18.01.2024

Justice at a huge price

The Horizon scandal was about faulty software, individual gullibility and corporate indifference, writes Mike Macnair. But the fundamental problem dates back to the 1660s and the selling of legal services to the highest bidder

Mr Bates vs the Post Office had a wide impact, which took the government and the other actors in the Horizon scandal by surprise. Since then, Rishi Sunak and the Tory media have been struggling to get back control of the narrative.

As readers in the UK will know, between 1999 and 2015 over 900 sub-postmasters and sub-postmistresses were falsely prosecuted for theft, false accounting and fraud, even though the account shortfalls were due to errors of the Post Office’s Horizon accounting software. The announcement of a bill to reverse the convictions attempted to draw a line under the story - even if some lawyers complained this was inconsistent with the ‘rule of law’ and would open the door to ‘bills of attainder’ (acts of parliament to convict a person of high treason when they could not be convicted by an ordinary trial, as was employed against the Earl of Strafford in 1640, and against the Jacobite alleged plotter, Sir John Fenwick, in 1696).¹ A better analogy would be the ‘bills of pains and penalties’, which were used in the 18th and into the 19th century to convict people by act of parliament of crimes less than treason.²

Meanwhile, press coverage has shifted towards finding scapegoats: Paula Vennells, Post Office CEO 2012-19, has agreed to give up her CBE and been urged to repay her bonuses; Fujitsu software engineer Gareth Jenkins has been blamed for seeking immunity from criminal prosecution; the public inquiry cross-examination of Post Office investigator Stephen Bradshaw has been widely reported with strong spin against him; there have been calls for Lib Dem leader Sir Ed Davey (minister responsible for the post office during part of the scandal) to resign, and attention has been drawn to Tony Blair’s links in 1996-98 to Fujitsu.³ With Paul Patterson, Fujitsu’s manager for Europe, admitting that the firm has a “moral responsibility” to compensate the sub-postmasters, this may be a big enough scalp to limit the scope of the ongoing blame game.⁴

But the Horizon scandal was the natural and probable consequence of a policy choice - the hope to make radical productivity gains in ‘service industries’ through information technology, and for Britain to be a ‘world leader’ in this field - combined with an institutional feature of the constitution: that the king sells and denies justice, in violation of Magna Carta chapter 29, through the free market in legal services. The regime has no intention of giving up either the policy or judicial corruption through the free market in legal services. The parliamentary sticking-plaster, and the individualised media (and public inquiry) blame-game, serve to divert attention from these.

The Horizon contract was a public-private partnership (PPP) originally awarded under the John Major government, in 1996, to International Computers Ltd (ICL) at a time when ICL was increasingly dependent on collaboration with the Japanese firm, Fujitsu, but had not yet been actually taken over by it.

British champions

ICL was in origin one of the ‘British champions’ created by the promotion of takeovers and mergers under the 1964-70 Wilson Labour government, which hoped to create British firms large enough to compete globally with those based in the US (another example is the car manufacturer, British Leyland). The policy failed, in essence because of a misunderstanding of the causes of the relative decline of British industry. It was not that British firms were too small, but that Britain’s past global hegemon role led to continuing high returns in finance; this, in turn, forced industrial companies to pay high returns to shareholders and bondholders and high commercial rents, squeezing productive investment as a share of profits (while in the post-1945 settlement Britain lost the ability to protect its imperial markets).

The Thatcher and Major administrations largely, but not completely, accepted Britain’s future as an offshore financial centre, and hence considerable de-industrialisation. But there were efforts to import tech by collaboration with overseas firms: thus British Leyland, renamed as Rover Group, became an assembler for Honda cars from the 1980s. And thus ICL’s collaboration with Fujitsu from 1981, ending with the former as a mere subsidiary of the latter from 1999. ICL was already dependent on public-sector contracts at the time when the Horizon PPP was awarded.⁵

This was a large part of the context in which the Blair government decided in 1998 to go ahead with Horizon in spite of being alerted to concerns that the software was defective.⁶ Peter Mandelson warned that cancelling the contract would cause “political fallout” from Post Office closures and damage relations with Fujitsu, which he described as a “major investor in the UK over the past decade”.⁷ In substance, cancelling Horizon would threaten the viability of ICL, and the whole policy of attracting inward investment to acquire new technology, since the UK would appear to have sold Fujitsu a business dependent on government contracts, only to pull the plug shortly afterwards. The USA can and does get away with selling busted companies to foreign investors and then pulling the plug; the UK cannot.

