LEGALLY PRIVILEGED AND CONFIDENTIAL
POST OFFICE LTD
GENERAL REVIEW
A. INTRODUCTION
1. The Post Office Ltd (“POL”) has commissioned me to review past practice
and make recommendations as to the future approach to the conduct of
prosecutions.
Z. Terms of Reference received from Bond Dickinson LLP (“BD”), solicitors
advising POL, invited from me the following:
A. Written Reports
1 To prepare by 5 August 2013 an interim review of Cartwright King's
current review process (as reflected principally in CK's document entitled
"Draft Paragraphs for Insertion into Reply by CCRC" and to a lesser
extent in Simon Clarke's Advice of 15 July 2013) (to be a non-publishable
report as legally privileged and confidential).
N
To review, and advise POL in writing on or by a date to be agreed (but
currently intended to be no later than 15 October 2013) — (also to be a non-
publishable report as legally privileged and confidential) on:
(a) Its strategy and process for reviewing past and current criminal
prosecutions in light of Second Sight's Interim Report of 8 July
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
(b)
(©)
2013 and/or on the role of Dr Gareth Jenkins (sic) ' and his impact
on any possible appeals;
Its response to the CCRC, and any subsequent action required by it
in dealing with, or responding to, any actual or potential appeals
and/or in reaching the appropriate resolution of any CCRC
investigation, and to advise POL about any further steps that may be
required as regards any actual or potential appeals against
conviction; and
The identification of any flaws in the process of, or from the
evidence arising from, the review of a statistically significant
number of past prosecutions in which Horizon has been an issue in
the proceedings.
B. Meeting/Reporting to the Post Office Audit Committee / Board
1
At the first available opportunity to:
Explain the background to the criminal appeal process including
appeals, and how you intend to fulfil your remit, and
On the efficacy of the process set out in A(1) above.
On or by a date to be agreed (but currently intended to be no later than 15
October 2013) to report on the efficacy of past prosecutions
including the preparation and conduct of past prosecutions set out in
A(2) above.
To meet with Post Office’s criminal law team to gather further background
and contextual information.
Refine the scope of and finalise these Terms of Reference.
@)
Gi)
2
Process
1
2
2
To fully understand the Horizon system — reviewing training and
materials.
'T have found Gareth Jenkins frequently referred to as “Doctor” e.g. in Simon Clarke’s Advice of 15
July 2013 and elsewhere, and therefore I had assumed up to now that he was appropriately referred to
as “Doctor Jenkins”. However, on closer analysis, he does not hold himself out as having a doctorate.
and when he gave evident
in the Misra trial in late 2010 (Transcript of 14 October 2010 at pages 2
3F), he did not hold himself out in evidence as having a doctorate either, and so I have assumed he is
not entitled to use of the title “Doctor”
2
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
4 To meet and interview as a fact-finding exercise anyone else you or POL
consider relevant to the process of the investigation and commencement of
prosecutions.
5 To review a statistically significant number of past prosecutions in which
Horizon was an issue.
3. My Interim Review document of CK’s current review process, dated 2
August 2013,” was submitted to POL through BD by email of the same date.
In the Interim Review document, I made a series of interim recommendations
to which Cartwright King (“CK”) responded in writing on 13 August 2013.
4. Regarding the process by which I have been asked to conduct my review,
and by reference to each in the above process list in chronological order:
Point 5: on 6 September 2013, I received a number of files from BD
containing material in advance of the meeting on 9 September 2013. Among
them there was a file containing 24 full case reviews performed by counsel
employed by CK of cases passing the initial sift process, which I shall return
to below. (On 30 September 2013, I received 6 more full reviews, and on 9
October 2013 I received one more by email.)
My Terms of Reference do not require me to review CK’s decision-making
or the individual judgments about the reviewed cases, merely to review their
review process.
Point 1: on 9 September 2013, I met with POL (Susan Crichton, Rodric
Williams and Jarnail Singh), CK (Simon Clarke, Harry Bowyer and Martin
Smith) and BD (Gavin Matthews and Andy Parsons) at my chambers at 2
Bedford Row, London, WC1.
2 Section A(1) of the Terms of Reference
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
Point 3: on 19 September 2013, I attended Guildford Classroom Training
Office (“CTO”) where I received a day’s training on the Horizon system.
Chris Gilding (Network Support Team Leader) trained me. Andy Holt
(Business Relationship Manager) was on hand to assist and answer
questions.
Point 2: on 23 September 2013, I received from Gavin Matthews of BD the
final Terms of Reference by email, which I reviewed and agreed by return
email the following day, having on 2 August 2013 refined an earlier draft in
writing.
Point 4: at the time of writing I have not identified (or had identified to me)
any other persons to meet and interview as part of a fact-finding exercise as
relevant to the process of the investigation and commencement of
prosecutions.
LEGALLY PRIVILEGED AND CONFIDENTIAL
LEGALLY PRIVILEGED AND CONFIDENTIAL
B. EXECUTIVE SUMMARY
5. The following is a summary of the main findings and recommendations in
the body of this review.
@
Gi)
(iii)
(iv)
(vy)
(vi)
The 1 January 2010 start date for Cartwright King’s review of
Horizon Online disclosure is logical, proportionate and practicable
in light of all the known circumstances.
The scope of the review properly takes account, not just of those
cases in which defendants have ben convicted following a trial, but
also where they have pleaded guilty, as well as where disclosure
might have been material to the sentencing exercise.
The scope of the review, which is to consider those cases, both past
and present, in which disclosure of the Second Sight and Helen
Rose reports should be made, appropriately also considers in past
conviction cases the stance Post Office Ltd might take to any
application for permission to appeal, or to any substantive appeal,
should permission to appeal be granted, the question of the “safety”
of the conviction being inexorably linked to the question of
disclosure.
I advise that individual sift reviews and full reviews should not be
undertaken by anyone who was professionally involved at any stage
with the case being sifted or reviewed, in order to ensure and
preserve independence and objectivity.
The statistical position of the number of sifts and reviews presents a
confused picture, which needs to be rectified in order to ensure an
audit trail of reviews that is robust.
The Wednesday hub conference call meetings are a sensible next
step, and may benefit further from consideration being given to the
establishment of a dedicated mailbox for the sharing of information
among the participants at the meetings, as well as from the
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
(vii)
(viii)
(ix)
(x)
(xi)
(xii)
identification of a person who is nominated to take responsibility for
its management.
Should any new information emanating from the hub meetings be
considered to affect, or possibly affect, any previously sifted or
reviewed cases, consideration will have to be given to broadening
the criteria for review, and to re-sifting and re-reviewing cases
already considered.
Any future successful defence challenge in any Horizon Online-
based case will have to be considered for future possible general
disclosure.
Cartwright King should exercise supervisory control over the
dissemination of information and material during the mediation
process, and should remain alive to the possibility of having to make
case-specific disclosure in the course of that process, or even
making additional general disclosure in the course of its current
review depending on the nature of any new information.
I agree that Gareth Jenkins is tainted and his position as an expert
witness is untenable. Thus, a new expert should be identified as
soon as is practicable.
Post Office Ltd has responded to the Criminal Cases Review
Commission appropriately in my opinion, but should the
Commission continue to show interest in these cases there might
have to come a time when Post Office Ltd considers sharing
Cartwright King’s review findings with the Commission, and
cooperating with the Commission.
Overall, my view, as expressed in my Interim Review document is
that Cartwright King’s review is fundamentally sound, and I have
not detected any systemic or fundamental flaws in the review
process, or in the evidence arising from it, but because the review is
a continuing process, and Post Office Ltd has a continuing duty of
disclosure (not only in cases subject to the Criminal Procedure and
6
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
(xiii)
(xiv)
(xv)
(xvi)
(xvii)
Investigations Act 1996 but in practice should also adopt a similar
or identical approach to past conviction cases falling within the
current review), Post Office Ltd and Cartwright King must be
prepared to keep under review, and reconsider, past case reviews
and disclosure decisions.
Where, as part of the review process, Cartwright King considers the
stance that should be adopted in any possible appeal process, they
should not adopt any over-rigid or robust approach, and must
remain alive to changing circumstances and the need always to
reconsider their stance.
Any new or revised Post Office Ltd prosecution policy document
must consider the retention and treatment of case files.
Post Office Ltd must ensure that Fujitsu Services Ltd retains all
Horizon data.
There should be an analysis for reconciliation purposes of all
Second Sight’s spot reviews with those Cartwright King cases that
have been subject to sift and full review.
If there is any Horizon-related civil litigation between any present or
former sub-postmaster and Post Office Ltd, Cartwright King should
be given complete visibility of the litigation in case this should
affect any decisions they are making about criminal cases.
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
G. BASIS OF THE REVIEW
6. Before coming to the focal point of this review document, it is necessary,
simply by way of introduction, for me to deal with the following relevant
discrete topics:
(i) POL’s prosecution role;
(ii) The Horizon system;
(iii) Horizon training and support;
(iv) Second Sight’s inquiry;
(v) Gareth Jenkins; and
(vi) The Helen Rose report.
ji) POL’s prosecution role
7. On 1 April 2012, the shares in POL were transferred from the Royal Mail
Group Ltd (“RMG”) to Royal Mail Holdings ple (“RMH”), since which
time POL has had an existence independent of RMG. Now, POL is a wholly
owned subsidiary of RMH. The Secretary of State for Business, Innovation
and Skills (“BIS”) holds a special share in POL. RMH and BIS, through the
Shareholder Executive (“ShEx”), have however no involvement in POL’s
day-to-day operations.
8. Prior to POL’s separation from RMG, RMG conducted the prosecution of
criminal offences allegedly committed by sub-postmasters and/or staff in
their employ. Following the separation out of POL’s business from RMG,
POL has assumed and retained the prosecution function hitherto enjoyed by
RMG prior to separation, although I am told RMG still retains a prosecution
function.
Annual Reports and Financial Statements 2012-2013, page 42
LEGALLY PRIVILEGED AND CONFIDENTIAL
LEGALLY PRIVILEGED AND CONFIDENTIAL
12.
