WITNO9890200
WITNO09890200
Witness Name: Simon Recaldin
Statement No.: WITN09890200
Dated: 13 July 2023
POST OFFICE HORIZON IT INQUIRY
Second Witness Statement of Simon Recaldin in the Post
Office Horizon IT Inquiry
1. 1, Simon Recaldin, of 100 Wood Street, London, EC2V 7ER, say as follows:
Introduction
2. Except where I indicate to the contrary, the facts and matters contained in this
witness statement are within my own knowledge. Where any information is not
within my own personal knowledge, I have identified the source of my
information or the basis for my belief. In particular, my knowledge contained
within this statement is informed by Peters & Peters Solicitors LLP (“P&P”), who
are instructed by Post Office Limited (“POL”). I have been assisted in preparing
this witness statement by Herbert Smith Freehills (‘HSF”) who represent POL in
the Post Office Horizon IT Inquiry (the “Inquiry”), and P&P, who advise and
support POL and HSF in respect of the Inquiry. The facts in this witness
statement are true to the best of my knowledge and belief.
3. I am employed by POL as the Remediation Unit (which was previously known
as the Historical Matters Unit and was sometimes also referred to as the
Historical Matters Business Unit) ("RU") Director, a role I have held since 10
January 2022. My role includes managing the delivery of the Horizon Shortfall
Scheme (previously known as the Historical Shortfall Scheme), criminal appeals
process and consequent civil liability and compensation issues, internal and
external communications and oversight of all matters relating to historic matters
WITNO9890200
WITNO09890200
(save for the Inquiry itself). The RU was established to centralise compensation
and criminal appeal related work emerging from the outcomes of the Group
Litigation Order judgments and settlement under my role as Director of RU. As
RU Director, I chair the Historical Matters Committee ("HMC") which reports into
the Remediation Committee (“RC”) (a sub-committee of the Board and
previously known as the Historical Remediation Committee), and together, we
are responsible for decision making in respect of RU within scope of the
delegated authority from the POL Board. I brief POL's Group Executive and
Board on matters considered by the HRC. Since 9 June 2023 my line manager
has been POL's Chief Executive Officer.
. This witness statement has been prepared in response to a request made by
the Inquiry pursuant to Rule 9 of the Inquiry Rules 2006 dated 23 June 2023
("Rule 9 Request") (“Request No. 47”). This statement addresses information
that POL has previously provided in response to various requests for
prosecution and conviction statistics, including information provided during the
Hamilton proceedings, the letter sent by HSF dated 15 December 2021 in
response to the Rule 9 Request dated 3 December 2021 (“Request No. 6”) (the
"Request No. 6 Response"), and five responses to requests made under the
Freedom of Information Act 2000 (“FOIA”)', and contains updated statistics
based on the available data as at 12 July 2023. This statement is based on my
understanding and belief, following an analysis of the requests, responses, and
information provided to me by P&P.
. Where I refer to specific documents in this statement, copies of those documents
are exhibited to this statement (as set out in the index below).
Request No. 47
6. Request No. 47 asked POL:
1 FO1201608250921 (29 September 2016), FOI2020/00186 (22 May 2020), FOI2021/00353 (20 May
2021), FOI2022/00304 (25 May 2022) and FO1I2023/00096 (24 March 2023). The requests and
responses are set out in full from paragraph 35 below.
WITNO9890200
WITNO09890200
(i) to set out in a statement the substance of the Request No. 6 Response
regarding prosecution and conviction statistics (addressing both the
statistics provided and the definition and limitations that apply);
(ii) to confirm whether the Request No. 6 Response and its responses to the
five FOIA requests are consistent with one another, and if not, why not;
(iii) I to confirm whether the responses detailed at paragraph 6(ii) are
consistent with information given to the Court of Appeal on 23 March 2021
(and if not, why not);
(iv) to confirm whether the table provided in response to FOI2020/00186 is
accurate and if it is not, to provide an equivalent table that is accurate (as
far as is possible, with an explanation as to any limitations); and
(v) to confirm whether, as the table provided in response to FOI2020/00186
appears to suggest, the number of prosecutions increased significantly in
the year(s) shortly before the roll out of Horizon or whether the number
increased after the introduction of Horizon.
Summary
7. POL has been asked on a number of occasions and in a number of contexts to
provide statistics in relation to the numbers of its historical prosecutions and/or
convictions. This has included providing information in response to FOIA
requests’, in the Hamilton & Others proceedings (the “Hamilton
Proceedings”) and to the Inquiry.‘ I will refer to these requests collectively as
“the Information Requests”.
8. The responses POL has given in response to each Information Request were
based on the data available at the time, and bar one, subject to caveats about
the completeness and reliability of the information provided.
2 This witness statement concerns the FOIA responses listed at footnote 1 above.
3 The information given to the Court of Appeal on 23 and 25 March 2021.
4 The information provided to the Inquiry as part of the Request No. 6 Response.
WITNO9890200
WITNO09890200
9. I understand that the information provided has derived from casework
spreadsheets that were manually created and maintained by the POL Security
Team during the period when prosecutions were taking place and, since 2020,
have in some, but not all, cases been verified by P&P for the purposes of the
Post-Conviction Disclosure Exercise (“PCDE”).
10.The figures provided have varied over time. I have set out in the statement
where there are differences in the figures and the reasons for those differences
where the information is known. Based on the material I have reviewed in
preparing this statement, it is my understanding that the differences in figures
provided in response to comparable Information Requests are attributable to
three main causes.
a. Firstly, the parameters of the Information Requests and responses were
different. This did not cause the responses to be inaccurate but does
mean that, in some cases, the figures provided are different.
b. Secondly, as a result of the work done for the PCDE and the Inquiry,
there were developments in the extent, quality and classification of the
available data.
c. Thirdly, it appears, following analysis conducted by P&P for the purpose
of assisting with this witness statement, that in respect of the response
to FOI2020/00186 (22 May 2020) there were inaccuracies in some of the
data provided. This has been corrected and an updated table containing
data (as at 12 July 2023) for 1990 to 2020 is incorporated, at Appendix
Il. The apparent inaccuracies in the May 2020 data are not understood
to have impacted the responses given to later Information Requests, and
appear to have stemmed from the way in which the data was captured
and understood at the time.
11.It appears from the data at Appendix II, that the conviction figures increased in
the years shortly before the roll-out of Horizon, with further increases after the
roll-out. However, I understand that the pre-2000 figures and post-1999 figures
are not directly comparable and that there remain limitations in the data,
especially the pre-1999 data.
WITNO9890200
WITNO09890200
Background
12.From 2020 onwards, POL’s criminal law solicitors, P&P have provided
assistance to POL with its responses to the Information Requests owing to its
knowledge of POL’s various repositories of information as a result of its work on
the PCDE. My First Witness Statement contains a detailed explanation of the
PCDE.
