Victims of British Postmaster Scandal set to be cleared

Miscellaneous and Personal Stuff I Must Share
21

Parliament’s Business Committee is hearing from various witnesses on precisely the subject of Post Office compensation today.

Firstly, Alan Bates, former sub-postmaster who lost the £100,000 worth of life savings he invested in his Post Office business when Horizon falsely claimed there was money missing from his accounts, and the driving force behind 25 years of campaigning against these injustices:

Post Office compensation scheme not getting any better, Alan Bates tells MPs

At the business committee three former post office operators are now giving evidence. They are Alan Bates, founder of the Justice for Subpostmasters Alliance and the person whose story formed the basis of the ITV drama that turbocharged political interest in the scandal; Tony Downey, another former post officer operator; and Tim Brentnall, another post officer operator.

Asked if there has been any improvement in the speed at which compensation is being dealt with, Bates said there had not been in his case. He said his claim was being refused.

Asked if he was happy with the way things were going, he replied:

It’s very disappointing, it’s been going on for years, and I can’t see an end to this.

He said it would be better if responsibility for handling the compensation scheme were removed from the Post Office.

In “don’t back down, double down” news, the current CEO apparently wrote to the government to say that the PO was not convinced that everyone convicted was innocent, and that there were c.350 cases where they would contest any appeal. He also claimed, somewhat in contradiction to the foregoing, that Post Office culture was changing.

But the main point of interest ws the evidence of Henry Staunton, former PO chair (but a recent one, appointed in the wake of the scandal, not the guy responsible). He was fired for reasons that remain a little murky - he says because they needed someone to take the rap for slow payments, the Minister who fired him (Kemi Badenoch) says it was because of poor governance and questions about his conduct. Whatever the rights and wrongs of that, Staunton also said that he was advised by a senior civil servant to go slow making compensation payments so the government could limp into the next election. This was in Dec 22, before the TV dramatisation that turned this into a live political issue. Badenoch of course denies he was ever told any such thing.

Henry Staunton, the former Post Office chair, is giving evidence now. He is also asked to swear on oath he will tell the truth.

Here is a summary of what he alleged in his Sunday Times interview, and here is a summary of Kemi Badenoch’s response.

Q: Where you ever told to slow down compensation payments?

Staunton says he wants to use the phrase “a nod and a wink”.

He says he met Sarah Munby, permanent secretary at th business department, in January last year.

Summarising all the problems at the Post Office, he said it would take three to five years to turn things round – probably five years.

He says Munby told him this was not a time for long-term planning, because money was tight.

He says there were only three spending levers they could change: the inquiry costs, the compensation costs, and the need for a new Horizon system.

He says he told Munby that the inquiry costs could not be changed, the compensation had to be paid, and the Post Office desperately needed a new IT system.

He says Munby told him again money was tight, and this was “not the time to rip off the band aid’.

He says this conversation was so unusual he made a note of it.

He says he discussed this with Nick Read. Read says '“they live in a different world”, he says.

Liam Byrne, the chair, is questioning Staunton.

Q: Your note of your conversation with Sarah Munby does not specifically refer to compensation?

Staunton says his note was not a full record of what was said.

Q: Kemi Badenoch says the money for compensation was ring fenced, and so the Post Office had no incentive to slow compensation payments.

Staunton says, if you read the Post Office accounts, they will show that there is not a “hard ring fence” for the compensation payments.

Q: In her record of the conversation, Munby gives a different account.

Staunton says the version published by Munby last week was written a year after the conversation. He says his note was contemporaneous.

Q: Do you think she is lying?

Staunton says he does not want to get into that. He is just setting out what happened from his point of view.

Q: Could Munby have had a different interpretation of the conversation with you?

Staunton suggests that’s unlikely. He goes on:

When you’re talking about three levers – this is not a PhD in accounting. This is three very simple issues that we’re talking about.

22

And for a bonus, remember how the Minister said Staunton was fired in part because of investigations into his personal conduct?

Staunton says it’s Nick Read who is subject to misconduct inquiry, not him, contrary to what impression given by Badenoch

Asked to explain why Kemi Badenoch told MPs that he was under investigation over serious allegations, Staunton said this related to Nick Read, the chief executive, falling out this his HR director. He said she produced an 80-page report setting out complaints about Read. He said one paragraph related to him, alleging he made a politically incorrect comment. He said this did not mean that he personally was the subject of an investigation.

He said Read found dealing with this complaint very stressful, and considered resigning.

Asked if the investigation into Read was still underway, Read said it was.

Jonathan Gullis (Con), who was asking the questions, said the committee did not realise this. He told Read he had '“made news”.

23

And this is why you ought to do this sort of stuff by email, not telephone, and end up getting into a ‘he said, she said’ thing.

24

Update-

There’s to be a new head of the Post Office. Hopefully, he can unphuck the whole mess.

25

I just finished watching the documentary on this on PBS. To say that I found it infuriating would be an understatement.

I do have a couple of probably stupid questions.

I’m not an accountant and, although I was a developer, I’ve never worked with accounting software. Wouldn’t there be a way to double check whether the “missing” money was or was not from actual sales? Couldn’t the amounts be checked against actual sales of stamps and such?

What about Fujitsu? Don’t they have any liability? Obviously the post office has a lot of responsibility in this, but Fujitsu provided the faulty software. Also, they claimed that remote access wasn’t possible while in reality they had people who had remote access to the terminals and who were constantly making “corrections” in a manner that wasnt obvious to the end users.