13.
POL is the only commercial organisation (albeit Government owned) I can
think of (apart from RMG who retains a residual prosecuting function) that
has a prosecution role, and it is, to that extent, exceptional if not unique.
Separate Terms of Reference invite me to consider the future of POL’s
prosecutorial role.
In England and Wales POL’s prosecutorial role is exercised by POL’s in-
house legal department and through the instruction of CK, CK’s in-house
advocates as well as by external counsel and agents. Prosecutions are
conducted both in the Magistrates’ Courts and the Crown Courts, and appeals
are prosecuted in the Court of Appeal.
In Northern Ireland the Public Prosecution Service (“PPS”) prosecutes POL
cases albeit with input and assistance from POL investigators. In Scotland it
is the Procurator Fiscal who prosecutes POL cases. Representatives of CK
were in Scotland the week before we met in conference in order to discuss
with the Procurator Fiscal’s office the recent interim findings by Second
Sight Support Services Ltd (“SS”). Currently, cases stand adjourned in
Scotland, where, as I understand it, POL has been granted special agency
status.
I understood in our meeting that CK had not yet spoken to the PPS in
Northern Ireland. CK acknowledged the need for them to visit the PPS.
However, to date there had been only two prosecutions in Northern Ireland,
neither of which involved allegations surrounding Horizon.
(ii) The Horizon system
Horizon is the name given to the computer system provided to POL under
contract by Fujitsu Services Ltd (“FSL”), formerly ICL. It is the system used
in all POL Crown offices, sub-post offices and agencies.
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
14. I take the brief history of the Horizon contract with POL and the Horizon
system from part of the content of a witness statement prepared by the
Horizon expert Gareth Jenkins dated 15 January 2013. The witness statement
was prepared for the case of Khayyam Ishaq, the case file of which I have
read.
15. Much, perhaps all, of the parts of the witness statement I quote from below
appears based upon the content of two documents entitled “Horizon Data
Integrity” dated 2 October 2009 and “Horizon Data Integrity for Post Office
Ltd” dated 2 April 2012, which are marked respectively as Mr Jenkins’
exhibits GIJ/1 and 2, and are exhibited to his 15 January 2013 statement. 4
16. Although there are issues about Mr Jenkins’ independence and objectivity
(with which I deal below), I am content to rely on Mr Jenkins’ witness
statements (based as they appear to be in whole or in part on the Horizon
integrity reports) for these purposes as providing a reasonably adequate, and
almost certainly accurate, summary of the Horizon system.
17. According to Mr Jenkins, FSL was originally awarded a contract in 1996 to
provide a Horizon system to POL. The most recent rollout of Horizon was,
says Mr Jenkins, Horizon Online (internally known as HNG-X) in 2010,
which was, according to Mr Jenkins (who was one of its architects), “a
complete re-implementation of Horizon of the business functionality at the
counter and utilised a central database to hold details of all transactions
rather than the MessageStore used by the original Horizon system.”
18. Mr Jenkins says that all POL branches migrated from the original Horizon
system to Horizon Online between January and September 2010. Again,
according to Mr Jenkins, “The Horizon system was designed to store all data
* At page 11
10
LEGALLY PRIVILEGED AND CONFIDENTIAL
LEGALLY PRIVILEGED AND CONFIDENTIAL
20.
21.
locally on the counter’s hard disk in what is referred to as the MessageStore.
Once the data has been successfully stored there it is then replicated (copied)
to the hard disks of any other counters in the branch (and in the case of a
single counter branch to the additional external storage on the single
counter). Data is also passed on from the gateway counter to the Horizon
data centre using similar mechanisms where it is stored in the
MessageStore.”
Mr Jenkins then describes how the replication process is designed to capture
data that fails to copy immediately (for instance due to a failure of the local
IT network in the branch or another counter being switched off or the branch
being disconnected from the data centre). In such a case further attempts are
made to replicate the data at regular intervals until it is finally copied
successfully. Once the data reaches the data centre a further copy is taken by
the “audit agent”, which writes it to an audit file, which is added into the
audit trail where it is available for retrieval for up to 7 years. Moreover, the
data in the audit trail is “sealed” with a secure checksum (“CRC”) that is
held separately to ensure that it has not been tampered with or corrupted.
Mr Jenkins adds that every record written to the transaction log has a unique
incrementing sequence number so it is possible to detect if any transaction
records have been lost, and while customer sessions are in progress,
transactions for that customer session are normally held in the computer’s
memory (known as the “stack”) until the session is settled. When that
happens all the transaction details (including method of payment) are written
to the local hard disk and replicated (as described before).
Any failures to write to the hard disk (after appropriate retries) will result in
the counter failing and needing to be restarted, and so any such failures
would be visible to the user.
11
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
22.
23.
24.
25.
Where data is retrieved for audit enquiries a number of checks are carried
out: (1) that the audit files have not been tampered with; (2) the individual
transactions have their CRCs checked to ensure they have not been
corrupted; and (3) a check is made that no records are missing.
Mr Jenkins then deals with Horizon Online data flows, describing how the
system is designed to store all data in an online database known as the branch
database (“BRDB”). Transactions are carried out locally on counters, and a
basket is built up during a customer session. Once the transactions have been
processed, so that the value of the basket is zero, the entire basket is sent to
the data centre as a “branch access layer” (“BAL”) message where the BAL
processes the message and all the accounting lines are recorded and
committed to the BRDB as part of a single Oracle commit (ie. all or none of
the transactions in the basket are written). Once successfully committed, a
response is returned to the counter indicating the transactions have been
successfully committed, which then allows any receipt to be printed.
The Oracle commit also includes an audit of the data originally transmitted
from the counter to the BRBD, which is the record used to provide extracts
of transactions used for litigation support.
Any auditable message from the counter is stored in the “audit table” (known
as the “message journal”) in the BRBD. After midnight, files from the
previous day are copied from the BRBD to a number of serial files, which,
following certain checks, are then copied to an audit system where they are
digitally sealed and held for 7 years for retrieval purposes. The audit record
may also include application events at the counter since the last auditable
message was sent to the data centre; data regarding other major activities
affecting the branch are also audited.
12
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
26.
27.
28.
The transactions in a basket are constructed using double-entry bookkeeping
meaning that in addition to accounting lines relating to business transactions,
separate accounting lines are generated also for the tender items such as cash
or credit or debit cards, resulting in all accounting lines in the basket totalling
zero. If the net value of all accounting lines does not total zero when the
basket is written to the BRBD an alert is raised and the basket is discarded
and an error Tesponse sent to the counter.
(iii) Horizon training and support
Because Horizon training is one issue that has been raised I requested sight
of Horizon training materials. However, what I was provided with was
information and materials regarding the training currently received on
Horizon, and, because it is current, and may have been revised given recent
events, it is important to exercise a degree of caution about it, as the
materials may not reflect the precise nature and extent of the training that
was offered to those who have fallen within SS’s review, or fall within CK’s
review. Consequently, I cannot conclude that the training and follow-up
support detailed below is necessarily reflective of that provided to those
falling within the review period.
According to the materials I have been provided with, the training given to a
Horizon operator depends on the operator’s current level of experience and
the nature of the post office to be run. I have been provided with a number of
so-called “Training Offers”, which show, for instance, the training offered to
a new post office “local” operator,’ involving total training and/or support of
between 15 and 17 days, in broad terms, comprising:
5 [understand from Andy Holt that a “local” post office is a smaller branch not offering a full range of
products as distinct from a “main” post office, which might offer a full range of goods. I am told by
Andy Holt that training offers 1 and 2 are relevant to new sub-postmasters
13
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
29.
30.
Distance learning in the operator or the operator’s staff's own time
prior to transfer or conversion,° including a written foundation
module;
CTO for classroom training;
On-site induction of 2 x 3 hour sessions for all members of staff;
On-site training and support lasts from 6 days to 8 days depending
on the number of members of staff.
The support thereafter consists of:
Follow-up schedule at a balance one week following the on-site
training;
Post-transfer call from a Field Team Leader within 1-2 weeks of the
transfer for feedback on the Field Support Advisor or on the training
and support;
Post-transfer visits after 1 month and 3 months, respectively;
Post-transfer audit 6-9 months (the trainee operator is not informed
of this).
The new entrant will receive a variety of hand-outs as part of the classroom
training, covering subjects taught during the course attended. Copies of such
hand-outs sent to me include “Horizon Cash Declarations — Top Tips Guide”,
“Daily Procedures and Horizon Reports”, “Priority Service Despatch”, and
“Balance Procedure”. New operators also receive a list of useful telephone
numbers including the number for the Network Business Support Centre
(“NBSC”) and the options from the main menu, which includes, I note,
“Option 1” for Horizon Online system queries or fault reporting.
© Andy Holt has informed me that “transfer” and “conversion” are terms relating to moving branches
on to the new models presently being introduced by POL, and it is before or after transfer or conversion
that progress calls are made; I am told that although this is a fairly new addition to the formal support
process, many Field Support Advisers would contact the agent under the former regime
14
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
31. A Performance Standards Assessment (“PSA”), also copied to me, is a
record of what the new entrant has been trained on and how the individual is
performing. It is begun during classroom training (or on-site if no classroom
training is given) and is completed during the on-site training. An action
plan is produced, which is discussed and agreed with the new entrant. The
new entrant also receives a copy of it.
(iv) Second Sight’s inquiry
32. Following discussions in June and July 2012 between POL senior
management, Rt Hon James Arbuthnot MP and Alana Bates and Kay Linnell
representing the Justice for Sub-postmasters Alliance (“JFSA”), SS was
engaged to perform a review into alleged problems with POL’s Horizon
system.
33. SS’s remit was “to consider and to advise on whether there [was] any
systemic issue and/or concerns with the “Horizon” system, including training
and support processes ...”.’ For this purpose, SS took POL’s Information
Manager’s definition of “Horizon” as embracing the software, computer
hardware and communications equipment installed in branch and central data
centres, as well as software used to control and monitor the systems, and
testing and training systems.* The inquiry was open to sub-postmasters, POL
employees, contractors and agency staff; it was therefore open to all groups.”
