent that there were any, I suspect that they would have been with POL’s in-
house lawyers rather than with management.
DISCLOSURE AND THE USE OF LEGAL PROFESSIONAL PRIVILEGE
52.
53.
I note that, in Request 2.4, I am asked to describe the nature and extent of my
role in relation to POL’s case in relation to disclosure and the use of legal
professional privilege. The only documents with which I have been provided
having relevance to disclosure and privilege during the period between my initial
instructions and the Generic Defence and Counterclaim concerned the Swift
review, which I consider above. However, as leading counsel for POL I would
have advised POL on any issues of disclosure and privilege that WBD raised
with me, if I was available. And I represented POL at several Case
Management Conferences (“CMCs") at which disclosure was dealt with
(including the first CMC, considered in paragraphs 134 and 135 below). Where
issues of disclosure and/or privilege were addressed at the hearings at which I
appeared, I would have advised on those issues and I would have commented
on any witness statements that were served in relation to those issues.
According to my diary, I had two conferences at POL's offices in relation to
CMC strategy and directions. These took place on 14 September 2017 (under
the entry “CMC Strategy”) and 29 September 2017 (under the entry “CMC /
Directions etc”). I have no recollection of the particular points that were
discussed at those conferences or who attended them.
WITN10500100
WITN10500100
THE COMMON ISSUES
54.
55.
56.
I did not represent POL at the Common Issues trial, which took place between 7
November 2018 and 6 December 2018. David Cavender QC was instructed on
that phase of the GLO Proceedings. I believe that he become involved towards
the end of 2017. In addition, a junior barrister called Gideon Cohen was brought
in to work alongside Owain Draper.
As I recall, David Cavender took the lead on the Common Issues from around
the time he was instructed. He also provided some help on a few aspects of the
Horizon Issues. I had a trial on another substantial matter between January and
March 2018 and I would have had limited capacity to work on the GLO
Proceedings in the months leading up to my trial and during the trial itself. But I
was still copied in on at least some of correspondence, and I recall
subsequently being involved in the drafting of individual Defences in the claims
that were prepared for the purposes of the Common Issues Trial (the Defences
bear my name and those of Owain Draper and Gideon Cohen). I also recall
representing POL at a CMC at which an application was made to strike out
some of the evidence which the claimants wished to rely on at the Common
Issues trial. But I do not have a clear recollection of how matters developed
during the course of the Common Issues phase of the litigation.
In Request 13.1, I am asked to provide an account of the nature and extent of
my role in advising POL on the Common Issues in respect of the development
of POL’s case on the settle centrally button:
(1) Ihave no recollection of POL’s case on this point developing in the
Common Issues phase, or of having performed a role in the development
of that case.
57.
(2)
WITN10500100
WITN10500100
I see from POL00364016 that, on 22 February 2019 (in the run up to the
Horizon Issues trial), I sent an email to Owain Draper raising a question
about settling centrally. In that email, I noted that the claimants’ expert for
the Horizon Issues (Jason Coyne) had asserted that, at the Common
Issues trial, POL’s case had been that a disputed shortfall which was
settled centrally was recorded as a debt due to POL. I asked for Owain
Draper's thoughts (he had appeared at the Common Issues trial).
Owain Draper replied by telling me that, having checked the document
which Mr Coyne relied on in support of this claim, the document did not
say what Mr Coyne said it said. In fact, the document indicated that the
amount settled centrally would be treated as a debt due to POL unless it
was disputed by the SPM.
I have no direct recollection of this email exchange or of what was
subsequently said about the settle centrally button. But I believe that this is
the case that was asserted in the Generic Defence and Counterclaim.
In Request 13.2.1, I am asked to provide an account of the nature and extent of
my role in advising POL on the Common Issues in respect of POL’s case on the
question whether there should be an implied term that POL was not to suspend
or terminate SPM contracts (a) without reasonable and proper cause or (b) in
circumstances where POL was in material breach of duty itself:
(1)
In its Generic Defence and Counterclaim of July 2017, POL denied that
the SPM contracts contained an implied term constraining either its right to
suspend SPM contracts or its right to terminate SPM contracts (see
paragraphs 99, 100, 105 to 106). As far as I recall, this remained POL’s
case in the Common Issues phase of the proceedings.
58.
59.
WITN10500100
WITN10500100
(2) AsImention in paragraph 35 above, in May 2018 POL’s counsel team
produced a joint opinion in which we explained our views on this question
(POL00103462). In September 2018, we produced a joint update to this
opinion (POL00022669).
(3) 1 do not believe that it is necessary for me to set out the views expressed
at length in these opinions. Amongst other things, we explained our view
that the express rights to suspend and terminate were absolute
contractual rights which POL was entitled to exercise in its own interests,
not contractual discretions which POL was required to exercise in the
interests of both parties. I confirm that these opinions reflected my
considered views.
We also explained our view on the question whether the claimants could rely on
an implied term to the effect that POL’s contractual discretion would not be
exercised dishonestly or in an arbitrary, capricious or irrational manner (see
Request 13.2.2). We recognised that such a term is frequently implied so as to
constrain contractual discretions, but we also noted that, in their Generic
Particulars of Claim, the claimants did not seek to apply such a term to any true
contractual discretions.
The joint opinion was dated 10 May 2018. I have been provided with unsigned
minutes of a meeting of the Postmaster Litigation Subcommittee of POL 5 days
later, on 15 May 2018 (POL00006754). I have some recollection of attending,
with David Cavender, a meeting with members of POL’s board at or around this
time. As I recall, David Cavender did much of the talking at this meeting, and he
spent some time explaining in simple terms the Common Issues and our views
as to the merits on those issues. I do not have a detailed recollection of the
meeting, however.
60.
61.
62.
WITN10500100
WITN10500100
The meeting was with members of POL’s board, and I note the reference in the
minutes to it being a meeting of the “Postmaster Litigation Subcommittee”. This
was not the only meeting we attended at POL’s offices in relation to the
Common Issues: according to my diary, we also attended such a meeting on 29
June 2018. However, I do not recall whether this was another meeting with POL
board members or with POL’s in-house lawyers. Nor do I recall what was
discussed.
As I explain further below, I subsequently attended some further meetings with
board members in relation to the Horizon Issues. It would thus have been clear
to me that, at least on important matters, the persons responsible for decision
making were these board members. However, I cannot say what demarcation
was applied between points that were decided on by these board members and
points that were decided on by POL’s in-house lawyers.
I do not remember having any telephone calls with POL on the Common Issues.
No such calls are identified in my diary, but this does not mean that none took
place.
COMMON ISSUES DISCLOSURE
63.
64.
In Request 13.3, I am asked to provide an account of the nature and extent of
my role in advising POL on the Common Issues in respect of disclosure. I
repeat paragraphs 54 and 55 above. I recall representing POL at more than one
CMC where questions of disclosure were dealt with, but I do not have a clear
recollection of the relevant hearings, and I do not have any CMC judgments on
disclosure to hand which could refresh my memory.
I believe POL was represented by David Cavender QC at one or more CMCs in
2018, but I do not recall whether any of these CMCs dealt with disclosure or, if
WITN10500100
WITN10500100
they did, whether the disclosure related to the Common Issues or the Horizon
Issues. If and to the extent that any disclosure issues were addressed at the
hearings attended by David Cavender, I may also have advised on those issues
and commented on any relevant witness statements. But I do not remember.
COMMON ISSUES WITNESS STATEMENTS
65.
I do not recall performing any role in relation to the POL witness statements that
were prepared for the Common Issues trial. I have no recollection of attending
any conferences or significant discussions with POL’s witnesses to discuss the
form or content of their evidence. The only role which I remember performing in
relation to the evidence at the Common Issues trial was representing POL at
the hearing of its application to strike out some of the claimants’ witness
evidence.
RECUSAL
66.
I had no role in advising POL in respect of the decision to apply for an order that
Fraser J be recused. The Common Issues judgment was handed down on 15
March 2019, and the Horizon Issues trial started on 11 March 2019. During the
relevant period, I was extremely busy. I did not have the time to read the
Common Issues judgment, still less to advise on its implications. POL and WBD
ensured that I was not involved in the process of considering the Common
Issues judgment and determining what action should be taken in relation to it.
CONFERENCES WITH POL’S LEGAL DEPARTMENT OR THE BOARD ON THE
COMMON ISSUES
67.
In Request 20, I am asked to describe any conferences or significant
discussions I had with POL’s legal department or the board or POL’s decision
WITN10500100
WITN10500100
makers on the matters set out in Request 13. The only Common Issues
conferences or discussions of which I am specifically aware are those I refer to
in paragraph 60 above. However, it is possible that the conferences I refer to in
paragraph 53 above addressed matters relevant to the trial of the Common
Issues.
THE ADEQUACY OF MY INSTRUCTIONS ON THE COMMON ISSUES
68.
Regarding the adequacy of my instructions on the Common Issues, I repeat
paragraph 26 above. As I recall, the Common Issues largely concerned matters
which were within POL’s own knowledge and documents which were within its
control. In relation to the Common Issues matters I dealt with, I do not recall
feeling any concern about the adequacy of my instructions at the time.
However, I did not take part in the Common Issues trial and, when the Common
Issues judgment was handed down, I did not have time to read the judgment (I
was hard at work on the Horizon Issues). Not least for these reasons, hindsight
has not given me a new perspective on the adequacy of my instructions on the
Common Issues.