13.1 understand that the statistics provided in response to the Information Requests
derive, in particular, from over 100 casework spreadsheets that were manually
created and maintained by members of the POL Security team during the period
when prosecutions were taking place, and provided by POL Security to P&P for
the purposes of the PCDE or identified by P&P during a review of POL Security’s
SharePoint site’. These spreadsheets are referred to by POL and P&P as
‘Casework Spreadsheets’.
14.1 am told that the Casework Spreadsheets were used by POL Security to record
various details about investigations between 1986 and 2020, including but not
limited to: the date the investigation commenced; branch details; suspect's
name; suspect's job title; whether there was a decision to prosecute; date of
charge; date of conviction; conviction court; and sentence details. There are
multiple different spreadsheets and versions of each of the spreadsheets, which
to some extent overlap. In the course of the collation and analysis of this data,
P&P identified instances where data in the Casework Spreadsheets was
incomplete, incorrect or inconsistent. Whilst P&P have attempted to verify and
reconcile the data, this has not always been possible.
° A copy of POL Security's Sharepoint site was collected as part of the PCDE. The spreadsheets
provided to P&P each included “Casework” within the document title. To ensure P&P had identified all
spreadsheets, they conducted a search across all material collated as part of the PCDE for
documents named “Casework” and included these within their review.
6 I am informed by P&P that these Casework Spreadsheets contain details of all investigations,
prosecutions and convictions in cases where POL (or RMG pre-separation) was the prosecuting
authority. A number of cases prosecuted by other prosecuting authorities in which POL was the
investigator do not appear to have been recorded on the Casework Spreadsheets, and there have
been inconsistencies identified in the recording of information. For example, the Casework
Spreadsheets sometimes recorded co-defendants’ cases separately but sometimes recorded multi-
defendant prosecutions as one case.
WITNO9890200
WITNO09890200
15.From February to May 2020, P&P undertook a reconciliation exercise of the
Casework Spreadsheets as described from paragraph 20 onwards of my First
Witness Statement. The purpose of the reconciliation exercise was to identify
and collate all cases that were potentially within the scope of PCDE into a single
consolidated spreadsheet. This consolidated spreadsheet was referred to as the
‘Master Search Terms List’ (“MSTL”). The reconciliation exercise extracted all
cases, post-1999, that:
a. Appeared to have resulted in a conviction; or
b. Recorded a decision to prosecute, but there was no further information
as to whether the case resulted in a conviction.”
16.P&P then categorised the extracted cases for the purpose of prioritising the
collection and review of data. The categories are set out at paragraph 21 of my
First Witness Statement.
17.P&P then commenced a document collation exercise and disclosure review for
all cases considered to be within scope of the PCDE. Throughout that review,
and by reference to the documents collated in the PCDE, P&P took steps to
verify the information in the MSTL. The verification process is a continuing
process as and when new material is identified for a case within scope of the
PCDE.
18.Cases were removed from the PCDE if it was confirmed from the review of the
underlying case papers or court records that those cases no longer met the
PCDE criteria, i.e. in circumstances where: (i) there was no conviction; (ii)
neither POL nor the Royal Mail Group (“RMG”) pre-separation was the
prosecutor; (iii) the prosecution did not rely on Horizon data wholly or in part or
(iv) the prosecution was not undertaken between 1999 and 20138.
7 This means that the MSTL records some cases which may not have resulted in a conviction, but the
actual result is unclear from the available data. To err on the side of caution these casesare included
in POL's conviction figures until they are verified as non-conviction cases.
8 POL only prosecuted three cases after 2013, only one of which (involving two defendants) relied on
evidence from Horizon. Those co-defendants’ cases, from 2015, were also included in the PCDE.
WITNO9890200
WITNO09890200
19.During the course of the PCDE additional repositories of data have been
identified. These have contained additional Casework Spreadsheets, tabs on
existing spreadsheets containing additional data of which POL had been
previously unaware and other documents such as branch files and court records
that, in certain instances, provided further relevant information.
20.New repositories of data are continually reviewed to see whether they contain
further documentation which can assist P&P in verifying the details of a case
and determining whether it is outside the scope of the PCDE, particularly in early
cases (i.e. where the prosecution took place between 1999 to 2004), where POL
holds limited information.
21.The verification process has resulted in:
a. Are-classification of some cases that were previously treated as ‘possible
conviction’ cases for the purposes of the PCDE as ‘non-conviction cases’;
b. Some cases in which initially it was unclear who the prosecuting authority
was being classified as definitively prosecuted by POL or re-classified as
being prosecuted by another agency;
c. Some cases being confirmed as unrelated to Horizon (i.e. they could not
have relied, in whole or in part, on Horizon data). For example, it became
evident during the verification process that many of the cases from 2000
and 2001 which had previously been recorded as within the scope of the
PCDE because they post-dated the roll out of Horizon, were not in fact
PCDE cases because the records showed that Horizon had not been
installed in the specific branch when the criminal investigation
commenced. These cases were therefore removed from the PCDE.
22.Case papers were also reviewed to verify conviction dates and job roles. Where
the material provided a conviction date or job role which was inconsistent with
that on the MSTL, the MSTL was updated to the verified date and role. This
meant the number of convictions per year, broken down by role, were subject to
change. Due to the absence of documentation retained by POL, RMG and the
courts, particularly in early cases, it was not always possible to verify these
details. In particular, in some cases, the date of conviction could not be verified.
WITNO9890200
WITNO09890200
23.The PCDE did not include pre-1999 (i.e. pre-Horizon) cases, and therefore the
MSTL did not record information about these cases. I am told that the limited
information that does exist in respect of pre-1999 cases is held in ‘early
casework spreadsheets’. These spreadsheets overlap in time periods and there
are instances where they contain inconsistent data. For example, one
spreadsheet may provide the job role of a convicted individual as ‘SPMR’?,
however on another spreadsheet, the job role field for the same individual is left
blank, so may be recorded as ‘unknown’. There are also inconsistencies in how
cases involving co-defendants are recorded, with some spreadsheets recording
a conviction of three co-defendants as one case and other spreadsheets
recording the same case as three cases.
24. By definition, the PCDE did not include cases where it could be confirmed that
they did not result in a conviction (‘non-conviction cases’). Therefore, the
verification and reconciliation processes described above were not initially
conducted in respect of known non-conviction cases, although P&P kept a list
of any cases removed from the PCDE on the basis that they were confirmed as
not having resulted in a conviction. The data for non-conviction cases was not
initially considered sufficiently complete to rely on. However, further analysis of
the Casework Spreadsheets was conducted by P&P in advance of the Request
No. 6 Response, so that a figure for prosecutions which did not result in a
conviction could be provided to the Inquiry (subject to caveats), consisting of all
cases marked as e.g. “prosecution withdrawn” or “acquittal” in the Casework
Spreadsheets, as well as any prosecutions previously included within the PCDE
which had been removed because the underlying case documents confirmed
that they did not result in a conviction. Further reconciliation and analysis of non-
conviction cases was done in 2022 for a discrete review conducted as part of
the PCDE.'° Accordingly, there is now more data available in respect of non-
conviction cases than previously, although the data available for pre-1999 non-
conviction cases is very limited and incapable of being verified as accurate.