34. SS had been asked to investigate 47 cases submitted to it either by James
Arbuthnot MP or by the JFSA.'° All were said to be highly critical of the
7 Appendix to SS report (“The Second Sight Inquiry — the Detail”, page 1)
® §S report paragraph 1.6
° See page I of “Raising Concerns with Horizon”, annexed to the report
report paragraphs 2.1-2.3: 29 cases were submitted through James Arbuthnot MP’s constituency
office and 18 cases of a total of 60 were submitted via the JFSA, which were considered suitable for
investigation with a cut-off date of 28 February 2013 (“Raising Concerns with Horizon”, paragraph
(B)2)
15
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
35.
36.
37.
Horizon system operated by POL, and the way in which POL had dealt with
the matters reported.'!
By way of context, in its report, SS observed that the Horizon system had
approximately 68,000 users, processing over 6 million transactions a day.
The whole population of over 11,800 branches had been notified about SS’s
proposed investigation resulting in only a further 14 cases being accepted for
notification.'* I have also received a PowerPoint presentation, entitled “The
Post Office — An Insight”, which provides invaluable insight into the scale of
Horizon transactions. One of the slides states that Horizon handles 47 million
transactions a week, 22 million banking transactions every month, and 2.5
billion transactions a year with a cash value of £100 billion. Given the scale
of the transactions Horizon is handling, it seems to me that the complaints
and issues around Horizon are comparatively small.
The provisional result of the inquiry resulted in SS’s Interim Report together
with four sample spot reviews,'* which were annexed to the main report, an
accompanying document entitled “Raising Concems with Horizon” (a
document setting out the steps to be taken by current or past sub-postmasters
who felt concerned or affected by issues with Horizon), and annexed to that
was a document setting out further detail of SS’s inquiry.
SS’s interim report, which is dated 8 July 2013, was posted to POL’s website
the same day at http://www.postoffice.co.uk/post-office-statement-horizon
and is accompanied by a statement from POL’s Chief Executive, Paula
Vennells.
" SS report paragraph 1.10
SS report paragraph 1.11
13 Spot reviews 1, 5,21 and 22
16
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
38.
39.
40.
41.
The central conclusion to result from SS’s review was that there was “no
evidence of system wide (systemic) problems with the Horizon software.”'*
The report identified only two “defects” with Horizon Online, which,
according to SS, POL had disclosed it had discovered in 2011 and 2012, and
which had impacted 76 branches:'* first, the so-called “Receipts and
Payments Mismatch Problem”, which had impacted 62 branches, and
amounted to a total discrepancy of £9,029 (the largest single shortfall being
£777 and the largest surplus being £7,044. POL had addressed each of them
with no loss to any sub-postmaster; the second, referred to as the “Local
Suspense Account Problem”, affected 14 branches only, and involved total
discrepancies of £4,486, with a temporary shortfall of £9,800 at one branch
and a surplus of £3,200 at another. Amounts of less than £161 affected the
remaining 12 branches. POL only became aware of this defect a year after its
first occurrence in 2011, when it reoccurred. Because the cause of the defect
could not be identified, POL wrote off the amount. FSL investigated the
issue in 2013 and corrected the defect.
It seems that it was Gareth Jenkins who had informed SS of the existence of
the two defects, although, in their report, SS do not attribute the information
to him."
Several other issues with Horizon had, however, been reported to SS by
multiple sub-postmasters as being of particular concern. Training and
inadequate user / Helpdesk support were identified as two such concerns.'”
The SS report expresses itself to be interim, expresses its conclusions as
“preliminary” and states there is “much work still to be done”.'*1 must
™ Paragraph 8.2(a)
'S Paragraphs 6.4-6.9
*© See paragraph 1 of Simon Clarke’s 8 July 2013 Advice, and paragraphs 30-31 of Simon Clarke’s 15
July 2013 Advice
1” Paragraph 7.2(e) and (h)
18 Paragraphs 8.1 and 8.2
17
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
42.
43.
44,
45.
proceed therefore on the basis that the SS report remains an interim report,
and, therefore, the conclusions I arrive at in this document are necessarily
subject to further consideration of any additional or different conclusions SS
might reach in the future.
In addition to SS’s report and appendices, I have received under separate
cover the underlying material to spot reviews 1, 5, 6, 10, 11, 12, 13, 21, 22
and 23, four of which (1, 5, 21 and 22) are reproduced, albeit anonymised, in
the spot reviews annexed to the SS report. Each review file contains a pro
forma sheet for completion by SS, providing the detail of the review
reference number, sub-postmaster name, any loss to the sub-postmaster, the
issue, the branch name, the status of the review, the date, the FAD (branch)
code, and the category (one of 4 check boxes) which had been intended by
SS to categorise where, apparently, the fault for the loss lay, but I am told the
box categories fell into disuse and are no longer relied upon.
The pro forma sheet then sets out what the sub-postmaster says happened,
the investigative work done by SS, POL’s response to the claim (which is
enclosed with the review file) and occasionally SS’s preliminary conclusions.
Some of the review files (by no means all) contain documentary evidence
and exchanges of correspondence.
Clearly, I have only seen a fraction of the spot reviews done by SS.
(v) Gareth Jenkins
Any challenge to the integrity of the Horizon data underpinning any given
prosecution case has required expert evidence to defeat the challenge, the
burden and standard of proof being on the prosecution throughout.
18
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
46.
47.
© Misra Trans
Gareth Jenkins, who has been involved in the Horizon project for many
years, has been the sole expert instructed to speak to the integrity of the
system. In his sample witness statement, to which I have made reference
above, Mr Jenkins deploys (as he must) his expert credentials in order for
him to be received by the court as an expert witness. In essence those
credentials are that he graduated from Cambridge University in mathematics
in 1973 and was awarded an MA in 1997; he says he is employed by FSL
who are contracted to POL to provide Horizon systems operating in post
offices nationwide; he was first employed by ICL in 1973, the forerunner to
FSL before a name change; in the late 1990s he became a “Distinguished
Engineer” with ICL; he provides detail of his memberships and associations -
he is a member of the British Computer Society (MBCS), a Chartered
Engineer (CEng) and a Chartered IT Professional (CITP); he says that since
1996 he has been working on the Horizon project in association with POL,
and details his involvement with it; he states he understands his role as an
expert, despite being in the employ of FSL.
Mr Jenkins was accepted as an expert, without apparent challenge, in the
Misra trial (which pre-dates the current review, the theft having occurred
between June 2005 and January 2008)."” It appears that this was the first and,
I am informed, only time he has ever given evidence in support of a POL
prosecution, but has attended court in preparedness to give evidence.”
However, Mr Jenkins has made several witness statements in various cases.
Simon Clarke’s Advice of 15 July 2013 lists a sample few he has made in
cases ranging between October 2012 and April 2013.7'1 shall return to
Gareth Jenkins below.
ipt of 11 October 2010 pages 27H-28A
°° Misra Transcript of 14 October 2010 page 3F
21 Paragraph 16
19
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
48.
49.
50.
(vi) The Helen Rose report
The Helen Rose report is a draft report by Helen Rose, a POL security fraud
analyst, of 12 June 2013, which deals with one Horizon issue concerning
unexplained reversals at the Lepton sub-post office.”
In summarising the facts in her report, Ms Rose describes a transaction
occurring at Lepton on 4 October 2012 at 10.42 (sic) for a British Telecom
bill payment for £76.09, ”> which was paid for by a Lloyds TSB cash
withdrawal of £80. Change of £3.91 was given. At 10.37 the same day the
British Telecom bill payment was reversed out to a cash settlement. The
branch was issued with a transaction correction for £76.09, which was
settled.
The draft report is based on an exchange of emails between Gareth Jenkins
and Helen Rose during the period 30 January 2013 to 13 February 2013.
Helen Rose’s recommendation was that the system had functioned as it
should, but that the available data was capable of misinterpretation, requiring
the data to be presented differently so as to ensure that all system-created
reversals were clearly identifiable. Again, I shall have to return to this.
» This is SS’s spot review SROI
7 note that SS’s spot review SRO1, which relates to this reversal, describes a different sequence of
events with the payment of the bill at 10.32, and the session receipts not being printed until 10.36
There is support for this timing in POL’s response to the spot review, which gives a recovery time as
10.37 (09.37 GMT), which suggests to me that Helen Rose cannot be right when she says that the first
transaction took place at 10.42
20
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
Sl.
52.
THE TERMS OF REFERENCE
It is against that background that I now confront the Terms of Reference set
out above,”* namely:
(a)
(b)
(©)
(a)
POL’s strategy and process for reviewing past and current criminal
prosecutions in light of Second Sight's Interim Report of 8 July
2013 and/or on the role of Dr Gareth Jenkins (sic) and his impact on
any possible appeals;
Its response to the CCRC, and any subsequent action required by it
in dealing with, or responding to, any actual or potential appeals
and/or in reaching the appropriate resolution of any CCRC
investigation, and to advise POL about any further steps that may be
required as regards any actual or potential appeals against
conviction; and
The identification of any flaws in the process of, or from the
evidence arising from, the review of a statistically significant
number of past prosecutions in which Horizon has been an issue in
the proceedings.
POL’s strategy and process of review
POL’s strategy and process was the subject of my Interim Review document
of 2 August, at which time I had a relatively imperfect knowledge and
understanding of CK’s review process. I had, however, seen and read two
full case files in the cases of Khayyam Ishaq and Lynette Hutchings. Since
that time I have had the benefit of reading Harry Bowyer’s response to my
Interim Review document, at the time of writing I have read 31 full reviews,
* Section A(2)(a)-(c)
21
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
53.
54.
35.
56.
and, as noted above, I have seen in conference POL senior management and
representatives of CK, and BD, in which I was able to seek further
particulars of POL and CK’s review process.
Since the Horizon issue has arisen, CK has conceived a review strategy of all
suitable cases meeting certain criteria. This applies both to past prosecutions
resulting in conviction, after a contested trial, and to cases following pleas of
guilty.