THE HORIZON ISSUES
69.
70.
Before giving an account of the nature and extent of my role in advising and
representing POL in the Horizon Issues phase of the GLO Proceedings, it may
be helpful to say a few words about the procedure pursuant to which those
issues were tried.
The Horizon Issues covered a wide range of questions, but as I saw it at the
time, the critical questions were 1) the incidence of bugs in Horizon that were
capable of adversely affecting branch transaction data and thereby creating
71.
72.
WITN10500100
WITN10500100
false shortfalls for SPMs, 2) whether the controls operated in relation to Horizon
were adequate to identify such bugs and fix them, 3) whether the controls
operated were adequate to identify the branch accounts which were affected by
these bugs and to correct those branch accounts, and 4) whether there was a
risk of Fujitsu or POL remotely accessing branch accounts so as to edit or
delete transaction data and thereby to create false shortfalls for SPMs.
The directions given in relation to the Horizon Issues trial were in my experience
unusual. The trial dealt with issues that covered a long period of time, nearly 20
years, during the course of which the Horizon system changed substantially. On
the basic question whether the Horizon was reliable, POL had to wait and see
what points were raised in the claimants’ expert evidence (by Mr Coyne). A
direction had been made requiring the claimants to serve an outline document
setting out the nature of their allegations in relation to the Horizon Issues, but as
I recall this was of little help.
My purpose in saying this is not to complain about it: there were valid reasons
for organising the GLO Proceedings in this way. For example, it would not have
been possible for the claimants to plead a proper case on the critical issues until
after disclosure had been given and after Mr Coyne had reviewed the
disclosure. But the chosen procedure meant that POL could not know the case
it had to meet until it saw Mr Coyne’s first expert report (“Coyne 1”). Directions
had been made for sequential reports, but the directions did not allow for much
time between the claimants’ expert report and POL’s expert report (by Robert
Worden), and supplemental reports. And Coyne 1 was long and complex,
raising innumerable issues and exhibiting innumerable documents, many of
which were highly technical and difficult for understand. All the experts’ reports
were long and complex.
26
73.
74.
75.
76.
WITN10500100
WITN10500100
The timetable would have caused difficulties for both sides. Looking at it from
the perspective of POL’s legal team, it meant that there was little time for us to
understand what Mr Coyne’s key criticisms were, what practical significance
they had, whether they were well founded and what the answers to them were.
In seeking to understanding these things, we needed help from Dr Worden. The
need for his help was reinforced by the fact that, on their own, many of Fujitsu’s
documents were incomprehensible to an outsider. These included thousands of
Peaks and KELs and many other documents of which we subsequently became
aware, such as OCPs, OCRs and MSCs. It was necessary to spend a great
deal of time with Dr Worden in order to gain a proper understanding of the
critical parts of Mr Coyne’s and his expert reports.
Quite apart from the expert reports, the court also directed witness statements
to be served, with an initial exchange of statements before Coyne 1 was due to
be served and supplemental statements being served during the period in which
that report was being assimilated by POL and Dr Worden’s first report
(‘Worden 1”) was being prepared.
In these circumstances, the last few months of 2018 and the first few months of
2019 were a very intensive period. I and the rest of the counsel team were
working very hard in order to get ready for trial. By that stage, a senior junior IT
specialist, Simon Henderson, had been brought in to work with me and Owain
Draper on the Horizon Issues. At some later point, it was recognised that we
needed more support and a further junior from Simon Henderson’s chambers
was also instructed (Rebecca Keating). We were very busy indeed. There was
no time to reflect on matters, or to read the vast number of documents that were
in the trial bundles. I did not even have time to read all the documents exhibited
to Coyne 1 and Coyne 2.
WITN10500100
WITN10500100
BUGS AND REMOTE ACCESS
Tt.
78.
J:
In Request 22.1, I am asked about my role in advising and representing POL in
respect of the development of POL’s case on the existence of bugs in Horizon
and on remote access.
It is not possible for me to provide a comprehensive answer to this question
because such an answer would require a detailed consideration of many
documents, including the Judgment of Fraser J on the Horizon Issues, the
expert reports and joint memoranda served by the parties and POL’s written
submissions for the Horizon Issues trial. I have not had time to do this. Indeed, I
have not had time to look at most of these documents at all. In the following
paragraphs, I set out my recollection of the position. These paragraphs give an
overview which of necessity cannot include everything that might be considered
relevant. A thorough account of the position would take an extremely long time
to write.
Bugs
In order to develop POL’s case on relevant bugs (i.e. bugs which were capable
of adversely affecting branch transaction data), we (the counsel team) needed a
clear account of what bugs had arisen in Horizon, how they had been detected
and what had been done about them. There were a small number of bugs which
the parties had known about for some time, and POL was able to obtain from
Fujitsu documents and information about those bugs and how they were dealt
with. Its case on those bugs was straightforward. But as for other bugs, my
understanding was that POL's case was based on what Fujitsu told it. This was
that there were likely to be other bugs, but in the normal course of things these
were detected and fixed and their consequences were identified and remedied.
80.
81.
82.
83.
WITN10500100
WITN10500100
It was also that Horizon system was designed and operated by Fujitsu in a way
that ensured that there were no systematic flaws that remained hidden in the
system and so had a lasting adverse effect on branch data.
The expectation was that these points would be substantiated once disclosure
was given of the relevant Fujitsu documents and parties’ experts had gone
through those documents.
The experts went through the disclosed documents and each of them identified
a number of bugs that were capable of affecting branch transaction data. These
bugs were listed in a table jointly prepared by the experts. That table became a
primary focus of analysis and argument at the trial. It covered 28 or 29 bugs
(some of which included more than one variant). For each bug, the table
identified the year of the bug, identified the evidence supporting the existence of
the bug and summarised the experts’ respective views on whether it was a bug,
its nature and effect, and how it had been dealt with.
As I recall, Dr Worden’s view was that not all the bugs in the bug table were
bugs, or bugs which could be said to have an adverse effect on branch
transaction data. But many were, and in relation to those he essentially took the
view that the systems in operation were effective at identifying such bugs, fixing
them, identifying their consequences and remedying those consequences.
My recollection is that, once the bug table was produced, it and the documents
it referred to became the principal battleground between the parties. But there
was also some witness evidence on bugs, including from Torstein Godeseth,
Fujitsu’s Chief Architect on the Post Office Account, and Stephen Parker, who
was Fujitsu’s head of Post Office Application Support.
29
84.
85.
86.
87.
88.
WITN10500100
WITN10500100
Thus, in relation to most of the bugs that were in issue, POL’s case was
developed in the course of the expert evidence processes. As I recall, the case
became very much more specific, but its essential thrust remained the same.
Once the issues between the parties on these bugs were identified, my role was
to put POL’s case on the relevant bugs to Mr Coyne in cross examination, to
oversee (and take part in) the process of drafting written closings which
addressed the evidence and to address the Judge in oral closings. For these
purposes, I had the assistance of detailed notes prepared by WBD containing a
detailed analysis of relevant bugs. I may also have had the assistance of call
with Mr Parker, although I see from FUJ00155196 that the call may have been
about remote access only. I discuss this call below.
WBD's notes were of particular help to me when preparing my cross
examination of Mr Coyne. My brief review of WBD’s note on remote access
(considered below) reminds me that these notes were based on analyses of the
documents on which Mr Coyne relied for his opinions and of other evidence in
the case, together with insights and explanations provided by a team of people
at Fujitsu and also some notes from Dr Worden. I see from POL00140306 that it
included comments from Gareth Jenkins, whom I discuss below.
I relied heavily on these notes in formulating the points that were put to Mr
Coyne in cross examination and to the Judge in closing. I found them
persuasive. My view was that the information and analyses they contained,
together with the other material referred to above, provided substantial support
for the detailed case that was asserted in POL’s written and oral closings.
Remote Access
In relation to remote access, POL’s case was also developed in the course of
the witness and expert evidence processes. As with POL’s case on bugs, its
30
89.
90.
WITN10500100
WITN10500100
case on remote access did not change radically, although it became more
detailed, particularly in relation to the controls applied to privileged Fujitsu users
who had administrator access and the audit records of the occasions on which
they exercised such rights. My recollection is that Dr Worden’s evidence was
consistent with the case pleaded in the Generic Defence and Counterclaim. It
may be worth noting that, even if privileged users could alter branch transaction
data remotely without any risk of detection, POL contended that it would never
have been in their interests to do so. POL’s case was that it was unreal to
suggest that, over two decades, some of Fujitsu’s senior employees had
abused their privileged access to create large (or small) shortfalls in the branch
accounts of a large number of SPMs.
From my perspective, during the run up to the Horizon Issues trial the most
striking development that occurred in relation to remote access was the
claimants’ service of Mr Roll’s witness statement, which made claims about the
widespread use of remote access to alter branch transaction data during the
early years of Horizon. These claims were impossible to reconcile with anything
that I had previously been told or any of the documents that I had seen. Not
surprisingly given the lapse of time, Mr Roll’s statement was expressed in very
general terms. This made it difficult to deal with, but it did have to be dealt with.
It was therefore necessary to call evidence from Mr Parker, who had worked
with Mr Roll in those early years. Mr Parker fundamentally disagreed with Mr
Roll’s account and suggested that he might be confusing other remote
processes with remote access to alter branch data.
Again, my role was to put POL’s case (including Mr Parker's evidence) on
remote access to Mr Roll and to Mr Coyne in cross examination, to oversee
(and take part in) the process of drafting written closings and to address the
31
91.