® This is a reference to sub-postmaster.
10 This review was referred to in the note provided to the Inquiry accompanying the letter dated 26
August 2022.
WITNO9890200
WITNO09890200
The Request No. 6 Response
25.Request No. 6 asked for confirmation of (a) the number of prosecutions POL
brought (in total) between 2000 and 2015, and (b) the number of those
prosecutions brought that resulted in a conviction.
26.As part of the Request No. 6 Response, HSF explained that they were instructed
that POL brought a total of 844 prosecutions" between 2000 and 2015, which
resulted in 705 convictions’, on the basis of the following understanding of the
scope of the request and limitations of the available data. The role of POL’s
criminal law solicitors, P&P, in assisting with its response to the Inquiry, was
explained.
27.1n relation to the scope of the request, the Request No. 6 Response explained
that POL understood Request No. 6 to be concerned with private prosecutions
brought by POL (or RMG (pre-separation) (referred to collectively as “POL
Prosecutions’) and not those brought in relation to POL matters by:
a. other enforcement authorities such as the Crown Prosecution Service
(‘CPS’), the Department for Work and Pensions (“DWP”), the Public
Prosecution Service for Northern Ireland (“PPSNI”) and the Scottish
Crown Office of the Procurator Fiscal Service (“COPFS"); or
b. other corporate entities (e.g. RMG post-separation).
28.Given the Inquiry’s Terms of Reference, it was understood that Request No. 6
concerned POL prosecutions brought (entirely or in part) on the basis of
evidence from the Horizon system and that, “Unless stated otherwise, the details
of the prosecution statistics and prosecution files have been confined
accordingly”.
‘1 “Prosecutions” is defined as including any case in which there was a decision to charge (including
acquittals and cases which were discontinued post-charge).
‘2 It is now understood that there were 700 convictions within the parameters ofRequest No. 6
(including three cases, treated as convictions for the purposes of the PCDE, which have not yet been
confirmed as convictions. The figure for confirmed convictions is 697.
29.
30.
31
WITNO9890200
WITNO09890200
It was explained that in order to respond to Request No. 6 assistance had been
provided by P&P, who had been assisting POL with appeals arising from its
historical private prosecutions, and that this had involved an extensive post-
conviction disclosure exercise covering cases where:
a. the individual was convicted of a criminal offence;
b. POL (or RMG for prosecutions relating to POL that took place pre-
separation) was the prosecutor;
c. the prosecution was based wholly or partly on data derived from the Horizon
IT system; and
d. the prosecution was undertaken between 2000 and 2013.1%
The Request No. 6 Response set out the limitations in respect of the data on
which the prosecution and conviction figures were based: the first was that the
main source of information for the data was the Casework Spreadsheets (with
the limitations described at paragraph 14 above).
. The second limitation related to steps that had been taken by P&P to verify some
of the information in the spreadsheets, in particular: whether POL/RMG was the
relevant prosecutor (or whether this was a public/other prosecutor); whether the
prosecutions were based in whole or in part on data derived from the Horizon
system; and the date of conviction. These steps included reviewing any
available underlying material and writing to the relevant court for each case to
request copies of a certificate of conviction and/or court records sheets. It was
not possible to verify these matters for all cases due to a lack of material having
been retained by POL/RMG or the courts. In particular, in some cases, the date
of conviction could not be verified.1* Where it was not possible to verify
information, it was assumed that the prosecutions fell within the PCDE.?>
'S See footnote 8.
14 I am informed by P&P that for older cases, the courts often require the date of conviction to
ascertain what records it holds and cannot search for information on convictions using wider search
terms.
‘5 Prosecutions brought in Scotland or Northern Ireland were excluded from the PCDE as P&P.
understands that POL (and RMG pre-separation) did not bring private prosecutions in those
jurisdictions. P&P have informed me that nothing has been identified by P&P in the course of their
review to suggest that this understanding is incorrect.
10
WITNO9890200
WITNO09890200
32. The letter explained that the limitations meant that the figures may not be entirely
accurate although neither POL nor P&P was aware of any specific inaccuracies.
Specifically, P&P had not identified any prosecutions that were in the scope of
the PCDE that had not been reflected in the Casework Spreadsheets.
33.As I set out in my First Witness Statement (at paragraph 27) in connection with
Potential Future Appellants (“PFAs”), the number of PFAs was subject to change
as cases were reviewed as part of the PCDE.'® At paragraph 31 of that
statement, I explained that the number had reduced to 701; I have since been
informed by P&P that it is now 700.1”
34.As the prosecution and conviction data is based on the same PCDE dataset it
is similarly liable to change. Accordingly, since the Request No. 6 Response,
there have also been changes to the prosecution and conviction statistics as a
result of the ongoing verification process. Based on the same parameters as the
Request No. 6 Response, the current data indicates that there were 700
convictions'® and 81 prosecutions which did not result in a conviction’? (i.e. 781
16 As explained at paragraph 34 of my First Witness Statement, the number of PFAs represents all the
individuals formerly convicted by POL to whom POL owed a post-conviction disclosure duty, as a
private prosecutor in England & Wales. It includes all cases, including those who have already
appealed or applied to the CCRC (whether successfully or unsuccessfully) and those whose appeals.
are pending and those who are yet to appeal. Cases are not removed from the PFA list on the basis
that an appeal has concluded or been abandoned.
‘7 The PFA number reduced to 700 because POL received information in respect of an applicant to
the Criminal Cases Review Commission (“CCRC") that their case was prosecuted by the Department
for Social Security (now the DWP) and not POL or RMG pre-separation. Therefore POL has notified
the DWP about the case, and has removed the case from its own PFA list.
18 N.B. This is the “unconfirmed” conviction figure based on the cases which are treated as possible
convictions for the purposes of the PCDE. The figure of confirmed convictions is 697, although it is not
yet clear for some of these whether they were Horizon-related cases and whether they were
prosecuted by POL (or RMG pre-separation) or another prosecuting authority.
18 The figure for prosecutions which did not result in a conviction includes all casesin England and
Wales where a decision to commence a prosecution is recorded on Casework Spreadsheets and
there is a recorded outcome of a prosecution withdrawal or acquittal. It also includes cases which
were previously included within the PCDE but subsequently removed if they were verified as not
meeting the PCDE criteria. Given some of the data has not gone through the same extent of
verification as the conviction data, the figure is not entirely reliable and is liable to change (for
example, on the basis that a case was prosecuted by a non-POL prosecutor). I understand one
reason for the change in the figure for non-conviction cases since the Request No. 6 Response on 15
December 2021 is that a significant number of cases were subsequently identified as Scottish and
Northern Irish cases.
11
WITNO9890200
WITNO09890200
prosecutions altogether) between 2000 and 2015 that may have relied on
Horizon data in whole or in part.