The other side of the policy issue was the belief, which formed a large part of the policy of the Blair government, that IT could provide solutions to many economic and governmental problems. “Realising the potential of new technology” was a subhead of the 1997 Labour manifesto (though the content was not that extensive). Horizon - like many other failed public-sector IT projects of the period - was a part of this story.⁸ Also part of it were the Electronic Communications Act 2000, providing among other things for ‘digitally signed’ documents; the Land Registration Act 2002, which attempted (but has so far failed) to create fully-online registration and transfer of rights in land; and a whole range of other initiatives.

One of the ones which passed under the political radar (unlike the failed attempt to introduce identity cards) was the 1997 Law Commission report on Evidence in criminal proceedings: hearsay and related topics. This report recommended the repeal of section 69 of the 1984 Police and Criminal Evidence Act (PACE), which required the prosecution, if it sought to introduce as evidence documents produced by IT, to prove that the machine was working correctly when the document was produced. Section 69 was regarded as unduly onerous. The Law Commission asserted, on the basis of a 1984 decision in the Divisional Court of the Queen’s Bench Division about an ‘Intoximeter’ electronic breathalyser, that repeal would leave a common-law presumption that the machine was working properly.⁹

The problem with this approach was already identified in 1997, but then dismissed. What is the weight which this presumption is to be given at a trial, if anything which might call into question the functioning of the machine is offered against it? And how can defendants in criminal proceedings possibly challenge the presumption, given that they have no access to the underlying technical information about the functioning of the machine? The more complex the machine gets, the more severe the problem.

The issue was central to the actual miscarriages of justice in the Horizon case. The software was buggy and unreliable. The Post Office resisted admitting the fact, and prosecutors failed to disclose material which defendants could have used to raise the issue. Too much weight was given to the presumption. Serious lawyers and tech people have argued, after the miscarriages of justice and before Mr Bates vs the Post Office was broadcast, that something analogous to PACE section 69 needs to be restored.¹⁰

But the government is plainly reluctant to act on this - or indeed even to admit that the issue is worth discussing. The explanation is that this opens a huge can of worms in relation to the whole agenda of “realising the potential of new technology”, which has continued to be part of the agendas of governments, including those of David Cameron and his successors. Anything which demands more human (at the expense of automated) decision-making implies that less can be gained in productivity by the introduction of new tech than its proponents claim - and implies that the new tech will not justify the expected costs saving by sacking people.

In fact, this is not terribly surprising: new tech is commonly not producing the expected productivity gains.¹¹ If we think about this a little abstractly, the problem is that the machines are being asked to do increasingly complex tasks. And even the artificial intelligence devices do not learn in the same way that a human learns; with the result that additional information, for the machine, though not for a human, tends to degrade existing understanding.¹² The result is that the increasing complexity of the machines inevitably means increasing the ‘bugginess’ of the hardware and software. So, for example, jobs interfacing with customers are replaced with those fixing the problems of machines; and nearly every big government IT project - Horizon included - fails. The problem is not that governments are uniquely incapable of buying IT: it is that the equivalent failures in the private sector are - private.

IT can be a very valuable aid to human decision-making. But to try to use it as a complete substitute for human decision-making is to set up the conditions for endless repeats of variants of the Horizon scandal, as well as vast waste.

Selling justice

The human decision-making in this affair has been also marked by institutional bias. The Post Office was eventually forced to waive (for the purposes of the public inquiry) large numbers of non-disclosure agreements, which had been forced on sub-postmasters and sub-postmistresses as conditions of settlements of litigation.¹³ Of the £42 million agreed as a settlement of litigation once the Post Office finally lost in court, £31 million turned out to be paid to lawyers and litigation funders in the case. Numerous sub-postmasters/postmistresses have told stories of being intimidated by the high costs risks, and so on, into abandoning claims or settling them disadvantageously. There are good reasons to believe that the Post Office has pursued a tactic of dragging out negotiations in the hope that the sub-postmasters/mistresses will die before they are due to be paid.¹⁴

If these phenomena were unique to this case, they could be dealt with by a real ‘bill of pains and penalties’, requiring senior Post Office management (and the lawyers acting for the Post Office) to pay up, and seizing their assets into the hands of parliamentary sequestrators by way of enforcement. But in reality they are not unique or even unusual. They are examples of ‘scorched earth litigation tactics’. I have written about this before in connection with the battle over the Labour Party, in 2016, and in connection with the Tories ‘free speech tsar’ last year.¹⁵ They are, in fact, utterly normal in accident litigation, as was shown in the 1980s by the systematic study of settlement negotiations in Hazel Genn’s book Hard bargaining.¹⁶

This aspect of the Horizon case is a story of grinding the faces of a section of the lower middle classes for the sake of financial advantages which are relatively marginal, compared to the Post Office’s annual revenue of £957 million. The accident litigation is, similarly, a matter of grinding the faces of the relatively poor for the sake of marginal financial advantages to the insurance industry, whose annual revenues are in the tens of billions.¹⁷ Of course, as far as the insurers are concerned these calculations of marginal advantage have been essential to the business model since the beginning of the industry in late medieval Italy.