92.
WITN10500100
WITN10500100
Judge in oral closings. Again, I had the assistance of a detailed note prepared
by WBD on remote access. I also had the assistance of a call with Mr Parker in
May 2019. I see from FUJ00155195 that, on 22 May 2019, Mr Parker was sent
an email which attached a document in anticipation of that call (FUJ00155196).
This document was the note that WBD had prepared to help me on remote
access.
FUJ00165648 indicates that my call with Mr Parker was postponed. My diary
suggests that it in fact took place on 24 May 2019. The call would have been to
help me prepare for my cross examination of Mr Coyne on remote access. It
took place during the adjournment of the Horizon Issues trial resulting from
POL’s recusal application. By that stage, POL’s factual evidence had all been
given: Mr Parker was POL'’s last factual witness, and he gave evidence on 11
April 2019 (day 12 of the trial).
Again, my view was that the information and analyses contained in WBD’s
notes and the other material referred to above provided substantial support for
the detailed case on remote access that was asserted in POL’s written and oral
closings.
HORIZON ISSUES WITNESS EVIDENCE
93.
94.
Turning to my role in the preparation of POL’s witness evidence, this was not a
case in which counsel gave detailed advice on who should be giving evidence
on POL’s behalf and the points that should be covered in their witness
statements. However, counsel did advise on the question whether POL should
call Gareth Jenkins as a witness.
For reasons which will not need to be explained, Gareth Jenkins was an
obvious candidate to give evidence for POL. However, I was aware from WBD
95.
96.
97.
WITN10500100
WITN10500100
that there were doubts about his reliability: he had given expert evidence about
the Horizon system in several criminal trials which was now said to be false. I
recall Andy Parsons telling me that he had set up a meeting with a solicitor who
had acted for POL in one or more of these trials and that I needed to hear what
this solicitor had to say. I do not remember the solicitor’s name or his firm, but I
see from Andy Parsons’ email to me and Simon Henderson of 7 September
2018 that two solicitors came to my chambers, Simon Clarke and Martin Smith
(WITN10500105). I did not know either of them.
My recollection of this meeting is not clear — indeed, until I saw the above email
I had thought that we had a telephone call. But the upshot was that I was told in
emphatic terms that Mr Jenkins was not a reliable witness. The solicitors said
that Mr Jenkins had given misleading evidence. They suggested in no uncertain
terms that I should be very cautious about calling him as a witness.
POL was a defendant to a substantial civil claim in which the reliability of
Horizon was in issue. This was adversarial litigation. POL’s case was that
Horizon was reliable and, in order to prove that case, the witness evidence that
it called needed to be reliable. As counsel for POL, it was my duty to promote
POL’s interests by all proper means, and it would not have been consistent with
that duty for me to advise POL to call an unreliable witness.
My conclusion was that, if POL needed a witness from Fujitsu, it needed a
witness who could be relied upon, not a witness who could not. If POL did not
call Mr Jenkins, the claimants would obviously turn this point to their advantage,
inviting the Judge to draw adverse inferences against POL on important issues.
But on the basis of what I was told, it was clearly in POL’s interests not to call
him.
98.
99.
100.
WITN10500100
WITN10500100
Fortunately, Mr Godeseth was available. His knowledge of Horizon across the
entire relevant period was not as great as Mr Jenkins’, but I was informed that it
was substantial.
I see from POL00134909 that, after the end of the Horizon Issues trial and
before judgment, Simon Henderson and I had a conference with WBD and
Herbert Smith Freehills (“HSF”) in which I addressed various questions asked
by Alex Lerner, an assistant at HSF. I have some recollection of a call or
meeting of this sort but not of what was said. POL0034909 is HSF’s attendance
note of the conference. As to that note:
(1) I see from page 1 that I was asked whether, at the Horizon Issues trial, the
claimants advanced a case that POL suppressed evidence regarding
bugs.
(2) Pages 9 and 10 record my answer to this question. I see that part of my
answer related to the decision not to call Mr Jenkins as a witness.
Essentially, I said that the claimants had asserted that the fact that POL
did not call Mr Jenkins was suppression, and I explained why he was not
called.
(3) Although the terms in which I spoke are rather more colourful than I would
use in a witness statement, the reasons I gave for not calling Mr Jenkins
were the reasons I have set out above. He would have been an unreliable
witness and at trial the claimants would have been able to undermine his
credibility.
While on the subject of Mr Jenkins, I note that Request 29 asks me to provide
an account of what I was told about him and what my impression of him was.
My recollection of what I was told about Mr Jenkins is set out above. I do not
recall having my own impression of him.
34
WITN10500100
WITN10500100
101. Turning to the witness statements on which POL relied for the purposes of the
Horizon Issues trial, these were served in several rounds: the first round of
evidence was in late September 2018, and the second was in mid November
2018. POL also served some additional evidence in January and February
2019, I think without permission. The further rounds of evidence reflected the
fact that new points were raised in the claimants’ witness and expert evidence,
and also the fact that POL had discovered that some of its existing evidence
needed to be corrected or clarified.
102. For example, following receipt of the claimants’ first round of witness evidence
and Coyne 1, it was necessary to respond to Mr Roll’s first witness statement
(‘Rolls 1”) and also to provide Fujitsu’s comments on a number of KELs which
Mr Coyne had identified in his report. The natural person to do that was Mr
Parker. He therefore made his first witness statement (“Parker 1”), which was
included in POL’s second round of evidence (POLO0000692).
103. All the witness statements on which POL relied were drafted by WBD, and their
drafts were circulated to counsel for our comments. Our comments covered a
wide variety of matters, including: asking questions; requesting clarification;
identifying obvious errors; querying points which appeared not to be consistent
with other statements, documents, or our understanding of the position; drawing
attention to the implications of the drafting which may not have been intended,
asking that sources of information be identified; suggesting deletions of
unnecessary text; and identifying additional points that the witnesses might
address.
104. In relation to the statements that were served in September, I believe that drafts
were shared with me and Simon Henderson. However, they were shared only a
few days before they were due to be served. I was asked to comment on them
35
105.
106.
107.
108.
WITN10500100
WITN10500100
and I did so as best I could in the time I had. I believe that the draft Godeseth
statement on which I was asked to comment was sent to me the day before it
was due to be served and my comments on it were hurried.
In relation to the statements that were subsequently served, my recollection is
that Simon Henderson and I often provided joint comments, and that we were
generally acting under significant time pressure. As I explain in paragraphs 70
to 76 above, during this period, the whole of POL’s legal team was extremely
busy.
It is also my recollection that more than a few of the draft statements with which
were provided to counsel were unclear, difficult to follow and/or raised new
questions. These drafts included passages which were poorly explained or
incomplete and passages which failed to make it clear that what was being said
was based on information provided by other parties. They sometimes appeared
to contain contradictions. And in relation to points on which I believed that I had
an understanding, they sometimes said things which threw my understanding
into disarray and appeared to require me to reorganise my thoughts.
This was frustrating, not least because we did not have a great deal of time in
which to deal with problems or bottom questions out.
To the best of my recollection, the biggest difficulties were with Fujitsu. My
perception from the draft statements I was seeing was that they sometimes
found it hard to give clear answers, and sometimes appeared to express
themselves in ways that raised more questions than they answered. In the
normal way, all interviews with actual or potential witnesses were conducted by
WBD. But the problems encountered with the Fujitsu witnesses were such that,
rather than spending time that we did not have to send draft statement back and
forth between Fujitsu, WBD and counsel, there was an occasion on which we
36
109.
110.
111.
WITN10500100
WITN10500100
felt it necessary for Simon Henderson to meet with Fujitsu with a view to
achieving clarity. I imagine that this is the meeting that Jonny Gribben refers to
in his email of 15 November 2018 (POL00363816).
These are the sorts of problems which would have been what I had in mind
when, in my email to Andy Parsons dated 18 November 2018 (POL00363851), I
referred to facing “striking difficulties” in getting clear instructions from POL and
Fujitsu. I see that I was there talking about the difficulties in obtaining the factual
instructions that were needed to produce POL’s witness statements in
accordance with the tight timetable within which we were working.
The documents with which I have been provided illustrate the role I played in
the preparation of POL’s witness statements. In the following paragraphs, I
provide some examples in relation to the two witnesses with whom I have been
particularly asked to deal, namely Mr Godeseth and Mr Parker.
In relation to Mr Godeseth:
(1) I see that, on 23 February 2019, I queried whether Mr Godeseth might
need to correct some of the things that he had said about remotely
deleting transaction data in his first statement (“Godeseth 1”)
(POL00000682). In an email from Andy Parsons (POL00367005), I was
told that Mr Godeseth had confirmed that Fujitsu did not remotely delete
transaction data (see also POL00364020). Andy Parsons suggested that
the point be clarified by the giving of an explanation of the distinction
between transaction data and other data. I agreed. I believe that this
explanation was given in Mr Godeseth’s third statement (POLO0000686)
(‘Godeseth 3”).
(2) I see from POL00363955 that, on 18 February 2019, I sent Andy Parsons
two emails attaching some notes, the first of which identified points on
37
WITN10500100
WITN10500100
which the counsel team wanted supplemental evidence to be prepared
and the second (POL00363956) identified points which we thought that Mr
Godeseth might need to correct or consider in the light of Mr Coyne’s
second report (“Coyne 2”) and of the claimants’ supplemental evidence.