Responses to the other Information Requests
35.In this section, I address five FOIA responses provided by POL between 2016
and 2023 and information given to the Court of Appeal during the Hamilton
Proceedings. Where the Information Requests and responses are comparable,
I have sought to explain why the figures may have differed. These explanations
are based on my understanding and belief, following an analysis of the
Information Requests, responses, and information provided to me by P&P.
Given the period of time the responses cover, and the developments in the
PCDE during this time, and the fact I was not personally involved in preparing
each of these responses, I cannot rule out the possibility that other factors of
which I am unaware contributed to the differences set out below. All responses,
bar one, were provided subject to caveats, including in respect of the
completeness and reliability of the data.
36.Based on the information I have been given when preparing this statement, it is
my understanding that the differences in figures provided in response to
comparable Information Requests are attributable to three main causes. Firstly,
the parameters of the Information Requests and responses were different. This
did not cause the responses to be inaccurate but does mean that, in some
cases, the figures provided are different. I exhibit at Appendix I a schedule
setting out the differences in the parameters of each response. Secondly, as a
result of the work done for the PCDE and the Inquiry, there were developments
in the extent, quality and classification of the available data which resulted in
some cases being removed from the numbers at various times over the last
three and a half years. Thirdly, it appears that in respect of the response to
FOI2020/00186 (22 May 2020) there were inaccuracies in some of the data
provided.
12
FOIA Response 1 — FOIA 201608250921 (26 September 2016)
WITNO9890200
WITNO09890200
37.On 25 August 2016, a FOIA request”° included the following question: “How
many subpostmasters have had charges filed against them for false accounting
and inflating figures and or theft? Since 2010.”
38.On 29 September 2016, POL responded.’ In response to this question, POL
stated “Please find a table below containing the information you requested with
respect to prosecutions brought by Post Office Limited”. The table contained the
following information:
Financial Number of
Year Prosecutions
2010/2011 31
2011/2012 38
2012/2013 42
2013/2014 2
2014/2015 i'}
2015/2016 1
2016/2017 0
39. This FOIA response was provided before the instruction of HSF and P&P. I am
informed that the data was likely compiled by a member of the ‘Intelligence
Team’ which had inherited the workload of the Casework Team and associated
documents and databases which included Casework Spreadsheets. The person
who is understood to have compiled the data has confirmed that the source of
the data is likely to have been the Casework Spreadsheets referred to at
paragraph 13 above”. It is therefore subject to the same limitations described
at paragraph 14.
20 Post office suspensions - a Freedom of Information request to Post Office Limited -
WhatDoTheyKnow
21 Nic de gaia FOIA A Response 26 09 16.pdf (whatdotheyknow.com)
22 Given the number and variety of locations in which the totality of the Casework Spreadsheets
ultimately identified by P&P were found, it is not known whether this FOIA was compiled from some or
all of the individual spreadsheets.
13
WITNO9890200
WITNO09890200
40.1 have been unable to confirm the exact parameters and limitations of this data
with the person understood to have compiled it, although he confirmed that he
would have filtered based on the parameters of the request. A colleague of the
person who compiled the data believes it is likely to relate to: prosecutions
(including both conviction and non-conviction cases); SPMs only (i.e. not
assistants, Crown Office employees or others), all offence types (although
noting this would most commonly involve theft and false accounting); cases
prosecuted by POL (or RMG pre-separation) and not other prosecutors; and that
where relevant parameters were unknown, such cases would likely have been
excluded from the figure provided. I am told that the figures were likely provided
without caveat because the limitations in the Casework Spreadsheets had not,
at that time, been fully understood by the Intelligence Team, as this pre-dated
the GLO and the PCDE.
FOIA Response 2 — FO12020/00186 (22 May 2020)
41.On 9 April 2020, a FOIA request asked POL to supply, “by year, the number of
prosecutions of: (a) Subpostmasters, (b) their assistants or (c) Post Office
workers (eg those in Crown Post Offices) brought by the Post Office / Royal Mail
since 1990. Please also supply, by year, the number of criminal convictions
which resulted from these prosecutions’ .?3
42.On 22 May 2020, POL responded” as follows:
“We do not hold all of the information which pertains to your request. The
information we do hold is set out in the table below. However, this data is
likely to be incomplete and therefore should be treated with caution, for the
following reasons:
23 Prosecutions and convictions since 1990 - a Freedom of Information request to Post Office Limited
- WhatDoTheyKnow
24 FOIA 2020 00186 FOIA Request Response. pdf (whatdotheyknow.com)
14
WITNO9890200
WITNO09890200
1. The table below is based solely on information held by Post Office
Limited, and does not therefore reflect information (if any) that may be held
by Royal Mail, to which your request also refers.
2. The request covers a very lengthy period of time extending back 30 years.
3. Data relating to convictions between 1991-2006 in particular has been
compiled from extremely limited available material.
4. The table excludes cases where the data held by Post Office is
inconclusive as to whether the prosecution was brought by any other
organisation, or whether it resulted in a conviction.
In addition, although you have asked for information on both prosecutions
and convictions, the table contains information on convictions only. This is
because Post Office does not hold a sufficiently complete set of data for
prosecutions which did not result in a conviction, again reflecting the length
and aged period of the request.
The below table sets out the number of prosecutions brought for which we
hold records and have been able to determine that there was a conviction.
To date, we have been unable to locate any records before 1991. There are
104 cases where Post Office holds information recording that the defendant
was convicted, but not the date of that conviction, for which the year has
been listed as “Unknown”.
15
Year Postmaster I Assistant I Employee I Unknown I Total
Unknown 36 29 20 19 104
1991 2 2
1992
1993 2 2 4
1994 2 2 1 5
1995 1 1
1996 3 i 2 6
1997 5 3 2 10
1998 8 7 4 1 20
1999 25 20 4 1 50
2000 25 20 2 37
2003 36 29 FE 78
2002 27, 20 9 56
2003 25 8 18, 51
2004 is. il 7 2 38.
2005 22 i FE} 46
2006 35 i 5 Ei
2007 20 14 6 40
2008 22 15 8 45
2009 45 20 7 2 74
2010 38 i 5 54
2011 38 1 6 3 48
2012 36 5 5 1 47
2013 22 2 3 27
2014
2015
2016
2017
2018
2019
2020
Total
242
152
30
918
WITNO9890200
WITNO9890200
In accordance with its duties as former prosecutor, Post Office is conducting a post-
conviction disclosure exercise to identify and disclose material in all historical cases
relating to Horizon data, not just those reviewed by the CCRC, that might cast doubt
on the safety of convictions. Some of the convictions referred to in the table above
from 1999 onwards will be within the scope of this exercise.
We continue to make full disclosure of documents to the CCRC and will be, similarly,
assisting the appeal courts regarding the cases that have now been referred to
them.”