POL tends to prosecute sub-postmasters for three or four types of offence:
theft and false accounting under sections 1 and 17 of the Theft Act 1968, and
fraud by false representation or by abuse of position under sections 2 and 4
of the Fraud Act 2006. Typically, the investigation and eventual prosecution
of such cases is dependent on the data output of Horizon. Naturally, the data
supporting any prosecution needs to be explained and presented to a bench of
justices or a jury, and it must not lack integrity, and, so far as it can be, the
evidence about it must be unimpeachable.
Given the adverse publicity about Horizon thus far, it would be unsurprising
if a “bandwagon” effect were soon to be evident (if not already so) and even
in those cases where Horizon was not in issue at trial or before a plea of
guilty may, following a process of post-rationalisation, suddenly become
Horizon issue cases.
CK’s experience of prosecuting these cases suggests that Horizon cases
involve challenges, either expressly or by implication, to the Horizon system
itself. Where the defendant admits fraud or false accounting, they do so
because Horizon failed in some way, and all they had done was to cover up
shortfalls that could not be explained by them. Moreover, defendants often
complain also about inadequate Horizon training and support and rely on
their own incompetence to explain what has happened.
22
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
57.
58.
59.
60.
POL, having accepted that in light of SS’s interim inquiry report appropriate
action had to be taken by it towards cases that might be impacted by their
findings, has instructed CK to review past and indeed present cases.
Nature and scope of CK’s review
The nature and scope of the review was defined in Simon Clarke’s 8 July
2013 Advice, which coincided with the day of the publication of SS’s report.
The advice was that cases affected included RMG cases pre-separation, and
the areas of concern were the proper functioning of Horizon, Horizon
training and customer support.
Mr Clarke sought to distinguish between two issues:
qd) The impact of the preliminary findings on past successful RMG
prosecutions and on past and pending (or indeed future) POL
successful prosecutions (whether on conviction or by guilty plea) as
regards prosecution duties of disclosure; and
(2) Issues of the “safety” of convictions (which is the test applied by the
Court of Appeal).
Mr Clarke had no doubt that the SS report fell to be considered for
disclosure. Consequently, Mr Clarke’s advice was, “Where it is determined
that the Second Sight interim report and contents fits within the scope of the
CPIA” as disclosable to the defence, defence solicitors will be so informed.
Thereafter, it is a matter for those defendants as to what course they take,
POL’s duty being satisfied by the conduct of a review process and the
informing of those whom ought to have been informed had the matters
25 Criminal Procedure and Investigations Act 1996
23
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
61.
62.
63.
64.
contained within the Second Sight interim report been available when
prosecuted.” He advised that similar considerations applied to cases pending
prosecutions but had yet to reach the trial or plea stage.
There was also consideration to the start date for this review. Horizon Online
was migrated into all post office branches between January and September
2010. Thus, the advice given was that the start date for the review process
should be 1 January 2010, that is to say the earliest date for the Horizon
Online rollout. Mr Clarke advised that in advising upon this date, he had in
mind issues of proportionality, resourcing, transparency and POL’s
reputation.
Simon Clarke’s view was that any sub-postmasters prosecuted under the
former Horizon data regime would have served any sentence of
imprisonment, or performed any unpaid work requirement or paid a fine; and
at all events the publicity from SS’s report would put those defendants on
notice.
Indeed, the SS report and the appendices to it are posted on POL’s website
and so it and its content are available to any member of the public without
the need for recourse to a formal approach to POL. But I am afraid I do not
see that those who have served their sentences, or those who had imposed on
them community-based or financial sentences, should, for that reason, be
excluded from the review. They have an interest if their conviction was
unsafe, and there must be people who fall within CK’s current review who
have been released from their sentences or had non-custodial sentences
imposed on them. Resourcing and POL’s reputation are also beside the
point.
When I queried the rationale behind the cut-off date, I was told, and entirely
accept, that, prior to each branch rollout, a cash audit was done so that each
24
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
65.
66.
67.
branch balanced. I advised in the conference and repeat here that although
POL has no positive duty to seek out individuals before the 1 January 2010
start date for a review of their case, nonetheless if POL was approached it
would need to make ad hoc case-specific decisions about the need for
disclosure.
The Misra case is an example of the type of case I have in mind that might
surface. Although the Misra case was tried in October 2010, the allegation
related to events between 2005 and 2008, long before the rollout of Horizon
Online. However, the issues raised in the case, which were made late by the
defendant in one or more defence statements, were very similar to those
generally being raised currently in relation to the Horizon Online system: the
defendant, Seema Misra, was eventually to claim the approximate £75,000
deficit in her post office was due to a technical error or her own
incompetence, having initially sought to blame her employees for the theft.
There is no information that Mrs Misra is now seeking to appeal her
conviction on the basis of the Horizon technical fault she claimed at her trial,
though she is, I am informed, one of those seeking to apply to participate in
the mediation process, which has recently been set up and permits those even
convicted of crime to participate in the process, subject to a decision about
suitability.?°
I am informed that there are currently 12 or 13 conviction case applicants of
a total of 38 people who have registered their interest so far for the mediation
process (the closing date being 18 November 2013). Some, if not many, of
these cases pre-date Horizon Online; some may even be cases that were
discontinued or where there were acquittals following a trial.
26 See Part I: Application Phase of the draft Mediation Pack
25
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
68.
69.
70.
Ts
In this regard, I have considered SS’s inquiry, and I note that SS’s inquiry is
not time-limited. The only date given is 28 February 2013, now long past, as
the latest date for any submission to SS’s inquiry.
The definition of “Horizon” that SS relied upon is in fact widely drawn, and,
despite the fact that that definition was provided to SS in May 2011, and on
one interpretation appears to be focused on the system then installed in
branch and the central data centres, the definition was not expressly limited
to Horizon Online, and evidently SS did not take it that way.’ Indeed, of the
several spot reviews I have seen, many of them involve Horizon issues pre-
dating the 2010 rollout of Horizon Online. The earliest I have seen (SRO12
Hamilton) goes as far back as 2004-2006.
Despite the open-fronted nature of SS’s inquiry, it is important to recognise
that SS has so far only discovered and reported upon two Horizon defects
respectively occurring in September 2010 (which I assume to be a defect
with Horizon Online) and in 2011.78
In my judgment, the 1 January 2010 start date for CK’s review is both a
logical and practicable approach to take. That is not to say however that if a
case pre-dating the rollout of Horizon Online presents itself POL and CK
should exclude it from consideration. There may be cases that raise genuine
thorny technical issues, which are not unrelated to issues concerning Horizon
Online (which after all is next generation Horizon), and, if they arise, they
will have to be dealt with as required on a piecemeal basis. If it got to the
stage where the floodgates of pre-Horizon Online cases began to open, then
POL and CK will have to remain alive to the possibility of commencing a
subsidiary review.
2” See paragraph 1.6
28 See paragraphs 6.4-6.9 and 8.2(b)
26
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
72.
73.
74.
75.
76.
During a telephone conference held on 4 October 2013, in which
representatives of POL, BD, CK and myself participated, it was agreed by all
that the 1 January 2010 start date for the CK review was logical and
proportionate, and there should be no change to it.
In terms of the geographical scope of the review, as I have set out above,
POL and CK prosecute in England and Wales, whereas the PPS of Northern
Ireland conduct POL prosecutions, whereas in Scotland it is the Procurator
Fiscal’s office, which conducts POL prosecutions. There have been
discussions between CK and Scottish prosecutors about SS’s findings.
Current Scottish prosecutions remain adjourned. CK is yet to discuss the
issues with the PPS but there had only been two prosecutions in Northern
Ireland, neither involving Horizon issues. Thus the real focus of CK’s review
has been England and Wales.
CK see the review issue as one of disclosure under the CPIA, the Code of
Practice made thereunder, the Protocol for the Control and Management of
Unused Material in the Crown Court, and the Attorney General’s Guidelines
. 2
on Disclosure.”
The question they have posed is: “Had POL been possessed of the material
contained within the Second Sight interim report during the currency of any
particular prosecution, should / would we have been required to disclose
some or all of that material to the defence?” It is that question that CK state
has defined their approach to the issue.
The scheme CK has devised and adopted for their review involves two
principal stages — a sift review and a full review. The original sifts are now
2° See CK’s “Draft Paragraphs for Insertion into Reply to CCRC” at page 2
27
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
Th
78.
79.
80.
also subject to a second sift process, which has been written into CK’s
scheme, as a sensible control of the original sift review process.*°
Essentially, the scheme involves CK identifying every case within the review
period in which the primary or main evidence against the defendant is based
on Horizon data. Once such a case is identified, then CK’s senior in-house
prosecutors (from the reviews I have read they are exclusively Simon Clarke
and Harry Bowyer) review the case fully for a determination of the question
that CK has posed.
CK has deliberately set the threshold low for a determination of the posed
question, so as to ensure that those affected receive proper disclosure of the
material. CK has been identifying cases (1) where the only evidence against
the defendant is the Horizon data, such as unexplained transaction reversals
or shortages; (2) where the main evidence is Horizon-based but there is some
other supportive evidence, such as admissions or even demonstrable lies in
interview or transfers of money into non-POL or personal accounts; and (3)
any case in which CK has served expert evidence relating to Horizon. This is
said to be a non-exhaustive list.
CK has also been identifying cases where the defendant has explicitly or
implicitly raised a Horizon issue. This will include situations where the
defendant suggests that he was only asked to do as he did by POL or a
customer, or does not directly criticise Horizon but cannot explain the
shortfall, and it embraces those defendants who allege a lack of, or
inadequate, Horizon training, and a lack of any or any sufficient Horizon
customer support.
CK have made clear, correctly, that the review process is not limited to those
found guilty following a trial but includes those who have pleaded guilty
*° See generally the “Initial Sift Protocol” devised and employed by CK for the review process
28
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
81.
82.
83.
84.
also. It also includes consideration of sentence and the impact of non-
disclosure, as has been advised in one reviewed case.*!