WBD would have raised these points with Mr Godeseth when drafting
Godeseth 3.
(3) I see from POL00364056 that, on 27 February 2019, I provided comments
on the current draft of Godeseth 3 to Jonny Gribben of WBD. I also see
that Jonny Gribben provided some responses to my comments and that I
provided some responses to his responses in capitals. I note that, in my
comments, I indicated that if Mr Jenkins was to be the source of any
information which Mr Godeseth relied on, Mr Jenkins needed to be
identified in the witness statement. I added that I would prefer his
statement not to be based on such information (for the reasons indicated
above). I believe that the relevant paragraphs of this draft (paragraphs
26.1 and 26.2) were omitted from the final version of Godeseth 3. Mr
Parker dealt with these points without needing to rely on any source of
information (see paragraphs 17.1 and 17.2 of Parker 3 at POLO00006839).
(4) I also see from POL00364056 that, on 28 February 2019, Jonny Gribben
told me that Mr Godeseth now considered that, with enough access rights
a privileged user could inject transaction data and asked me whether this
should be made clear in Godeseth 3. I imagine that I advised that the point
should be made clear and I see that it was, in paragraph 14 of Godeseth
3.
112. In relation to Mr Parker:
WITN10500100
WITN10500100
(1) As I have already explained, Parker 1 responded to Mr Roll’s statement,
provided some information on KELs and Peaks and attached a table
addressing a number of KEL’s which Mr Coyne had relied on in Coyne 1.
As to this table:
(a)
(b)
(c)
(d)
I see from paragraph 66 of Parker 1 that the table was said to contain
the initial explanations produced by a team from Fujitsu’s SSC.
I also see that, on 12 November 2018 (four days before Parker 3 was
finalised and served), Jonny Gribben sent an email to Simon
Henderson and me in which he said that the relevant KELs were
being analysed by Mr Parker’s team plus Mr Jenkins
(POL00363775). In my reply, I balked at this, reminding him that as
we had decided that Mr Jenkins should not be a witness, he should
also not be a source of information. I pointed out that where Mr
Jenkins was acting as a source, the claimants would know this (i.e.
because sources have to be identified). I expressed dissatisfaction at
the fact that Mr Jenkins kept popping up on technical questions and I
asked that his involvement be limited as much as possible.
Jonny Gribben replied by confirming that WBD were limiting Mr
Jenkins’ involvement as much as possible but that if Mr Godeseth or
Mr Parker covered the bugs they still needed to speak to Mr Jenkins.
Parker 1 did not say that, in the course of producing the table, Mr
Parker had spoken to Mr Jenkins. If explanations contained in the
table had been produced by Mr Jenkins rather than the team from
the SSC, this should have been made clear. But it was not made
clear. On the basis of these documents, I believe that I bear
responsibility for this. I sincerely regret that it happened.
39
WITN10500100
WITN10500100
(2) I see from email exchanges at POL00366967, POL00366968,
POL00133074 that, shortly after Mr Parker’s second witness statement
was served (“Parker 2”, at POL00000687), Fujitsu informed WBD that the
statement contained an error, in a footnote to paragraph 35 dealing with
Giro payments. Jonny Gribben raised this with counsel, asking whether a
short letter should be written correcting the relevant footnote. Simon
Henderson and I both expressed concern. I was particularly concerned
because it appeared me that this error undermined the point that Mr
Parker was making in paragraph 35. I expressed the view that, if I was
right, the point was “horrifying”. However, I did not know whether I was
right, and I advised that we needed to know the true position. I also said
that this should be escalated to the highest level of Post Office and Fujitsu.
(3) I see from the email exchanges at POL00364052 and POL00364056 that,
in late February 2019, Jonny Gribben sent me a draft of Mr Parker's third
statement for review (“Parker 3”). I responded with an email setting out my
comments on that draft. These included a comment on paragraph 20,
which explained the true position in relation to Giro payments. I said that
the paragraph was confusing and “frankly evasive”. I explained that we
had discussed this before and that, if my understanding of the position
was correct, paragraph 20 should clearly state that the SSC could use
remote access powers to make payments in normal bank accounts.
(4) Inhis reply, Jonny Gribben simply said that the footnote to para 35 was
incorrect and paragraph 20 explained why. He added that Andy Parsons
was going to speak to me about this.
40
113.
114.
115.
WITN10500100
WITN10500100
(5) 1 do not have any recollection of this incident or of what Andy Parsons
would have said when he spoke to me. But I see that, in Parker 3
(POLO0000689), paragraph 20 was not altered.
In these emails, I discern exasperation on my part. This is consistent with my
recollection. At this time, we (counsel) had a great many things to do. I would
not say that fire-fighting issues of this kind was a distraction, since the
preparation of witness evidence is important. But the trial was looming and, in
order to get ready for it, we needed to master a vast number of
contemporaneous documents and substantial expert evidence, to draft our
opening submissions and to prepare for cross examination. Particularly in
January and February 2019, the need to deal with numerous issues arising in
the preparation of yet more witness evidence took me away from these tasks.
The fact that witnesses needed to correct statements that they had previously
made was particularly troubling. I see from an internal Fujitsu email that the day
before Mr Parker gave evidence, Simon Henderson, I and WBD met with him
with a view to checking that that he was comfortable with everything that was
said in his three statements (FUJ00205178). I do not recall this meeting, but it is
worth noting that it is not my practice to meet witnesses for reasons of this sort.
It would have reflected a concern on my part that further points might need to
be corrected and a determination to ensure that, if there were such points, they
could be corrected by Mr Parker when giving evidence in chief. I see that some
points did need to be corrected, and I imagine that these are the points set out
in the corrections document at POL00000698.
For completeness, I note that Simon Henderson and I attended a site visit at
Fujitsu’s offices in Bracknell on 4 February 2019. I believe that the purpose of
the visit was for Fujitsu to provide us with a technical briefing on the Horizon
41
WITN10500100
WITN10500100
documents and how to use them. We may also have used the opportunity to
discuss questions arising on the Coyne 2 expert report that had just been
served. I do not have a clear recollection of the meeting, but I am sure that it
was not to discuss the form or content of any evidence to be given by any
Fujitsu witnesses.
HORIZON ISSUES EXPERT EVIDENCE
116.
117.
118.
119.
As for my role in the preparation of POL’s expert evidence, I believe that I first
met Dr Worden on 27 April 2018. This was at an early stage of his
investigations, and at this stage the primary focus was on procedural matters
such as requests for further information, protocols governing how the experts
should interact with each other and so on.
On 13 June 2018 (WITN10500106 to WITN10500108), Andy Parsons sent
Simon Henderson and me two documents produced by Dr Worden: a
“Foundations Report” containing the basic elements which he proposed to build
out into a full report and also a document entitled “Quantitative Approach to
Horizon Bugs”, in which he explained some ideas he had about how to arrive at
estimates, for each error identified in Horizon, of the possible net impact on the
claimants’ branch accounts and, for all possible errors in Horizon, of their
largest aggregate impact on those branch accounts.
The Quantitative Approach document consisted of a proposal as to how to
estimate these numbers from the sort of evidence that was available. Dr
Worden indicated that he would appreciate feedback on this proposal.
We had a meeting to discuss the Foundation Report and the Quantitative
approach document. I do not have a clear recollection of this meeting, but we
would have had questions to ask and comments to make about various aspects
120.
121.
WITN10500100
WITN10500100
of these documents and about Horizon. I recall being intrigued by Dr Worden’s
proposed Quantitative Approach, but also a little anxious. Dr Worden was
enthusiastic about the idea of approaching the case as engineer would, using
statistical techniques. My anxiety was based on the fact that this approach was
not what the parties or the Judge had in mind when giving directions for expert
evidence. What they had in mind was a qualitative analysis of the sort indicated
in the Foundation Report, in which explanations and analyses were given of the
functionality, use and architecture of the Horizon system, the checks built into
the system, and so on. I was doubtful whether the Judge would pay any regard
to Dr Worden’s proposed quantitative approach. But on the basis that the
qualitative analysis remained the primary focus of Dr Worden’s work, I had no
objection to his including the quantitative analysis as a helpful back up to his
qualitative analysis.
The extent to which and way in which Dr Worden should rely on the quantitative
approach became a matter of debate between counsel and Dr Worden. As I
recall, he was enthusiastic about his approach, and seemed to think that it
should be put at the forefront of his analysis because it provided a direct way of
demonstrating that the claimants’ claims about the losses they had suffered
through Horizon could not be right. On the other hand, we (counsel) took the
view that this was not one of the questions for which permission to give expert
had been given. We thought that, if it was to be included, it should be included
as a back-up, in the way I describe above.
On 15 July 2018 (WITN10500109 and WITN10500110), Andy Parsons sent us
a summary which Dr Worden had prepared of the opinions that he had by then
been able to form. During July and August, we had two meetings with Dr
Worden in which we discussed how his views were developing and our queries
43
122.
WITN10500100
WITN10500100
and also various procedural questions arising such as information requests and
whether the Horizon issues should be amended.
My recollection is that the reliance which Dr Worden wanted to place on his
quantitative analysis was a running issue between us, at these and subsequent
meetings. Our discussions on this issue are alluded to in a document which has
been provided to me (POL0000647 71). I do not recall seeing this document
before, but I see that it is described as a “Noting Paper” to POL’s Steering
Group Meeting and that it is dated 28 November 2018 (9 days before Worden 1
was ultimately served). As to this paper:
(1) Paragraph 1.4 refers to “lengthy debates” between Dr Worden and the
legal team about how best to communicate his central conclusion to the
Judge. The only lengthy debate I recall concerned the nature and extent of
Dr Worden’s reliance on his quantitative analysis. I see from Section 3 of
the paper that this was the debate that the paper was alluding to.