43.The statistics used for the response were taken from the then current version of
the MSTL spreadsheet and the ‘early case data spreadsheets’. The figures
provided related to:
16
WITNO9890200
WITNO09890200
a. Prosecutions which had definitively resulted in conviction and not
prosecutions which did not result in conviction (or where it was unclear
whether a prosecution had resulted in a conviction). At this stage, some
of the data had not been fully reconciled and verified for the purposes of
the PCDE.
b. Cases in which the prosecuting authority was believed to be POL or (for
cases pre-separation and where information was available) RMG. Cases
involving different prosecution agencies (for example the CPS, COPFS,
PPSNI, DWP) or where it was unclear who the prosecuting agency was,
were excluded;
c. Cases which appeared to be based on a shortfall in the branch accounts
and those which did not (for example robbery and burglary) because the
request did not confine itself to ‘shortfall’ / Horizon cases;
d. Cases of defendants who were sub-postmasters, their assistants, or
‘Post Offices workers’ (eg Crown Office employees), and cases where
the job title of the individual was ‘unknown’ (this is because it was
thought that unknown individuals might fall within the parameters of the
request, but individuals with job titles such as ‘engineer’ or ‘cleaner’ were
not included as they were understood not to fall within the request.25
44.Following analysis conducted by P&P for the purpose of assisting with this
witness statement, it has been identified that some of the information provided
in this FOIA response appears to have been inaccurate at the time. It was
incorrect to state that POL held no records before 1991 (there was some data
prior to 1991, albeit it was very limited), and it appears that during the
consolidation of the data, some cases that should have been included in the
table were not. For example, P&P have identified a case from 1992 which should
have been included in the table but was not.
25 Although the headings in the table referred to “postmasters” and “employees”, these were
supposed to be synonyms for the categories requested, i.e. sub-postmasters and “Post Office workers
(e.g. those in Crown Post Offices)” rather than changes to the categories requested.
17
WITNO9890200
WITNO09890200
45.This would have contributed to differences between the data sets, but I am
informed this did not impact on the accuracy of data subsequently provided in
response to later Information Requests. Certain differences in the May 2020
table and figures provided in subsequent years are attributable to the different
parameters of the responses (see Appendix I which sets out the different
parameters of each response) and the changes to extent, quality and
classification of the available data as a result of work done in connection with
the PCDE.
46.For example, in respect of the parameters, the information provided in response
to FOI2020/00186 (22 May 2020) included all offence types, whereas the
Request No. 6 Response (as explained at the time) was confined to cases where
the prosecution was based wholly or partly on data derived from the Horizon IT
system. The information provided in response to FOI2020/00186 (22 May 2020)
included confirmed conviction cases only (and cases where the conviction date
was unknown were included in a row entitled ‘unknown’), while the figures
provided in the Request No. 6 Response (insofar as it related to convictions)
adopted the approach of the PCDE and treated a small number of unconfirmed
convictions as convictions (subject to caveats about the limitations of the data).?°
Additionally, cases in which the prosecutor was unknown were not included
within the 22 May 2020 FOIA response but were included in the figures provided
in the Request No. 6 Response.
47.The extent, quality and classification of available data had also evolved between
22 May 2020 and 15 December 2021 because of the work conducted for the
PCDE and Inquiry. This included:
a. The removal of a number of cases (dating from 2000 to 2002) from the
PCDE because it was identified that Horizon had not been installed in the
branches concerned when the criminal investigation had opened;
26 I am informed by P&P that there were three cases in which it was not possible to confirm whether
they were convictions at the time of the Request No. 6 Response but they were treated as convictions
for the purposes of the PCDE and that approach was followed inthe Request No. 6 Response. These
three cases remain unconfirmed.
18
WITNO9890200
WITNO09890200
b. The response to FOI2020/00186 (22 May 2020) was prior to the steps
taken to identify the prosecutor and conviction date of many of the cases
deemed to be within the scope of the PCDE so, as explained at
paragraphs 21 and 22 above, cases in which the prosecutor was
unknown were not included within the response (although cases in which
the conviction date was unknown were included in a row entitled
‘unknown’), whereas the Request No. 6 Response, which adopted the
approach taken for the PCDE, included cases where it was unclear
whether POL (or RMG pre-separation) was the relevant prosecutor or
whether this was a public/other prosecutor;
c. Identification of further inconsistencies in the recording of the data in the
multiple Casework Spreadsheets;
d. Additional analysis done in respect of cases which did not result in a
conviction, meaning the Request No. 6 Response could take into account
these cases (for 2000 to 2015), whereas it had not been possible to
include this information at the time of the 22 May 2020 FOIA response.
48.As a result of the verification work done as part of the PCDE subsequent to the
22 May 2020 FOIA response and for the Inquiry, more reliable data is now
available but it is still subject to change as and when more information becomes
available, particularly for early cases. A table based on the data available as at
12 July 2023 is provided at Appendix II. The numbers of convictions in 1999,
2000 and 2001 (i.e. around the time that Horizon was installed in branches) have
remained broadly consistent in respect of the figures available in 2020 and the
figures included in Appendix II.
Information provided during the Hamilton Proceedings (23 and 25 March 2021)
49.In the course of the Hamilton Proceedings, on 22 March 2021 the Court asked
counsel for some of the appellants whether he wished to argue that the number
of prosecutions rose immediately after Horizon was rolled out. In response,
19
WITNO9890200
WITNO09890200
counsel, apparently relying on numbers taken from the 22 May 2020 FOIA
response, told the Court that “there were maybe two or three prosecutions pre-
Horizon, and it went up to between 40 or 50 once Horizon came in.”
50.On 23 March 2021, the following information was provided orally to the Court by
counsel for POL:
“In 1996, there were 11 prosecutions of sub-postmasters; in 1997, 22;
in 1998, 41. None of those could have relied on Horizon...In 1999,
there were 60 in total. It is not believed any were Horizon based. In
2000 once rollout had started there were six prosecutions; in 2001, 41;
in 2002, 64; in 2003, there were 56....In 2004, 59; 2005, 68; 2006, 69;
2007, 50; 2008, 48; 2009, 70; 2010, 55; 2011, 44; 2012, 50. Finally,
2013, 56."27
51.On 25 March 2021, a note was provided to the Court to explain and clarify the
figures that had been provided on 23 March 2021 and why they differed from
those referred to by counsel for some of the appellants, which he had explained
were based on POL’s FOIA response dated 22 May 2020.8
52.In summary, the note explained that the figures relied on by POL in the Court of
Appeal were based on an analysis which took account of information gained
during the PCDE. The pre-1999 figures given to the Court by POL were based
on P&P’s analysis of POL’s early casework spreadsheets at that time: these
cases had not gone through the PCDE verification process described at
paragraphs 17 to 22 above because pre-1999 cases did not fall within the
PCDE. They related only to cases which had resulted in convictions and
included cases which appeared to be based on a shortfall in the branch accounts
as well as those which did not. The note provided on 25 March 2021 also
provided the number of convictions for pre-2000 cases once offences such as
robbery or burglary had been excluded.”° This resulted in no change for 1996,
27 Counsel also caveated that the figures may not be reliable.
28 Joint Note RE: Numbers of Prosecutions / Convictions
29 This figure included cases described as ‘other’ on the basis that these may include shortfall cases.
In later responses to requests for information regarding ‘shortfall’ cases (i.e. cases which related to a
20
WITNO9890200
WITNO09890200
18 cases for 1997 (rather than 22), 31 cases for 1998 (rather than 41) and 50
cases for 1999 (rather than 60), none of which relied on Horizon.
