Read the Article: The Post Office Made Them Pay
Faulty accounting data became demands for real money. The people who challenged the system then had to fight for the means to recover what it had cost them.
Case file
The Record: Post Office Horizon
Companion to: The Post Office Made Them Pay
Record reviewed: September 20, 2026
Scope: Historical court findings, Volume 1 of the public inquiry's final report, the government's published response and dated redress figures. This is a documentary record assembled from published primary sources. It contains no new interviews.
What the documents establish
The Post Office required operators to meet apparent shortfalls generated by Horizon and used Horizon data in criminal prosecutions. In the cases defined as Horizon cases in the 2021 Hamilton appeal, the Court of Appeal found no independent evidence of an actual shortfall, serious failures of investigation and disclosure, and an effective reversal of the burden of proof. Thirty-nine appeals succeeded on both grounds of abuse of process. Three appeals failed because Horizon reliability was not essential to those prosecution cases. Hamilton judgment
The 2019 Common Issues judgment addressed the contracts and the treatment of disputed trading statements. It did not determine individual damages. The 2025 inquiry report separately addressed human harm and redress, including the inadequacy of the group litigants' recovery compared with their losses. Common Issues judgment; Inquiry report
The government reports substantial redress payments. The payment total, the number of people receiving any payment, accepted offers and administratively settled claims are different measures. The July 2026 figures must be read with their definitions. Government redress figures
Chronology
Horizon pilot followed by rollout to branches | H, paragraph 9
Hughie Thomas pleads guilty to false accounting and is sentenced to nine months' imprisonment | H, paragraphs 149 to 155
Civil judgment against Lee Castleton, followed by substantial legal costs | V, paragraph 3.215
Hamilton enters fourteen false accounting pleas and later receives a community sentence | H, paragraphs 142 to 148
Justice for Subpostmasters Alliance formed | V, paragraph 1.12
Ismay report addresses system integrity and implications of independent review | H, paragraphs 24 to 25
Common Issues judgment delivered | B
Major group litigation judgments and settlement | V, paragraphs 1.12, 4.3 to 4.7 and 6.130
Horizon Shortfall Scheme opens | V, paragraph 4.3
Thirty-nine convictions quashed in Hamilton appeals; three appeals dismissed | H, paragraph 447
Additional redress funding announced for group litigants | V, paragraph 4.7
Separate GLO redress scheme launches, a year after announcement | V, paragraph 4.7
Remaining Overturned Convictions claims move to HCRS following closure of the former scheme | D, notes to table 1
Inquiry publishes Volume 1 | V
Government publishes response to Volume 1 | G
Chair explains that remaining report volumes require further work and responses to proposed criticism | U
Cutoff for redress figures used in the article | D
Source index
| ID | Primary source | Date and pinpoint | What it supports | Limits |
|---|---|---|---|---|
| H | Hamilton and others v Post Office Ltd, full judgment | April 23, 2021; [2021] EWCA Crim 577 | System operation, disclosure failures, prosecution conditions, judicial findings and the forty-two appeal outcomes | These are the cases before that court. Their exact findings do not extend to every Post Office prosecution. |
| B | Bates and others v Post Office Ltd, Common Issues judgment | March 15, 2019; [2019] EWHC 606 (QB); especially paragraphs 1111 to 1122 | Contract interpretation, good faith, unreasonable terms, disputed accounts and suspension | Paragraph 1112 expressly reserves breach, causation and loss. This is not the later Horizon Issues judgment. |
| V | Post Office Horizon IT Inquiry, Final Report Volume 1 | Published July 8, 2025; evidence considered through June 6, 2025 | Human impact, scheme development, redress assessment, recommendations and the Castleton material | Historical cutoff. Not a criminal conviction or determination of civil liability. Later volumes were still being prepared in the chair's July 2026 update. |
| G | Government response to Volume 1 | First published October 9, 2025; updated February 12, 2026 | Official explanations, acceptance and rejection of recommendations, stated changes to redress | An accepted recommendation or announced measure is not independent proof of successful delivery in every case. |
| D | Financial redress data as of July 31, 2026 | Updated September 11, 2026 | Payment totals, claim stages, definitions and scheme targets | Includes interim payments. Settled can include accepted offers awaiting payment and specified other closures. Monetary components are rounded. |
| U | Chair's progress update | July 8, 2026 | Status and intended structure of remaining five report volumes; response process for proposed criticism | Does not reveal unpublished conclusions or establish individual criminal liability. |
These are outbound source links. Unchecked does not host or independently authenticate these files.
Financial snapshot
All figures below are from the July 31, 2026 government dataset. Payment values are rounded to the nearest £1 million. Rounding means the displayed scheme components need not sum exactly to the displayed overall total.
| Scheme | Paid to date | Full claims received in headline table | Claims classified as settled |
|---|---|---|---|
| Horizon Shortfall Scheme | £1,026 million | 14,609 | 12,260 |
| Group Litigation Order scheme | £235 million | 486 | 457 |
| Horizon Convictions Redress Scheme | £319 million | 535 | 505 |
| Horizon Shortfall Scheme Appeals | £17 million | 120 complete claims, reported separately | 26 |
| Closed Overturned Convictions scheme | £68 million | No directly comparable active headline cohort | See historical scheme data |
| Government aggregate | Approximately £1,666 million | Rows are not unique people and are not summed | Rows are not unique people and are not summed |
Definitions: Paid includes interim redress. Settled includes accepted offers paid or soon to be paid. HSS additionally includes ninety-four alternative closures, of which twenty-four subsequently settled through HSSA. Appeals and original claims can refer to the same claimant. Branch claims and individual fixed offers can also differ in unit of count. Government definitions
Remaining HSS stages: 1,023 offers awaiting acceptance and 1,326 claims under assessment. Neither stage alone identifies the reason for the delay or establishes wrongdoing.