But the exploitation of ‘scorched earth litigation tactics’ has been an institutional instrument of the political power of the capitalist class since the 1650s, intensifying as time has gone on, albeit with episodes of retreat, like that of the period of ‘legal aid’ between 1949 and the 1980s.

Class constitution

Before the 1640s, the English legal system was characterised by the existence of a group of central courts (King’s Bench, Common Pleas, Exchequer, etc) and a large mass of local courts, both in boroughs in towns and in manors in the countryside, which dealt with very many small claims. The procedural forms of the common law were designed to provide rough and ready justice without excessive delay and cost (as contrasted with the church ‘canon law’, which allowed endless procedural manoeuvres).

Thus lord chancellor Bromley, inducting Sir Edmund Anderson as Chief Justice of Common Pleas in 1582, said, among other points, that in judging points of law,

there are four requisites for judges to observe. The first: that they do not introduce a general inconvenience in order to avoid a private mischief. The second: that they do not open any gap through which fraud can have passage. The third: that they not insist so greatly on form that they neglect substance. The fourth: that they prefer the intention before the letter.¹⁸

And a little later, he had this to say about cost and delay:

The fourth matter is diligence, which is to be exercised in speeding judicial causes, in order to avoid intolerable expenses to the suitors, and to prevent the danger which can arise from delay, as death of the parties, decrease of wealth, loss and lack of profit to the true owners; by which oppressions it can often be seen that he who sues for his right would much rather have a brief judgment pass against him than a long judgment for him.

A “general inconvenience” was a rule which would increase cost or delay for 90% of litigants; a “private mischief” was an injustice which might happen to 10% because of the effects of the rules designed to reduce cost and delay (as, for example, that a document might be lost, or a jury might be misled).

In the 1640s, the civil war caused a collapse of the central legal system at Westminster - not completely, but with a very radical practical reduction in caseload. People seem to have made do with local courts and remedies, which continued to operate. And in the 1650s, law reformers coming from the lower middle classes began to argue more forcibly for radical decentralisation of the judicial system.¹⁹

For the gentry and for the Westminster-based elite of the legal profession, this was the world turned upside-down. And, as the gentry got back political control from the lower orders - especially after the 1660 Restoration - the judges overthrew the principle that they must not “introduce a general inconvenience in order to avoid a private mischief”. Instead, avoiding “private mischiefs” led to a radical expansion of the possibilities of going to Westminster to attack local court judgments on the basis of very small technical errors in stating the record, and to attack jury verdicts on the ground that they were against the evidence or affected by the improper admission or exclusion of evidence or ‘misdirection’ by the judge, and endless other such procedural devices. These developments radically facilitated scorched earth litigation tactics.

By the early-mid 18th century the result was that the local courts of common law had been crushed and the lower middle classes had been effectively driven out of civil litigation. The process was by no means complete, and certain sorts of litigation and other legal business remained controlled by fixed fees into the 19th century. On the other hand, the Court of Chancery used procedures modelled on the ‘canon law’, which allowed very extensive possibilities for cost and delay: hence Charles Dickens’s portrayal of the court in Bleak House (1853). Bizarrely, but symptomatically, when Chancery and the ‘common law’ courts were unified in 1873-75, it was Chancery’s dilatory and expensive procedure which was adopted - “introduc[ing] a general inconvenience in order to avoid a private mischief” on the largest possible scale.

By then, the memory of the common law system as a means of producing rough and ready, but cheap and not too dilatory, justice had been forgotten. The ‘principle’ of the free market in legal services had become normalised. Justice is sold to the highest bidder and altogether denied to the lower orders.

It is fundamental then not to imagine that the Horizon/Post Office scandal is a matter of personal misconduct by small groups of individuals. It is the natural and probable consequence of the combination of government policy decisions with the institutional forms of the capitalist constitution. The government wants to give up neither the policy decisions - to expand IT decision-making in order to save labour costs - nor the institutional corruption that is the ‘free market in legal services’. Hence, even if the sub-postmasters/mistresses are in the end exonerated by statute and adequately compensated, this scandal will inevitably be followed by others that have the same dynamics.

To overcome the recurrent problems would mean starting from opposite principles: principles of human decision-making, as opposed to delusions in robotised decision-making.²⁰ It would mean localising as far as possible, and having steadily in our minds in relation to law and litigation the principle that judges should “not introduce a general inconvenience in order to avoid a private mischief.”