53.The note explained that during the PCDE, POL was provided with additional
information and so was able to exclude the non-Horizon-related cases from the
post-1999 figures. In respect of some branches it was possible to ascertain the
date that Horizon was installed and exclude the conviction if it could not have
been Horizon related (i.e. as the criminal investigation had commenced prior to
Horizon being installed in branch).5° Where the date of Horizon installation could
not be ascertained, the conviction was included in the post-1999 figures. The
post-2000 figures therefore related to convictions of any individual (not just sub-
postmasters ("SPMs") which did or may have relied on Horizon data. The note
clarified that in addition to the six cases referred to for the year 2000 that ‘likely
relied on Horizon data’, there were 53 non-Horizon related shortfall convictions.
The post-1999 figures were based on the data verified for the purposes of the
PCDE (as described at paragraphs 17 to 22 above). Cases that might have
relied on Horizon evidence (whether or not the defendant was an SPM) were
included in the figures, while cases which did not rely on Horizon evidence were
excluded.
54. The conviction figure (705) provided in the Request No. 6 Response in respect
of convictions between 2000 and 2015, is lower than the figure given in the
Hamilton Proceedings for the same period (736 excluding 53 cases for the year
2000 where it had been identified that Horizon had not been installed in
branch)*! even though the parameters of the information provided were the
same. I am informed by P&P that this is because of additional information being
received and reviewed during the course of the PCDE between March 2021 and
December 2021, as follows:
shortfall discrepancy in branch accounts), offences described as ‘other’ were not treated as shortfall
cases because the cases were not considered likely to be shortfall cases.
30 The first tranche of Horizon installation dates was provided to P&P on 1 March 2021 so the number
of cases treated as Horizon-related for the purposes of the PCDE and information requests
decreased after this date.
31 Had these figures been included, the figure in Hamilton would have been 789 for the period 2000 to
2015.
21
WITNO9890200
WITNO09890200
a. lam informed that the conviction data provided for the years 2000 to 2002
and 2013 respectively were different in the Hamilton Proceedings to the
figures for those years in the data on which the Request No. 6 Response
was based as a result of the additional verification work conducted for the
purposes of the PCDE between March and December 2021. For
example, I understand that P&P received a second tranche of Horizon
installation dates in early April 2021 which resulted in a number of cases
previously treated as possible convictions dating from 2000 to 2002 being
removed from the PCDE (and the subsequent statistics provided in
Request No. 6 Response) because they were not cases which relied in
whole or in part on Horizon.
b. Additionally, it was also identified that the ‘date of conviction’ on the
Casework Spreadsheets had sometimes been recorded inaccurately by
the POL Security Team, who had been using the date of sentence as the
date of conviction, rather than the actual date of conviction. This caused
the dates of conviction to move slightly to accommodate this error.
c. There was a significant drop in the figures for 2013 because the detail
recorded on the Casework Spreadsheets was limited for this period, often
recording the fact a prosecution had commenced but not the outcome.
These cases were originally included within the PCDE (and treated as
cases that may have been convictions for the purposes of the PCDE and
the information given to the Court in the Hamilton Proceedings) but when
the case material was reviewed as part of the PCDE, many cases were
removed because it was confirmed they were, in fact, non-conviction
cases.
55. The information provided in the Hamilton Proceedings did not address any post
2013 cases because the PCDE period was 1999 to 2013. There is a 2015 case
involving co-defendants in the same prosecution®? which was included in the
table in response to FOI2020/00186 (22 May 2020) and the data on which the
Request No. 6 Response was based (it was also referred to specifically in
2 Both defendants are included in the list of Potential Future Appellants (PFAs). See also footnote 8.
22
WITNO9890200
WITNO09890200
Request No. 6 Response) as these responses were not confined to 1999 to
2013, and this case relied in whole or in part on Horizon data.
FOIA Response 3 - FOI2021/00353 (20 May 2021)
56.On 23 April 2021, POL received a FOIA request saying, “! would like to know
when the last time Post Office attempted a prosicution [sic]. I would also like to
know the last time a Sub-postmasters unwilling or unable to make good any
reported shortfalls was prosecuted.”**
57.On 20 May 2021, POL responded.** It stated that, “Post Office Limited last
prosecuted an individual in 2019. This prosecution was not branch related and
did not relate to or rely on any evidence from Horizon.” It further stated:
“Post Office Limited’s prosecution records are incomplete. It is not
possible to determine in any given case whether the person who was
prosecuted was unwilling or unable to make good a shortfall, nor when
the last such circumstances arose. The last time a postmaster was
prosecuted by Post Office Limited in relation to a shortfall identified
within their branch was 2015. Post Office Limited does not know,
however, whether any postmasters have been prosecuted in relation to
shortfalls at their branch by any other prosecuting entity (e.g. the Crown
Prosecution Service) subsequent to 2015.”
58. The caveated answers given in this response are still understood to be accurate
and are consistent with the latest information in respect of POL prosecutions. I
have been informed that a postmaster’s assistant was prosecuted by the PPSNI
in Northern Ireland and pleaded guilty in March 2017 in relation to shortfalls at
a branch (this case has since been overturned by the Court of Appeal in
Northern Ireland)*° and that there have been prosecutions in relation to shortfalls
33 When did Post Office stop prosicutions? - a Freedom of Information request to Post Office Limited -
WhatDoTheyKnow
+4 FOIA Response FOI2021 00353.pdf (whatdotheyknow.com)
% This case related to an SPM's assistant rather than an SPM and so did not fall within the
parameters of the FOIA request.
23
WITNO9890200
WITNO09890200
by the CPS post the installation of HNG-A.*° There may also be other cases
post-2015 prosecuted by non-POL prosecutors that have not yet been appealed,
and in relation to which POL has not undertaken an independent analysis to
ascertain whether they are shortfall related or not.
59.The 2019 case was not included in the table given in response to
FOI2020/00186 (22 May 2020) because the case involved an individual who
was not considered to fall within the parameters of “(a) Subpostmasters, (b) their
assistants or (c) Post Office workers (eg those in Crown Post Offices)”.°’ The
case was not included in the figures for the Request No. 6 Response because
the prosecution did not rely in whole or in part on Horizon.
FOIA Response 4 — FO12022/00304 (25 May 2022°*)
60.On 26 April 2022, POL received a FOIA request saying, “Please confirm how
many criminal prosecutions took place against Postmasters and Postmistresses
for fraud and/or false accounting in the 10 years prior to the introduction of the
‘Horizon IT Computer System’.