(2) Section 3 refers to numerous calls and conferences and to over 20 hours
of debate We did have a number of meetings with Dr Worden and we
would probably have had some calls also: as his work progressed, Dr
Worden’s report became quite long and complicated. There was a lot of
detail for us to digest and discuss.
(3) Section 5 reminds me that we had other anxieties in relation to Dr
Worden's quantitative approach. However, Dr Worden stuck to his guns.
He is a strong-minded individual who has faith in his own judgment and is
determined to take his own way. This is a commendable attitude in an
expert, although it caused POL’s legal team some anguish when, in the
second half of the Horizon trial, he insisted on producing to the Court a
third report (“Worden 3”). He did this in the knowledge that I and the rest
44
123.
124.
125.
126.
WITN10500100
WITN10500100
of POL’s legal team would have preferred him not to do this and in spite of
the fact that, on our advice, POL did not seek the Judge’s permission to
rely on the report.
In the event, our anxieties about Dr Worden’s reliance on the quantitative
approach were vindicated: as I recall, Mr Coyne refrained from engaging with
the quantitative analysis and, in the Horizon Issues judgment, the Judge did not
accept the relevant parts of Dr Worden’s reports or find them helpful.
I see from POL00142397 that, on 7 September 2018, Andy Parsons sent us an
outline of the report that he intended to produce and that, on 12 September
2018, I responded in an email which set out my and Simon Henderson’s
combined thoughts on this outline. I note the comments made about the
quantitative approach in paragraphs 5.e. and 5.f. of the email.
I also note that paragraph 8 of the email asks how and where remote access
would be addressed. My recollection is that Dr Worden thought that the
claimants’ case on remote access was unreal. I do not wish to put words into
his mouth but my perception of his view was that, in the real world, there was no
possibility of Fujitsu privileged users engaging in a scheme to evade Fujitsu's
controls so as to create false shortfalls in SPM accounts. I sympathised with
that view, but my recollection is that I would have preferred him to have
provided a fuller analysis of the theoretical possibility of such rights being
abused in that way, of the controls guarding against such abuse, and of the
records that would be created if it had been done.
We had a further meeting with Dr Worden on 21 September 2016 and I see
that, on 20 September 2018, Lucy Bremner sent us an agenda for that meeting
(WITN10500111). The email included a document containing Dr Worden’s
responses to our comments (WITN10500112). These were helpful in enabling
45
WITN10500100
WITN10500100
us to understand his approach and to ensure that he would cover all the points
that we thought relevant, if and to the extent that he also thought them relevant.
127. Thereafter, Dr Worden would have been busy drafting his report. My diary
suggests that we had two further meetings with him before he finalised Worden
1 in early December 2018. In the meantime, he provided drafts of sections of
his report for Simon Henderson and me to comment on. Andy Parsons’ Noting
paper at POL00006471 reminds me that he provided us with sections 6 and 7 of
his report first.
128. Turning to Worden 2, this was served on 1 February 2019. I have had not have
time to skim through my many emails during this period with a view to reminding
myself of the role that Simon Henderson and I played in relation to that report.
We would have commented on drafts provided by Dr Worden, and may have
met with him, either in person or on the telephone. My diary suggests that we
had a meeting with him on 19 December 2018.
129. My diary also suggests that we (counsel) had the following meetings with Dr
Worden and WBD during 2019: 7 February 2019, 11 February 2019, 28 March
2019, 4 April 2019, 16 April 2019 and 24 April 2019. As to these meetings:
(1) The February meetings took place after Worden 2 had been served but
before the trial started. I suspect that the meetings would have been to
help us understand the significance of and deal with Coyne 2, which had
been served on 1 February. But there may have been other issues to
discuss, such as a request for further information by Mr Coyne.
(2) I suspect that the primary purpose of the March and April meetings was to
help me prepare for my cross examination of Mr Coyne. There would also
have been discussions about Dr Worden’s proposed Worden 3 (see
paragraph 122(3) above).
46
WITN10500100
WITN10500100
In addition to these meetings, my diary indicates that, on 1 May 2019, I hada
“call with HSF and Rodric re discuss Worden @ HSF Offices”. I have no
recollection of that call.
HORIZON ISSUES DISCLOSURE
130.
131.
132.
Regarding my role in advising and representing POL in respect of Horizon
Issues disclosure, I had to respond to numerous problems that blew up in
relation to this disclosure. These included more than one occasion when, on the
basis of instructions which I had been given, I made factual submissions to the
Court which I subsequently discovered to be false. Fraser J's Horizon Issues
judgment has a section dealing with disclosure. I have looked at that section
with a view to refreshing my memory about these problems. I refer to Section F
of his judgment (paragraph 559 to 653) for a summary of the underlying facts.
As is usually the case in litigation of this sort, the disclosure process in the GLO
Proceedings was managed on POL’s behalf by WBD. This would have entailed
a substantial level of cooperation, not only between WBD and POL but also
between WBD, POL and Fujitsu. Counsel was not involved in these
arrangements. However, we did become involved once problems had arisen. In
these situations, we would help WBD draft letters to Freeths in which problems
would be revealed and solutions proposed. This would sometimes include
making suggestions as to the sort of solutions to propose. In providing help of
this sort, we proceeded on the basis of the facts as we understood them. Our
understanding would have been based on our instructions
Request 22.2 asks in particular about the disclosure of the PinICL, Peak and
KEL databases in particular. I discuss the disclosure of KELs and Peaks below.
I have no recollection specifically in relation to PinICLs.
47
WITN10500100
WITN10500100
KELs
133. The story of POL’s disclosure of KELs is summarised in paragraphs 573 to 614
of the judgment. It is an extraordinary story.
134. My original instructions were that the KEL database (the “Known Error Log’)
was irrelevant and not within POL’s control. Its irrelevance was asserted more
than once in WBD’s correspondence with Freeths. And in the Generic Defence
and Counterclaim, we (Owain Draper and I!) pleaded both that it was irrelevant
and that it was not in POL’s control. These were my instructions at the first
CMC, which is discussed in paragraphs 585 to 591 of the judgment. I see from
paragraph 586 that Andy Parsons had made a witness statement for the CMC.
which maintained both of these points. That reflected my understanding of the
position at that time.
135. I see from paragraph 591 that, at the CMC, I suggested that the parties’ experts
be permitted to inspect the Known Error Log so that they could determine
whether it was relevant. I do not recall making that suggestion, but I do
remember the outcome: the experts determined that the Known Error Log was
definitely relevant.
136. I remember being very surprised when I learned about this. My previous
instructions would have been based on what Fujitsu had told POL. It was hard
to understand how Fujitsu came to give such a misleading account on such a
fundamental point.
137. Once it was established that the Known Error Log was relevant, arrangements
were made to ensure that it was disclosed to the claimants.
138. From that point, I do not recall any objection to disclosure being suggested on
the basis that it was not within POL’s control. I do not remember what was said
about control around this time, either within POL’s legal team or to Freeths. But
48
WITN10500100
WITN10500100
POL and its legal team clearly took the view that, if it was relevant to the
Horizon Issues, it would be wrong to resist its disclosure. Fujitsu did not object.
139. I see from the judgment that POL’s previous claim that the Known Error Log
was not in POL’s control were raised at the trial and that, after the trial, the
Judge invited the parties to lodge further written submissions on whether, when
POL initially took that point, it had been entitled to do so. By that stage, the
governing contract between Fujitsu was available to the parties and the Court.
This provided for POL to have quite wide rights of inspection of Fujitsu
documents. However, we (counsel) took the view that there was a respectable
argument that those rights did not extend to the sort of trivial and irrelevant
document that POL had believed the Known Error Log to be at the relevant time
(i.e. during the period leading up to and including the first CMC). The Judge was
not impressed by that argument. But in any event, this was long after the event.
As I explain above, at my suggestion, the Known Error Log had been inspected
by the experts in the previous year and, as a result, many thousands of KELs
had been disclosed.
140. I see from paragraphs 612 of the judgment that further KELs were disclosed in
respect of specific bugs in November 2018, and that further KELs were
disclosed during 2019. Subject to what I say below, I have no recollection as to
why the November KELs were disclosed then, or as to why further KELs were
disclosed during the trial (although I see a reference to deleted (meaning
archived) KELs being disclosed in January 2019).
141. When documents were disclosed late, it was usually because POL/WBD only
became aware of their existence late. When this happened, WBD would inform
Freeths of their discovery in correspondence and arrangements would be
agreed for their disclosure. Counsel was generally asked to comment on the
49
142.
143.
144.
145.
WITN10500100
WITN10500100
drafting of the relevant correspondence. In relation to matters of this sort, our
role was largely reactive.
After the trial, it became necessary to give disclosure of further documents that
had been discovered. Some 5,000 KELs were disclosed after the trial. The
circumstances surrounding this disclosure is described in paragraphs 627 to
630 of the judgment. It is another extraordinary story.
As can be seen from those paragraphs, these were previous versions of KELs
which had been superseded by later versions. The later versions had been
disclosed but Fujitsu had informed POL/WBD that it did not retain the earlier
iterations. Only after the trial did POL/WBD discover from Fujitsu that it did
retain the earlier versions after all.