1. Announcement: www.gov.uk/government/news/government-to-quash-wrongful-post-office-convictions. For examples of criticism see www.lawgazette.co.uk/news-focus/news-focus-po-scandal-government-sets-legal-and-constitutional-precedent/5118396.article; www.lawgazette.co.uk/news/bill-to-exonerate-post-office-victims-an-exception-says-government/5118372.article; and www.theguardian.com/uk-news/2024/jan/11/legal-experts-post-office-horizon-exoneration-bill. There is a convenient short reference in Wikipedia to the ‘Bill of attainder’, though some of the later examples are of people who were dead or had escaped, rather than, as in the Strafford and Fenwick cases and the Horizon cases, ‘undue process’ in the ordinary legal system preventing just results.↩︎

2. See the discussion in O MacDonagh, ‘The last bill of pains and penalties: the case of Daniel O’Sullivan, 1869’ Irish Historical Studies Vol 19, 1974, pp136-37.↩︎

3. For Vennells see www.theguardian.com/uk-news/2024/jan/09/former-post-office-boss-paula-vennells-to-return-cbe-amid-horizon-scandal; Jenkins: www.telegraph.co.uk/news/2024/01/09/horizon-expert-gareth-jenkins-post-office-inquiry-immunity; Davey: www.standard.co.uk/news/politics/ed-davey-post-office-scandal-horizon-liberal-democrat-leader-b1132382.html. Blair: www.dailymail.co.uk/news/article-12960715/Sir-Tony-Blair-close-links-Fujitsu-Prime-Minister-signed-900m-Post-Office-contract-firm-1999-despite-warned-Horizon-accounting-software-looked-increasingly-flawed.html.↩︎

4. www.bbc.co.uk/news/business-67993493. But note that this is not an admission of legal responsibility.↩︎

5. www.footmanjames.co.uk/blog/the-golden-years-of-rover; see also archivesit.org.uk/icl.↩︎

6. www.theguardian.com/uk-news/2024/jan/12/tony-blair-was-warned-about-flawed-horizon-system-memo-shows.↩︎

7. news.sky.com/story/post-office-horizon-scandal-tony-blair-was-warned-system-could-be-flawed-when-he-was-prime-minsiter-13047150.↩︎

8. www.labour-party.org.uk/manifestos/1997/1997-labour-manifesto.shtml. On the fetishism of tech and its problems see, for instance, www.theguardian.com/technology/2006/mar/09/idcards.insideit.↩︎

9. Law Commission Report No245, part XIII; Castle v Cross (1984), available online at anticyberforensics.wordpress.com/2010/12/03/computer-print-out-through-human-mind-castle-v-cross-queens-bench-division-1985-1-all-er-87-1984-1-wlr-1372-1984-crim-lr-682-1985-rtr-62.↩︎

10. Eg, P Marshall et al, ‘Recommendations for the probity of computer evidence’ November 12 2020; N Bohm et al, ‘Briefing note: the rule that computers are presumed to be operating correctly - unforeseen and unjust consequences’ June 2022. As for now, see www.lawgazette.co.uk/law/it-experts-call-for-review-of-computer-is-always-right/5118414.article (January 15).↩︎

11. Eg, www.forbes.com/sites/adigaskell/2022/10/20/why-is-technology-not-producing-productivity-improvements.↩︎

12. www.ox.ac.uk/news/2024-01-03-study-shows-way-brain-learns-different-way-artificial-intelligence-systems-learn.↩︎

13. www.postofficehorizoninquiry.org.uk/news/non-disclosure-agreements-will-not-be-enforced-post-office-says-inquiry-chair.↩︎

14. www.scottishlegal.com/articles/horizon-scandal-authorities-accused-of-purposely-delaying-as-victims-die-without-payouts (April 27 2023); www.lawgazette.co.uk/commentary-and-opinion/post-office-scandal-lawyers-in-the-frame/5118335.article (January 5 2024).↩︎

15. ‘Scorched earth litigation’ Weekly Worker August 4 2016 (weeklyworker.co.uk/worker/1118/scorched-earth-litigation); ‘Knavery and folly’ Weekly Worker June 8 2023 (weeklyworker.co.uk/worker/1446/knavery-and-folly).↩︎

16. H Genn Hard bargaining: out of court settlement in personal injury actions Oxford 1987.↩︎

17. www.statista.com/statistics/1012113/post-office-annual-turnover; see also www.statista.com/topics/4511/insurance-industry-uk/#dossier-chapter2.↩︎

18. F Moore Cases collect and report per Sir Fra Moore, Chevalier 2nd edition London 1688, p116 (my translation from the ‘Law French’).↩︎

19. See, for example, D Veall The popular movement for law reform, 1640-1660 Oxford 1970; NL Matthews William Sheppard, Cromwell’s law reformer Cambridge 1985.↩︎

20. Cf S Meder Legal machines: of subsumption automata, artificial intelligence, and the search for the ‘correct’ judgment trans New Jersey 2023.↩︎