61.On 25 May 2022 POL responded:
“Ican confirm that Post Office holds some information in relation to this.
The Horizon IT system was introduced in 1999. For the ten years priors
to that (i.e. between 1989 & 1999), Post Office cannot confirm the
number of criminal prosecutions that took place. This is because the age
and manner in which information was recorded during this period, before
3° The version of Horizon in use since February 2017 is known as HNG-A. It is an updated version of
HNG-X which was introduced in 2010 as a replacement for the first version of Horizon (now known as
‘Legacy Horizon’). HNG-X and HNG-A are together known as ‘Horizon Online’. However, the Horizon
Issues Judgment found HNG-A to be more robust than its predecessors and so the findings in that
judgment do not apply to HNG-A at least from the date of that judgment, as explained at paragraph
964: “Findings in this judgment as to the performance and robustness of Legacy Horizon from 2000 to
2010, and then of Horizon Online (in both its forms, HNG-X and HNG-A) from 2010 to 2018 are not
findings on the Horizon system as it exists at the date of this judgment. These findings cannot be
routinely applied to the way that HNG-A operates as at December 2019. It is agreed by the experts
that the Horizon System in its HNG-A form is now relatively robust. This judgment is a historical
analysis of the Horizon System as it relates to the period in question in the group litigation, not a
judgment upon Horizon HNG-A as it is today.”
37 The 2015 case was included in the 22 May 2020 table.
38 N.B. The letter was dated 17 May 2022 but sent on 25 May 2022
24
WITNO9890200
WITNO09890200
widespread use of computerised records, means it is not possible to
verify the accuracy of Post Office's records. The surviving historical
records for pre-1999 cases are extremely limited and the information
cannot be relied on as definitive.
A search of these records provides information about 57 prosecutions
that Post Office appears to have carried out that resulted in convictions.
These were initiated after a shortfall was identified in the relevant
individual's branch which, for the purposes of this data, usually relate to
cases recorded as theft, fraud, cash loss, audit shortage or false
accounting and does not include convictions which appear not to have
relied on Horizon data in whole or in part (e.g. burglary or robberies)...”
62.1 believe that there must be a mistake in the final sentence of the response above
(‘and does not include convictions which appear not to have relied on Horizon
data in whole or part...”) because any conviction before 2000 cannot have relied
on Horizon data in whole or in part because Horizon had not been rolled out. I
assume the intention was to explain that the figure of 57 convictions excluded
cases that did not appear to arise from a shortfall in the branch accounts, even
though evidence from Horizon was irrelevant to these cases.
63. The 57 figure included convictions of individuals regardless of their specific role
(even though the request referred to SPMs) because this had been the approach
taken in response to other preceding Information Requests (including in the
Hamilton Proceedings and Request No. 6 Response), and the figure included
so-called ‘shortfall cases’ (excluding cases such as burglaries and robberies and
cases Classified as ‘other’).°°
64.It is likely that the figure of 57 (which included convictions of individuals
regardless of their specific role) was lower than the figures provided for the
equivalent time period in the response to FOI2020/00186 on 22 May 2020, and
39 Although cases categorised as ‘other’ had been included as ‘shortfall cases’ in the figures given to
the Court of Appeal in the Hamilton Proceedings on 23 and 25 March 2021, this was on the basis that
they “may relate to shortfall cases”. At the time of this FOIA response they were not considered likely
to be shortfall cases and were not included in the figure of 57. This FOIA response is not directly
comparable to the pre-1999 figures provided in the Hamilton Proceedings since those figures only
went back as far as 1996.
25
WITNO9890200
WITNO09890200
FOI2023/00096 (24 March 2023) because of the different parameters of the
responses: while the responses to FOI2020/00186 on 22 May 2020 covered all
offence types, the 57 figure was limited to shortfall cases only (excluding cases
categorised as other).
FOIA Response 5 — FO12023/00096 (24 March 2023)
65.On 2 February 2023 POL received a request saying, “Can you say how many
SPM's you prosecuted between 1990 and 1999? Can you say how many SPM
you prosecuted between 2000 and 2009? What is the % increase in the latter
figure? What % of 1990-1999 convictions were overturned? What % of the latter
figure were overturned?”*°
66.On 24 March 2023, POL responded, saying that:
“...We can confirm that Post Office does hold the information you have
requested. For the period of 1990-1999, Post Office cannot verify the accuracy
of this information due to the age and way information was recorded before the
widespread use of computerised systems. The surviving historical records for
pre-1999 cases are extremely limited, and the information cannot be relied on
as definitive.
A search of these records provides information has provided the following
information [sic]:
1. Can you say how many SPM's you prosecuted between 1990 and 1999? —
Due to the limited records (please note the caveat above regarding
accuracy), we can confirm there were 65 convictions for postmasters**
2. Can you say how many SPM you prosecuted between 2000 and 2009? 380
3. What is the % increase in the latter figure? 484% but please note caveat in
relation to limitation of records as above
40 Prosecutions of SPM by Post Office - a Freedom of Information request to Post Office Limited -
WhatDoTheyKnow
41] understand that this figure includes figures for SPMs only (and not individuals in other roles), and
relates to all offence types.
26
WITNO9890200
WITNO09890200
4. What % of 1990-1999 convictions were overturned? 0%
5. What % of the latter figure [2000-2009] were overturned? 8%
You may wish to note that more detailed information regarding historical
prosecutions is available on our website including data about the numbers
of both successful and unsuccessful appeals of cases and the current
numbers progressing through the Courts...”
67.POL was able to provide figures for prosecutions that did not result in convictions
in response to this request in respect of the post-1999 period because of the
additional work done for the Inquiry and for other discrete reviews for the PCDE
(described at paragraph 24 above). It had not been able to provide this
information in response to FOI2020/00186 in May 2020 because that work had
not yet been necessary to undertake.
68.The approach taken in responding to this FOIA request was limited to SPMs
only (and not individuals in other roles), and related to all offence types as the
request was not confined to specific types of offences. This resulted in the figure
of 65 being provided. I understand that this differed to the figure of 57 provided
in May 2022 as that figure had been confined only to cases that appeared to
arise from a shortfall in the branch accounts and related to a slightly different
time period (1989 to 1999). It is understood that the figure is different to the
figure of 47 provided in respect of SPMs during the equivalent time period in
response to FOI2020/00186 in May 2020 owing in part to the inaccuracies,
described at paragraph 44 above, in the May 2020 data.
Consistency of responses to Information Requests
69.As set out above, the figures POL has given in response to the Information
Requests have changed over time. Based on the information I have reviewed in
preparing this statement, I am informed that the differences relate to differences
in the parameters of the responses (as set out in the attached schedule),
42 FOIA Response FO12023 00096.pdf (whatdotheyknow.com)
27
WITNO9890200
WITNO09890200
changes to the extent, quality and classification of data as more material
became available during the course of the PCDE (as set out above), and certain
inaccuracies in the May 2020 data referred to at paragraph 44 above.