Two different uses of 492: The GLO scheme covers 492 original litigants who were not convicted; 457 had settled in that scheme. Separately, 492 of all 555 original litigants had settled across relevant schemes. The two denominators are not interchangeable.
First-offer target: Ninety percent within forty working days after a complete claim. Reported performance was 88.0 percent for GLO, 88 percent for HSSA and 93.8 percent for HCRS. These measures stop at a first offer and start after completion of the claim. They do not measure total time since harm or final payment.
Claim and evidence ledger
| Claim in article | Evidentiary status | Pinpoint | Wording boundary |
|---|---|---|---|
| Operators had to use Horizon and could not dispute figures inside the system | Court's description | H, paragraphs 11 to 14 | A helpline existed. There was a channel for reporting a problem. |
| Apparent shortfalls were required to be funded personally or deducted from income | Court's description | H, paragraphs 13 to 22 | The court's description does not establish that every operator used the same payment route. |
| The Common Issues judgment rejected treating notified disputed balances as settled accounts | Judicial finding | B, paragraph 1116 | Retain the qualification concerning a dispute notified to the helpline. |
| Earlier and newer contracts had materially different loss provisions | Judicial finding | B, paragraphs 1114 and 1122 | The finding does not establish that all contractual terms were void or that all losses were irrecoverable. |
| Fujitsu could amend branch data without the operator's knowledge | Finding described by appeal court | H, paragraph 16 | Capability does not prove that every disputed balance resulted from remote alteration. |
| The Post Office could obtain relevant supplier information | Court's description and conclusion | H, paragraphs 18 and 131 | Cost did not excuse disclosure failures. No price per request is established. |
| Hamilton's investigator reported no evidence of theft | Case-specific judicial account | H, paragraph 145 | Use the precise evidentiary statement. No claim that every employee knew an exact software cause. |
| Repayment and restrictions on criticizing Horizon accompanied Hamilton's plea arrangement | Judicial account and Post Office concessions | H, paragraphs 142 and 145 to 148 | Court said the arrangement gave the impression that prosecution was used to enforce repayment. Attribute that assessment. |
| Thomas faced a plea condition endorsing Horizon | Judicial account and concessions | H, paragraphs 149 to 155 | His conviction was quashed despite the earlier guilty plea. |
| Thirty-nine of forty-two Hamilton appeals succeeded | Disposition | H, paragraph 447 | Not all forty-two succeeded, and not all historical prosecutions were decided here. |
| The Post Office effectively reversed the burden of proof | Judicial conclusion | H, paragraph 137 | Applies to the Horizon cases defined in the judgment. |
| Settlement totaled £57.75 million, but claimants' receipts were inadequate compared with true losses | Inquiry conclusion | V, paragraphs 4.6 and 6.130 | No invented uniform payout or unsupported division of settlement proceeds. |
| Castleton's civil principal and agreed legal costs were £25,858.95 and £270,995.78, plus interest | Inquiry's historical account | V, paragraph 3.215 | This is a civil case. Volume 1 did not decide his proposed fraud claim. |
| Costs exceeded principal by more than ten times | Article's arithmetic | £270,995.78 divided by £25,858.95 is approximately 10.48 | Label as comparison, not a separate finding. |
| Government broadly accepted advice recommendation but interpreted timing and route | Official response | G, recommendation 4 | It is not an identical implementation of the inquiry proposal. |
| Government rejected recommendation to close existing HSS dispute procedure | Official response | G, recommendation 13 | Include the government's claimant-choice explanation. |
| Approximately £1.666 billion had been paid by July 31, 2026 | Administrative report | D, overview and section 1 | Includes interim payments and all specified schemes. Not total proven loss or total final awards. |
| Remaining inquiry volumes were under preparation in July 2026 | Chair's dated statement | U |
Linked reporting
What this record does not establish
- It does not reconstruct every erroneous transaction or prove that every Post Office shortfall was fictitious.
- It does not establish that every Horizon-related conviction was quashed in the Hamilton judgment.
- It does not decide individual criminal liability for Post Office or Fujitsu personnel.
- It does not establish the present legal disposition of Castleton's proposed fraud proceedings.
- It does not independently audit the government's redress data or measure whether each claimant's payment was adequate.
- It does not verify the operational status of every reform or every later family-redress measure.
- It does not contain fresh replies from the Post Office, Fujitsu, government, Hamilton, Thomas or the Castleton family.
Published institutional positions included
The article includes the Post Office's concessions as recorded in the Hamilton judgment and notes that three appeals failed. It includes the government's explanation for its approach to legal advice, its rationale for retaining a choice of dispute routes, substantial reported redress payments and performance exceeding the first-offer target in HCRS. The inquiry's favorable historical assessment of prompt HCRS payments is also included.
These are published positions and findings. They are not responses to UNCHECKED. Current comment has not been sought for this article. No inference should be drawn from that absence.
Method
The article combines findings in published judgments, the inquiry's own assessment, official policy responses and government data with clearly identifiable financial analysis. It does not reconstruct scenes, private conversations or unreported motives. Monetary comparisons and reporting proposals are the author's analysis, not additional judicial findings.
Court paragraph references and inquiry paragraph references allow readers to locate the relevant passage. Government and inquiry material is published under the Open Government Licence where indicated by the originating source. Source availability does not authorize the use of unrelated photographs or third-party illustrations.