This was obviously a serious matter. Prompt steps were taken by POL to inform
the claimants about what had happened, to inform the Judge, and to ensure
that earlier iterations of relevant KELs were disclosed as quickly as possible.
POL Counsel would have advised on the steps to be taken and the letter that
was sent to Freeths and the email that was sent to the Judge. I see that, in
paragraph 631 of the judgment, the Judge explains that POL essentially left it to
the claimants to decide whether there should be any further submissions or
evidence. I also see that he described this approach as pragmatic and sensible.
I see from POL00043141 that, soon after it discovered the existence of these
additional KELs, POL was considering whether to commission Deloitte to audit
the completeness and accuracy of all the documents which had been provided
by Fujitsu. Simon Henderson and I were asked to advise on the merits and risks
of doing so and on the question whether litigation privilege would apply to
Deloitte’s audit work. In a note, we indicated what we saw as the benefits and
30
146.
147.
148.
149.
WITN10500100
WITN10500100
the risks of undertaking any audit. We also advised that we saw the question
whether the audit would be privileged as being of little practical importance.
Peaks
The judgment discuss Peaks in paragraphs 615 to 626. I see from paragraphs
615 and 616 that the claimants requested these documents in July 2018, that
218,0000 Peaks were disclosed on 27 September 2018, that a further 3,866
were disclosed on 25 October 2018 and that other Peaks were disclosed in May
2019.
I do not recall much about the Peaks that were disclosed late. There were many
occasions when additional Fujitsu documents were found after the relevant
documents should have been disclosed. To say that these problems were
frustrating would be an understatement, but this was not something that was
within counsel's control. As I explain above, we became involved after the
documents were found.
I see from paragraph 616 that, when the May 2019 Peaks were disclosed, the
Judge was told in submissions that Fujitsu had discovered an old database that
had been copied more than ten years previously, and that POL had speedily
provided the contents of this database to the claimants. I would have made
those submissions. They would have been based on instructions from WBD. I
would have believed my instructions.
Paragraph 617 indicates that, soon after these Peaks were disclosed, it was
revealed that one of them was dated 21 August 2019. It was thus clear that, on
instructions, I had unintentionally misled the court. As will be clear from the rest
this statement, this was not the first occasion on which such a thing happened,
and nor was it the last. It is a horrifying experience.
31
150.
151.
152.
153.
154.
155.
WITN10500100
WITN10500100
Subsequently, Andy Parsons made a witness statement explaining what had
happened. In the time available, I have not had an opportunity to read that
statement.
Other Documents
Paragraphs 616 and 617 of the judgment discusses the disclosure of certain
OCPs in January 2019 and certain OCPs in April 2019. I have no recollection of
the disclosure of these documents.
Paragraph 622 of the judgment criticises POL’s approach to Release Notes. I
do not have a recollection as to what (if any) directions were given in relation to
Release Notes, or of what documents were available to POL or of what was
said or done in this regard.
Two matters that are within my recollection are the matter of redactions and the
disclosure of Royal Mail audit documents. These are dealt with in paragraph
565 of the judgment.
When conducting disclosure WBD applied redactions to documents for reasons
such as privilege. At the trial, it appeared that similar documents had been
redacted in different ways. When one compared documents which had been
redacted more heavily with those which had been redacted less heavily, it
seemed that some redactions were not justified. My recollection is that, when
this issue became apparent, junior counsel reviewed the relevant documents to
ensure that they were properly redacted. After that, I was asked to perform a
review, and I duly did so, also providing a short note which explained to the
judge and to the claimants the approach that I had adopted.
During the trial, a question arose as to POL’s failure to disclose some Royal
Mail audit reports that the claimants wanted. On instructions, I informed the
court that the Royal Mail had refused to produce these reports to POL for
52
WITN10500100
WITN10500100
onward disclosure to the claimants. I then discovered that the Royal Mail had
not even been asked to provide them. When I discovered this, I made haste to
correct the position and to apologise. Not surprisingly, POL was ordered to
provide a witness statement explaining how this had come about.
POL’S DECISION MAKERS
156. As I explain in paragraph 24 above, during the GLO Proceedings, it was my
impression that the importance of this claim meant that decisions on significant
issues were considered by senior management, rather than just by POL’s in-
house lawyers. As the Horizon trial approached, I attended some conferences
with members of the board. I discuss these below.
CONFERENCES WITH POL’S LEGAL DEPARTMENT OR THE BOARD ON THE
HORIZON ISSUES
157. In Request 31, I am asked to describe any conferences or significant
discussions I had with POL’s legal department or the board or POL’s decision
makers on the issues set out in Request 22. According to my diary, I had the
following conferences on the following dates:
(1) On7 February 2019, Owain Draper, Simon Henderson and I attended a
conference in my chambers with several WBD people, Jane MacLeod,
Rod Williams, Angela Ven den Bogerd and Mark Underwood. I do not
recall this conference, but presume that we would have been discussing
things such as how we saw the merits now that supplemental expert
reports had been exchanged, how our preparations for trial were going,
and what further steps could be taken in order to be ready for the
forthcoming PTR on 14 February 2019.
33
WITN10500100
WITN10500100
(2) On 21 February 2019, I had a conference with the “Board GLO Sub
Committee”. This is the meeting for which I have been provided with a
minute (POL00006753). My diary states that it was a call but the minute
suggests that I was physically present and this is consistent with my
recollection. I do not have a detailed recollection of the meeting but such
recollection as I do have is consistent with the minute. I provided a high
level briefing. It was my view that, on the question of Horizon’s reliability,
POL had good arguments and, although some aspects of the system
would probably be criticised, I still felt cautiously optimistic. As I recall, the
committee members were concerned about the impact of an adverse
judgment on its ability to operate its business and were anxious to do what
they could to mitigate the risks to the business if such a judgment was
given.
158. My diary does not identify any further meetings with POL but I recall attending a
meeting with the same committee after the trial had finished and I see from
POL00006752 that this meeting took place on 20 June 2019. My recollection is
that I provided another high level briefing: I explained that POL’s witnesses had
not performed well but that, in my view, Mr Coyne had accepted important parts
of POL’s case in cross examination. My belief was that the evidence showed
that Horizon was a robust system, but I was concerned about the view that the
Judge appeared to take. My briefing was short. Indeed, I recall feeling that it
was cut short by the chair before I had finished discussing my impression of the
Judge’s attitude.
34
WITN10500100
WITN10500100
THE ADEQUACY OF MY INSTRUCTIONS
159.
160.
161.
By this time, I had a negative view of the adequacy of my instructions. I
summarise the reasons for this in the following paragraphs.
First, as I explain in paragraphs 93 to 114 above, I found the process by which
the witness statements were produced unsatisfactory. Quite apart from the
problems I discuss in those paragraphs, the draft witness statements with which
counsel were provided often failed to grapple with the issues properly, or even
to do so in a way which was clear. WBD and counsel were both acting under
considerable time pressure, and we did our best to provide WBD with
comments on the drafts produced as quickly as possible. However, our
comments were sometimes not addressed in the way one might have hoped. I
remember being particularly irritated by the response to my desire to ensure
that, where the witnesses relied on information provided by other persons, the
sources of that information were properly identified. At times, it seemed almost
as if WBD had forgotten the need to identify such sources. And several of the
statements that were ultimately served contained evidence that was difficult to
reconcile with other evidence, appeared contradictory and/or begged significant
questions. Examples of such problems are given in my emails to WBD of 18
November 2018 (POL00363851) and 17 March 2019 (POL00268606).
Furthermore, my recollection is that, when several of POL’s witnesses gave
evidence (including Mr Godeseth), in cross examination they quickly accepted
propositions which were inconsistent with the witness statements which they
had signed. In view of the time that has elapsed since the trial, I cannot now
give specific examples. But I have a strong memory of being astonished by
some of the things that they said. I found it hard to understand how witness
35
162.
163.
164.
WITN10500100
WITN10500100
statements could have been prepared which appeared not to reflect the
evidence they gave at trial.
It should be recognised that POL’s evidence was drafted under a tight timetable
and what was being sought was clear evidence on a large number of quite
narrow and in some cases very technical questions whose full significance may
not always have been appreciated. This made the task of producing witness
statements very difficult. With the benefit of hindsight, I think that the WBT team
dealing with witness statements would have benefitted from having some more
experienced people on the team. But for me, the experiences I describe above
were unprecedented.
So were my experiences in relation to disclosure. Again, the task of undertaking
disclosure was a very difficult one. It covered a wide array of documents over a
long period of time and I believe that, over that time, the relevant documents
had been stored in different places and different ways, which made them hard
to find. These difficulties were compounded by the fact that large numbers of
important documents were held by Fujitsu. POL was dependent on Fujitsu to
know what documents were relevant, whether they existed and where they
could be found and, as the case progressed, it appeared that Fujitsu had let
POL down in all these categories.
In these circumstances, my suspicion is that POL would have found it difficult to
avoid many of the problems with disclosure that are identified in the Horizon
Issues judgment. But that is by no means true of all of them.
MATTERS WHICH POL APPEARED TO CONSIDER IMPORTANT
165.
In Request 33, I am asked to set out my views on what matters POL’s board
and its legal department appeared to consider important when determining its
36
166.
167.
168.
WITN10500100
WITN10500100
strategy, and to describe any divergence of views between legal
representatives and/or member of the board.