70.All of the responses bar one were subject to clear caveats: the figures in the
table provided in response to FOI2020/00186 (22 May 2020) indicated that the
information was “incomplete” and “to be treated with caution”; counsel for POL
in the Hamilton Proceedings caveated that the figures may not be reliable and
a note of clarification was provided to the Court; the response to FOI2021/00353
(20 May 2021) stated that “Post Office Limited’s prosecution records are
incomplete”; the Request No. 6 Response was subject to the caveats referred
to at paragraphs 30 and 31 above; the response to FOI2022/00304 (25 May
2022) stated that a search of extremely limited records, the accuracy of which
could not be verified, provided information that “should not be relied on”; and the
response to FOI2020/00096 (24 March 2023) stated that “Post Office cannot
verify the accuracy of this information due to the age and way information was
recorded before the widespread use of computerised systems. The surviving
historical records for pre-1999 cases are extremely limited, and the information
cannot be relied on as definitive.”
Updated Conviction Statistics
71.Appendix II contains the available data, as at the date of this statement, of
convictions in England and Wales (by date of conviction, where known) between
1990 and 2020. It is subject to the following important caveats:
a. The post-1999 figures are based on the criteria applied in the PCDE
because this data has, where possible, been verified as accurate. These
figures relate to cases where it is confirmed (or in some cases assumed
where it has not been possible to verify) that: (a) the individual was
convicted of a criminal offence**; (b) POL (or RMG for prosecutions
48 Three of these cases are assumed to be conviction cases unless they can be verified as non
conviction cases.
28
WITNO9890200
WITNO09890200
relating to POL that took place pre-separation) was the prosecutor“; and
(c) the prosecution was based wholly or partly on data derived from the
Horizon IT system;
b. In respect of 2000, 2001 and 2002, data for cases which did not rely in
whole or in part on data derived from the Horizon IT system has also been
provided in notes in Appendix II;
c. The pre-2000 data relates to all offence types where a conviction has
been confirmed or is assumed, prosecuted by POL (or RMG pre-
separation) where the defendant was an SPM, assistant or POL
employee, including cases in which it has been presumed, based on the
offence type, that the prosecution involved a shortfall in the branch
accounts and those where it has been presumed, based on offence type,
that the prosecution did not involve a shortfall in the branch accounts,
(e.g. robberies and burglaries, and ‘other’ cases). A figure has also been
provided in brackets with the subset of cases likely to involve a shortfall
based on the offence type*5;
d. Post-2013 there was only one POL prosecution (involving two
defendants) resulting in convictions in which Horizon evidence was relied
on in whole or part;
e. All of the data provided derives from the POL Security Team’s manually
created and maintained Casework Spreadsheets. There are over 100 of
44 There are 135 convictions for which the prosecutor has not been verified.
45 I understand that for pre-2000 convictions, it is difficult to determine the offence type as POL does
not usually hold the relevant prosecution material. Instead, whether a case is likely to involve a
shortfall in branch accounts has been inferred by reference to the categories that were historically
used on the Casework Spreadsheets to describe a case. The various categories were: Theft; False
Accounting; Audit Shortage; Cash Loss; Giro Suppression; P&A fraud; P&A overclaims; P&A
laundering; Stock loss; Allegations affecting the character and ‘Other’. The categories considered
likely to involve a shortfall in branch accounts are Theft; False Accounting; Audit Shortage and Cash
Loss. As such, these are the cases included in these figures. Convictions which are not recorded as
falling within these categories have been excluded from the figures provided on the basis that these
are unlikely to be cases involving a shortfall in branch accounts. Some cases included in the PCDE
and overturned on appeal have involved pension and allowance (P&A) frauds but I understand that
this category includes such a wide variety of potential criminal conduct that a decision was taken to
exclude it from the list of categories likely to involve a shortfall in branch accounts.
29
WITNO9890200
WITNO09890200
these spreadsheets and they are understood to contain inaccuracies and
inconsistencies, and may not be complete. While no POL-prosecuted
cases within the PCDE criteria have been identified as missing from these
spreadsheets, the possibility that case(s) have been missed cannot be
conclusively ruled out;
f. The data for pre-2000 cases should be treated with particular caution
given it has not been verified and, in many cases, is not verifiable given
the passage of time and absence of records. It should not be relied upon
as accurate or complete;
g. None of this data includes cases confirmed to have been prosecuted by
a prosecuting authority other than POL (or RMG pre-separation), such as
the CPS, DWP or RMG post-separation, PPSNI or COPFS. However, in
cases where it has not been possible to confirm the prosecutor, the case
has been included in order to avoid missing any cases;
h. Cases involving co-defendants have been separated. For example, a
case where three co-defendants appear to have been convicted is
recorded as three convictions.*¢
72.\t appears from the data at Appendix II, that the conviction figures increased in
the years shortly before the roll-out of Horizon (per the figures included in the
notes in Appendix II), with further increases after the roll-out. However, I
understand that this data should continue to be treated with caution given the
limitations of the pre-2000 data which has not been verified for the purposes of
the PCDE or the Inquiry, and given the pre-2000 data and post-1999 data is not
directly comparable.
46 Please note that there are instances where a cases involving co-defendants have been recorded on
the casework spreadsheets as one conviction, but when the underlying papers have been reviewed, it
has transpired that only one of the individuals was convicted. For the purposes of these figures for the
pre-PCDE numbers, if a conviction is recorded on the spreadsheets for a co-defendant case, it is
presumed that all individuals were convicted.
30
WITNO9890200
WITNO09890200
Statement of truth
I believe the content of this statement to be true.
Simon Recaldin
Dated: 13th July 2023
31
WITNO9890200
WITNO09890200
Index to the Second Witness Statement of Simon Recaldin
URN Document Description (Control Number
WITNO09890201 I HSF letter to the Inquiry dated 15 WITN09890201
December 2021
WITNO9890202 POL's response to I WITN09890202
FOI201608250921 dated 29
September 2016
WITN09890203 I POL's response to FOI2020/00186 I WITN09890203
dated 22 May 2020
WITN09890204 I POL's response to FOI2021/00353 I WITN09890204
dated 20 May 2021
WITNO09890205_ I POL's response to FOI2022/00304 I WITN09890205
dated 17 May 2022
WITN09890206 I POL's response to FOI2023/00096 I WITN09890206
dated 24 March 2023
WITNO9890207 I Appendix I — Schedule setting out WITNO09890207
the difference in parameter of each
response
WITNO09890208 I Appendix Il — Updated Conviction WITNO09890208
Statistics 1990 to 2020
32
WITNO9890200
WITNO09890200
Witness Name: Simon Recaldin
Statement No.: WITN09890200
Dated: 13 July 2023
THE POST OFFICE
HORIZON IT INQUIRY
SECOND WITNESS STATEMENT OF
SIMON RECALDIN
Herbert Smith Freehills LLP
Exchange House
Primrose Street
London
EC2A 2EG
Ref: 2066/31048180
33