Once identified for review, the case is considered and the disclosure decision
documented in a full review in writing by those CK counsel.
From my reading of the full reviews, which have been made available to me,
counsel do not always limit themselves to making decisions on disclosure. In
teviews of pending or current cases, in effect, they review the test, which
POL adopts from the Code for Crown Prosecutors, namely, whether, in light
of the disclosure, there remains “a realistic prospect of conviction”.
At the time of writing, reviews of pending or current cases have thus resulted
in the termination of four cases: Jishaan Patel, Brown, Wylie and Knight
(typically due to the nature of the issue raised by the defendant, and the fact
no replacement expert has yet been identified for Gareth Jenkins, the Horizon
technical expert). While some of the decisions might, objectively, be
regarded as generous, I do not suggest that these were decisions that no
reasonable prosecutor could have made, applying his mind to the relevant
test.
In reviews of past convictions counsel tend also to provide advice about what
POL or CK’s stance should be to possible appeals by offenders to the Court
of Appeal, which must mean consideration of the “safety” of the conviction,
thus the likely stance to any application for permission to appeal the
conviction based on the disclosed material and/or to any substantive appeal,
if permission is granted.” I made it clear to CK in conference that they must
be alive to changing circumstances. They must therefore not adopt an over-
rigid approach; each case must be approached on a case-by-case basis.
*! Case review of Sefton and Nield in light of the judge's sentencing remarks
* T have noted, however, that, typically, case reviews expressly assert that the purpose of the review is
not to determine whether or not the conviction is unsafe
29
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
85.
86.
87.
88.
The process intentionally encompasses both Magistrates’ Court and Crown
Court cases, but the process is a different one in the Magistrates’ Court,
where in the first instance, any appeal on conviction is to the Crown Court.
In order to standardise their approach, CK have compiled a document
entitled “Initial Sift Protocol”, which explains the review process through its
stages, sets the parameters for the process, and is designed to direct the
sifter’s attention to the appropriate questions to be asked when deciding
whether a given case should go to full review. The document annexes four
appendices. In summary, Appendix I describes the sifting process, Appendix
2, the process of full review, Appendix 3, the required qualifications of the
reviewer, and Appendix 4, the single page “Initial Sift Result Sheet”, which
synthesises the relevant questions to be asked by the sifter with boxes for the
details of the decision to be recorded; this document is to be attached to the
front of the relevant file. I have also seen and read through three sample files,
which were sifted but did not go to full review.
The two essential questions asked in the sift review are:
I. “Was or might Horizon reasonably have been more than just the
information provider?”
2. Irrespective of the plea, did the defence raise at any stage:
a) Alleged or implied Horizon failings, however expressed,
general, nebulous or ill-defined; or
b) — Alleged or implied lack or inadequacy of training by POL; or
c) Alleged or implied lack or inadequacy of Horizon customer
support (i.e. sub-postmaster) by POL.
If the answer to both questions is or may be “Yes” then the case is submitted
for a full review. The protocol sets out the type of considerations and criteria
30
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
89.
90.
91.
92.
that sifters should be addressing in seeking to answer the questions in any
given case.
I did find the first question to be somewhat cumbersome. I take it to direct
the sifter’s attention to the significance of the Horizon data as underlying the
particular case under review. However, there are adequate checks and
balances in the scheme, including the second control sift, to satisfy me that
the focus of the question may be sufficiently understood so that deserving
cases do not fall through the net.
Appendix 1 “Sift Reviews” addresses, among other things, the issue of guilty
pleas, the fact that the threshold for moving from an initial sift to a full
review has been set very low and the proper approach in borderline cases.
Importantly, the Appendix ends by stating that at the conclusion of the sift
review process, in all cases where the sifter has advised there should be no
full review, senior in-house counsel will themselves sift those cases so as to
ensure and preserve a uniformity of approach and to check the correctness of
the original reviewer’s decision. As indicated above, this process has now
begun.
Appendix 2 “Full Reviews” addresses the analysis that is expected at this
stage of the process. The sole question is said to be: “Would the Second
Sight and Helen Rose material have been disclosable during the currency of
the prosecution?” The instruction provided that it is not necessary to consider
whether or not a conviction may be said to be “safe”, which is a
consideration for the Court of Appeal, appears to me generally to be ignored.
Appendix 3 “Reviewer Qualification” addresses the level of reviewer for the
sift process and for the full review. Among other things, those solicitors and
counsel involved in the process were obliged to attend an induction meeting
during which the protocol document was explained in full.
31
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
93.
94.
95.
96.
On this topic, I have been told that the sifters include CK solicitors Martin
Smith and Andrew Bolc, who have been involved in sifting some of their
own cases. Some of the barrister reviewers have occasionally had some input
into cases, which were prosecuted by external counsel, such as advising on
charge or evidence.
At the conference I did make the observation, which was foreshadowed in
my interim review,® that lawyers should not be engaged in sifting or
reviewing a case if they were responsible for conducting the case at trial, so
as to ensure a proper level of objectivity and independence in the process.
In his response to my Interim Review document, Harry Bowyer suggests that
if I felt the process was, to that extent, flawed then those cases that were not
put up for a full review might be re-sifted by lawyers wholly independent of
the case, and those full reviews that were reviewed by in-house counsel, who
had some input into the case, can be re-reviewed, the numbers affected being
relatively small.
I have considered the position. It seems to me that the control sift by senior
counsel of all the original sift reviews, which were not recommended for full
review, ought to ensure an independence (and indeed standardisation) of
approach, so far as the sifting process is concerned. However, it would be
better if those cases were not re-sifted by counsel if they were involved in
prosecuting the case at trial or advising on any aspect of it. The same
approach ought to be adopted towards negative full reviews. In my view, this
recommendation is sufficient and proportionate to resist any later challenge
to a decision not to disclose as based on bias.
%3 Paragraph 6
52
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
97. I did hesitate at suggesting that any counsel who had any input at an earlier
stage of a case, but did not later prosecute it at court, should not be involved
in the full review, but, on balance, feel that it must be a logical and necessary
step to ensure an entirely independent approach to the case that such a person
should not be involved in the review.
98. I recognise also that there is always going to be the possibility of the
suggestion of a commercial conflict of interest, given CK’s professional
relationship with POL and the fact that the very counsel and solicitors
making decisions about POL cases are those who rely on CK and POL for
this work.
99. I have considered this issue with some care and, having met with
tepresentatives of CK, and having considered the many Advices and other
material I have seen emanating from CK representatives, I have seen no
evidence other than a professional and independent approach to this review.
Consequently, on the material available to me, I would reject any suggestion
that CK’s solicitors and counsel cannot act, or have not acted, with an
independent and professional approach to the Horizon issues, which have
arisen, and to their review.
100. Appendix 4 “Initial Sift Result Sheet” is an essential means by which the
result of the sift review can be completed and attached to the case file. Like
the full case review document, which goes with the case file, these are
absolutely necessary to provide a full audit trail of every action taken as
regards these cases.
101. One other issue occurs to me, and that is how long such files ought to be
retained. I am unaware of POL or CK policy for the archiving of prosecution
files. I understand that POL prosecution and enforcement policy is/are in
draft, and has not been approved by POL’s Board, and that hitherto POL has
33
LEGALLY PRIVILEGED AND CONFIDENTIAL
LEGALLY PRIVILEGED AND CONFIDENTIAL
simply relied upon RMG policy as informing their prosecution policy.
Indeed, a document entitled “Post Office Prosecution Policy”, effective from
1 April 2012 and reviewable on 1 April 2013, (although I know not if it was
reviewed on that date), appears to be largely derived from the wording of
RMG’s prosecution policy document of April 2011.
102. Equally, I understand that CK is in the throes of advising upon, and
redrafting, the policy for POL’s purposes. I have read through the several
(sometimes overlapping) draft policy documents that have been sent to me.
None I have seen provides for any POL policy about the retention and
treatment of case files. POL must have a policy dealing with such matters, if
none is already written. POL is subject to the Data Protection Act 1998, and
it is a public authority by virtue of Part VI of Schedule 1 to the Freedom of
Information Act 2000. Thus it must comply with the data protection
principles. It is also open to FOI requests.
103. POL may envisage liability under these statutes as applying only in the
course of its ordinary business, but they may apply equally to its prosecution
tole, which places an onus on it to have policies in place that inform, among
other things, its case file retention policy.**
The case reviews
104. I have been sent a table setting out the statistics of the review thus far, which
shows that as at 26 September 2013, CK had sifted 197 case files, second
sifted 17 cases, fully reviewed 28 cases (in which disclosure was advised in 9
cases), and has discontinued four cases (as mentioned above). CK has, I am
“T have not overlooked the fact that in its recent draft Protocol on disclosure, CK has provided, at
clause 4.5.6.iv, for a retention period, but this is not the same as an internal retention policy for all case
files, because, by the definition provided, it applies only to “all material to which clause 6 applies”,
which, by clause 6, is limited to the “identification of equipment and material subject to this protocol”
[My emphasis]
34
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
105.
106.
107.
told, 18 cases yet to be fully reviewed. It is unclear to me whether any cases
are still subject to a first sift.
I have had to query these statistics, because I was sent 24 full case reviews
ahead of the conference I had with POL and CK on 9 September 2013. I have
recently received an extra 6 reviews by email of 30 September 2013 (some of
them being RMG cases) but all of those 6 case reviews post-date the 9
September 2013 date of the conference and they pre-date 26 September
2013, which means I had by on or about 30 September seen and read a total
of 30 full reviews.
However, at the conference I was told by CK that by that date CK had in fact
conducted 27 full case reviews. I had seen only 24 by then and so if that was
accurate there were 3 case reviews missing from the number I had seen and
read by the time of the conference, but those missing 3 could not be any of
the 6 case reviews I have recently received because none of them (the 6) pre-
dates 23 September 2013 or post-dates 26 September 2013. This suggests to
me that CK must have reviewed at least 30 cases by 26 September 2013, and
if there were 3 other case reviews missing from those sent to me in advance
of the conference then they should have conducted 33 case reviews, not the
28 mentioned in the table.