To the best of my recollection, a matter which was of concern to POL in the
early days of my involvement was the fact that it had made a number of false
statements about the impossibility of remote access. As I was instructed, the
relevant individuals at POL had made those statements on the basis of
information provided by Fujitsu, and they only discovered the true position later.
This had been discussed at my initial meeting with POL in May 2016. My advice
was that POL should be open about the fact that, despite its previous
statements to the contrary, remote access by Fujitsu was possible. WBD
agreed with this approach and, as far as I recall, so did Jane MacLeod. But as I
explain above, when it came to approve the letter of response, it appeared that
POL's senior management was reluctant to be as overt on this point as its
lawyers suggested.
I do not believe that this divergence in views between POL and its lawyers was
one of substance or that it had any impact on POL’s litigation strategy. The
point was more fully ventilated in subsequent correspondence. Moreover, the
GLO Proceedings proceeded on the basis that remote access was possible and
that POL’s previous statements to the contrary were wrong.
I do not recall other occasions in which POL failed or refused to take the advice
of its legal representatives. In saying this, I am not suggesting that there were
no such occasions over the whole period of my involvement in the GLO
Proceedings. It would not surprise me if there were some occasions, even if
only on small points. But on the limited number of documents that I have been
able to review in the time available, I cannot bring any such occasions to mind.
37
WITN10500100
WITN10500100
169. As to other matters which POL appeared to consider important, I recall a
general concern that, if the claimants’ case on the true effect of the SPM
contracts and on the reliability of Horizon was correct, this could make it very
difficult for POL to operate its business, to deal effectively with any shortfalls
arising in any SPM accounts and to exercise its rights of termination and
suspension when it felt that it needed to do so. From my perspective, this
concern was not surprising. However, I am not sure what impact this concern
had on POL’s strategy in the GLO Proceedings. As far as I could tell, POL
believed in its case as to the true effect of the SPM contracts, it believed that
Horizon was reliable and it believed that, although remote access was possible,
there was no realistic possibility of Fujitsu's privileged users having manipulated
it so as to create false shortfalls in SPM accounts.
MY REFLECTIONS
170. In Request 34, I am asked to set out in detail my reflections regarding the
171.
advice I gave to POL and my involvement in this matter. And I am also asked
whether, with the benefit of hindsight, I would have done anything differently.
Regarding the first question, I advised POL on innumerable points during the
course of the GOL Proceedings and, given the limited time I have had to
prepare this statement and the limited documents I have been able to review in
this time, it is not possible for me to review all the advice I gave or the
circumstances in which it was given.
172. At the times I gave any advice, I would have believed it to be the right advice.
For example, as set out in the joint opinions at POL00103462 and
POL00022669, the advice that I and my colleagues gave as to the proper
meaning and effect of the SPM contracts represented my considered view. And
38
173.
174.
WITN10500100
WITN10500100
at the end of the Horizon trial, I believed that the evidence showed that the
Horizon system was robust and that there was no realistic possibility of Fujitsu
privileged users abusing their access rights so as to create false shortfalls in
SPM accounts: the counsel team drafted lengthy closing submissions on these
points and I believed that those submissions were justified by the evidence. In
his Common Issues and the Horizon Issues judgments, the Judge took a very
different view. But that was not my view.
Turning to the second question, it is important to be clear about what hindsight
is being referred to. I now know the outcome of the Common Issues trial and of
the Horizon trial. Had I had the benefit of this knowledge when I was first
instructed, I would have advised POL to settle the claims on a generous basis
and to provide a full apology. However, I doubt that is what the question is
aimed at.
I have the following reflections about the advice I gave and my involvement
more generally.
(1) First, in relation to the early advice I gave on whether POL should continue
implementing the recommendations of the Swift review:
(a) I remain of the view expressed in my emails to Andy Parsons of 8
June 2016 (POL00242402) that, as a defendant in substantial
litigation, POL should protect its interests by ensuring that any
investigations which were relevant to the litigation were undertaken
pursuant to the litigation and under the protection of litigation
privilege.
(b) However, I see that, in his letter to POL of 21 June 2016
(POL00006601), Andy Parsons indicates that my advice was that Mr
Swift's recommendations numbered 4, 5, 6 and 8 should be
39
(2)
(3)
WITN10500100
WITN10500100
implemented as part of the litigation. It might be thought that my
advice was that the recommendations numbered 1, 2, 3 and 7 should
not be implemented. As I have said, I do not recall giving any advice
on this subject. But on the material now available to me, if I took the
view that the recommended advice should not be obtained
(recommendation 1), and the recommended reviews and analyses
should not be undertaken (recommendations 2, 3 and 7), it is not
obvious to me why I would have done so. However, it may be that
some or all of this work had already been done (as Andy Parker's
letter suggested on the first page), or it may be that I took the view
that they should be done at a later stage, or it may be that the letter
does not give a complete account of my advice.
Second, if I had known how unreliable Fujitsu’s information would be
about the Known Error Log and about the retention of previous iterations
of KELs, and if I had known how unreliable Fujitsu's efforts would be in
locating and providing relevant documents in a timely fashion, at the
earliest possible stage I would have advised POL to instruct a firm such as
Deloitte to undertake a detailed review of the documents held by Fujitsu
which were relevant to the litigation, of their significance and of the
locations at which and form in which they were held. Had that been done,
some of the problems I describe in paragraphs 133 to 152 above might
have been avoided. But these problems did not become apparent to me
until much later.
Third, although Deloitte were instructed to investigate various matters,
including remote access, they were not instructed to undertake a full scale,
in depth review of Fujitsu’s relevant documents with a view to ascertaining
60
(4)
(5)
WITN10500100
WITN10500100
what relevant bugs had arisen in Horizon during its life and how those
bugs were dealt with. Had that been done, POL would not have had to rely
on Fujitsu's assurances in relation to these matters (as I think it did) in
pleading its case in the Generic Defence and Counterclaim. Moreover, as
the Horizon Issues trial approached, it would have been in a much better
position to determine what factual witnesses should be called and what
points they could address. However, this would have been an extremely
expensive undertaking and, as I recall, Freeths were already complaining
that POL was using a “shadow expert”. And in any event, on the
information that I had in 2016, I was not in a position to justify advice that it
should instruct Deloitte to do a job of this magnitude, which would
duplicate the work that would be done by POL’s litigation expert when he
was appointed.
Fourth, in relation to the advice I gave on the meaning and effect of POL’s
SPM contracts, a significant part of my analysis was based on the view
that the SPM contracts represented the true agreement between the
parties and that they were business to business contracts and should be
interpreted accordingly. I have reflected on the question whether I should
have taken a different view on these issues. However, it was how I saw
the matter. I was very surprised when I discovered that, in his Common
Issues judgment, the Judge had found that the contracts did not represent
the true agreement and that it was necessary to imply numerous terms,
even including a term constraining the way in which POL could exercise its
right to give notice of termination.
Fifth, had I anticipated the problems that I describe in relation to POL’s
witness statements, I would have advised that the team of people which
61
(7)
(8)
WITN10500100
WITN10500100
was responsible for identifying the witnesses to interview and on what
points, and for undertaking those interviews and drafting the relevant
statements, should include some more senior solicitors and possibly one
or more experienced junior barristers as well. In relation to the first round
of evidence, I would also have advised that, if possible, POL’s witnesses
be identified and interviewed and draft statements produced much earlier
than they were.
Sixth, as I explain above, I think that I should have ensured that Parker 1
indicated that Mr Jenkins had contributed in some way to the information
contained in the table that was exhibited. And there may have been other
points in relation to which Mr Parker’s statements were not sufficiently
clear about the sources of his information. I regret that I did not take a
stronger stand on this.
Seventh, this brings me to Mr Jenkins himself. Given what I had been told
about his reliability, not relying on him as a witness seemed the obvious
thing to do. However, as matters transpired, it turned out that Mr Godeseth
was not as well placed as I had thought to give relevant evidence. Had I
known that sooner, I would have advised that another witness or
witnesses be called instead. As I recall, I was informed that there was a
distinct lack of witnesses at Fujitsu who were willing and able to speak to
the relevant issues, but by the time the problem became apparent, witness
statements were about to be served and it would have been very difficult
to change tack.
I see that, in an email to Andy Parsons on 18 November 2018
(POL00363851), I wondered in the light of this problem whether the
WITN10500100
WITN10500100
decision not to call Mr Jenkins should be reviewed. I find myself wondering
the same thing now.
ANYTHING FURTHER
175. In Request 35, I am asked whether there is anything further of which I think the
Chair ought to be aware. In this witness statement, I have done my best to
answer the many questions I have been asked in the time available to me. In
relation to the matters covered by those questions, I cannot think of anything to
add of which the Chair ought to be aware. But I wish to take the opportunity to
say how much I sympathise with those SPMs who have suffered as a result of
flaws in Horizon. During the period in which that I acted for POL, my instructions
were that the Horizon system was reliable. As I saw it, POL’s case was
advanced on substantial grounds. It was supported by an independent expert. It
was my duty to present that case to the best of my ability, and that is what I
sought to do.
I believe the content of this statement to be true.