26

AFAIK, the “missing” money was all digital. There would be no physical cash or
postage stamps or whatever involved.

Yes.
This document gives a good account of the saga :-

27

I know that the money was digital, I didn’t mean to imply otherwise. But aren’t there records of whatever was purchased? Stamps, in particular are physical objects that must be accounted for.

So if £100 worth of stamps are recorded as purchased then £100 worth of stamps should be gone from inventory. If not tthen that should indicate that there is no missing money and there’s some sort of computer error since there hasn’t been that many stamps sold. And stamps are likely to be the.main product sold at a post office.

28

As well as providing postal services, post offices in in the UK provide a wide range of other services. They are agents for nearly all the major banks and you can transact banking business there — lodgment cash or cheques, withdraw cash, check your balance, etc. You can pay utility bills at the post office — gas, electricity, water; pay local taxes; pay rent if you are a tenant of the local government or of a housing association; buy transport tickets or top up prepaid travel cards; pay a phone bill or top up a prepaid mobile phone; buy vouchers and gift card for a variety of services and merchants — Apple, Google, Netflix; buy travel insurance; buy foreigh currency or a prepaid foreign currency debit card; apply for or renew a passport; apply for or renew a driving licence; claim a lottery prize; buy or manage your car, home, life or pet insurance; apply for a loan or an investment product; collect your social security benefits if you don’t have a bank account; etc, etc.

Not all these services are available at every post office, but the number of post offices at which the sale of postage stamps is one of the primary transactions, in terms of value, is not likely to be large.

29

Okay. That’s very different from here in the US.

30

I’ve worked in small-team software development of accounting systems and the really weird thing about the problems of distributed transactions is that developers often seem to just assume they need to invent the solution themselves, completely from-scratch, without consulting the existing body of knowledge that exists on managing the problems they’re trying to solve.

I worked for a guy who was trying to develop a multi-user system for inventory management and he just couldn’t get his head around the concept of race conditions (where two different things are trying to access and/or modify the same value).
He initially just assumed that electronic transactions were instantaneous, so it would never matter and even after I patiently explained that, even if that were true, the process he had written was something where users had to click through several steps, so it was anything but instantaneous.
His answer to that was that it was just incredibly unlikely that two different users would ever be trying to process a transaction for the same product simultaneously, so there was just no need to worry about it.
But some of our clients (users of the software) only had one product and thus, if they needed a multi-user system to process sales of that product, collisions were not only likely, they were inevitable.

I don’t know what it is that makes people think “Everyone else says this is a complex problem, but I think it’s really simple”, and never question it beyond that.

31

Software tends to be written by people who don’t use it, nor consult people who use it, and so they just imagine scenarios instead. These are inherently limited, and are either too broad or too specific to represent real world usage.

In my job I use machines that are almost, but not quite, fit for the purpose we put them to. But it means the steps we take are convoluted, limited, and have features and options that are obscure and opaque, but seem to be fundamental to how the machine was designed so can’t be avoided. We just fill in some random value and ignore it.

Anyway, that means software and hardware can cause unforeseen problems that they couldn’t predict, because they knew that if we just follow their 27 steps it would work every time.

32

Good point. Also, software is often written around the limitations of the technology used to develop it (as well as the limitations you mention - limitations of of the understanding of those doing the job).
In my last job, the development framework† for the core business software just had some weird quirks - such as: when you deleted an item from a table/array, all the higher-numbered items in the array shuffle down to close the space‡ - changing their indices in the process. This meant that if something went wrong in the process of deleting a whole product record (which required the deletion of correspondingly-indexed records from a collection of different tables), half the database would suddenly have the wrong descriptions and prices and other data.

† Unidata/Unibasic - one of two competing database solutions that was owned by IBM back in the day - the other one being a little thing called SQL - there was a winner and a loser in this competition (SQL), but for some reason, the loser never completely went away.

‡ How to shoot yourself in the foot, in Unidata: You try to shoot your foot, but only succeed in shooting off your big toe; your remaining toes move over to fill the gap.

33

Race conditions and collisions should be a basic consideration in design. Updates to multiple tables need to be written as a single transaction, which is possible in SQL ( and no doubt in other technologies, especially since I’ve retired).

Another consideration is avoiding duplicate updates. I worked on a system that sold streaming video by the minute. I was asked to go over and cleanup the payments functionality that had been written by another coder. I discovered a problem with the landing page which confirmed the payment and showed the updated minutes of use remaining. The problem was that simply reloading the page increased the number of minutes without any extra.charge. So, for example, if a user purchased 60 minutes, they could repeatedly add 60 more minutes to their time at no extra charge just by repeatedly reloading the landing page. The other coder was shown the door.

34

The first volume of the final report from the official Post Office Horizon IT Inquiry, chaired by Sir Wyn Williams, was published on July 9, 2025 . This report focuses on the human impact of the scandal and the issue of compensation for victims. It documents the profound suffering endured by sub-postmasters, including financial ruin, emotional distress, broken families, wrongful prosecutions, and even suicides as a result of the faulty Horizon IT system:

35

Just like the executive in a country with a codified constitution can’t deport people without due process?

36

This is a crazy story. Glad to see some justice.

37

“Mr Bates vs The Post Office” will soon be available to stream from Disney.

Well worth a watch IMO.

38

A technical question: I’d naïvely assume that the collisions and race conditions involved in the faulty system would produce approximately equal numbers and amounts of spurious cash shortfalls and spurious cash overages. Wouldn’t that have been a red flag from the beginning?