Since then I have received information in an email from Martin Smith to
Jarnail Singh of 2 October 2013 that over and above the 24 case reviews I
was sent and reviewed in advance of the conference, 3 other cases had indeed
been reviewed by CK by the time of the conference. They are Brown,
Chrystal and Knight, all apparently submitted to POL in July and apparently
Chrystal and Knight were submitted to BD on 5 September 2013. Knight and
Brown have been discontinued (as mentioned above, assuming it to be the
same “Brown’”). At the time of writing I have not seen these reviews. The
35
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
108.
109.
110.
1.
case of Patel was reviewed and terminated verbally in light of the urgency of
the case.
To complete the picture, on 9 October 2013, the case review of Robinson,
dated 27 September 2013, was emailed to me. That means I have now seen
and read 31 full reviews.*®
I have also been provided with an Excel spreadsheet, which I am told is
updated, and submitted to POL weekly, which purports to show current to 26
September 2013 the cases that have been reviewed. That spreadsheet,
however, appears to me to show a total number of 175 reviews, and,
according to Martin Smith, shows that as of that date CK has fully reviewed
28 cases, which is the number which has been inputted into the table I was
originally sent. If I have understood Martin Smith’s email correctly, the 6
further reviews are yet to be added to that weekly Excel spreadsheet (correct
to 26 September) and therefore once added should make 34 full case reviews
but I have no confidence that my assumptions are correct.
Thus, the statistical picture is confusing and I have been unable to reconcile
the number of cases reviewed by CK with those seen by me. This needs
rectification, if CK’s audit trail is to be robust.
Other initiatives
I have been provided with an Advice written by Simon Clarke dated 2
August 2013, in which he advised on the question of disclosure and POL’s
duty to record and retain material. In it, he confirmed advice he had given
POL at a meeting at POL’s head office on 3 July 2013 that, “... there ought
to be a single, central hub, the function of which was to act as the primary
Since committing these statistics to writing I have received a new table, now current to 10 October
2013, which shows there have been 278 initial sifts, 44 second sifis, 35 fully reviewed cases, 11 in
which disclosure has been advised, and four cases discontinued (the same four identified above)
36
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
112.
113.
114.
repository for all Horizon-related issues. The hub would collate, from all
sources into one location, all Horizon-related defects, bugs, complaints,
queries and Fujitsu remedies, thereby providing a future expert witness, and
those charged with disclosure duties, with recourse to a single information-
point where all Horizon issues could be identified and considered.”
The rationale behind it is to put in place a mechanism to protect POL from
future such issues. In the result a weekly Wednesday conference-call meeting
has been established to meet the requirement. Attendees are expected to
bring all Horizon-related issues to the meeting and minutes are kept.
However, early teething and “cultural” problems arose as highlighted in
Simon Clarke’s 2 August 2013 Advice,*° and indeed to me in Harry
Bowyer’s response to my interim review.*”
However, the meetings have also produced information, which requires
further investigation, and they and other future issues may highlight other
Horizon-based issues, which POL was previously unaware of. CK must keep
an open mind to any new Horizon issues as they arise and if it is considered
that any information emanating from the hub meetings affects, or might
affect, any of the cases previously sifted or fully reviewed, then CK will have
to remain alive to the possibility of broadening the criteria for the review and
having to re-sift or re-review cases already considered, both past and pending
cases.
More recently, I have received from, and authored by, CK a document
headed “Protocol”, which is a protocol for the Wednesday morning
telephone conferences held for the purpose stated above. The document has
been sent to me for any suggestions I may make as to its content and possible
revision or alteration.
% Paragraph 5
*7 Paragraph v)
52
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
115.
116.
117.
°8 Paragraph 6.2. 1.v
The document sets out the fundamental principles of the prosecutor’s duty of
disclosure, the purpose of the hub meetings, its Chairman and members’
functions, their duties and responsibilities, the recording and retention of
relevant material, and the group’s constitution. The protocol is aimed at what
the authors call the “front-line in the information gathering exercise.”
I make two observations about the document:
qd)
(2)
In so far as the retention of material is concerned, as observed above
at footnote 34, at clause 4.5.6.iv, the protocol provides for a
retention period of not less than 6 years. The only observation I
would make is that because, according to Gareth Jenkins, the data
sent by the branch to the audit system of the data centre remain
retrievable for 7 years, then the period of retention ought to be not
less than 7 years;
The all-important clause 6, which sets out in great and
comprehensive detail the identification of equipment and material
subject to the protocol, defines the “Horizon Online system”
sufficiently widely to capture the categories of material that are
expressed to be subject to the protocol. The only further criterion
that might be considered is any information, which might indicate
that past or present Horizon training is inadequate or not fit for
purpose, further to the criterion that provides for any information
indicating there is “a requirement for further training of those
operating the Horizon Online system.”**
I have also advised in conference, and repeat, that it might be an idea for
those seeking to manage the flow of information to set up a confidential and
dedicated email inbox to be shared by the contributors to the hub meetings
38
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
118.
119.
120.
121.
for the central receipt of information from those contributing to the
Wednesday hub meetings, and as an adjunct to it, and to avoid the risk of
relevant information being lost, as well as for an individual to be nominated
to take responsibility for managing it.
As a cautionary warning, I have noted from the product of the hub meetings
that there appeared to be possibly greater focus on the fix to a problem rather
than focus on actioning the issue for the purposes of disclosure. While the
hub meetings may well serve a dual purpose, the central point of the hub
meetings must not be overlooked or marginalised.
It is important that now I emphasise and comment upon certain issues about
POL’s disclosure duties.
POL’s duty of disclosure
On 2 August 2013, Simon Clarke produced his Advice setting out fairly
extensively, and accurately, POL’s duties to record, retain and disclose
material under the CPIA etc.
I would however add this: POL’s relationship with FSL is, in the context of
its prosecution role, unique. The prosecution of the typical case of theft,
fraud or false accounting tends to rely wholly or primarily on data from
Horizon. At its simplest, FSL is contracted to POL to provide the Horizon
system, but FSL is not part of POL. They are nothing more than contracting
partners, and FSL is therefore a third party for disclosure purposes. I advised
at the conference that I had considered whether or not FSL should be invited
to participate in the Wednesday hub meetings, but upon mature reflection I
considered they should not be, and should be kept at arm’s length as a third
party.
39
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
122,
123.
124.
125.
Not unnaturally, FSL is likely to regard its Horizon system as commercially
sensitive, and would not be likely to wish it exposed to challenge, but despite
that natural inclination, I was told by Rodric Williams at the conference that
FSL has not been resistant to cooperation with POL’s review but they have
been slow.
Thus, because FSL is a third party for disclosure purposes, POL, as
prosecutor, has, in the ordinary case, an obligation to take such steps as it
regards as appropriate to obtain third party material that it regards as possibly
relevant to the prosecution case, and, if the third party fails to cooperate, the
prosecutor’s duty is to seek access to the material if need be enforced by a
witness summons.”
Moreover, it is important always to bear in mind that under the head of
“General Responsibilities” the Code of Practice made under the CPIA
provides that in conducting an investigation the investigator should pursue all
reasonable lines of enquiry whether they point towards or away from the
suspect.*°
There are two other issues as regards FSL, which I need mention. First, I
understand that POL is currently re-procuring its IT system, and, second, that
FSL was considering migrating their system to new servers. The first is a
business issue for POL but, in tandem with the second, as the review process
continues, it is imperative that POL remain alive to the fact that past
convictions as well as pending or current prosecutions, which are Horizon-
based, require it to ensure that the data is retained in order that POL can
defend past convictions and prosecute current charges to conviction, and to
receive assurances from FSL, whatever the future for their contractual
relationship, to that effect.
® Attorney General’s Guidelines on Disclosure, paragraphs 52 and 52
“ Paragraph 3.5
40
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
126.
127.
128.
Following the conclusion of the proceedings, POL has a general common
law duty to act fairly and to assist in the administration of justice. The post-
trial duty of disclosure falls outside the provisions of the CPIA but the duty
of disclosure continues so long as proceedings remain, whether at first
instance or on appeal. The situation here is different as there are (so far as I
know) no appeal proceedings outstanding that relate to Horizon issues but
POL has, in the special circumstances obtaining here, very properly
acknowledged its duty to consider cases for disclosure, which is inevitably
interlinked with considerations of the safety of convictions.*!
Although the test for disclosure in past conviction cases is not that under the
CPIA, at common law the issue is one of “materiality”, which is not a very
different thing, and if CK has been conducting itself in the disclosure review
exercise by applying the CPIA test even to past convictions, it is hardly
likely to be criticised for doing so. The duty under the CPIA is a continuing
duty of review, and it would be wise for POL and CK to apply a similar if
not identical approach to past conviction cases falling within its current
review, although, strictly, they fall outside any post-trial period criminal
proceedings, where there remains a common law duty to disclose material
that might cast doubt on the safety of the conviction.
Finally, in this regard, it must be borne in mind that any successful challenge
to the Horizon system, defined in its broadest sense, whether in an on-going
trial or on appeal, risks undermining other on-going cases as well as
successful past convictions. Indeed, one successful challenge to the system
on any basis will require CK to consider disclosure of that failure across the
board and to re-review all previous decisions.
I Makin [2004] EWCA Crim 1607; and see paragraphs 59-60 of the Attorney General’s Guidelines
C
if material comes to light after the conclusion of the proceedings, which might cast doubt on the
safety of the conviction, there is a duty to consider disclosure.”)
41
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
129.
130.
131,
The mediation scheme
During the telephone conference of 4 October 2013, the main topic of
discussion was the extent to which CK should be involved in exercising a
supervisory function over the criminal cases going to mediation, such as Mrs
Misra’s case. There is understandable concern that offenders might use the
mediation scheme to gain information as a platform from which to launch a
fresh or new appeal, and so CK wish to exercise a measure of control over
the dissemination of information and material during the process.