14 May 2024
63
Index to First Witness Statement of Anthony John de Garr Robinson
WITN10500100
WITN10500100
URN
Document description
Date
Control Number
P140216
Email chain from Andrew
Parsons to Anthony de
Garr Robinson Cc Paul
Loraine Tom Porter and
Others RE Bates v POL -
Commentary on
Counsels Bundles [BD-
4A.FID26859284]
01/06/2016
POL-BSFF-
0195824
POL00242402
Email chain from Andrew
Parsons to Anthony de Garr
Robinson and others re
Jonathan Swift
08/06/2016
POL-BSFF-
0080465
POL00006601
Letter to PO from Bond
Dickinson re group litigation
21/06/2016
POL-0017859
POL00110507
Letter from Andrew Parsons
to James Hartley
(Freeths LLP) Re Bates &
Others v Post Office
Limited
28/07/2016
POL-0108242
POL00003340
Letter from Andrew Parsons
to James Hartley, re:
Bates & Others -v- Post
Office Llmited - Generic
Defence and Counterclaim
18/07/2017
VIS00004354
POL00031502
‘Bramble’ — Draft Report
31/10/2016
POL-0028404
WITN10500101
Email from Andrew Parsons
to Anthony de Garr
Robinson, Owain Draper and
CC Amy Prime re. Draft
Executive Summary
19/06/2017
WITN10500101
WITN10500102
Alan Bates & Others v Post
Office Limited Privileged
Users -— Draft Executive
Summary
19/06/2017
WITN10500102
POL00041491
Deloitte - Bramble - Draft
Report.
01/09/2017
POL-0037973
10
POL00028928
Deloitte "Bramble" - Draft
Report
19/01/2018
POL-0025410
11
POL00103462
Alan Bates & Others and
Post Office Limited - Opinion
on the Common Issues
10/05/2018
POL-0103045
12
WITN10500103
Email from Andrew Parsons
to Anthony de Garr Robinson
27/07/2017
WITN10500103
WITN10500100
WITN10500100
ccing Amy Prime re: Remote
Access Wording
13
WITN10500104
Letter from Andrew Parsons
to Mr K Hartley re Bates &
Others -v- Post Office
Limited - Claim Number:
HQ16X01238.
30/11/2016
WITN10500104.
14
POL00337340
Email from Amy Prime to
Anthony de Garr Robinson,
Owain Draper, Jonathan
Gribben and others re
Deloitte Report [BD-
4A.FID26896945
14/11/2016
POL-BSFF-
0163551
15
POL00024971
Meeting of 14/11/2016 -
Agenda - Group Litigation
Strategy
14/11/2016
POL-0021450
16
POL00364016
Email from Jonathan
Gribben to Anthony de Garr
Robinson, Andrew Parsons
and others re: draft third
Witness Statement — of
Stephen Parker
22/02/2019
POL-BSFF-
0195824
17
POL00022669
Post Office Group Litigation
between Alan Bates &
Others -and- Post Office
Limited - Update to the
opinion on the common
issues
28/09/2018
POL-0019148
18
POL00006754
Meeting Minutes of the
Postmaster Litigation
Subcommittee of POL
15/05/2018
POL-0018012
19
FUJ00155196
Alan Bates and others v Post
Office Limited
Coyne 2 Report- Remote
Access Paragraphs
3.221- 3.287
21/05/2019
POINQ0161391F
20
POL00140306
Email from Johnathan
Gribben to Anthony de
Garr Robinson Ce Andrew
Parsons & Others RE
Remote Access queries
06/03/2019
POL-0141542
21
FUJ00155195
Email from Katie Simmonds
to Anthony de Garr Robinson
and ParkerSP. Re:
Telephone conference with
Leading Counsel
22/05/2019
POINQ0161390F
22
FUJ00165648
Email from Katie Simmonds
to Anthony de Garr Robinson
and SP Parker re: Cancelled:
Telephone conference with
Leading Counsel
22/05/2019
POINQ0171826F
65
WITN10500100
WITN10500100
23
WITN10500105
Email from Andrew Parsons
to Anthony de Garr
Robinson, Simon
Henderson, Gavin Matthews
and others re: Scope of
Horizon [WBDUk-
AC.FID26896945].
07/09/2018
WITN10500105
24
POL00134909
Attendance notes for The
Post Office Group Litigation
04/10/2019
POL-0139362
25
POL00000692
Witness Statement _— of
Stephen Paul Parker
16/11/2018
VIS00001706
26
POL00363816
Email from = Jonathan
Gribben To: Anthony de
Garr Robinson, Simon
Henderson cc Andrew
Parsons and others re Steve
Parker
[WBDUKAC.FID27032497]
15/11/2018
POL-BSFF-
0195624
27
POL00363851
Email from Anthony de Garr
Robinson To:
Andrew Parsons CC: Simon
Henderson and
Jonathan Gribben Re:
Robert conclusions
[WBDUK-AC.FID26896945]
18/11/2018
POL-BSFF-
0195659
28
POL00000682
Witness Statement _—of
Torstein Olav Godeseth
27/09/2018
VISO0001696
29
POL00367005
Email from Anthony de Garr
Robinson to Andrew
Parsons, RE: Data Deletion
23/02/2019
POL-BSFF-
0198826
30
POL00364020
Email from Anthony De Garr
Robinson to Andrew
Parsons, RE: Data Deletion
23/02/2019
POL-BSFF-
0195828
31
POL00000686
Alan Bates & Others v POL -
Third Witness
Statement of Torstein Olav
Godeseth for The
Post Office Group Litigation
28/02/2019
VIS00001700
32
POL00363955
Email from Andrew de Garr
Robinson to Andrew
Parsons, Jonathan Gribben
and others re: Further
comments from FJ - TG
paragraphs requiring
correction or consideration
attached
18/02/2019
POL-BSFF-
0195763
33
POL00363956
Torstein Olav Godeseth
Witness Statement 1 -
paragraphs requiring
correction
18/02/2019
POL-BSFF-
0195764
34
POL00364056
Email from Jonathan
Gribben to Anthony de Garr
28/02/2019
POL-BSFF-
0195864
66
WITN10500100
WITN10500100
Robinson: Godeseth 3 and
Parker 3
35
POL00000689
Alan Bates and Others and
Post Office Limited, Third
Witness Statement of
Stephen Paul Parker
28/02/2019
VIS00001703
36
POL00363775
Email from Jonathan
Gribben to Anthony De Garr,
RE: Catherine Hamilton
14/11/2018
POL-BSFF-
0195583
37
POL00366967
Email from Simon
Henderson to Jonathan
Gribben, RE: Parker 2
30/01/2019
POL-BSFF-
0198788
38
POL00366968
Email from Anthony De Garr
Robinson to Simon
Henderson, RE: Parker 2
30/01/2019
POL-BSFF-
0198789
39
POL00133074
Email chain from Jonathan
Gribben to Anthony de Garr
Robinson and cc'd Simon
Henderson re:
Parker 2 [WBDUK-
AC.FID27032497]
30/01/2019
POL-0136446
20
POL00000687
Second witness statement
of Stephen Paul Parker
29/01/2019
VIS00001701
41
POL00364052
Email from Jonathan
Gribben to Anthony de Garr
Robinson: Parker 3
27/02/2019
POL-BSFF-
0195860
42
FUJ00205178
Email from Beth Durkin to
Rachel Roberts cc
David Barker RE: Tentative:
Call - Project
Bramble - legally privileged
and confidential
[PM-AC.FID3854255]
10/04/2019
POINQ0210899F
rr)
POL00000698
Alan Bates and Others and
Post Office Limited,
Corrections to the Third
Witness Statement of
Stephen Paul Parker
01/01/2019
VIS00001712
44
WITN10500106
Email from Andrew Parsons
to Anthony De Garr
Robinson, Simon Henderson
and others Re: Expert Report
13/06/2019
WITN10500106
45
WITN10500107
Note on Quantitative
Approaches to Horizon Bugs
19/09/2018
WITN10500107
46
WITN10500108
Foundation Report - Alan
Bates and others v Post
Office Limited Dr Robert
Worden and Chris Emery
31/05/2018
WITN10500108
47
WITN10500109
Email from Andrew Parsons
to Anthony de Garr
Robinson, Simon Henderson
ccing Jonathan Gribben and
15/07/2018
WITN10500109
or
WITN10500100
WITN10500100
others re: Expert Requests
for Information
48
WITN10500110
Report on general issues
and summary of opinions.
01/01/1900
WITN10500110
49
POL00006471
Steering Group Noting Paper
- Expert Report of Dr Robert
Worden
28/11/2018
POL-0017776
50
POL00142397
Email from Anthony De Garr
Robinson to Andrew
Parsons: Structure for expert
report
12/09/2018
POL-0141681
51
WITN10500111
Email from Lucy Bremner to
Anthony de Garr Robinson,
Simon Henderson, Andrew
Parsons and others re:
Agenda for call 21 Sept 10-
11am [WBDUK-
AC.FID27032497].
20/09/2018
WITN10500111
52
WITN10500112
Document headed
‘Counsels Comments on
Report Plan’
01/01/1900
WITN105001 12
53
POL00043141
Note on Potential Audit of
Fujitsu Disclosure
17/10/2019
POL-0039623
54
POL00006753
Meeting Minutes of the
Group Litigation
Subcommittee of POL
21/02/2019
POL-0018011
55
POL00006752
Draft Meeting Minutes of the
Postmaster Litigation
Subcommittee of POL
20/06/2019
POL-0018010
56
POL00268606
Email from Angela Van-Den-
Bogerd to Andrew Parsons
cc Jonathan Gribben
18/03/2019
POL-BSFF-
0106669