As a precursor to the telephone conference, a question had been floated about
the Falkirk issue, as one example, as potentially bearing on the mediation
scheme. The Falkirk event was raised in the Misra case. In the course of
Gareth Jenkins’ evidence in Misra, Mr Jenkins gave evidence about a
Horizon event, which had occurred at Callendar Square post office in Falkirk
in 2005, whereby information recorded on one terminal was not being
correctly passed to another terminal within the branch, creating a receipts and
payments mismatch.*” A software fix of the problem was distributed into the
system in March 2006 as part of a fairly major functional change, and,
therefore, post-March 2006, the Falkirk defect was no longer an issue. In my
view it represents an isolated instance, which has no relevance to events
falling within CK’s review but if during mediation an individual complained
of, for instance, the identical issue during the period before the 2006 fix, then
the Falkirk event would almost certainly fall for disclosure as being a highly
material issue in that case.
What I advised is that in exercising control over disclosure during the
mediation scheme, CK must consider the approach it is to take with
® Misra Transcript of 14 October 2010 at pages 46G-53A,
42
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
132.
133.
134.
applicants. Mediation is not a formalised court process and the CPIA rules of
disclosure do not apply. However, the mediation process in a particular case
could give rise to the view that, mindful of its common law duties, case-
specific disclosure ought to be made, and POL and CK must keep an open
mind to this. The mediation process might even give rise to consideration of
making further general disclosure within the current review, depending on
the nature of the new information.
I recall being informed that SS is to be directly involved in the mediation
process, which adds yet another dimension of possible uncontrolled
dissemination of information and material. I have suggested that CK consider
writing up a protocol or policy document to direct this part of their task so to
ensure a uniformity of approach and to avoid too arbitrary an approach to
cases going through the mediation process. That being said, I acknowledge
that CK may have to take ad hoc decisions to cases, which fall outside the
current review. I suspect however that such cases will be rare.
Disclosure letter
In those past conviction cases where, following a full review, CK advises
that disclosure of the SS report and the Rose report be made, they send a pro
forma letter, one copy of which I have seen in the Ishaq file. In on-going
cases I suspect that CK simply sends the usual letter under section 3 of the
CPIA or makes further disclosure in light of the evidence that is known to
them, and in light of the content of any defence statement.
Focusing on the question CK has consistently posed itself, they write in the
pro forma letter, “We have formed the view that had the prosecution been
possessed of the material contained within the two reports during the
currency of the prosecution of your client, we should and would have
disclosed that material to you in compliance with our disclosure duties”, and
43
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
135.
136.
137.
138.
they inform the recipient that they have reconsidered their duties under the
CPIA 1996, the Code of Practice and the Attorney General Guidelines.
As I have observed above, although not strictly accurate, I do not think that
the application of the test under the CPIA etc is necessarily misplaced in
post-conviction cases. However, because technically the duty is at common
law and not derived from the provisions of the CPIA, the reference to the
CPIA might be regarded as otiose. In conference I suggested that the letter be
revised to include a reference also to POL’s duty to consider the disclosure
of material, which might cast doubt on the safety of the conviction. I have
seen from recent examples I have been sent that that is now done.
Gareth Jenkins and his impact on possible appeals
In his Advice dated 15 July 2013, Simon Clarke advised about the future use
of Gareth Jenkins as an expert witness in support of POL prosecutions. In the
body of the Advice document, Mr Clarke sets out the relevant principles
regarding the duty of an expert.**
That Jenkins is an expert on Horizon issues cannot be seriously doubted. His
independence was unsurprisingly called into question in cross-examination in
Misra when it was suggested to him that he was employed by FSL who were
under contract to POL.*
Although there is no impediment to an expert giving evidence even where he
is employed by the party calling him, there is an inevitable risk of challenge
to an expert like Gareth Jenkins when he is being called to give evidence
about his area of expertise, especially, when, as here, he is employed by the
company manufacturing and supporting the integrity of the IT system he is to
Paragraphs 6-9
“ See Misra Transcript of 14 October 2010 at page 60B-C; and see pages 61A-C and 64E
44
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
139.
140.
141.
142.
give expert evidence about, and his employer is contracted to the very
organisation prosecuting the case in which he is a witness.
But a conflict of interest on the part of an expert witness does not operate so
as automatically to disqualify him from giving evidence in the proceedings;
the key question is whether his expression of opinion is independent of the
parties and the pressures of litigation. It is the independence and objectivity
of his opinion that is important.
Such a conflict will not infringe the requirement in Article 6 of the European
Convention on Human Rights for an “independent and impartial tribunal”,
which is concerned with the integrity of the tribunal, and does not require
that the same tests of independence applicable to a judge be satisfied by an
expert witness.
Of serious consequence is where the expert fails in his overriding obligation
to provide an unbiased and objective opinion, as provided by Part 33 of the
Criminal Procedure Rules 2013, or fails to disclose material within his
knowledge, which might cast doubt on the correctness of the prosecution
case or might assist the defence case.
Of grave concem is that Mr Jenkins informed the SS inquiry of the two
defects, which they reported at section 6 of the report, suggesting that he
knew of them in a period Mr Clarke argues in his 15 July 2013 Advice to be
between 5 October 2012 and 3 April 2013, which are the dates of essentially
five Jenkins witness statements Mr Clarke sampled. Yet in none of them is
there to be found any reference to those two system defects.*° On the
contrary, his reports speak to the general integrity of the system.
*S Which, according to Simon Clarke in his 8 July 2013 and 2 August 2013 Advices, are technically
characterised as “B14” and “B63”, although the numbering and their precise meaning remain obscure
45
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
143.
144.
145.
146.
It was during a telephone conference CK representatives had with Mr
Jenkins on 28 June 2013 that Mr Jenkins informed CK that it was he who
had informed SS of the two defects.*°
Support for the proposition that Mr Jenkins was aware of integrity issues as
long ago as February 2013 is found in Helen Rose’s report of 12 June 2013
in which she sets out an exchange of email correspondence between her and
Mr Jenkins regarding the Lepton sub-post office reversal issue. In an email to
him dated 13 February 2013, she says, “I know you are aware of all Horizon
integrity issues ...” In isolation, this may not mean much, but coupled with
the fact that it was Mr Jenkins who furnished the information about the two
defects to SS, it lends itself to the reasonable interpretation that his true level
of knowledge about the integrity of the system in general, and two defects in
particular, was far greater than he was prepared to reduce to writing in his
several witness statements during the material period of time.
Of course, it may be argued, possibly correctly, that the two defects that Mr
Jenkins has divulged to SS do not amount to very much in terms of the
overall integrity of the system, and why otherwise divulge them to SS if his
intention had been to suppress them? But that would be to misunderstand the
purpose of disclosure and the expert’s duties. It is not for the prosecutor to
make qualitative judgments about the utility or indeed the admissibility of
the information to be disclosed, and it is not for the expert to make secretive
judgments about such issues. So long as the material passes the test for
disclosure it must be disclosed.
I agree that had this information been known about at the material time as
regards any defendant who had properly raised the issue, it would, at the
least, have been considered for disclosure, and in all probability disclosure of
some kind would have followed. Because Mr Jenkins did not reveal his
“ Paragraph 30
46
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
147.
148.
149.
knowledge to the prosecution no consideration to the question of disclosure
could be made.
Mr Clarke concludes that all this means that Mr Jenkins was in breach of his
duty as an expert, that his credibility as an expert is accordingly fatally
undermined, and that he could no longer be relied upon to give expert
evidence. The consequence of this advice has been to lead to POL
commissioning the current CK review of past and pending or current
prosecutions, and indeed my review of CK’s strategy and process.
I am not clear whether Mr Jenkins was challenged about the non-disclosure
to POL and, if so, what the explanation was for it. But given the SS inquiry,
based in part on his revelations, has led to the current review, Gareth Jenkins
is to that extent tainted and his future role as an expert is untenable. It should
be remembered that POL had been unaware of the existence of the second of
the two defects revealed to SS by Mr Jenkins until a year after its first
occurrence.*”
As yet no new expert has been identified, far less appointed, to replace Mr
Jenkins, who is and was uniquely placed to give evidence about Horizon,
which is an unhappy state of affairs about which little can be done. The
Jenkins problem, even when a new witness is found, may not be at an end,
because he will doubtless still remain employed by FSL, yet not be asked to
report on, or be called by POL as a witness to speak to, Horizon’s integrity.
How much real capital may be made of the fact that Mr Jenkins will always
be a background figure in the Horizon story is impossible to predict. But
what I think I can predict with a degree of confidence is that in the hands of
capable counsel, more is bound to be made of the non-disclosure issue than
the mere instruction of a new expert will resolve for future trials.
“7 SS report paragraph 6.7
47
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
150.
151.
152.
lam asked to consider the impact of Gareth Jenkins on possible appeals. As I
have pointed out above, he has only ever given evidence once in any case,
and that was the pre-Horizon Online case of Misra, whose case is outside the
scope of CK’s review. However, he has made witness statements in several
cases. I have only seen and read two only of his reports and those were for
the Ishaq case; one was his principal report and the other was supplementary
report. For the remainder, I have taken my knowledge of the general nature
of his principal case reports from Simon Clarke’s analysis in his Advice of
15 July 2013 of the sample number of Jenkins reports ranging between
October 2012 and July 2013, which I assume to be a sufficiently
representative cross-section.
Given the current review is predicated on the basis of Horizon non-
disclosure, should any case go to the Court of Appeal on the basis that the
conviction or the plea of guilty was unsafe in light of the non-disclosure, the
Court may require some investigation of the circumstances behind the non-
disclosure, which could require Mr Jenkins to provide a full explanation for
not mentioning the two defects he revealed to SS for the purposes of their
inquiry, and any other undisclosed issues that ought to have been revealed as
relevant to any issues raised in the appeal.
For the purposes of any appeal, or of any application for permission to
appeal, the Court has the power to “order the production of any document,
exhibit or other thing connected with the proceedings, the production of
which appears to them necessary for the determination of the case.” The
Court is also empowered for the same purposes to “order any witness to
attend for examination and be examined before the Court, (whether or not he
was called in the proceedings from which the appeal lies).”**
*S Section 23(1)(a) and (b) of the Criminal Appeal Act 1968
48
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
153.
154.
155.
156.
157.
It would also be open to any such applicant or appellant to instruct an expert
of their own to support the appeal, and to seek to introduce the report as fresh
evidence.”
In such cases the Court of Appeal in arriving at the decision whether the
conviction was unsafe will ask itself “whether there is a real possibility that
the jury would have arrived at a different verdict had the necessary disclosure
50
been made.””” Moreover, non-disclosure can lead to the quashing of a
conviction even where there has been a plea of guilty.*'
POL will simply have to confront any potential appeal cases as they arise on
their own facts and on a case-by-case basis. There is no one-size-fits-all
approach to cases, and POL will have to be mindful that any concessions
made in one case may have a knock-on effect to others.
Finally, there is one further curiosity over which I have puzzled; this is
paragraph 6.10 in SS’s report, which reads, “POL has informed us that it has
disclosed, in witness statements to English Courts, information about one
other subsequently-corrected defect or “bug” in the Horizon software.” I am
unclear if this is a reference, albeit an inaccurate one, to the Helen Rose
report or perhaps a reference to Gareth Jenkins’ witness statement in the case
of Grant Allen where he conceded there had been an unusual “non-polling
event” at Winsford Post Office over a 12 day period.
(b) The response to the Criminal Cases Review Commission
(“CCRC”)
Ms Vennells the Chief Executive of POL received a letter from Mrs Berlin of
the CCRC of 12 July 2013 enquiring about the number of criminal
® Section 23(1)(c) of the Criminal Appeal Act 1968
® McInnes (Paul) v HM Advocate [2010] HRLR 17, SC
5! Rv Smith [2004] EWCA Crim 1626 (where the information could have had a causative effect on an
abuse of process argument)
49
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
158.
159.
160.
ection
% Section
* Section
convictions that might be affected by the Horizon issue they had become
aware of through media reports and what POL’s proposed course of action
was. She noted that the Attorney General had been asked to set up an
enquiry, and the CCRC had been in touch with his office.
The CCRC has a wide jurisdiction and can even refer cases without
application to it.°? The test the CCRC has to apply in making a reference is
whether they consider that there is a “real possibility” that the conviction
would not be upheld on account of some argument, or evidence not raised in
the proceedings which led to it, and there must have been an appeal, or an
application for permission to appeal, which has been refused. Failing these
conditions, the CCRC may still make a reference if there appear to be
“exceptional circumstances” for so doing.**
The CCRC may also refer a conviction from the Magistrates’ Court to the
Crown Court whether or not the person pleaded guilty.**
A response under the authorship of Susan Crichton, General Counsel, was
sent to the CCRC on 24 July 2013, informing them that the case was under
investigation. As I understand it, this was followed up by Ms Crichton in a
letter dated 26 July 2013 setting out the steps POL had thus far taken, which
included informing them of my appointment and my terms of reference in
broad terms, as well as the instruction of CK to review potentially affected
cases. Mrs Berlin responded on 30 July 2013, informing POL that the CCRC
had not itself identified any impacted cases, and expressing her contentment
to await my review before seeking further details. A holding response to the
CCRC may have to be considered in due course.
14 of the Criminal Appeal Act 1995
13 of the Criminal Appeal Act 1995
11 of the Criminal Appeal Act 1995
50
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
161.
162.
163.
164.
After consideration, my remit was not broadened to encompass advice upon
the “safety” of any of the Horizon-based convictions, and CK’s review has
essentially been limited to the review of the question of disclosure in past
convictions and in present and on-going prosecutions. Indeed, paragraph 2 of
Appendix 2 to the “Initial Sift Protocol” document states, under the heading
“Full Review”, “It is not necessary to consider whether or not a conviction
may be said to be ‘safe’ — that is a decision for the Court of Appeal if the
case gets there.”
As I have pointed out above, despite this, CK has terminated four pending
cases and, in the course of full reviews of past indictable cases, where CK
has advised disclosure must be made, CK has tended also to advise on its
likely stance to any application for permission to appeal, or to any
substantive appeal, should permission to appeal be granted. That amounts to
consideration of the safety of the conviction.
Inevitably, the issue of non-disclosure does affect the safety of a conviction,
ot the “reliability” of it, as Simon Clarke put it in his 8 July 2013 Advice.**
Inevitably, and sensibly also, in reviewing the case and the issue of non-
disclosure, counsel has been asking the question whether the conviction is
arguably not unsafe in light of all the other facts of the case, as informing
their likely stance to any future appeal process; in other words “whether there
is a real possibility that the jury would have arrived at a different verdict had
the necessary disclosure been made.”
Against that background, it seems to me that there may have to come a time,
if the CCRC maintains its interest, when POL, through CK, feels bound to
share CK’s review findings with the CCRC and cooperate with it. POL’s
approach and reaction at that stage will be better informed by the number of
55 Paragraph ii.(iv)
51
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
165.
166.
167.
cases, which have been fully reviewed and their outcome, and it will have a
better understanding of the scale of the problem.
If the CCRC was minded to consider and refer any cases, then the matter
would be out of POL’s hands. However, I emphasise that no reference may
be made by the CCRC unless there has been an appeal, or an application for
permission to appeal, which has been refused, which, unless there appears to
the CCRC to be “exceptional circumstances” for making a reference, must
limit dramatically the number of Horizon-based convictions that would be
amenable to CCRC intervention.
Recently, I was made aware of POL’s consideration to appointing Sir
Anthony Hooper to the Chair of the mediation working party. I was sent the
notes of his interview with POL on 24 September 2013. I noted in particular
that, at paragraph 4c, he suggested (quite firmly) that it might be more
appropriate for cases that have been through the courts to be referred to the
CCRC rather than go through the mediation scheme. It seems the mediation
process and CK’s review was explained to him, and it may be that his initial
suggestion was tempered by the explanation that was given.
I should add that during my conference with POL on 9 September 2013, I
made my views clear about permitting persons who had been convicted of
crime against POL to engage in a mediation process with POL. I thought
there lurked real dangers in it. But I understood the policy reasons for it. If a
policy decision has been taken to permit those convicted of crime against
POL to participate in the mediation process, then there is no case to refer
convicted cases wishing to engage in mediation to the CCRC. I simply
reiterate my advice above that POL through CK must exercise a measure of
control over the dissemination of further information and material to guard
against participants using the process as a platform to launch an appeal out of
time.
2
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
168.
169.
170.
171.
172.
(c) The identification of flaws in the review process
As I have stated, I have now seen and read 31 full reviews, albeit all 31 were
not past prosecutions, and not all of those that were past prosecutions raised
Horizon issues requiring disclosure, as is clear by the decisions on disclosure
that was made by counsel reviewing them. However, I am satisfied that I
have seen for myself a statistically significant number of past prosecutions
where Horizon was deemed to be an issue so as to determine whether or not
the review process is flawed in any way.
Except in the cases of Khayyam Ishaq and Lynette Hutchings I have not seen
the original case files, but I have seen the full case review documents written
up by Simon Clarke and Harry Bowyer.
As observed in my interim review,*° CK’s approach to the review process is,
as I have concluded, fundamentally sound. Thus, I have not detected any
systemic or fundamental flaws in the process of review, or in any of the
evidence arising from it.
From what I have seen, CK has initiated an organised and efficient regime of
disclosure, which is being considered at an appropriate level within CK, and
it is addressing the right tests.
The decisions made by counsel are inevitably fact-specific and are
necessarily subjective judgments. I do not pretend that I would always have
arrived at the same conclusion on all cases, but that is none to the point. So
long as POL through CK recognise it is their continuing duty in pending
cases to keep disclosure under review in light of new or different
information, then POL will have complied with its duty, subject always to
56 Paragraph 15
53
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
173.
174.
175.
176.
the intervention of the court. The same must apply to their consideration of
past prosecutions, as I have suggested above.
Although this is no answer to POL’s duty of disclosure, the practical reality
is that anyone who wants access to the SS report can gain it through POL’s
website. The Helen Rose report adds very little, it seems to me, other than to
point to a particular issue at Lepton, and the implication from the report that
as early as February 2013 Gareth Jenkins was aware of integrity issues with
Horizon, none of which he revealed. The SS report and the Rose report are
the limit of the current disclosure regime. I am unaware of anyone being
provided with anything more than this.
The threshold that CK has set itself has been set deliberately very low, and
CK must be prepared to reconsider disclosure decisions as and when it is
appropriate to do so, as it must as regards its stance to possible appeals.
Thus, the review must remain an on-going process in which POL and CK
keep their mind open to reconsidering its past disclosure decisions, as it is in
second sifting the cases, which have gone through a first sift but not to a full
Teview.
Although CK points out that it is unconcerned with the question of the safety
of convictions, there is an inexorable link between the disclosure decisions it
makes and the view it might take towards possible appeals, based on its view
of the strength overall of the other evidence in the case. It is right to observe
that even where there has been non-disclosure in a given case that does not
mean that any appeal based on it is likely to succeed. But CK must not adopt
any over-rigid or overly robust approach to any possible appeals, and should
be prepared to adapt to the circumstances of individual cases.
Finally, I make two further recommendations:
54
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581
LEGALLY PRIVILEGED AND CONFIDENTIAL
qd) That there should be a reconciliation of all SS’s spot reviews and
those CK cases that have been subject to sift and full review, in
order to check that there is no factual or other inconsistency between
them as would affect CK’s review. (I have been told that although
there are around 30 SS spot reviews, only about 12 were ever sent to
POL); and
(2) If there is any Horizon-related civil litigation between any present or
former sub-postmaster and POL related to any of the Horizon issues
that are the subject of CK’s review, then CK should be given
complete visibility of the litigation, in case this affects any decisions
they are making about Horizon criminal cases.
Brian Altman QC
2 Bedford Row
London WC1R 4BU.
baltman@2bedfordrow.co.uk
15 October 2013
Fe)
LEGALLY PRIVILEGED AND CONFIDENTIAL
POLO00006581
POL00006581